slide 1:
9 781292 023359
ISBN 9781292023359
Business Statistics
A DecisionMaking Approach
Groebner Shannon Fry
Ninth Edition
Business Statistics Groebner Shannon Fry 9eslide 2:
Business Statistics
A DecisionMaking Approach
Groebner Shannon Fry
Ninth Editionslide 3:
Pearson Education Limited
Edinburgh Gate
Harlow
Essex CM20 2JE
England and Associated Companies throughout the world
Visit us on the World Wide Web at: www.pearsoned.co.uk
© Pearson Education Limited 2014
All rights reserved. No part of this publication may be reproduced stored in a retrieval system or transmitted
in any form or by any means electronic mechanical photocopying recording or otherwise without either the
prior written permission of the publisher or a licence permitting restricted copying in the United Kingdom
issued by the Copyright Licensing Agency Ltd Saffron House 6–10 Kirby Street London EC1N 8TS.
All trademarks used herein are the property of their respective owners. The use of any trademark
in this text does not vest in the author or publisher any trademark ownership rights in such
trademarks nor does the use of such trademarks imply any afﬁ liation with or endorsement of this
book by such owners.
British Library CataloguinginPublication Data
A catalogue record for this book is available from the British Library
Printed in the United States of America
ISBN 10: 129202335X
ISBN 13: 9781292023359
ISBN 10: 129202335X
ISBN 13: 9781292023359slide 4:
Table of Contents
PEARSON C U S T OM LIBRA R Y
I
1. The Where Why and How of Data Collection
1
1 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
2. Graphs Charts and Tables  Describing Y our Data
33
33 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
3. Describing Data Using Numerical Measures
87
87 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
4. Special Review Section I
143
143 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
5. Introduction to Probability
151
151 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
6. Discrete Probability Distributions
197
197 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
7. Introduction to Continuous Probability Distributions
243
243 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
8. Introduction to Sampling Distributions
277
277 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
9. Estimating Single Population Parameters
319
319 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
10. Introduction to Hypothesis Testing
363
363 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
11. Estimation and Hypothesis Testing for Two Population Parameters
417
417 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
12. Hypothesis Tests and Estimation for Population Variances
469
469 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
13. Analysis of Variance
497
497 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smithslide 5:
II
14. Special Review Section II
551
551 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
15. GoodnessofFit Tests and Contingency Analysis
569
569 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
16. Introduction to Linear Regression and Correlation Analysis
601
601 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
17. Multiple Regression Analysis and Model Building
657
657 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
18. Analyzing and Forecasting TimeSeries Data
733
733 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
19. Introduction to Nonparametric Statistics
797
797 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
20. Introduction to Quality and Statistical Process Control
831
831 David F. Groebner/Patrick W. Shannon/Phillip C. Fry/Kent D. Smith
861
861 Indexslide 6:
The Where Why and How
of Data Collection
Quick Prep Links
Locate a recent copy of a business periodical
such as Fortune or Business Week and take
note of the graphs charts and tables that are
used in the articles and advertisements.
Recall any recent experiences you have
had in which you were asked to complete
a written survey or respond to a telephone
survey.
Make sure that you have access to Excel
software. Open Excel and familiarize yourself
with the software.
What Is Business Statistics
Procedures for Collecting
Data
Populations Samples and
Sampling Techniques
Data Types and Data
Measurement Levels
A Brief Introduction to
Data Mining
Outcome 1. Know the key data collection methods.
Why you need to know
A transformation is taking place in many organizations involving how managers are using data to help improve their
decision making. Because of the recent advances in software and database systems managers are able to analyze
data in more depth than ever before. A new discipline called data mining is growing and one of the fastestgrowing
career areas is referred to as business intelligence. Data mining or knowledge discovery is an interdisciplinary field
involving primarily computer science and statistics. People working in this field are referred to as “data scientists.”
Doing an Internet search on data mining will yield a large number of sites talking about the field.
In today’s workplace you can have an immediate competitive edge over other new employees and even
those with more experience by applying statistical analysis skills to realworld decision making. The purpose of this
text is to assist in your learning process and to complement your instructor’s efforts in conveying how to apply a
variety of important statistical procedures.
The major automakers such as GM Ford and Toyota maintain databases
with information on production quality customer satisfaction safety records and
much more. Walmart the world’s largest retail chain collects and manages mas
sive amounts of data related to the operation of its stores throughout the world.
Its highly sophisticated database systems contain sales data detailed customer
data employee satisfaction data and much more. Governmental agencies amass
extensive data on such things as unemployment interest rates incomes and
education. However access to data is not limited to large companies. The rela
tively low cost of computer hard drives with 100gigabyte or larger capacities
makes it possible for small firms and even individuals to store vast amounts of
Outcome 2. Know the difference between a population and
a sample.
Outcome 3. Understand the similarities and differences
between different sampling methods.
Outcome 4. Understand how to categorize data by type and
level of measurement.
Data Mining
The application of statistical techniques and
algorithms to the analysis of large data sets.
Business Intelligence
The application of tools and technologies for
gathering storing retrieving and analyzing data
that businesses collect and use.
Outcome 5. Become familiar with the concept of data mining
and some of its applications.
Anton Foltin/Shutterstock
From Chapter 1 of Business Statistics A DecisionMaking Approach Ninth Edition. David F. Groebner
Patrick W. Shannon and Phillip C. Fry. Copyright © 2014 by Pearson Education Inc. All rights reserved.slide 7:
The Where Why and How of Data Collection
data on desktop computers. But without some way to transform the data into useful information the data these compa
nies have gathered are of little value.
Transforming data into information is where business statistics comes in—the statistical procedures introduced
in this text are those that are used to help transform data into information. This text focuses on the practical applica
tion of statistics we do not develop the theory you would find in a mathematical statistics course. Will you need to use
math in this course Yes but mainly the concepts covered in your college algebra course.
Statistics does have its own terminology. You will need to learn various terms that have special statistical mean
ing. You will also learn certain dos and don’ts related to statistics. But most importantly you will learn specific meth
ods to effectively convert data into information. Don’t try to memorize the concepts rather go to the next level of
learning called understanding. Once you understand the underlying concepts you will be able to think statistically.
Because data are the starting point for any statistical analysis this text is devoted to discussing various aspects
of data from how to collect data to the different types of data that you will be analyzing. You need to gain an under
standing of the where why and how of data and data collection.
What Is Business Statistics
Articles in your local newspaper news stories on television and national publications such
as the Wall Street Journal and Fortune discuss stock prices crime rates governmentagency
budgets and company sales and profit figures. These values are statistics but they are just
a small part of the discipline called business statistics which provides a wide variety of
methods to assist in data analysis and decision making.
Descriptive Statistics
Business statistics can be segmented into two general categories. The first category involves
the procedures and techniques designed to describe data such as charts graphs and numeri
cal measures. The second category includes tools and techniques that help decision makers
draw inferences from a set of data. Inferential procedures include estimation and hypothesis
testing. A brief discussion of these techniques follows.
BUSINESS APPLICATION DESCRIBING DATA
INDEPENDENT TEXTBOOK PUBLISHING INC. Independent Textbook Publishing
Inc. publishes 15 collegelevel texts in the business and social sciences areas. Figure 1
shows an Excel spreadsheet containing data for each of these 15 textbooks. Each column
Business Statistics
A collection of procedures and techniques
that are used to convert data into meaningful
information in a business environment.
Excel 2010 Instructions:
1. Open File: Independent
Textbook.xlsx.
FIGURE 1 
Excel 2010 Spreadsheet
of Independent Textbook
Publishing Inc.slide 8:
The Where Why and How of Data Collection
in the spreadsheet corresponds to a different factor for which data were collected. Each
row corresponds to a different textbook. Many statistical procedures might help the owners
describe these textbook data including descriptive techniques such as charts graphs and
numerical measures.
Charts and Graphs Other text will discuss many different charts and graphs—such as the
one shown in Figure 2 called a histogram. This graph displays the shape and spread of the
distribution of number of copies sold. The bar chart shown in Figure 3 shows the total num
ber of textbooks sold broken down by the two markets business and social sciences.
Bar charts and histograms are only two of the techniques that could be used to graphi
cally analyze the data for the textbook publisher.
BUSINESS APPLICATION DESCRIBING DATA
CROWN INVESTMENTS At Crown Investments a senior analyst is preparing to present
data to upper management on the 100 fastestgrowing companies on the Hong Kong Stock
Exchange. Figure 4 shows an Excel worksheet containing a subset of the data. The columns
correspond to the different items of interest growth percentage sales and so on. The data
for each company are in a single row. The entire data are in a file called Fast100.
Number of Books
Under 50000 50000 100000 100000 150000 150000 200000
Number of Copies Sold
Independent Textbook Publishing Inc. Distribution of Copies Sold
0
1
2
3
4
5
6
7
8
FIGURE 2 
Histogram Showing the
Copies Sold Distribution
0 100000 200000 300000 400000 500000 600000 700000 800000
Market Classifcation
Total Copies Sold
Total Copies Sold by Market Class
Social
Sciences
Business
FIGURE 3 
Bar Chart Showing Copies
Sold by Sales Categoryslide 9:
The Where Why and How of Data Collection
In addition to preparing appropriate graphs the analyst will compute important numeri
cal measures. One of the most basic and most useful measures in business statistics is one
with which you are already familiar: the arithmetic mean or average.
Arithmetic Mean or Average
The sum of all values divided by the number of
values.
–99 indicates missing data
Excel 2010 Instructions:
1. Open file: Fast100.xlsx.
FIGURE 4 
Crown Investment Example
Average
The sum of all the values divided by the number of values. In equation form:
Average
a
N
i 1
x
i
N
Sum of all data values
Number of data values
1
where:
N Number of data values
x
i
ith data value
The analyst may be interested in the average profit that is the average of the col
umn labeled “Profits” for the 100 companies. The total profit for the 100 companies
is 3193.60 but profits are given in millions of dollars so the total profit amount is
actually 3193600000. The average is found by dividing this total by the number of
companies:
Average
+3193600000
100
+31936000 or +31.936 million
The average or mean is a measure of the center of the data. In this case the ana
lyst may use the average profit as an indicator—firms with aboveaverage profits are
rated higher than firms with belowaverage profits.
The graphical and numerical measures illustrated here are only some of the many
descriptive procedures that will be introduced elsewhere. The key to remember is that the
purpose of any descriptive procedure is to describe data. Your task will be to select the proce
dure that best accomplishes this. As Figure 5 reminds you the role of statistics is to convert
data into meaningful information.slide 10:
The Where Why and How of Data Collection
Inferential Procedures
Advertisers pay for television ads based on the audience level so knowing how many viewers
watch a particular program is important millions of dollars are at stake. Clearly the networks
don’t check with everyone in the country to see if they watch a particular program. Instead
they pay a fee to the Nielsen company http://www.nielsen.com/ which uses statistical
inference procedures to estimate the number of viewers who watch a particular television
program.
There are two primary categories of statistical inference procedures: estimation
and hypothesis testing. These procedures are closely related but serve very different
purposes.
Estimation In situations in which we would like to know about all the data in a large data
set but it is impractical to work with all the data decision makers can use techniques to esti
mate what the larger data set looks like. The estimates are formed by looking closely at a
subset of the larger data set.
BUSINESS APPLICATION STATISTICAL INFERENCE
NEW PRODUCT INTRODUCTION Energyboosting drinks such as Red Bull Go Girl
Monster and Full Throttle have become very popular among college students and young
professionals. But how do the companies that make these products determine whether they
will sell enough to warrant the product introduction A typical approach is to do market
research by introducing the product into one or more test markets. People in the targeted
age income and educational categories target market are asked to sample the product
and indicate the likelihood that they would purchase the product. The percentage of people
who say that they will buy forms the basis for an estimate of the true percentage of all
people in the target market who will buy. If that estimate is high enough the company will
introduce the product.
Hypothesis Testing Television advertising is full of product claims. For example
we might hear that “Goodyear tires will last at least 60000 miles” or that “more doctors
recommend Bayer Aspirin than any other brand.” Other claims might include statements
like “General Electric light bulbs last longer than any other brand” or “customers prefer
McDonald’s over Burger King.” Are these just idle boasts or are they based on actual data
Probably some of both However consumer research organizations such as Consumers
Union publisher of Consumer Reports regularly test these types of claims. For example
in the hamburger case Consumer Reports might select a sample of customers who would
be asked to blind taste test Burger King’s and McDonald’s hamburgers under the hypoth
esis that there is no difference in customer preferences between the two restaurants. If the
sample data show a substantial difference in preferences then the hypothesis of no differ
ence would be rejected. If only a slight difference in preferences was detected then Con
sumer Reports could not reject the hypothesis.
Statistical Inference Procedures
Procedures that allow a decision maker to reach
a conclusion about a set of data based on a
subset of that data.
FIGURE 5 
The Role of Business
Statistics
Information Data Descriptive
Inferential
Statistical Proceduresslide 11:
The Where Why and How of Data Collection
MyStatLab
Journal. Find three examples of the use of a graph to
display data. For each graph
a. Give the name date and page number of the
periodical in which the graph appeared.
b. Describe the main point made by the graph.
c. Analyze the effectiveness of the graphs.
112. The human resources manager of an automotive supply
store has collected the following data showing the number
of employees in each of five categories by the number of
days missed due to illness or injury during the past year.
Missed Days 0–2 days 3–5 days 6–8 days 8–10 days
Employees 159 67 32 10
Construct the appropriate chart for these data. Be sure
to use labels and to add a title to your chart.
113. Suppose Fortune would like to determine the average
age and income of its subscribers. How could statistics
be of use in determining these values
114. Locate an example from a business periodical or
newspaper in which estimation has been used.
a. What specifically was estimated
b. What conclusion was reached using the estimation
c. Describe how the data were extracted and how they
were used to produce the estimation.
d. Keeping in mind the goal of the estimation discuss
whether you believe that the estimation was
successful and why.
e. Describe what inferences were drawn as a result of
the estimation.
115. Locate one of the online job Web sites and pick several
job listings. For each job type discuss one or more
situations in which statistical analyses would be used.
Base your answer on research Internet business
periodicals personal interviews etc.. Indicate whether
the situations you are describing involve descriptive
statistics or inferential statistics or a combination of both.
116. Suppose SuperValue a major retail food company
is thinking of introducing a new product line into
a market area. It is important to know the age
characteristics of the people in the market area.
a. If the executives wish to calculate a number that
would characterize the “center” of the age data
what statistical technique would you suggest
Explain your answer.
b. The executives need to know the percentage of
people in the market area that are senior citizens.
Name the basic category of statistical procedure
they would use to determine this information.
c. Describe a hypothesis the executives might wish to
test concerning the percentage of senior citizens in
the market area.
Skill Development
11. For the following situation indicate whether the
statistical application is primarily descriptive or
inferential.
“The manager of Anna’s Fabric Shop has collected data for
10 years on the quantity of each type of dress fabric that
has been sold at the store. She is interested in making a
presentation that will illustrate these data effectively.”
12. Consider the following graph that appeared in a company
annual report. What type of graph is this Explain.
45000
40000
35000
30000
25000
20000
15000
10000
5000
Fruit
Vegetables
Meat and
Poultry
FOOD STORE SALES
Monthly Sales
Canned Goods
Department
Cereal and
Dry Goods
Other
0
13. Review Figures 2 and 3 and discuss any differences
you see between the histogram and the bar chart.
14. Think of yourself as working for an advertising firm.
Provide an example of how hypothesis testing can be
used to evaluate a product claim.
15. Define what is meant by hypothesis testing. Provide
an example in which you personally have tested a
hypothesis even if you didn’t use formal statistical
techniques to do so.
16. In what situations might a decision maker need to use
statistical inference procedures
17. Explain under what circumstances you would use
hypothesis testing as opposed to an estimation
procedure.
18. Discuss any advantages a graph showing a whole set of
data has over a single measure such as an average.
19. Discuss any advantages a single measure such as an
average has over a table showing a whole set of data.
Business Applications
110. Describe how statistics could be used by a business
to determine if the dishwasher parts it produces last
longer than a competitor’s brand.
111. Locate a business periodical such as Fortune or Forbes
or a business newspaper such as The Wall Street
1 Exercises
END EXERCISES 11slide 12:
The Where Why and How of Data Collection
Procedures for Collecting Data
We have defined business statistics as a set of procedures that are used to transform data
into information. Before you learn how to use statistical procedures it is important that you
become familiar with different types of data collection methods.
Data Collection Methods
There are many methods and procedures available for collecting data. The following are con
sidered some of the most useful and frequently used data collection methods:
● Experiments
● Telephone surveys
● Written questionnaires and surveys
● Direct observation and personal interviews
BUSINESS APPLICATION EXPERIMENTS
FOOD PROCESSING A company often must conduct a specific experiment or set of
experiments to get the data managers need to make informed decisions. For example Lamb
Weston McCain and the J. R. Simplot Company are the primary suppliers of french fries to
McDonald’s in North America. At its Caldwell Idaho factory the J. R. Simplot Company
has a test center that among other things houses a mini french fry plant used to conduct
experiments on its potato manufacturing process. McDonald’s has strict standards on the
quality of the french fries it buys. One important attribute is the color of the fries after
cooking. They should be uniformly “golden brown”—not too light or too dark.
French fries are made from potatoes that are peeled sliced into strips blanched partially
cooked and then freezedried—not a simple process. Because potatoes differ in many ways
such as sugar content and moisture blanching time cooking temperature and other factors
vary from batch to batch.
Simplot employees start their experiments by grouping the raw potatoes into batches
with similar characteristics. They run some of the potatoes through the line with blanch time
and temperature settings set at specific levels defined by an experimental design. After
measuring one or more output variables for that run employees change the settings and run
another batch again measuring the output variables.
Figure 6 shows a typical data collection form. The output variable for example percent
age of fries without dark spots for each combination of potato category blanch time and
temperature is recorded in the appropriate cell in the table.
Experiment
A process that produces a single outcome
whose result cannot be predicted with certainty.
Experimental Design
A plan for performing an experiment in which
the variable of interest is defined. One or
more factors are identified to be manipulated
changed or observed so that the impact or
influence on the variable of interest can be
measured or observed.
Chapter Outcome 1.
FIGURE 6 
Data Layout for the French Fry
Experiment
Potato Category
1 234 Blanch Temperature Blanch Time
100
110
120
10 minutes
100
110
120
15 minutes
100
110
120
20 minutes
100
110
120
25 minutesslide 13:
The Where Why and How of Data Collection
BUSINESS APPLICATION TELEPHONE SURVEYS
PUBLIC ISSUES Chances are that you have been on the receiving end of a telephone
call that begins something like: “Hello. My name is Mary Jane and I represent the XYZ
organization. I am conducting a survey on …” Political groups use telephone surveys to poll
people about candidates and issues. Marketing research companies use phone surveys to
learn likes and dislikes of potential customers.
Telephone surveys are a relatively inexpensive and efficient data collection procedure.
Of course some people will refuse to respond to a survey others are not home when the
calls come and some people do not have home phones—only have a cell phone—or cannot
be reached by phone for one reason or another. Figure 7 shows the major steps in conducting
a telephone survey. This example survey was run a few years ago by a Seattle television sta
tion to determine public support for using tax dollars to build a new football stadium for the
National Football League’s Seattle Seahawks. The survey was aimed at property tax
payers
only.
Because most people will not stay on the line very long the phone survey must be
short—usually one to three minutes. The questions are generally what are called closedend
questions. For example a closedend question might be “To which political party do you
belong Republican Democrat Or other”
The survey instrument should have a short statement at the beginning explaining the
purpose of the survey and reassuring the respondent that his or her responses will remain
confidential. The initial section of the survey should contain questions relating to the central
issue of the survey. The last part of the survey should contain demographic questions such
as gender income level education level that will allow you to break down the responses and
look deeper into the survey results.
ClosedEnd Questions
Questions that require the respondent to select
from a short list of defined choices.
Demographic Questions
Questions relating to the respondents’
characteristics backgrounds and attributes.
FIGURE 7 
Major Steps for a Telephone
Survey
Determine
Sample Size and
Sampling Method
Pretest
the
Survey
Defne the
Population
of Interest
Select Sample
and
Make Calls
Develop
Survey
Questions
Defne the
Issue
Do taxpayers favor a special bond to build a new football stadium
for the Seahawks If so should the Seahawks’ owners share the cost
Population is all residential property tax payers in King County
Washington. Te survey will be conducted among this group only.
Limit the number of questions to keep survey short.
Ask important questions frst. Provide specifc response options
when possible.
Establish eligibility. “Do you own a residence in King County”
Add demographic questions at the end: age income etc.
Introduction should explain purpose of survey and who is
conducting it—stress that answers are anonymous.
Try the survey out on a small group from the population. Check for
length clarity and ease of conducting. Have we forgotten anything
Make changes if needed.
Get phone numbers from a computergenerated or “current” list.
Develop “callback” rule for no answers. Callers should be trained to
ask questions fairly. Do not lead the respondent. Record responses
on data sheet.
Sample size is dependent on how confdent we want to be of our
results how precise we want the results to be and how much
opinions difer among the population members. Various sampling
methods are available.slide 14:
The Where Why and How of Data Collection
A survey budget must be considered. For example if you have 3000 to spend on calls
and each call costs 10 to make you obviously are limited to making 300 calls. However
keep in mind that 300 calls may not result in 300 usable responses.
The phone survey should be conducted in a short time period. Typically the prime
calling time for a voter survey is between 7:00 p.m. and 9:00 p.m. However some people
are not home in the evening and will be excluded from the survey unless there is a plan for
conducting callbacks.
Written Questionnaires and Surveys The most frequently used method to collect
opinions and factual data from people is a written questionnaire. In some instances the
questionnaires are mailed to the respondent. In others they are administered directly
to the potential respondents. Written questionnaires are generally the least expensive
means of collecting survey data. If they are mailed the major costs include postage to
and from the respondents questionnaire development and printing costs and data anal
ysis. Figure 8 shows the major steps in conducting a written survey. Note how written
surveys are similar to telephone surveys however written surveys can be slightly more
involved and therefore take more time to complete than those used for a telephone
survey. However you must be careful to construct a questionnaire that can be easily
completed without requiring too much time.
A written survey can contain both closedend and openend questions. Openend ques
tions provide the respondent with greater flexibility in answering a question however the
responses can be difficult to analyze. Note that telephone surveys can use openend ques
tions too. However the caller may have to transcribe a potentially long response and there is
risk that the interviewees’ comments may be misinterpreted.
Written surveys also should be formatted to make it easy for the respondent to provide
accurate and reliable data. This means that proper space must be provided for the responses
OpenEnd Questions
Questions that allow respondents the freedom to
respond with any value words or statements of
their own choosing.
FIGURE 8 
Written Survey Steps
Determine
Sample Size and
Sampling Method
Pretest
the
Survey
Defne the
Population
of Interest
Select Sample
and
Send Surveys
Design the
Survey
Instrument
Defne the
Issue
Clearly state the purpose of the survey. Defne the objectives. What
do you want to learn from the survey Make sure there is agreement
before you proceed.
Defne the overall group of people to be potentially included in the
survey and obtain a list of names and addresses of those individuals
in this group.
Limit the number of questions to keep the survey short.
Ask important questions frst. Provide specifc response options
when possible.
Add demographic questions at the end: age income etc.
Introduction should explain purpose of survey and who is
conducting it—stress that answers are anonymous.
Layout of the survey must be clear and attractive. Provide location
for responses.
Try the survey out on a small group from the population. Check for
length clarity and ease of conducting. Have we forgotten anything
Make changes if needed.
Mail survey to a subset of the larger group.
Include a cover letter explaining the purpose of the survey.
Include prestamped return envelope for returning the survey.
Sample size is dependent on how confdent we want to be of our
results how precise we want the results to be and how much
opinions difer among the population members. Various sampling
methods are available.slide 15:
The Where Why and How of Data Collection
and the directions must be clear about how the survey is to be completed. A written survey
needs to be pleasing to the eye. How it looks will affect the response rate so it must look
professional.
Y ou also must decide whether to manually enter or scan the data gathered from your writ
ten survey. The survey design will be affected by the approach you take. If you are adminis
tering a large number of surveys scanning is preferred. It cuts down on data entry errors and
speeds up the data gathering process. However you may be limited in the form of responses
that are possible if you use scanning.
If the survey is administered directly to the desired respondents you can expect a high
response rate. For example you probably have been on the receiving end of a written survey
many times in your college career when you were asked to fill out a course evaluation form
at the end of the term. Most students will complete the form. On the other hand if a survey
is administered through the mail you can expect a low response rate—typically 5 to 20.
Therefore if you want 200 responses you should mail out 1000 to 4000 questionnaires.
Overall written surveys can be a lowcost effective means of collecting data if you can
overcome the problems of low response. Be careful to pretest the survey and spend extra time
on the format and look of the survey instrument.
Developing a good written questionnaire or telephone survey instrument is a major chal
lenge. Among the potential problems are the following:
● Leading questions
Example: “Do you agree with most other reasonably minded people that the city
should spend more money on neighborhood parks”
Issue: In this case the phrase “Do you agree” may suggest that you should agree.
Also by suggesting that “most reasonably minded people” already agree the
respondent might be compelled to agree so that he or she can also be consid
ered “reasonably minded.”
Improvement: “In your opinion should the city increase spending on neighbor
hood parks”
Example: “To what extent would you support paying a small increase in your prop
erty taxes if it would allow poor and disadvantaged children to have food and
shelter”
Issue: The question is ripe with emotional feeling and may imply that if you don’t
support additional taxes you don’t care about poor children.
Improvement: “Should property taxes be increased to provide additional funding
for social services”
● Poorly worded questions
Example: “How much money do you make at your current job”
Issue: The responses are likely to be inconsistent. When answering does the
respondent state the answer as an hourly figure or as a weekly or monthly
total Also many people refuse to answer questions regarding their income.
Improvement: “Which of the following categories best reflects your weekly
income from your current job
Under 500
500–1000
Over 1000”
Example: “After trying the new product please provide a rating from 1 to 10 to
indicate how you like its taste and freshness.”
Issue: First is a low number or a high number on the rating scale considered a
positive response Second the respondent is being asked to rate two factors
taste and freshness in a single rating. What if the product is fresh but does not
taste good
Improvement: “After trying the new product please rate its taste on a 1 to 10 scale
with 1 being best. Also rate the product’s freshness using the same 1 to 10
scale.
Taste
Freshness”
The way a question is worded can influence the responses. Consider an example that
occurred in September 2008 during the financial crisis that resulted from the subprimeslide 16:
The Where Why and How of Data Collection
mortgage crisis and bursting of the real estate bubble. Three surveys were conducted on the
same basic issue. The following questions were asked:
“Do you approve or disapprove of the steps the Federal Reserve and Treasury Depart
ment have taken to try to deal with the current situation involving the stock market
and major financial institutions” ABC News/Washington Post 44 Approve — 42
Disappro v e —14 Unsure
“Do you think the government should use taxpayers’ dollars to rescue ailing private
financial firms whose collapse could have adverse effects on the economy and market
or is it not the government’s responsibility to bail out private companies with taxpayer
dollars” LA Times/Bloomberg 31 Use Tax Payers’ Dollars — 55 Not Government’s
Responsibility— 14 Unsure
“As you may know the government is potentially investing billions to try and keep
financial institutions and markets secure. Do you think this is the right thing or the wrong
thing for the government to be doing” Pew Research Center 57 Right Thing — 30
Wrong Thing—13 Unsure
Note the responses to each of these questions. The way the question is worded can affect
the responses.
Direct Observation and Personal Interviews Direct observation is another procedure
that is often used to collect data. As implied by the name this technique requires the pro
cess from which the data are being collected to be physically observed and the data recorded
based on what takes place in the process.
Possibly the most basic way to gather data on human behavior is to watch people. If you
are trying to decide whether a new method of displaying your product at the supermarket will
be more pleasing to customers change a few displays and watch customers’ reactions. If as
a member of a state’s transportation department you want to determine how well motorists
are complying with the state’s seat belt laws place observers at key spots throughout the state
to monitor people’s seat belt habits. A movie producer seeking information on whether a
new movie will be a success holds a preview showing and observes the reactions and com
ments of the movie patrons as they exit the screening. The major constraints when collecting
observations are the amount of time and money required. For observations to be effective
trained observers must be used which increases the cost. Personal observation is also time
consuming. Finally personal perception is subjective. There is no guarantee that different
observers will see a situation in the same way much less report it the same way.
Personal interviews are often used to gather data from people. Interviews can be either
structured or unstructured depending on the objectives and they can utilize either
openend or closedend questions.
Regardless of the procedure used for data collection care must be taken that the data
collected are accurate and reliable and that they are the right data for the purpose at hand.
Other Data Collection Methods
Data collection methods that take advantage of new technologies are becoming more prev
alent all the time. For example many people believe that Walmart is one of the best com
panies in the world at collecting and using data about the buying habits of its customers.
Most of the data are collected automatically as checkout clerks scan the UPC bar codes on
the products customers purchase. Not only are Walmart’s inventory records automatically
updated but information about the buying habits of customers is also recorded. This allows
Walmart to use analytics and data mining to drill deep into the data to help with its deci
sion making about many things including how to organize its stores to increase sales. For
instance Walmart apparently decided to locate beer and disposable diapers close together
when it discovered that many male customers also purchase beer when they are sent to the
store for diapers.
Bar code scanning is used in many different data collection applications. In a DRAM
dynamic randomaccess memory wafer fabrication plant batches of silicon wafers have
bar codes. As the batch travels through the plant’s workstations its progress and quality are
tracked through the data that are automatically obtained by scanning.
Unstructured Interview
Interviews that begin with one or more broadly
stated questions with further questions being
based on the responses.
Structured Interview
Interviews in which the questions are scripted.slide 17:
The Where Why and How of Data Collection
Every time you use your credit card data are automatically collected by the retailer
and the bank. Computer information systems are developed to store the data and to provide
decision makers with procedures to access the data.
In many instances your data collection method will require you to use physical measurement.
For example the Andersen Window Company has quality analysts physically measure the width
and height of its windows to assure that they meet customer specifications and a state Department
of Weights and Measures will physically test meat and produce scales to determine that customers
are being properly charged for their purchases.
Data Collection Issues
Data Accuracy When you need data to make a decision we suggest that you first see if
appropriate data have already been collected because it is usually faster and less expensive
to use existing data than to collect data yourself. However before you rely on data that
were collected by someone else for another purpose you need to check out the source to
make sure that the data were collected and recorded properly.
Such organizations as Bloomberg V alue Line and Fortune have built their reputations on
providing quality data. Although data errors are occasionally encountered they are few and
far between. You really need to be concerned with data that come from sources with which
you are not familiar. This is an issue for many sources on the World Wide Web. Any organiza
tion or any individual can post data to the Web. Just because the data are there doesn’t mean
they are accurate. Be careful.
Interviewer Bias There are other general issues associated with data collection. One of
these is the potential for bias in the data collection. There are many types of bias. For exam
ple in a personal interview the interviewer can interject bias either accidentally or on pur
pose by the way she asks the questions by the tone of her voice or by the way she looks
at the subject being interviewed. We recently allowed ourselves to be interviewed at a trade
show. The interviewer began by telling us that he would only get credit for the interview if we
answered all of the questions. Next he asked us to indicate our satisfaction with a particular
display. He wasn’t satisfied with our lessthanenthusiastic rating and kept asking us if we
really meant what we said. He even asked us if we would consider upgrading our rating How
reliable do you think these data will be
Nonresponse Bias Another source of bias that can be interjected into a survey data
collection process is called nonresponse bias. We stated earlier that mail surveys suffer from a
high percentage of unreturned surveys. Phone calls don’t always get through or people refuse
to answer. Subjects of personal interviews may refuse to be interviewed. There is a potential
problem with nonresponse. Those who respond may provide data that are quite different from
the data that would be supplied by those who choose not to respond. If you aren’t careful the
responses may be heavily weighted by people who feel strongly one way or another on an issue.
Selection Bias Bias can be interjected through the way subjects are selected for data
collection. This is referred to as selection bias. A study on the virtues of increasing the stu
dent athletic fee at your university might not be best served by collecting data from students
attending a football game. Sometimes the problem is more subtle. If we do a telephone sur
vey during the evening hours we will miss all of the people who work nights. Do they share
the same views income education levels and so on as people who work days If not the
data are biased.
Written and phone surveys and personal interviews can also yield flawed data if the inter
viewees lie in response to questions. For example people commonly give inaccurate data
about such sensitive matters as income. Lying is also an increasing problem with exit polls in
which voters are asked who they voted for immediately after casting their vote. Sometimes
the data errors are not due to lies. The respondents may not know or have accurate informa
tion to provide the correct answer.
Observer Bias Data collection through personal observation is also subject to problems.
People tend to view the same event or item differently. This is referred to as observer bias.
Bias
An effect that alters a statistical result by
systematically distorting it different from a
random error which may distort on any one
occasion but balances out on the average.slide 18:
The Where Why and How of Data Collection
One area in which this can easily occur is in safety check programs in companies. An impor
tant part of behavioralbased safety programs is the safety observation. Trained data collec
tors periodically conduct a safety observation on a worker to determine what if any unsafe
acts might be taking place. We have seen situations in which two observers will conduct an
observation on the same worker at the same time yet record different safety data. This is
especially true in areas in which judgment is required on the part of the observer such as the
distance a worker is from an exposed gear mechanism. People judge distance differently.
Measurement Error A few years ago we were working with a wood window manufac
turer. The company was having a quality problem with one of its saws. A study was devel
oped to measure the width of boards that had been cut by the saw. Two people were trained to
use digital calipers and record the data. This caliper is a Ushaped tool that measures distance
in inches to three decimal places. The caliper was placed around the board and squeezed
tightly against the sides. The width was indicated on the display. Each person measured 500
boards during an 8hour day. When the data were analyzed it looked like the widths were
coming from two different saws one set showed considerably narrower widths than the other.
Upon investigation we learned that the person with the narrower width measurements was
pressing on the calipers much more firmly. The soft wood reacted to the pressure and gave
narrower readings. Fortunately we had separated the data from the two data collectors. Had
they been merged the measurement error might have gone undetected.
Internal Validity When data are collected through experimentation you need to make sure
that proper controls have been put in place. For instance suppose a drug company such as
Pfizer is conducting tests on a drug that it hopes will reduce cholesterol. One group of test
participants is given the new drug while a second group a control group is given a placebo.
Suppose that after several months the group using the drug saw significant cholesterol reduc
tion. For the results to have internal validity the drug company would have had to make
sure the two groups were controlled for the many other factors that might affect cholesterol
such as smoking diet weight gender race and exercise habits. Issues of internal validity are
generally addressed by randomly assigning subjects to the test and control groups. However
if the extraneous factors are not controlled there could be no assurance that the drug was
the factor influencing reduced cholesterol. For data to have internal validity the extraneous
factors must be controlled.
External Validity Even if experiments are internally valid you will always need to be con
cerned that the results can be generalized beyond the test environment. For example if the
cholesterol drug test had been performed in Europe would the same basic results occur for
people in North America South America or elsewhere For that matter the drug company
would also be interested in knowing whether the results could be replicated if other subjects
are used in a similar experiment. If the results of an experiment can be replicated for groups
different from the original population then there is evidence the results of the experiment
have external validity.
An extensive discussion of how to measure the magnitude of bias and how to reduce bias
and other data collection problems is beyond the scope of this text. However you should be
aware that data may be biased or otherwise flawed. Always pose questions about the potential
for bias and determine what steps have been taken to reduce its effect.
Internal Validity
A characteristic of an experiment in which data
are collected in such a way as to eliminate the
effects of variables within the experimental
environment that are not of interest to the
researcher.
External Validity
A characteristic of an experiment whose results
can be generalized beyond the test environment
so that the outcomes can be replicated when
the experiment is repeated.
Skill Development
117. If a pet store wishes to determine the level of customer
satisfaction with its services would it be appropriate to
conduct an experiment Explain.
118. Define what is meant by a leading question. Provide an
example.
119. Briefly explain what is meant by an experiment and an
experimental design.
1 Exercises
MyStatLabslide 19:
The Where Why and How of Data Collection
day you receive an email containing a questionnaire
asking you to rate the quality of the experience. Discuss
both the advantages and disadvantages of using this form
of questionnaire delivery.
128. In your capacity as assistant sales manager for a large
office products retailer you have been assigned the
task of interviewing purchasing managers for medium
and large companies in the San Francisco Bay area.
The objective of the interview is to determine the office
product buying plans of the company in the coming
year. Develop a personal interview form that asks
both issuerelated questions as well as demographic
questions.
129. The regional manager for Macy’s is experimenting with
two new endofaisle displays of the same product. An
endofaisle display is a common method retail stores
use to promote new products. You have been hired
to determine which is more effective. Two measures
you have decided to track are which display causes
the highest percentage of people to stop and for those
who stop which causes people to view the display the
longest. Discuss how you would gather such data.
130. In your position as general manager for United Fitness
Center you have been asked to survey the customers
of your location to determine whether they want to
convert the racquetball courts to an aerobic exercise
space. The plan calls for a written survey to be handed
out to customers when they arrive at the fitness center.
Your task is to develop a short questionnaire with
at least three “issue” questions and at least three
demographic questions. You also need to provide the
finished layout design for the questionnaire.
131. According to a national CNN/USA/Gallup survey of
1025 adults conducted March 14–16 2008 63 say
they have experienced a hardship because of rising
gasoline prices. How do you believe the survey was
conducted and what types of bias could occur in the
data collection process
120. Refer to the three questions discussed in this section
involving the financial crises of 2008 and 2009 and
possible government intervention. Note that the
questions elicited different responses. Discuss the way
the questions were worded and why they might have
produced such different results.
121. Suppose a survey is conducted using a telephone
survey method. The survey is conducted from 9 a.m. to
11 a.m. on Tuesday. Indicate what potential problems
the data collectors might encounter.
122. For each of the following situations indicate what type
of data collection method you would recommend and
discuss why you have made that recommendation:
a. collecting data on the percentage of bike riders who
wear helmets
b. collecting data on the price of regular unleaded
gasoline at gas stations in your state
c. collecting data on customer satisfaction with the
service provided by a major U.S. airline
123. Assume you have received a class assignment to
determine the attitude of students in your school
toward the school’s registration process. What are the
validity issues you should be concerned with
Business Applications
124. According to a report issued by the U.S. Department
of Agriculture USDA the agency estimates that
Southern fire ants spread at a rate of 4 to 5 miles a
year. What data collection method do you think was
used to collect this data Explain your answer.
125. Suppose you are asked to survey students at your
university to determine if they are satisfied with the
food service choices on campus. What types of biases
must you guard against in collecting your data
126. Briefly describe how new technologies can assist
businesses in their data collection efforts.
127. Assume you have used an online service such as Orbitz or
Travelocity to make an airline reservation. The following
END EXERCISES 12
Populations Samples and
Sampling Techniques
Populations and Samples
Two of the most important terms in statistics are population and sample.
The list of all objects or individuals in the population is referred to as the frame. Each
object or individual in the frame is known as a sampling unit. The choice of the frame depends
on what objects or individuals you wish to study and on the availability of the list of these
objects or individuals. Once the frame is defined it forms the list of sampling units. The next
example illustrates this concept.
BUSINESS APPLICATION POPULATIONS AND SAMPLES
U.S. BANK We can use U.S. Bank to illustrate the difference between a population and a
sample. U.S. Bank is very concerned about the time customers spend waiting in the driveup
teller line. At a particular U.S. Bank on a given day 347 cars arrived at the driveup.
Population
The set of all objects or individuals of interest or
the measurements obtained from all objects or
individuals of interest.
Sample
A subset of the population.
Chapter Outcome 2.slide 20:
The Where Why and How of Data Collection
A population includes measurements made on all the items of interest to the data gath
erer. In our example the U.S. Bank manager would define the population as the waiting time
for all 347 cars. The list of these cars possibly by license number forms the frame. If she
examines the entire population she is taking a census. But suppose 347 cars are too many to
track. The U.S. Bank manager could instead select a subset of these cars called a sample. The
manager could use the sample results to make statements about the population. For example
she might calculate the average waiting time for the sample of cars and then use that to con
clude what the average waiting time is for the population.
There are tradeoffs between taking a census and taking a sample. Usually the main
tradeoff is whether the information gathered in a census is worth the extra cost. In organiza
tions in which data are stored on computer files the additional time and effort of taking a
census may not be substantial. However if there are many accounts that must be manually
checked a census may be impractical.
Another consideration is that the measurement error in census data may be greater than
in sample data. A person obtaining data from fewer sources tends to be more complete and
thorough in both gathering and tabulating the data. As a result with a sample there are likely
to be fewer human errors.
Parameters and Statistics Descriptive numerical measures such as an average or a pro
portion that are computed from an entire population are called parameters. Corresponding
measures for a sample are called statistics. Suppose in the previous example the U.S. Bank
manager timed every car that arrived at the driveup teller on a particular day and calculated
the average. This population average waiting time would be a parameter. However if she
selected a sample of cars from the population the average waiting time for the sampled cars
would be a statistic.
Sampling Techniques
Once a manager decides to gather information by sampling he or she can use a sampling
technique that falls into one of two categories: statistical or nonstatistical.
Both nonstatistical and statistical sampling techniques are commonly used by decision
makers. Regardless of which technique is used the decision maker has the same objective—
to obtain a sample that is a close representative of the population. There are some advantages
to using a statistical sampling technique as we will discuss many times throughout this text.
However in many cases nonstatistical sampling represents the only feasible way to sample
as illustrated in the following example.
BUSINESS APPLICATION NONSTATISTICAL SAMPLING
SUNCITRUS ORCHARDS SunCitrus Orchards owns
and operates a large fruit orchard and fruitpacking plant in
Florida. During harvest time in the orange grove pickers load
20pound sacks with oranges which are then transported to
the packing plant. At the packing plant the oranges are graded
and boxed for shipping nationally and internationally. Because
of the volume of oranges involved it is impossible to assign a
quality grade to each individual orange. Instead as each sack moves up the conveyor into the
packing plant a quality manager selects an orange sack every so often grades the individual
oranges in the sack as to size color and so forth and then assigns an overall quality grade to
the entire shipment from which the sample was selected.
Because of the volume of oranges the quality manager at SunCitrus uses a nonstatis
tical sampling method called convenience sampling. In doing so the quality manager is
willing to assume that orange quality size color etc. is evenly spread throughout the many
sacks of oranges in the shipment. That is the oranges in the sacks selected are of the same
quality as those that were not inspected.
There are other nonstatistical sampling methods such as judgment sampling and ratio
sampling which are not discussed here. Instead the most frequently used statistical sampling
techniques will now be discussed.
Census
An enumeration of the entire set of
measurements taken from the whole population.
Statistical Sampling Techniques
Those sampling methods that use selection
techniques based on chance selection.
Nonstatistical Sampling
Techniques
Those methods of selecting samples using
convenience judgment or other nonchance
processes.
Convenience Sampling
A sampling technique that selects the items
from the population based on accessibility and
ease of selection.
Elena Elisseeva/Shutterstockslide 21:
The Where Why and How of Data Collection
Statistical Sampling Statistical sampling methods also called probability sampling
allow every item in the population to have a known or calculable chance of being included in
the sample. The fundamental statistical sample is called a simple random sample. Other types
of statistical sampling discussed in this text include stratified random sampling systematic
sampling and cluster sampling.
BUSINESS APPLICATION SIMPLE RANDOM SAMPLING
CABLEONE A salesperson at CableOne wishes to estimate the percentage of people in a
local subdivision who have satellite television service such as Direct TV. The result would
indicate the extent to which the satellite industry has made inroads into CableOne’s market.
The population of interest consists of all families living in the subdivision.
For this example we simplify the situation by saying that there are only five families in
the subdivision: James Sanchez Lui White and Fitzpatrick. We will let N represent the pop
ulation size and n the sample size. From the five families N 5 we select three n 3
for the sample. There are 10 possible samples of size 3 that could be selected.
James Sanchez Lui James Sanchez White James Sanchez Fitzpatrick
James Lui White James Lui Fitzpatrick James White Fitzpatrick
Sanchez Lui White Sanchez Lui Fitzpatrick Sanchez White Fitzpatrick
Lui White Fitzpatrick
Note that no family is selected more than once in a given sample. This method is called sam
pling without replacement and is the most commonly used method. If the families could be
selected more than once the method would be called sampling with replacement.
Simple random sampling is the method most people think of when they think of ran
dom sampling. In a correctly performed simple random sample each of these samples would
have an equal chance of being selected. For the CableOne example a simplified way of
selecting a simple random sample would be to put each sample of three names on a piece of
paper in a bowl and then blindly reach in and select one piece of paper. However this method
would be difficult if the number of possible samples were large. For example if N 50 and
a sample of size n 10 is to be selected there are more than 10 billion possible samples. Try
finding a bowl big enough to hold those
Simple random samples can be obtained in a variety of ways. We present two examples
to illustrate how simple random samples are selected in practice.
BUSINESS APPLICATION RANDOM NUMBERS
STATE SOCIAL SERVICES Suppose the state director for a Midwestern state’s social
services system is considering changing the timing on food stamp distribution from once a
month to once every two weeks. Before making any decisions he wants to survey a sample
of 100 citizens who are on food stamps in a particular county from the 300 total food stamp
recipients in that county. He first assigns recipients a number 001 to 300. He can then use
the random number function in Excel to determine which recipients to include in the sample.
Figure 9 shows the results when Excel chooses 10 random numbers. The first recipient
sampled is number 115 followed by 31 and so forth. The important thing to remember is that
assigning each recipient a number and then randomly selecting a sample from those numbers
gives each possible sample an equal chance of being selected.
RANDOM NUMBERS TABLE If you don’t have access to computer software such as
Excel the items in the population to be sampled can be determined by using the random
numbers table. Begin by selecting a starting point in the random numbers table row and
digit. Suppose we use row 5 digit 8 as the starting point. Go down 5 rows and over 8 digits.
Verify that the digit in this location is 1. Ignoring the blanks between columns that are there
only to make the table more readable the first threedigit number is 149. Recipient number
149 is the first one selected in the sample. Each subsequent random number is obtained
from the random numbers in the next row down. For instance the second number is 127.
The procedure continues selecting numbers from top to bottom in each subsequent column.
Numbers exceeding 300 and duplicate numbers are skipped. When enough numbers are
Simple Random Sampling
A method of selecting items from a population
such that every possible sample of a specified
size has an equal chance of being selected.
Chapter Outcome 3.
Excel Tutorial
Excel
tutorialsslide 22:
The Where Why and How of Data Collection
found for the desired sample size the process is completed. Foodstamp recipients whose
numbers are chosen are then surveyed.
BUSINESS APPLICATION STRATIFIED RANDOM SAMPLING
FEDERAL RESERVE BANK Sometimes the sample size
required to obtain a needed level of information from a simple
random sampling may be greater than our budget permits. At other
times it may take more time to collect than is available. Stratified
random sampling is an alternative method that has the potential
to provide the desired information with a smaller sample size. The
following example illustrates how stratified sampling is performed.
Each year the Federal Reserve Board asks its staff to estimate the total cash holdings of
U.S. financial institutions as of July 1. The staff must base the estimate on a sample. Note that
not all financial institutions banks credit unions and the like are the same size. A majority
are small some are medium sized and only a few are large. However the few large institu
tions have a substantial percentage of the total cash on hand. To make sure that a simple
random sample includes an appropriate number of small medium and large institutions the
sample size might have to be quite large.
As an alternative to the simple random sample the Federal Reserve staff could divide the
institutions into three groups called strata: small medium and large. Staff members could
then select a simple random sample of institutions from each stratum and estimate the total
cash on hand for all institutions from this combined sample. Figure 10 shows the stratified
random sampling concept. Note that the combined sample size n
1
+ n
2
+ n
3
is the sum of
the simple random samples taken from each stratum.
The key behind stratified sampling is to develop a stratum for each characteristic of inter
est such as cash on hand that has items that are quite homogeneous. In this example the
size of the financial institution may be a good factor to use in stratifying. Here the combined
sample size n
1
+ n
2
+ n
3
will be less than the sample size that would have been required
if no stratification had occurred. Because sample size is directly related to cost in both time
and money a stratified sample can be more cost effective than a simple random sample.
Multiple layers of stratification can further reduce the overall sample size. For example
the Federal Reserve might break the three strata in Figure 10 into substrata based on type of
institution: state bank interstate bank credit union and so on.
Most largescale market research studies use stratified random sampling. The wellknown
political polls such as the Gallup and Harris polls use this technique also. For instance the
Gallup poll typically samples between 1800 and 2500 people nationwide to estimate how
more than 60 million people will vote in a presidential election. We encourage you to go to
the Web site http://www.gallup.com/poll/101872/howdoesgalluppollingwork.aspx to read
a very good discussion about how the Gallup polls are conducted. The Web site discusses
how samples are selected and many other interesting issues associated with polling.
Stratified Random Sampling
A statistical sampling method in which the
population is divided into subgroups called strata
so that each population item belongs to only
one stratum. The objective is to form strata such
that the population values of interest within each
stratum are as much alike as possible. Sample
items are selected from each stratum using the
simple random sampling method.
Excel 2010 Instructions:
1. On the Data tab click
Data Analysis.
2. Select Random Number
Generation option.
3. Set the Number of
Random Numbers to 10.
4. Select Uniform as the
distribution.
5. Defne range as between
1 and 300.
6. Indicate that the results
are to go in cell A1.
7. Click OK.
To convert numbers to integers
select the data in column A and
on the Home tab in the Number
group. Click the Decrease
decimal button several times
to remove the decimal places.
FIGURE 9 
Excel 2010 Output of Random
Numbers for State Social
Services Example
Jonathan Larsen/Shutterstockslide 23:
The Where Why and How of Data Collection
BUSINESS APPLICATION SYSTEMATIC RANDOM SAMPLING
STATE UNIVERSITY ASSOCIATED STUDENTS A few years ago elected student
council officers at midsized state university in the Northeast decided to survey fellow
students on the issue of the legality of carrying firearms on campus. To determine the opinion
of its 20000 students a questionnaire was sent to a sample of 500 students. Although simple
random sampling could have been used an alternative method called systematic random
sampling was chosen.
The university’s systematic random sampling plan called for it to send the question
naire to every 40th student 20000500 40 from an alphabetic list of all students. The
process could begin by using Excel to generate a single random number in the range 1 to
40. Suppose this value was 25. The 25th student in the alphabetic list would be selected.
After that every 40th students would be selected 25 65 105 145 . . . until there were
500 students selected.
Systematic sampling is frequently used in business applications. Use it as an alternative
to simple random sampling only when you can assume the population is randomly ordered
with respect to the measurement being addressed in the survey. In this case students’ views
on firearms on campus are likely unrelated to the spelling of their last name.
BUSINESS APPLICATION CLUSTER SAMPLING
OAKLAND RAIDERS FOOTBALL TEAM The Oakland Raiders of the National Football
League plays its home games at O.co formerly Overstock.com Coliseum in Oakland California.
Despite its struggles to win in recent years the team has a passionate fan base. Recently an
outside marketing group was retained by the Raiders to interview season ticket holders about
the potential for changing how season ticket pricing is structured. The Oakland Raiders Web site
http://www.raiders.com/tickets/seatingpricemap.html shows the layout of the O.co Coliseum.
The marketing firm plans to interview season ticket holders just prior to home games
during the current season. One sampling technique is to select a simple random sample of
size n from the population of all season ticket holders. Unfortunately this technique would
likely require that interviewers go to each section in the stadium. This would prove to be an
expensive and timeconsuming process. A systematic or stratified sampling procedure also
would probably require visiting each section in the stadium. The geographical spread of those
being interviewed in this case causes problems.
A sampling technique that overcomes the geographical spread problem is cluster
sampling. The stadium sections would be the clusters. Ideally the clusters would each
have the same characteristics as the population as a whole.
Systematic Random Sampling
A statistical sampling technique that involves
selecting every kth item in the population after a
randomly selected starting point between 1 and
k. The value of k is determined as the ratio of
the population size over the desired sample size.
Cluster Sampling
A method by which the population is divided into
groups or clusters that are each intended to be
minipopulations. A simple random sample of
m clusters is selected. The items chosen from
a cluster can be selected using any probability
sampling technique.
Large Institutions
Population:
Cash Holdings
of All Financial
Institutions in
the United States
Stratifed Population
Financial Institutions
MediumSize Institutions
Small Institutions
Stratum 1
Stratum 2
Stratum 3
Select n
1
Select n
2
Select n
3
FIGURE 10 
Stratified Sampling Exampleslide 24:
The Where Why and How of Data Collection
After the clusters have been defined a sample of m clusters is selected at random from
the list of possible clusters. The number of clusters to select depends on various factors
including our survey budget. Suppose the marketing firm randomly selects eight clusters:
104  142  147  218  228  235  307  327
These are the primary clusters. Next the marketing company can either survey all the
ticketholders in each cluster or select a simple random sample of ticketholders from each
cluster depending on time and budget considerations.
Skill Development
132. Indicate which sampling method would most likely be
used in each of the following situations:
a. an interview conducted with mayors of a sample of
cities in Florida
b. a poll of voters regarding a referendum calling for a
national valueadded tax
c. a survey of customers entering a shopping mall in
Minneapolis
133. A company has 18000 employees. The file containing
the names is ordered by employee number from 1 to
18000. If a sample of 100 employees is to be selected
from the 18000 using systematic random sampling
within what range of employee numbers will the first
employee selected come from
134. Describe the difference between a statistic and a
parameter.
135. Why is convenience sampling considered to be a
nonstatistical sampling method
136. Describe how systematic random sampling could be
used to select a random sample of 1000 customers
who have a certificate of deposit at a commercial bank.
Assume that the bank has 25000 customers who own a
certificate of deposit.
137. Explain why a census does not necessarily have to
involve a population of people. Use an example to
illustrate.
138. If the manager at First City Bank surveys a sample
of 100 customers to determine how many miles they
live from the bank is the mean travel distance for this
sample considered a parameter or a statistic Explain.
139. Explain the difference between stratified random
sampling and cluster sampling.
140. Use Excel to generate five random numbers between 1
and 900.
Business Applications
141. According to the U.S. Bureau of Labor Statistics the
annual percentage increase in U.S. college tuition
and fees in 1995 was 6.0 in 1999 it was 4.0 in
2004 it was 9.5 and in 2011 it was 5.4. Are these
percentages statistics or parameters Explain.
142. According to an article in the Idaho Statesman a poll
taken the day before elections in Germany showed
Chancellor Gerhard Schroeder behind his challenger
Angela Merkel by 6 to 8 percentage points. Is this a
statistic or a parameter Explain.
143. Give the name of the kind of sampling that was most
likely used in each of the following cases:
a. a Wall Street Journal poll of 2000 people to
determine the president’s approval rating
b. a poll taken of each of the General Motors GM
dealerships in Ohio in December to determine an
estimate of the average number of Chevrolets not
yet sold by GM dealerships in the United States
c. a qualityassurance procedure within a FritoLay
manufacturing plant that tests every 1000th bag of
Fritos Corn Chips produced to make sure the bag is
sealed properly
d. a sampling technique in which a random sample
from each of the tax brackets is obtained by the
Internal Revenue Service to audit tax returns
144. Your manager has given you an Excel file that contains
the names of the company’s 500 employees and has
asked you to sample 50 employees from the list. You
decide to take your sample as follows. First you assign
a random number to each employee using Excel’s
random number function Rand. Because the random
number is volatile it recalculates itself whenever
you modify the file you freeze the random numbers
using the Copy—Paste Special—Values feature. You
then sort by the random numbers in ascending order.
Finally you take the first 50 sorted employees as your
sample. Does this approach constitute a statistical or a
nonstatistical sample
Computer Applications
145. Sysco Foods is a statewide food distributor to
restaurants universities and other establishments that
prepare and sell food. The company has a very large
warehouse in which the food is stored until it is pulled
from the shelves to be delivered to the customers. The
warehouse has 64 storage racks numbered 164.
Each rack is three shelves high labeled A B and C and
each shelf is divided into 80 sections numbered 180.
1 Exercises
MyStatLabslide 25:
The Where Why and How of Data Collection
END EXERCISES 13
Products are located by rack number shelf letter and
section number. For example breakfast cereal is located
at 43A52 rack 43 shelf A section 52.
Each week employees perform an inventory for a
sample of products. Certain products are selected and
counted. The actual count is compared to the book count
the quantity in the records that should be in stock. To
simplify things assume that the company has selected
breakfast cereals to inventory. Also for simplicity’s sake
suppose the cereals occupy racks 1 through 5.
a. Assume that you plan to use simple random
sampling to select the sample. Use Excel to
determine the sections on each of the five racks to
be sampled.
b. Assume that you wish to use cluster random
sampling to select the sample. Discuss the steps you
would take to carry out the sampling.
c. In this case why might cluster sampling be
preferred over simple random sampling Discuss.
146. United Airlines established a discount airline named
Ted. The managers were interested in determining how
flyers using Ted rate the airline service. They plan to
question a random sample of flyers from the November
12 flights between Denver and Fort Lauderdale. A
total of 578 people were on the flights that day. United
has a list of the travelers together with their mailing
addresses. Each traveler is given an identification
number here from 001 to 578. Use Excel to generate
a list of 40 flyer identification numbers so that those
identified can be surveyed.
147. The National Park Service has started charging a user
fee to park at selected trailheads and crosscountry
ski lots. Some users object to this fee claiming they
already pay taxes for these areas. The agency has
decided to randomly question selected users at fee
areas in Colorado to assess the level of concern.
a. Define the population of interest.
b. Assume a sample of 250 is required. Describe the
technique you would use to select a sample from
the population. Which sampling technique did you
suggest
c. Assume the population of users is 4000. Use Excel
to generate a list of users to be selected for the
sample.
148. Mount Hillsdale Hospital has more than 4000 patient
files listed alphabetically in its computer system. The
office manager wants to survey a statistical sample of
these patients to determine how satisfied they were
with service provided by the hospital. She plans to use
a telephone survey of 100 patients.
a. Describe how you would attach identification
numbers to the patient files for example how many
digits and which digits would you use to indicate
the first patient file
b. Describe how the first random number would be
obtained to begin a simple random sample method.
c. How many random digits would you need for each
random number you selected
d. Use Excel to generate the list of patients to be
surveyed.
Data Types and Data
Measurement Levels
As you will see the statistical techniques deal with different types of data. The level of mea
surement may vary greatly from application to application. In general there are four types of
data: quantitative qualitative timeseries and crosssectional. A discussion of each follows.
Quantitative and Qualitative Data
In some cases data values are best expressed in purely numerical or quantitative terms
such as in dollars pounds inches or percentages. As an example a cell phone provider
might collect data on the number of outgoing calls placed during a month by its customers.
In another case a sports bar could collect data on the number of pitchers of beer sold weekly.
In other situations the observation may signify only the category to which an item
belongs. Categorical data are referred to as qualitative data.
For example a bank might conduct a study of its outstanding real estate loans and keep
track of the marital status of the loan customer—single married divorced or other. The
same study also might examine the credit status of the customer—excellent good fair or
poor. Still another part of the study might ask the customers to rate the service by the bank on
a 1 to 5 scale with 1 very poor 2 poor 3 neutral 4 good and 5 very good.
Note although the customers are asked to record a number 1 to 5 to indicate the service
quality the data would still be considered qualitative because the numbers are just codes for
the categories.
Quantitative Data
Measurements whose values are inherently
numerical.
Qualitative Data
Data whose measurement scale is inherently
categorical.
Chapter Outcome 4.slide 26:
The Where Why and How of Data Collection
TimeSeries Data and CrossSectional Data
Data may also be classified as being either timeseries or crosssectional.
The data collected by the bank about its loan customers would be crosssectional because
the data from each customer relates to a fixed point in time. In another case if we sampled
100 stocks from the stock market and determined the closing stock price on March 15 the
data would be considered crosssectional because all measurements corresponded to one
point in time.
On the other hand Ford Motor Company tracks the sales of its F150 pickup trucks on a
monthly basis. Data values observed at intervals over time are referred to as timeseries data.
If we determined the closing stock price for a particular stock on a daily basis for a year the
stock prices would be timeseries data.
Data Measurement Levels
Data can also be identified by their level of measurement. This is important because the higher
the data level the more sophisticated the analysis that can be performed.
We shall discuss and give examples of four levels of data measurements: nominal ordinal
interval and ratio. Figure 11 illustrates the hierarchy among these data levels with nominal
data being the lowest level.
Nominal Data Nominal data are the lowest form of data yet you will encounter this type
of data many times. Assigning codes to categories generates nominal data. For example a
survey question that asks for marital status provides the following responses:
1. Married 2. Single 3. Divorced 4. Other
For each person a code of 1 2 3 or 4 would be recorded. These codes are nominal data.
Note that the values of the code numbers have no specific meaning because the order of the
categories is arbitrary. We might have shown it this way:
1. Single 2. Divorced 3. Married 4. Other
With nominal data we also have complete control over what codes are used. For exam
ple we could have used
88. Single 11. Divorced 33. Married 55. Other
All that matters is that you know which code stands for which category. Recognize also
that the codes need not be numeric. We might use
S Single D Divorced M Married O Other
TimeSeries Data
A set of consecutive data values observed at
successive points in time.
Ordinal Data
Ratio/Interval Data
Categorical Codes
ID Numbers
Category Names
Rankings
Ordered Categories
Lowest Level
Basic Analysis
Higher Level
MidLevel Analysis
Highest Level
Complete Analysis
Measurements
Nominal Data
FIGURE 11 
Data Level Hierarchy
CrossSectional Data
A set of data values observed at a fixed point
in time.slide 27:
The Where Why and How of Data Collection
Ordinal Data Ordinal or rank data are one notch above nominal data on the measure
ment hierarchy. At this level the data elements can be rankordered on the basis of some
relationship among them with the assigned values indicating this order. For example a
typical market research technique is to offer potential customers the chance to use two
unidentified brands of a product. The customers are then asked to indicate which brand
they prefer. The brand eventually offered to the general public depends on how often it was
the preferred test brand. The fact that an ordering of items took place makes this an ordinal
measure.
Bank loan applicants are asked to indicate the category corresponding to their household
incomes:
Under 20000
20000 to 40000
over 40000
1 2 3
The codes 1 2 and 3 refer to the particular income categories with higher codes assigned to
higher incomes.
Ordinal measurement allows decision makers to equate two or more observations or to
rankorder the observations. In contrast nominal data can be compared only for equality.
You cannot order nominal measurements. Thus a primary difference between ordinal and
nominal data is that ordinal data can have both an equality and a greater than 7 or a
less than 6 relationship whereas nominal data can have only an equality relationship.
Interval Data If the distance between two data items can be measured on some scale and
the data have ordinal properties 7 6 or the data are said to be interval data. The best
example of interval data is the temperature scale. Both the Fahrenheit and Celsius tempera
ture scales have ordinal properties of “.” or “” and “5” In addition the distances between
equally spaced points are preserved. For example 32°F 7 30°F and 80°C 7 78°C. The dif
ference between 32°F and 30°F is the same as the difference between 80°F and 78°F two
degrees in each case. Thus interval data allow us to precisely measure the difference between
any two values. With ordinal data this is not possible because all we can say is that one value
is larger than another.
Ratio Data Data that have all the characteristics of interval data but also have a true zero
point at which zero means “none” are called ratio data. Ratio measurement is the highest
level of measurement.
Packagers of frozen foods encounter ratio measures when they pack their products by
weight. Weight whether measured in pounds or grams is a ratio measurement because it
has a unique zero point—zero meaning no weight. Many other types of data encountered in
business environments involve ratio measurements for example distance money and time.
The difference between interval and ratio measurements can be confusing because
it involves the definition of a true zero. If you have 5 and your brother has 10 he
has twice as much money as you. If you convert the dollars to pounds euros yen or
pesos your brother will still have twice as much. If your money is lost or stolen you
have no dollars. Money has a true zero. Likewise if you travel 100 miles today and
200 miles tomorrow the ratio of distance traveled will be 2/1 even if you convert the
distance to kilometers. If on the third day you rest you have traveled no miles. Dis
tance has a true zero. Conversely if today’s temperature is 35°F 1.67°C and tomorrow’s
is 70°F 21.11°C is tomorrow twice as warm as today The answer is no. One way to
see this is to convert the Fahrenheit temperature to Celsius: The ratio will no longer be
21 12.641. Likewise if the temperature reads 0°F 17.59°C this does not imply that
there is no temperature. It’s simply colder than 10°F 12.22°C Also 0°C 32°F is not
the same temperature as 0°F. Thus temperature measured with either the Fahrenheit or
Celsius scale an intervallevel variable does not have a true zero.
As was mentioned earlier a major reason for categorizing data by level and type is that
the methods you can use to analyze the data are partially dependent on the level and type of
data you have available.slide 28:
The Where Why and How of Data Collection
EXAMPLE 1 CATEGORIZING DATA
For many years U.S. News and World Report has published
annual rankings based on various data collected from U.S.
colleges and universities. Figure 12 shows a portion of the data
in the file named Colleges and Universities. Each column cor
responds to a different variable for which data were collected.
Before doing any statistical analyses with these data U.S.
News and World Report employees need to determine the type
and level for each of the factors. Limiting the effort to only those factors that are shown in
Figure 12 this is done using the following steps:
Step 1 Identify each factor in the data set.
The factors or variables in the data set shown in Figure 12 are
College State Public 1 Math V erbal appli. appli. new FT PT
Name Private 2 SAT SAT rec’d. accepted. stud. under under
enrolled grad grad
Each of the 10 columns represents a different factor. Data might be missing for
some colleges and universities.
Step 2 Determine whether the data are timeseries or crosssectional.
Because each row represents a different college or university and the data are
for the same year the data are crosssectional. Timeseries data are measured
over time—say over a period of years.
Step 3 Determine which factors are quantitative data and which are qualitative data.
Qualitative data are codes or numerical values that represent categories.
Quantitative data are those that are purely numerical. In this case the data for
the following factors are qualitative:
College Name
State
Code for Public or Private College or University
Data for the following factors are considered quantitative:
Math SAT Verbal SAT new stud. enrolled
appl. rec’d. appl. accepted
PT undergrad FT undergrad
FIGURE 12 
Data for U.S. Colleges and
Universities
Joe Gough/Shutterstockslide 29:
The Where Why and How of Data Collection
Step 4 Determine the level of data measurement for each factor.
The four levels of data are nominal ordinal interval and ratio. This data set
has only nominal and ratiolevel data. The three nominallevel factors are
College Name
State
Code for Public or Private College or University
The others are all ratiolevel data.
Skill Development
149. For each of the following indicate whether the data are
crosssectional or timeseries:
a. quarterly unemployment rates
b. unemployment rates by state
c. monthly sales
d. employment satisfaction data for a company
150. What is the difference between qualitative and
quantitative data
151. For each of the following variables indicate the level
of data measurement:
a. product rating 1 excellent 2 good 3 fair
4 poor 5 very poor
b. home ownership own rent other
c. college grade point average
d. marital status single married divorced other
152. What is the difference between ordinal and nominal
data
153. Consumer Reports in its rating of cars indicates
repair history with circles. The circles are either white
black or half and half. To which level of data does this
correspond Discuss.
Business Applications
154. Verizon has a support center customers can call to get
questions answered about their cell phone accounts.
The manager in charge of the support center has
recently conducted a study in which she surveyed
2300 customers. The customers who called the support
center were transferred to a third party who asked the
customers a series of questions.
a. Indicate whether the data generated from this study
will be considered crosssectional or timeseries.
Explain why.
b. One of the questions asked customers was
approximately how many minutes they had been
on hold waiting to get through to a support person.
What level of data measurement is obtained from
this question Explain.
c. Another question asked the customer to rate the
service on a scale of 1–7 with 1 being the worst
possible service and 7 being the best possible
service. What level of data measurement is achieved
from this question Will the data be quantitative or
qualitative Explain.
155. The following information can be found in the
Murphy Oil Corporation Annual Report to Share
holders. For each variable indicate the level of data
measurement.
a. List of Principal Offices e.g. El Dorado Calgary
Houston
b. Income in millions of dollars from Continuing
Operations
c. List of Principal Subsidiaries e.g. Murphy Oil
USA Inc. Murphy Exploration Production
Company
d. Number of branded retail outlets
e. Petroleum products sold in barrels per day
f. Major Exploration and Production Areas e.g.
Malaysia Congo Ecuador
g. Capital Expenditures measured in millions of
dollars
156. You have collected the following information on 15
different real estate investment trusts REITs. Identify
whether the data are crosssectional or timeseries.
a. income distribution by region in 2012
b. per share diluted funds from operations FFO for
the years 2006 to 2012
c. number of properties owned as of December 31
2012
d. the overall percentage of leased space for the 119
properties in service as of December 31 2012
e. dividends per share for the years 2006–2012
157. A loan manager for Bank of the Cascades has the
responsibility for approving automobile loans. To
assist her in this matter she has compiled data on
428 cars and trucks. These data are in the file called
2004Automobiles.
1 Exercises
MyStatLabslide 30:
The Where Why and How of Data Collection
END EXERCISES 14
Column A Column B Column C Column D Column E Column F
Account Number Caller Gender
Account Holder
Gender Past Due Amount
Current Amount
Due
Was This a Billing
Question
Unique Tracking 1 Male 1 Male Numerical V alue Numerical V alue 3 Yes
2 Female 2 Female 4 No
A small portion of the data is as follows:
Indicate the level of data measurement for each of the
variables in this data file.
158. Recently the manager of the call center for a large
Internet bank asked his staff to collect data on a
Account Number Caller Gender
Account Holder
Gender Past Due Amount
Current Amount
Due
Was This a
Billing Question
4348291 2 2 40.35 82.85 3
6008516 1 1 0 129.67 4
17476479 1 2 0 76.38 4
13846306 2 2 0 99.24 4
21393711 1 1 0 37.98 3
random sample of the bank’s customers. Data on the
following variables were collected and placed in a file
called Bank Call Center:
A Brief Introduction to Data Mining
Data Mining—Finding the Important
Hidden Relationships in Data
What food products have an increased demand during hurricanes How do you win baseball
games without star players Is my best friend the one to help me find a job What color
car is least likely to be a “lemon” These and other interesting questions can and have been
answered using data mining. Data mining consists of applying sophisticated statistical tech
niques and algorithms to the analysis of big data i.e. the wealth of new data that organiza
tions collect in many and varied forms. Through the application of data mining decisions
can now be made on the basis of statistical analysis rather than on only managerial intuition
and experience. The statistical techniques introduced in this text provide the basis for the
more sophisticated statistical tools that are used by data mining analysts.
WalMart the nation’s largest retailer uses data mining to help it tailor product selec
tion based on the sales demographic and weather information it collects. While WalMart
managers might not be surprised that the demand for flashlights batteries and bottled water
increased with hurricane warnings they were surprised to find that there was also an increase
in the demand for strawberry PopTarts before hurricanes hit. This knowledge allowed Wal
Mart to increase the availability of PopTarts at selected stores affected by the hurricane
alerts. The McKinsey Global Institute estimates that the full application of data mining to
retailing could result in a potential increase in operating margins by as much as 60. Source:
McKinsey Global Institute: Big Data: The Next Frontier for Innovation Competition and
Productivity May 2011 by James Manyika Michael Chui Brad Brown Jacques Bughin
Richard Dobbs Charles Roxburgh Angela Hung Byers.
Chapter Outcome 5.
a. Would you classify these data as timeseries or crosssectional Explain.
b. Which of the variables are quantitative and which are qualitative
c. For each of the six variables indicate the level of data measurement.slide 31:
The Where Why and How of Data Collection
Data are everywhere and businesses are collecting more each day. Accounting and sales
data are now captured and streamed instantly when transactions occur. Digital sensors in
industrial equipment and automobiles can record and report data on vibration temperature
physical location and the chemical composition of the surrounding air. But data are now
more than numbers. Much of the data being collected today consists of words from Internet
search engines such as Google searches and from pictures from social media postings on
such platforms as Facebook. Together with the traditional numbers comprising quantitative
data the availability of new unstructured qualitative data has led to a data explosion. IDC
a technology research firm estimates that data are growing at a rate of 50 percent a year.
All of these data—referred to as big data—have created a need not only for highly skilled
data scientists who can mine and analyze it but also for managers who can make decisions
using it. McKinsey Global Institute a consultancy firm believes that big data offer an oppor
tunity for organizations to create competitive advantages for themselves if they can under
stand and use the information to its full potential. They report that the use of big data “will
become a key basis of competition and growth for individual firms.” This will create a need
for highly trained data scientists and managers who can use data to support their decision
making. Unfortunately McKinsey predicts that by 2018 there could be a shortage in the
United States of 140000 to 190000 people with deep analytical skills as well as 1.5 million
managers and analysts with the knowhow needed to use big data to make meaningful and
effective decisions. Source: McKinsey Global Institute: Big Data: The Next Frontier for
Innovation Competition and Productivity May 2011 by James Manyika Michael Chui
Brad Brown Jacques Bughin Richard Dobbs Charles Roxburgh Angela Hung Byers. The
statistical tools you will learn in this course will provide you with a good first step toward
preparing yourself for a career in data mining and business analytics.slide 32:
The Where Why and How of Data Collection
1 What Is Business Statistics
Summary
The two areas of statistics descriptive statistics and inferential
statistics are introduced. Descriptive statistics includes visual tools such
as charts and graphs and also the numerical measures such as the
arithmetic average. The role of descriptive statistics is to describe data and
help transform data into usable information. Inferential techniques are those that
allow decisionmakers to draw conclusions about a large body of data
by examining a smaller subset of those data. Two areas of inference
estimation and hypothesis testing are described.
2 Procedures for Collecting Data
Summary
Before data can be analyzed using business statistics techniques the
data must be collected. The types of data collection reviewed are:
experiments telephone surveys written questionnaires and direct
observation and personal interviews. Data collection issues such as
interviewer bias nonresponse bias selection bias observer bias
and measurement error are covered. The concepts of internal validity
and external validity are defned.
Outcome 1. Know the key data collection methods.
3 Populations Samples and Sampling
Techniques
Outcome 2. Know the difference between a population and a sample.
Outcome 3. Understand the similarities and differences between different
sampling methods.
4 Data Types and Data Measurement
Levels
Summary
The important concepts of population and sample are defned and examples
of each are provided. Because many statistical applications involve samples
emphasis is placed on how to select samples. Two main sampling categories are
presented nonstatistical sampling and statistical sampling. The focus
is on statistical sampling and four statistical sampling methods are discussed:
simple random sampling stratifed random sampling cluster
sampling and systematic random sampling.
Summary
This section discusses various ways in which data are classifed.
For example data can be classifed as being either quantitative
or qualitative. Data can also be crosssectional or
timeseries. Another way to classify data is by the level of
measurement. There are four levels from lowest to highest:
nominal ordinal interval and ratio. Knowing the
type of data you have is very important because the
data type infuences the type of statistical procedures
you can use.
5 A Brief Introduction to Data Mining
Summary
Because electronic data storage is so inexpensive organizations
are collecting and storing greater volumes of data that ever
before. As a result a relatively new feld of study called data
mining has emerged. Data mining involves the art and science
of delving into the data to identify patterns and conclusions that
are not immediately evident in the data. This section briefy
introduces the subject and discusses a few of the applications.
Although data mining is not covered in depth in this text the
concepts presented throughout the text form the basis for this
important discipline.
Outcome 4. Understand how to categorize data by type and
level of measurement.
Outcome 5. Become familiar with the concept of data mining
and some of its applications.
Conclusion
Statistical analysis begins with data. You need to know how to
collect data how to select samples from a population and the
type and level of data you are using. Figure 13 summarizes
the different sampling techniques presented in this chapter.
Figure 14 gives a synopsis of the different data collection
procedures and Figure 15 shows the different data types
and measurement levels.
Business statistics is a collection of procedures and techniques used by
decisionmakers to transform data into useful information. This chapter
introduces the subject of business statistics. Included is a discussion of the
different types of data and data collection methods. This chapter also describes
the difference between populations and samples.slide 33:
The Where Why and How of Data Collection
TimeSeries
Data Levels
Data Type
Data Timing
Qualitative
Nominal Ordinal
Quantitative
Interval Ratio
CrossSectional
FIGURE 15 
Data Classification
Population
N items
Sampling Techniques
Sample
n items
Nonrandom
Sampling
Convenience Sampling
Judgment Sampling
Ratio Sampling
Simple Random Sampling
Stratifed Random Sampling
Systematic Random Sampling
Cluster Sampling
Random
Sampling
Sample
n items
Many possible
samples
FIGURE 13 
Sampling Techniques
Experiments
Data Collection
Method Advantages Disadvantages
Telephone Surveys
Direct Observation
Personal Interview
Mail Questionnaires
Written Surveys
Provide controls
Preplanned objectives
Costly
Timeconsuming
Requires planning
Timely
Relatively inexpensive
Poor reputation
Limited scope and length
Inexpensive
Can expand length
Can use openend questions
Expands analysis opportunities
No respondent bias
Low response rate
Requires exceptional clarity
Potential observer bias
Costly
FIGURE 14 
Data Collection Techniquesslide 34:
The Where Why and How of Data Collection
Key Terms
Arithmetic mean or average
Bias
Business intelligence
Business statistics
Census
Closedend questions
Cluster sampling
Convenience sampling
Crosssectional data
Data Mining
Demographic questions
Experiment
Experimental design
External validity
Internal validity
Nonstatistical sampling techniques
Openend questions
Population
Qualitative data
Quantitative data
Sample
Simple random sampling
Statistical inference procedures
Statistical sampling techniques
Stratified random sampling
Structured interview
Systematic random sampling
Timeseries data
Unstructured interview
MyStatLab
Chapter Exercises
Conceptual Questions
159. Several organizations publish the results of presidential
approval polls. Movements in these polls are seen as an
indication of how the general public views presidential
performance. Comment on these polls within the context
of what was covered.
160. With what level of data is a bar chart most appropriately
used
161. With what level of data is a histogram most
appropriately used
162. Two people see the same movie one says it was average
and the other says it was exceptional. What level of data
are they using in these ratings Discuss how the same
movie could receive different reviews.
163. The University of Michigan publishes a monthly
measure of consumer confidence. This is taken as a
possible indicator of future economic performance.
Comment on this process within the context of what was
covered.
Business Applications
164. In a business publication such as The Wall Street Journal
or Business W eek find a graph or chart representing
timeseries data. Discuss how the data were gathered
and the purpose of the graph or chart.
165. In a business publication such as The Wall Street Journal
or Business W eek find a graph or chart representing
crosssectional data. Discuss how the data were gathered
and the purpose of the graph or chart.
166. The Oregonian newspaper has asked readers to email
and respond to the question “Do you believe police
officers are using too much force in routine traffic
stops”
a. Would the results of this survey be considered a
random sample
b. What type of bias might be associated with a data
collection system such as this Discuss what options
might be used to reduce this bias potential.
167. The makers of Mama’s HomeMade Salsa are concerned
about the quality of their product. The particular trait of
concern is the thickness of the salsa in each jar.
a. Discuss a plan by which the managers might
determine the percentage of jars of salsa believed
to have an unacceptable thickness by potential
purchasers. 1 Define the sampling procedure to
be used 2 the randomization method to be used
to select the sample and 3 the measurement to be
obtained.
b. Explain why it would or wouldn’t be feasible or
perhaps possible to take a census to address this
issue.
168. A maker of energy drinks is considering abandoning
can containers and going exclusively to bottles because
the sales manager believes customers prefer drinking
from bottles. However the vice president in charge of
marketing is not convinced the sales manager is correct.
a. Indicate the data collection method you would use.
b. Indicate what procedures you would follow to apply
this technique in this setting.
c. State which level of data measurement applies to the
data you would collect. Justify your answer.
d. Are the data qualitative or quantitative Explain.slide 35:
The Where Why and How of Data Collection
Statistical Data Collection McDonald’s
Think of any wellknown successful business in your community.
What do you think has been its secret Competitive products or
services Talented managers with vision Dedicated employees
with great skills There’s no question these all play an important
part in its success. But there’s more lots more. It’s “data.” That’s
right data.
The data collected by a business in the course of running its
daily operations form the foundation of every decision made.
Those data are analyzed using a variety of statistical techniques
to provide decision makers with a succinct and clear picture of
the company’s activities. The resulting statistical information then
plays a key role in decision making whether those decisions are
made by an accountant marketing manager or operations spe
cialist. To better understand just what types of business statistics
organizations employ let’s take a look at one of the world’s most
wellrespected companies: McDonald’s.
McDonald’s operates more than 30000 restaurants in more
than 118 countries around the world. Total annual revenues
recently surpassed the 20 billion mark. Wade Thomas vice presi
dent of U.S. Menu Management for McDonalds helps drive those
sales but couldn’t do it without statistics.
“When you’re as large as we are we can’t run the business on
simple gut instinct. We rely heavily on all kinds of statistical data
to help us determine whether our products are meeting customer
expectations when products need to be updated and much more”
says Wade. “The cost of making an educated guess is simply too
great a risk.”
McDonald’s restaurant owner/operators and managers also
know the competitiveness of their individual restaurants depends
on the data they collect and the statistical techniques used to ana
lyze the data into meaningful information. Each restaurant has a
sophisticated cash register system that collects data such as indi
vidual customer orders service times and methods of payment to
name a few. Periodically each U.S.–based restaurant undergoes a
restaurant operations improvement process or ROIP study. A spe
cial team of reviewers monitors restaurant activity over a period of
several days collecting data about everything from frontcounter
service and kitchen efficiency to drivethru service times. The data
are analyzed by McDonald’s U.S. Consumer and Business Insights
group at McDonald’s headquarters near Chicago to help the res
taurant owner/operator and managers better understand what
they’re doing well and where they have opportunities to grow.
Steve Levigne vice president of Consumer and Busi
ness Insights manages the team that supports the company’s
decisionmaking efforts. Both qualitative and quantitative data are
collected and analyzed all the way down to the individual store
level. “Depending on the audience the results may be rolled up to
an aggregate picture of operations” says Steve. Software packages
such as Microsoft Excel SAS and SPSS do most of the number
crunching and are useful for preparing the graphical representa
tions of the information so decision makers can quickly see the
results.
Not all companies have an entire department staffed with spe
cialists in statistical analysis however. That’s where you come
in. The more you know about the procedures for collecting and
analyzing data and how to use them the better decision maker
you’ll be regardless of your career aspirations. So it would seem
there’s a strong relationship here—knowledge of statistics and
your success.
Discussion Questions:
1. You will recall that McDonald’s vice president of U.S. Menu
Management Wade Thomas indicated that McDonald’s
relied heavily on statistical data to determine in part if its
products were meeting customer expectations. The narrative
indicated that two important sources of data were the
sophisticated register system and the restaurant operations
improvement process ROIP. Describe the types of data
that could be generated by these two methods and discuss
how these data could be used to determine if McDonald’s
products were meeting customer expectations.
2. One of McDonald’s uses of statistical data is to determine
when products need to be updated. Discuss the kinds of data
McDonald’s would require to make this determination. Also
provide how these types of data would be used to determine
when a product needed to be updated.
3. This video case presents the types of data collected and used
by McDonald’s in the course of running its daily operations.
For a moment imagine that McDonald’s did not collect
these data. Attempt to describe how it might make a decision
concerning for instance how much its annual advertising
budget would be.
4. Visit a McDonald’s in your area. While there take note
of the different types of data that could be collected using
observation only. For each variable you identify determine
the level of data measurement. Select three different variables
from your list and outline the specific steps you would use to
collect the data. Discuss how each of the variables could be
used to help McDonald’s manage the restaurant.
videoslide 36:
The Where Why and How of Data Collection
1. Descriptive use charts graphs tables and numerical measures.
3. A bar chart is used whenever you want to display data that have
already been categorized while a histogram is used to display
data over a range of values for the factor under consideration.
5. Hypothesis testing uses statistical techniques to validate a
claim.
13. statistical inference particularly estimation
17. written survey or telephone survey
19. An experiment is any process that generates data as its
outcome.
23. internal and external validity
27. Advantages—low cost speed of delivery instant updating of
data analysis disadvantages—low response and potential
confusion about questions
29. personal observation data gathering
33. Part range
Population size
Sample size
18000
100
180
Thus the first person selected will come from employees 1
through 180. Once that person is randomly selected the second
person will be the one numbered 100 higher than the first and so
on.
37. The census would consist of all items produced on the line in
a defined period of time.
41. parameters since it would include all U.S. colleges
43. a. stratified random sampling
b. simple random sampling or possibly cluster random sampling
c. systematic random sampling
d. stratified random sampling
49. a. timeseries
b. crosssectional
c. timeseries
d. crosssectional
51. a. ordinal—categories with defined order
b. nominal—categories with no defined order
c. ratio
d. nominal—categories with no defined order
53. ordinal data
55. a. nominal data
b. ratio data
c. nominal data
d. ratio data
e. ratio data
f. nominal data
g. ratio data
61. interval or ratio data
67. a. Use a random sample or systematic random sample.
b. The product is going to be ruined after testing it. You
would not want to ruin the entire product that comes off
the assembly line.
Answers to Selected OddNumbered Problems
This section contains summary answers to most of the oddnumbered problems in the text. The Student Solutions Manual contains fully developed
solutions to all oddnumbered problems and shows clearly how each answer is determined.
Berenson Mark L. and David M. Levine Basic Business Sta
tistics: Concepts and Applications 12th ed. Upper Saddle
River NJ: Prentice Hall 2012.
Cryer Jonathan D. and Robert B. Miller Statistics for Busi
ness: Data Analysis and Modeling 2nd ed. Belmont CA:
Duxbury Press 1994.
DeVeaux Richard D. Paul F. Velleman and David E. Bock
Stats Data and Models 3rd ed. New York: AddisonWesley
2012.
Fowler Floyd J. Survey Research Methods 4th ed. Thousand
Oaks CA: Sage Publications 2009.
Hildebrand David and R. Lyman Ott Statistical Thinking for
Managers 4th ed. Belmont CA: Duxbury Press 1998.
John J. A. D. Whitiker and D. G. Johnson Statistical Thinking
for Managers 2nd ed. Boca Raton FL: CRC Press 2005.
Microsoft Excel 2010 Redmond WA: Microsoft Corp. 2010.
Scheaffer Richard L. William Mendenhall R. Lyman Ott and
Kenneth G. Gerow Elementary Survey Sampling 7th ed.
Brooks/Cole 2012.
Siegel Andrew F. Practical Business Statistics 5th ed. Burr
Ridge IL: Irwin 2002.
Referencesslide 37:
The Where Why and How of Data Collection
Nonstatistical Sampling Techniques Those methods of
selecting samples using convenience judgment or other
nonchance processes.
OpenEnd Questions Questions that allow respondents the
freedom to respond with any value words or statements of
their own choosing.
Population The set of all objects or individuals of interest or
the measurements obtained from all objects or individuals
of interest.
Qualitative Data Data whose measurement scale is inherently
categorical.
Quantitative Data Measurements whose values are inherently
numerical.
Sample A subset of the population.
Simple Random Sampling A method of selecting items from
a population such that every possible sample of a specified
size has an equal chance of being selected.
Statistical Inference Procedures Procedures that allow a
decision maker to reach a conclusion about a set of data
based on a subset of that data.
Statistical Sampling Techniques Those sampling methods
that use selection techniques based on chance selection.
Stratified Random Sampling A statistical sampling method
in which the population is divided into subgroups called
strata so that each population item belongs to only one stra
tum. The objective is to form strata such that the population
values of interest within each stratum are as much alike as
possible. Sample items are selected from each stratum using
the simple random sampling method.
Structured Interview Interviews in which the questions are
scripted.
Systematic Random Sampling A statistical sampling tech
nique that involves selecting every kth item in the popula
tion after a randomly selected starting point between 1 and
k. The value of k is determined as the ratio of the population
size over the desired sample size.
TimeSeries Data A set of consecutive data values observed at
successive points in time.
Unstructured Interview Interviews that begin with one or
more broadly stated questions with further questions being
based on the responses.
Arithmetic Average or Mean The sum of all values divided by
the number of values.
Bias An effect that alters a statistical result by systematically
distorting it different from a random error which may dis
tort on any one occasion but balances out on the average.
Business Intelligence The application of tools and technolo
gies for gathering storing retrieving and analyzing data
that businesses collect and use.
Business Statistics A collection of procedures and techniques
that are used to convert data into meaningful information in
a business environment.
Census An enumeration of the entire set of measurements
taken from the whole population.
ClosedEnd Questions Questions that require the respondent
to select from a short list of defined choices.
Cluster Sampling A method by which the population is
divided into groups or clusters that are each intended to
be minipopulations. A simple random sample of m clusters
is selected. The items chosen from a cluster can be selected
using any probability sampling technique.
Convenience Sampling A sampling technique that selects the
items from the population based on accessibility and ease of
selection.
CrossSectional Data A set of data values observed at a fixed
point in time.
Data Mining The application of statistical techniques and
algorithms to the analysis of large data sets.
Demographic Questions Questions relating to the respon
dents’ characteristics backgrounds and attributes.
Experiment A process that produces a single outcome whose
result cannot be predicted with certainty.
Experimental Design A plan for performing an experiment in
which the variable of interest is defined. One or more factors
are identified to be manipulated changed or observed so
that the impact or influence on the variable of interest can
be measured or observed.
External Validity A characteristic of an experiment whose
results can be generalized beyond the test environment so
that the outcomes can be replicated when the experiment is
repeated.
Internal Validity A characteristic of an experiment in which
data are collected in such a way as to eliminate the effects of
variables within the experimental environment that are not
of interest to the researcher.
Glossaryslide 38:
Graphs Charts and Tables—
Describing Your Data
Quick Prep Links
Review the definitions for nominal ordinal
interval and ratio data in Sections 1–4.
Examine the statistical software such as
Excel that you will be using during this
course to make sure you are aware of the
procedures for constructing graphs and
tables. For instance in Excel look at the
Charts group on the Insert tab and the
Pivot Table feature on the Insert tab.
Look at popular newspapers such as
USA Today and business periodicals such as
Fortune Business Week or The Wall Street
Journal for instances in which charts graphs
or tables are used to convey information.
Frequency Distributions and
Histograms
Bar Charts Pie Charts and
Stem and Leaf Diagrams
Line Charts and Scatter
Diagrams
Outcome 1. Construct frequency distributions both manually
and with your computer.
Outcome 2. Construct and interpret a frequency histogram.
Outcome 3. Develop and interpret joint frequency distributions.
Why you need to know
We live in an age in which presentations and reports are expected to include highquality graphs and charts that
effectively transform data into information. Although the written word is still vital words become even more power
ful when coupled with an effective visual illustration of data. The adage that a picture is worth a thousand words
is particularly relevant in business decision making. We are constantly bombarded with visual images and stimuli.
Much of our time is spent watching television playing video games or working at a computer. These technolo
gies are advancing rapidly making the images sharper and more attractive to our eyes. Flatpanel highdefinition
televisions and highresolution monitors represent significant improvements over the original technologies they
replaced. However this phenomenon is not limited to video technology but has also become an important part of
the way businesses communicate with customers employees suppliers and other constituents.
When you graduate you will find yourself on both ends of the data analysis spectrum. On the one hand
regardless of what you end up doing for a career you will almost certainly be involved in preparing reports and
making presentations that require using visual descriptive statistical tools presented in this chapter. You will be on
the “do it” end of the data analysis process. Thus you need to know how to use these statistical tools.
On the other hand you will also find yourself reading reports or listening to presentations that others have
made. In many instances you will be required to make important decisions or to reach conclusions based on the
information in those reports or presentations. Thus you will be on the “use it” end of the data analysis process.
You need to be knowledgeable about these tools to effectively screen and critique the work that others do for you.
Charts and graphs are not just tools used internally by businesses. Business periodicals such as Fortune and
Business Week use graphs and charts extensively in articles to help readers better understand key concepts. Many
advertisements will even use graphs and charts effectively to convey their messages. Virtually every issue of The
Wall Street Journal contains different graphs charts or tables that display data in an informative way.
Outcome 4. Construct and interpret various types of bar charts.
Outcome 5. Build a stem and leaf diagram.
Outcome 6. Create a line chart and interpret the trend in
the data.
Outcome 7. Construct a scatter diagram and interpret it.
MishAl/Shutterstock
From Chapter 2 of Business Statistics A DecisionMaking Approach Ninth Edition. David F. Groebner
Patrick W. Shannon and Phillip C. Fry. Copyright © 2014 by Pearson Education Inc. All rights reserved.slide 39:
Graphs Charts and Tables—Describing Your Data
Thus you will find yourself to be both a producer and a consumer of the descriptive statistical
techniques known as graphs charts and tables. You will create a competitive advantage for yourself through
out your career if you obtain a solid understanding of the techniques introduced in this text. This chapter
introduces some of the most frequently used tools and techniques for describing data with graphs charts and
tables. Although this analysis can be done manually we will provide output from Excel software showing that
software can be used to perform the analysis easily quickly and with a finished quality that once required a
graphic artist.
Frequency Distributions
and Histograms
In today’s business climate companies collect massive amounts of data they hope will be
useful for making decisions. Every time a customer makes a purchase at a store like Macy’s
or the Gap data from that transaction is updated to the store’s database. Major retail stores
like Walmart capture the number of different product categories included in each “market
basket” of items purchased. Table 1 shows these data for all customer transactions for a
single day at one Walmart store in Dallas. A total of 450 customers made purchases on the
day in question. The first value in Table 1 4 indicates that the customer’s purchase included
four different product categories for example food sporting goods photography supplies
and dry goods.
TABLE 1  Product Categories per Customer at the Dallas Walmart
42588 1014834113 4
14454449544 10 7 11 4
10267 10546462324 5
54 1114192466762 3
65345653 1065774 3
82265 1199556531 7
66538433444764 9
1655447566956 10 4
758447466442 10 4 5
4 11879564284266 6
64657169159 1055 10
54757695321555 5
59532572464444 4
65855555255646 5
57 1022683135633 6
5453379445 10 6 105 9
43871843136755 5
474 116637944297 5
16683844193934 2
9557 10534776224 4
47354923432164 6
1814355 10444692 7
94536553465736 8
36157754666369 5
45 1015578916566 4
1065551656479 10 2 6
446 11954435 4626 7
35674546943369 4
3756 11448428242 3
651 105954514954 4slide 40:
Graphs Charts and Tables—Describing Your Data
Although the data in Table 1 are easy to capture with the technology of today’s cash reg
isters in this form the data provide little or no information that managers could use to deter
mine the buying habits of their customers. However these data can be converted into useful
information through descriptive statistical analysis.
Frequency Distribution
One of the first steps would be to construct a frequency distribution.
The product data in Table 1 take on only a few possible values 1 2 3 c 11. The
minimum number of product categories is 1 and the maximum number of categories in these
data is 11. These data are called discrete data.
When you encounter discrete data where the variable of interest can take on only a rea
sonably small number of possible values a frequency distribution is constructed by count
ing the number of times each possible value occurs in the data set. We organize these counts
into a frequency distribution table as shown in Table 2. Now from this frequency distribu
tion we are able to see how the data values are spread over the different number of possible
product categories. For instance you can see that the most frequently occurring number of
product categories in a customer’s “market basket” is 4 which occurred 92 times. You can
also see that the three most common numbers of product categories are 4 5 and 6. Only a
very few times do customers purchase 10 or 11 product categories in their trip to the store.
Consider another example in which a consulting firm surveyed random samples of
residents in two cities Philadelphia and Knoxville. The firm is investigating the labor
markets in these two communities for a client that is thinking of relocating its corporate
offices to one of the two locations. Education level of the workforce in the two cities is a
key factor in making the relocation decision. The consulting firm surveyed 160 randomly
selected adults in Philadelphia and 330 adults in Knoxville and recorded the number of
years of college attended. The responses ranged from zero to eight years. Table 3 shows
the frequency distributions for each city.
Suppose now we wished to compare the distribution for years of college for Philadelphia
and Knoxville. How do the two cities’ distributions compare Do you see any difficulties in
making this comparison Because the surveys contained different numbers of people it is dif
ficult to compare the frequency distributions directly. When the number of total observations
differs comparisons are aided if relative frequencies are computed. Equation 1 is used to
compute the relative frequencies.
Table 4 shows the relative frequencies for each city’s distribution. This makes a com
parison of the two much easier. We see that Knoxville has relatively more people with
out any college 56.7 or with one year of college 18.8 than Philadelphia 21.9
Frequency Distribution
A summary of a set of data that displays
the number of observations in each of the
distribution’s distinct categories or classes.
Discrete Data
Data that can take on a countable number of
possible values.
Relative Frequency
The proportion of total observations that are in a
given category. Relative frequency is computed
by dividing the frequency in a category by
the total number of observations. The relative
frequencies can be converted to percentages by
multiplying by 100.
TABLE 2  Dallas Walmart Product
Categories Frequency Distribution
Number of Product
Catagories Frequency
125
229
342
492
583
671
735
819
929
10 18
11 7
Total 450
Chapter Outcome 1.slide 41:
Graphs Charts and Tables—Describing Your Data
TABLE 3  Frequency Distributions of Years
of College Education
Philadelphia Knoxville
Years of
College Frequency
Years of
College Frequency
0 35 0 187
121 1 62
224 2 34
322 3 19
431 4 14
513 5 7
66 6 3
75 7 4
8 3 8 0
Total 160 Total 330
Relative Frequency
Relative frequency 5
f
n
i
f f
1
where:
f
i
Frequency of the ith value of the discrete variable
nf
i
k
i
f f
1
∑
Total numbe m ro e e f observa r r ti a a ons
k The number of different values for the discrete variable
TABLE 4  Relative Frequency Distributions of Years of College
Philadelphia Knoxville
Years of
College Frequency
Relative
Frequency Frequency
Relative
Frequency
035 35160 0.219 187 187330 0.567
121 21160 0.131 62 62330 0.188
224 24160 0.150 34 34330 0.103
322 22160 0.138 19 19330 0.058
431 31160 0.194 14 14330 0.042
513 13160 0.081 7 7330 0.021
66 6160 0.038 3 3330 0.009
75 5160 0.031 4 4330 0.012
8 3 3160 0.019 0 0330 0.000
Total 160 330
and 13.1. At all other levels of education Philadelphia has relatively more people than
Knoxville.
The frequency distributions shown in Table 2 and Table 3 were developed from
quantitative data. That is the variable of interest was numerical number of product catego
ries or number of years of college. However a frequency distribution can also be developedslide 42:
Graphs Charts and Tables—Describing Your Data
TABLE 5  TV Source
Frequency Distribution
TV Source Frequency
DISH 80
DirectTV 90
Cable 20
Other 10
Total 200
when the data are qualitative data or nonnumerical data. For instance if a survey asked
homeowners how they get their TV signal the possible responses in this region are:
DISH DirectTV Cable Other
Table 5 shows the frequency distribution from a survey of 200 homeowners.
EXAMPLE 1 FREQUENCY AND RELATIVE FREQUENCY DISTRIBUTIONS
Real Estate Transactions In late 2008 the United States experienced a major economic
decline thought to be due in part to the subprime loans that many lending institutions made
during the previous few years. When the housing bubble burst many institutions experienced
severe problems. As a result lenders became much more conservative in granting home loans
which in turn made buying and selling homes more challenging. To demonstrate the mag
nitude of the problem in Kansas City the Association of Real Estate Brokers conducted a
survey of 16 real estate agencies and collected data on the number of real estate transactions
closed in December 2008. The following data were observed:
3001
1220
0210
2142
The real estate analysts can use the following steps to construct a frequency distribution and a
relative frequency distribution for the number of real estate transactions.
Step 1 List the possible values.
The possible values for the discrete variable listed in order are 0 1 2 3 4.
Step 2 Count the number of occurrences at each value.
The frequency distribution follows:
Transactions Frequency Relative Frequency
05 5/16 0.3125
14 4/16 0.2500
25 5/16 0.3125
31 1/16 0.0625
4 1 1/16 0.0625
Total 16 1.0000
Step 3 Determine the relative frequencies.
The relative frequencies are determined by dividing each frequency by 16
as shown in the righthand column above. Thus just over 31 of those
responding reported no transactions during December 2008.
END EXAMPLE
TRY PROBLEM 1
EXAMPLE 2 FREQUENCY DISTRIBUTION FOR QUALITATIVE DATA
Automobile Accidents State Farm Insurance recently surveyed
a sample of the records for 15 policy holders to determine the make
of the vehicle driven by the eldest member in the household. The
following data reflect the results for 15 of the respondents:
Ford Dodge Toyota Ford Buick
Chevy Toyota Nissan Ford Chevy
Ford Toyota Chevy BMW Honda
Example 1
Developing Frequency and
Relative Frequency Distributions
for Discrete Data
To develop a discrete data frequency
distribution perform the following
steps:
List all possible values of the
variable. If the variable is ordi
nal level or higher order the
possible values from low to
high.
Count the number of occur
rences at each value of the vari
able and place this value in a
column labeled “frequency.”
To develop a relative frequency
distribution do the following:
Use Equation 1 and divide each
frequency count by the total
number of observations and
place in a column headed “rela
tive frequency.”
Elena Elisseeva/Shutterstockslide 43:
Graphs Charts and Tables—Describing Your Data
The frequency distribution for this qualitative variable is found as follows:
Step 1 List the possible values.
For these sample data the possible values for the variable are BMW Buick
Chevy Dodge Ford Honda Nissan Toyota.
Step 2 Count the number of occurrences at each value.
The frequency distribution is
Car Company Frequency
BMW 1
Buick 1
Chevy 3
Dodge 1
Ford 4
Honda 1
Nissan 1
Toyota 3
Total 15
END EXAMPLE
TRY PROBLEM 8
BUSINESS APPLICATION FREQUENCY DISTRIBUTIONS
ATHLETIC SHOE SURVEY In recent years a status symbol for many students has been
the brand and style of athletic shoes they wear. Companies such as Nike and Adidas compete
for the top position in the sport shoe market. A survey was recently conducted in which 100
college students at a southern state school were asked a number of questions including how
many pairs of Nike shoes they currently own. The data are in a file called SportsShoes.
The variable Number of Nike is a discrete quantitative variable. Figure 1 shows the
frequency distribution output from Excel for the number of Nike shoes owned by those
Excel 2010 Instructions:
1. Open File: SportsShoes.xlsx.
2. Enter the Possible Values
for the Variable i.e. 0 1
2 3 4 etc.
3. Select the cells to contain
the Frequency values.
4. Select the Formulas tab.
5. Click on the f
x
button.
6. Select the Statistics—
FREQUENCY function.
7. Enter the range of data
and the bin range the cells
containing the possible
number of shoes.
8. Press CtrlShiftEnter to
determine the frequency
values.
Minitab Instructions for similar results:
1. Open fle: SportsShoes.MTW.3.In Variables enter data column.
2. Choose Stat Tables Tally 4. Under Display check Counts.
Individual Variables. 5. Click OK.
FIGURE 1 
Excel 2010 Output—Nike
Shoes Frequency Distribution
Excel
tutorials
Excel Tutorialslide 44:
Graphs Charts and Tables—Describing Your Data
surveyed. These frequency distributions show that although a few people own more than six
pairs of Nike shoes the bulk of those surveyed own two or fewer pairs.
Grouped Data Frequency Distributions
In the previous examples the variable of interest was a discrete variable and the number of
possible values for the variable was limited to only a few. However there are many instances
in which the variable of interest will be either continuous e.g. weight time length or dis
crete and will have many possible outcomes e.g. age income stock prices yet you want to
describe the variable using a frequency distribution.
BUSINESS APPLICATION GROUPED DATA FREQUENCY DISTRIBUTIONS
NETFLIX Netflix is one of the largest video rental companies in the United States. It rents
movies that are sent to customers by mail and also provides a video streaming service via
the Internet. Recently a distribution manager for Netflix conducted a survey of customers.
Among the questions asked on the written survey was “How many DVD movies do you
own” A total of 230 people completed the survey Table 6 shows the responses to the
DVD ownership question. These data are discrete quantitative data. The values range
from 0 to 30.
The manager is interested in transforming these data into useful information by con
structing a frequency distribution. Table 7 shows one approach in which the possible values
for the number of DVD movies owned is listed from 0 to 30. Although this frequency distribu
tion is a step forward in transforming the data into information because of the large number
of possible values for DVD movies owned the 230 observations are spread over a large range
making analysis difficult. In this case the manager might consider forming a grouped data
frequency distribution by organizing the possible number of DVD movies owned into discrete
categories or classes.
To begin constructing a grouped frequency distribution sort the quantitative data from
low to high. The sorted data is called a data array. Now define the classes for the variable
of interest. Care needs to be taken when constructing these classes to ensure each data
point is put into one and only one possible class. Therefore the classes should meet four
criteria.
Continuous Data
Data whose possible values are uncountable and
that may assume any value in an interval.
TABLE 6  DVD Movies Owned: Netflix Survey
9 4 13 10 51013141019
0 10 16 9 11 14 8 15 7 15
10 11976 12 12 14 15 16
15 14 10 13 91212101011
15 14 9 19 3 9 16 19 15 9
4 245623475
6 220083432
2 525226256
5 273516436
3 771627132
4 022462537
4 16 9 10 11 7 10 9 10 11
11 12 98979 17 8 13
14 13 10 6 12 5 14 7 13 12
9 6 10 157799 13 10
9 3 17 5 11969 15 8
11 13 41613 9 11 51213
0 333214020
3 715223213
2 333033311
13 24 24 17 17 15 25 20 15 20
21 23 25 17 13 22 18 17 30 21
18 21 17 16 25 14 15 24 21 15
Chapter Outcome 1.slide 45:
Graphs Charts and Tables—Describing Your Data
First they must be mutually exclusive.
Second they must be allinclusive.
Third if at all possible they should be of equal width.
Fourth avoid empty classes if possible.
Equalwidth classes make analyzing and interpreting the frequency distribution easier.
However there are some instances in which the presence of extreme high or low values makes
it necessary to have an openended class. For example annual family incomes in the United
States are mostly between 15000 and 200000. However there are some families with
much higher family incomes. To best accommodate these high incomes you might consider
having the highest income class be “over 200000” or “200000 and over” as a catchall for
the highincome families.
Empty classes are those for which there are no data values. If this occurs it may be
because you have set up classes that are too narrow.
Steps for Grouping Data into Classes There are four steps for grouping data such as
that found in Table 6 into classes.
Step 1 Determine the number of groups or classes to use.
Although there is no absolute right or wrong number of classes one rule of
thumb is to have between 5 and 20 classes. Another guideline for helping you
determine how many classes to use is the 2
k
Ú n rule where k the number
Mutually Exclusive Classes
Classes that do not overlap so that a data value
can be placed in only one class.
AllInclusive Classes
A set of classes that contains all the possible
data values.
EqualWidth Classes
The distance between the lowest possible value
and the highest possible value in each class is
equal for all classes.
TABLE 7  Frequency Distribution of DVD Movies Owned
DVD Movies Owned Frequency
08
18
222
322
411
513
612
714
85
919
10 14
11 9
12 8
13 12
14 8
15 12
16 6
17 7
18 2
19 3
20 2
21 4
22 1
23 1
24 3
25 3
26 0
27 0
28 0
29 0
30 1
Total 230slide 46:
Graphs Charts and Tables—Describing Your Data
of classes and is defined to be the smallest integer so that 2
k
Ú n where n is
the number of data values. For example for n 230 the 2
k
Ú n rule would
suggest k 8 classes 2
8
256 Ú 230 while 2
7
128 6 230. This latter
method was chosen for our example. Our preliminary step as specified previ
ously is to produce a frequency distribution from the data array as in Table 7.
This will enhance our ability to envision the data structure and the classes.
Remember these are only guidelines for the number of classes. There is
no specific right or wrong number. In general use fewer classes for smaller
data sets more classes for larger data sets. However using too few classes
tends to condense data too much and information can be lost. Using too many
classes spreads out the data so much that little advantage is gained over the
original raw data.
Step 2 Establish the class width.
The minimum class width is determined by Equation 2.
Class Width
The distance between the lowest possible value
and the highest possible value for a frequency
class.
For the Netflix data using eight classes we get
W
Largest value Smallest value
Number of classes
30 0
8
375 .
This means we could construct eight classes that are each 3.75 units wide to
provide mutually exclusive and allinclusive classes. However because our
purpose is to make the data more understandable we suggest that you round up
to a more convenient class width such as 4.0. If you do round the class width
always round up.
Step 3 Determine the class boundaries for each class.
The class boundaries determine the lowest possible value and the highest pos
sible value for each class. In the Netflix example if we start the first class at 0
we get the class boundaries shown in the first column of the following table.
Notice the classes have been formed to be mutually exclusive and allinclusive.
Step 4 Determine the class frequency for each class.
The count for each class is known as a class frequency. As an example the
number of observations in the first class is 60.
DVD Movies Owned Classes Frequency
093 60
497 50
8911 47
12915 40
16919 18
20923 8
24927 6
28931 1
Total 230
Another step we can take to help analyze the Netflix data is to construct a
relative frequency distribution a cumulative frequency distribution and a
cumulative relative frequency distribution.
Class Boundaries
The upper and lower values of each class.
Cumulative Frequency
Distribution
A summary of a set of data that displays the
number of observations with values less than or
equal to the upper limit of each of its classes.
Cumulative Relative Frequency
Distribution
A summary of a set of data that displays the
proportion of observations with values less than
or equal to the upper limit of each of its classes.
2
W
Largest value Smallest value
Number of clas sses
www.downloadslide.comslide 47:
Graphs Charts and Tables—Describing Your Data
The cumulative frequency distribution is shown in the “Cumulative
Frequency” column. We can then form the cumulative relative frequency
distribution as shown in the “Cumulative Relative Frequency” column.
The cumulative relative frequency distribution indicates as an example that
85.7 of the sample own fewer than 16 DVD movies.
EXAMPLE 3 FREQUENCY DISTRIBUTION FOR CONTINUOUS VARIABLES
Emergency Response Communication Links One of the major efforts of the United States
Office of Homeland Security has been to improve the communication between emergency respond
ers like the police and fire departments. The communications have been hampered by problems
involving linking divergent radio and computer systems as well as communication protocols. While
most cities have recognized the problem and made efforts to solve it Homeland Security recently
funded practice exercises in 72 cities of different sizes throughout the United States. The resulting
data already sorted but representing seconds before the systems were linked are as follows:
35 339 650 864 1025 1261
38 340 655 883 1028 1280
48 395 669 883 1036 1290
53 457 703 890 1044 1312
70 478 730 934 1087 1341
99 501 763 951 1091 1355
138 521 788 969 1126 1357
164 556 789 985 1176 1360
220 583 789 993 1199 1414
265 595 802 997 1199 1436
272 596 822 999 1237 1479
312 604 851 1018 1242 1492
Homeland Security wishes to construct a frequency distribution showing the times until
the communication systems are linked. The frequency distribution is determined as follows:
Step 1 Group the data into classes.
The number of classes is arbitrary but typically will be between 5 and 20
depending on the volume of data. In this example we have n 72 data items.
A common method of determining the number of classes is to use the 2
k
≥ n
guideline. We get k 7 classes since 2
7
128 Ú 72 and 2
6
64 6 72.
Step 2 Determine the class width.
W
Largest value Smallest value
Number of cl lasses
1 492 35
7
208 1429 225
. ⇒
Note we have rounded the class width up from the minimum required value of
208.1429 to the more convenient value of 225.
Example 3
Developing Frequency
Distributions for Continuous
Variables
To develop a continuous data
frequency distribution perform
the following steps:
Determine the desired number
of classes or groups. One rule of
thumb is to use 5 to 20 classes.
The 2
k
Ú n rule can also be
used.
Determine the minimum class
width using
W
Largest Value Smallest Value
Number of classes
5
Round the class width up to a
more convenient value.
Define the class boundaries
making sure that the classes that
are formed are mutually exclu
sive and allinclusive. Ideally
the classes should have equal
widths and should all contain at
least one observation.
Determine the class frequency
for each class.
DVD Movies Frequency
Relative
Frequency
Cumulative
Frequency
Cumulative Relative
Frequency
093 60 0.261 60 0.261
497 50 0.217 110 0.478
8911 47 0.204 157 0.683
12915 40 0.174 197 0.857
16919 18 0.078 215 0.935
20923 8 0.035 223 0.970
24927 6 0.026 229 0.996
28931 1 0.004 230 1.000
Total 230slide 48:
Graphs Charts and Tables—Describing Your Data
Step 3 Define the class boundaries.
0 and under 225
225 and under 450
450 and under 675
675 and under 900
900 and under 1125
1125 and under 1350
1350 and under 1575
These classes are mutually exclusive allinclusive and have equal widths.
Step 4 Determine the class frequency for each class.
Time to Link Systems
in seconds Frequency
0 and under 225 9
225 and under 450 6
450 and under 675 12
675 and under 900 13
900 and under 1125 14
1125 and under 1350 11
1350 and under 1575 7
This frequency distribution shows that most cities took between 450 and 1350
seconds 7.5 and 22.5 minutes to link their communications systems.
END EXAMPLE
TRY PROBLEM 5
Histograms
Although frequency distributions are useful for analyzing large sets of data they are presented
in table format and may not be as visually informative as a graph. If a frequency distribution
has been developed from a quantitative variable a frequency histogram can be constructed
directly from the frequency distribution. In many cases the histogram offers a superior format
for transforming the data into useful information. Note: Histograms cannot be constructed from
a frequency distribution in which the variable of interest is qualitative. However a similar graph
called a bar chart discussed later in this chapter is used when qualitative data are involved.
A histogram shows three general types of information:
1. It provides a visual indication of where the approximate center of the data is. Look for
the center point along the horizontal axes in the histograms in Figure 2. Even though the
shapes of the histograms are the same there is a clear difference in where
the data are centered.
2. We can gain an understanding of the degree of spread or variation in the data. The
more the data cluster around the center the smaller the variation in the data. If the
data are spread out from the center the data exhibit greater variation. The examples in
Figure 3 all have the same center but are different in terms of spread.
3. We can observe the shape of the distribution. Is it reasonably flat is it weighted to one side
or the other is it balanced around the center or is it bell shaped
BUSINESS APPLICATION CONSTRUCTING HISTOGRAMS
CAPITAL CREDIT UNION Even for applications with small amounts of data such as the
Netflix example constructing grouped data frequency distributions and histograms is a time
consuming process. Decision makers may hesitate to try different numbers of classes and
different class limits because of the effort involved and because the “best” presentation of the
data may be missed.
We showed earlier that Excel provides the capability of constructing frequency distribu
tions. It is also quite capable of generating grouped data frequency distributions and histograms.
Frequency Histogram
A graph of a frequency distribution with the
horizontal axis showing the classes the vertical
axis showing the frequency count and for equal
class widths the rectangles having a height
equal to the frequency in each class.
Excel Tutorial
Excel
tutorials
Chapter Outcome 2.
www.downloadslide.comslide 49:
Graphs Charts and Tables—Describing Your Data
Consider Capital Credit Union CCU in Mobile Alabama which recently began issuing
a new credit card. Managers at CCU have been wondering how customers use the card so a
sample of 300 customers was selected. Data on the current credit card balance rounded to the
nearest dollar and the genders of the cardholders appear in the file Capital.
As with the manual process the first step in Excel is to determine the number of classes.
Recall that the rule of thumb is to use between 5 and 20 classes depending on the amount of
data. Suppose we decide to use 10 classes.
Next we determine the class width using Equation 2. The highest account balance in the
sample is 1493.00. The minimum is 99.00. Thus the class width is
W
1 493 00 99 00
139 40
. .
.
10
which we round up to 150.00.
100 a
b
c
200 300 400 500 600 700 800
100 200 300 400 500 600 700 800
100 200 300 400 500 600 700 800
FIGURE 2 
Histograms Showing Different
Centers
FIGURE 3 
Histograms—Same Center
Different Spread
100 a
b
c
200 300 400 500 600 700 800
100 200 300 400 500 600 700 800
100 200 300 400 500 600 700 800
www.downloadslide.comslide 50:
Graphs Charts and Tables—Describing Your Data
Our classes will be
+90+239.99
+240+389.99
+390+539.99
etc.
The resulting histogram in Figure 4 shows that the data are centered in the class
from 690 to 839.99. The customers vary considerably in their credit card balances
but the distribution is quite symmetrical and somewhat bell shaped. CCU managers
must decide whether the usage rate for the credit card is sufficient to warrant the cost of
maintaining the credit card accounts.
Issues with Excel If you use Excel to construct a histogram as indicated in the instruc
tions in Figure 4 the initial graph will have gaps between the bars. Because histograms
illustrate the distribution of data across the range of all possible values for the quantitative
variable histograms do not have gaps. Therefore to get the proper histogram format you
need to close these gaps by setting the gap width to zero as indicated in the Excel instruc
tions shown in Figure 4.
FIGURE 4 
Excel 2010 Histogram of
Credit Card Balances
6. Put on a new worksheet ply and
include the Chart Output.
7. Right mouse click on the bars and use
the Format Data Series Options to set
gap width to zero and add lines to the
bars.
8. Convert the bins to actual class labels
by typing labels in Column A. Note:
The bin 239.99 is labeled 0–239.99.
Excel 2010 Instructions:
1. Open fle: Capital.xlsx.
2. Set up an area on the worksheet for the
bins defned as 239.99 389.99 etc. up
to 1589.99. Be sure to include a label
such as “Bins.”
3. On the Data tab click Data Analysis.
4. Select Histogram.
5. Input Range specifes the actual data
values as the Credit Card Account
Balance column and the bin range as the
area defned in Step 2.
Minitab Instructions
for similar results:
1. Open fle: Capital.MTW.
2. Choose Graph
Histogram.
3. Click Simple.
4. Click OK.
5. In Graph variables
enter data column.
6. Click OK.
www.downloadslide.comslide 51:
Graphs Charts and Tables—Describing Your Data
Emergency Response Time Distribution
Frequency
Emergency Response Times Seconds
300
250
200
150
100
50
0
0 30 60 90 120 150 180 210 240 300 270
This histogram indicates that the response times vary considerably. The center is
somewhere in the range of 120 to 180 seconds.
END EXAMPLE
TRY PROBLEM 10
Example 4
Constructing Frequency
Histograms
To construct a frequency histo
gram perform the following steps:
1–4. Follow the steps for
constructing a frequency
distribution see
Example 3.
5. Use the horizontal axis to
represent classes for the
variable of interest. Use
the vertical axis to repre
sent the frequency in each
class.
6. Draw vertical bars for
each class or data value so
that the heights of the bars
correspond to the frequen
cies. Make sure there are
no gaps between the bars.
Note if the classes do not
have equal widths the bar
height should be adjusted
to make the area of the
bar proportional to the
frequenc y .
7. Label the histogram
appropriately.
EXAMPLE 4 FREQUENCY HISTOGRAMS
Emergency Response Times The Paris France director of
Emergency Medical Response is interested in analyzing the time
needed for response teams to reach their destinations in emer
gency situations after leaving their stations. She has acquired
the response times for 1220 calls last month. To develop the
frequency histogram perform the following steps:
Steps 1–4 Construct a frequency distribution.
Because response time is a continuous variable measured in seconds
the data should be broken down into classes and the steps given in
Example 3 should be used. The following frequency distribution
with 10 classes was developed:
Response Time Frequency Response Time Frequency
0 and under 30 36 180 and under 210 145
30 and under 60 68 210 and under 240 80
60 and under 90 195 240 and under 270 43
90 and under 120 180 270 and under 300 31
120 and under 150 260 Total 1220
150 and under 180 182
Step 5 Construct the axes for the histogram.
The horizontal axis will be response time and the vertical axis will be
frequency.
Step 6 Construct bars with heights corresponding to the frequency of
each class.
Step 7 Label the histogram appropriately.
This is shown as follows:
Robert Wilson/Fotolia
www.downloadslide.comslide 52:
Graphs Charts and Tables—Describing Your Data
Relative Frequency Histograms and Ogives
Histograms can also be used to display relative frequency distributions and cumulative rela
tive frequency distributions. A relative frequency histogram is formed in the same manner as a
frequency histogram but relative frequencies rather than frequencies are used on the vertical
axis. The cumulative relative frequency is presented using a graph called an ogive. Example 5
illustrates each of these graphical tools.
EXAMPLE 5 RELATIVE FREQUENCY HISTOGRAMS AND OGIVES
Emergency Response Times Continued Example 4
introduced the situation facing the emergency response manager
in Paris. In that example she formed a frequency distribution for a
sample of 1220 response times. She is now interested in graphing
the relative frequencies and the ogive. To do so use the following
steps:
Step 1 Convert the frequency distribution into relative frequencies and cumulative
relative frequencies.
Response Time Frequency Relative Frequency
Cumulative Relative
Frequency
0 and under 30 36 361220 0.0295 0.0295
30 and under 60 68 681220 0.0557 0.0852
60 and under 90 195 1951220 0.1598 0.2451
90 and under 120 180 1801220 0.1475 0.3926
120 and under 150 260 2601220 0.2131 0.6057
150 and under 180 182 1821220 0.1492 0.7549
180 and under 210 145 1451220 0.1189 0.8738
210 and under 240 80 801220 0.0656 0.9393
240 and under 270 43 431220 0.0352 0.9746
270 and under 300 31 311220 0.0254 1.0000
1220 1.0000
Step 2 Construct the relative frequency histogram.
Place the quantitative variable on the horizontal axis and the relative frequencies
on the vertical axis. The vertical bars are drawn to heights corresponding to the
relative frequencies of the classes.
Ogive
The graphical representation of the cumulative
relative frequency. A line is connected to points
plotted above the upper limit of each class at a
height corresponding to the cumulative relative
frequency.
Robert Wilson/Fotolia
Emergency Response Time Relative Frequency Distribution
Relative Frequency
Response Times seconds
.25000
.20000
.15000
.10000
.05000
.00000
0 30 60 120 90 180 150 240 300 210 270
Note the relative frequency histogram has exactly the same shape as the
frequency histogram. However the vertical axis has a different scale.
www.downloadslide.comslide 53:
Graphs Charts and Tables—Describing Your Data
Step 3 Construct the ogive.
Place a point above the upper limit of each class at a height corresponding
to the cumulative relative frequency. Complete the ogive by drawing a line
connecting these points.
.00000
1.00000
.90000
.80000
.70000
.60000
.50000
.40000
.30000
.20000
.10000
Cumulative Relative Frequency
Response Times
Emergency Response Times
Ogive
0 30 60 90 120 180 240 300 150 210 270
END EXAMPLE
TRY PROBLEM 16
Joint Frequency Distributions
Frequency distributions are effective tools for describing data. Thus far we have discussed
how to develop grouped and ungrouped frequency distributions for one variable at a time.
For instance in the Capital Credit Union example we were interested in customer credit card
balances for all customers. We constructed a frequency distribution and histogram for that
variable. However often we need to examine data that are characterized by more than one
variable. This may involve constructing a joint frequency distribution for two variables. Joint
frequency distributions can be constructed for qualitative or quantitative variables.
EXAMPLE 6 JOINT FREQUENCY DISTRIBUTION
Miami City Parking Parking is typically an issue in many large cities like Miami Flor
ida. Problems seem to occur for customers and employees both in locating a parking spot and
in being able to quickly exit a lot at busy times. The parking manager for Miami City Parking
has received complaints about the time required to exit garages in the downtown sector. To
start analyzing the situation she has collected a small sample of data from 12 customers
showing the type of payment cash or charge and the garage number Garage Number 1 2
or 3. One possibility is that using credit card payments increases exit times at the parking
lots. The manager wishes to develop a joint frequency distribution to better understand the
paying habits of those using her garages. To do this she can use the following steps:
Step 1 Obtain the data.
The paired data for the two variables for a sample of 12 customers are obtained.
Customer Payment Method Parking Garage
1 Charge 2
2 Charge 1
3 Cash 2
4 Charge 2
5 Charge 1
Chapter Outcome 3.
www.downloadslide.comslide 54:
Graphs Charts and Tables—Describing Your Data
Customer Payment Method Parking Garage
6 Cash 1
7 Cash 3
8 Charge 1
9 Charge 3
10 Cash 2
11 Cash 1
12 Charge 1
Step 2 Construct the rows and columns of the joint frequency table.
The row variable will be the payment method and two rows will be used
corresponding to the two payment methods. The column variable is parking
garage number and it will have three levels because the data for this variable
contain only the values 1 2 and 3. Note if a variable is continuous classes
should be formed using the methods discussed in Example 3.
Parking Garage
123
Payment
Cash
Step 3 Count the number of joint occurrences at each row level and each column
level for all combinations of row and column values and place these
frequencies in the appropriate cells.
Parking Garage
1 2 3 Total
Charge 4 2 1 7
Cash 2 2 1 5
Total 6 4 2 12
Step 4 Calculate the row and column totals see Step 3.
The manager can now see that for this sample most people charged their
parking fee seven people and Garage number 1 was used by most people in
the sample used six people. Likewise four people used Garage number 1 and
charged their parking fee.
END EXAMPLE
TRY PROBLEM 12
BUSINESS APPLICATION JOINT FREQUENCY DISTRIBUTION
CAPITAL CREDIT UNION CONTINUED Recall that the Capital Credit Union discussed
earlier was interested in evaluating the success of its new credit card. Figure 4 showed the
frequency distribution and histogram for a sample of customer credit card balances. Although
this information is useful the managers would like to know more. Specifically what does the
credit card balance distribution look like for male versus female cardholders
One way to approach this is to sort the data by the gender variable and develop frequency
distributions and histograms for males and females separately. You could then make a visual
comparison of the two to determine what if any difference exists between males and females.
However an alternative approach is to jointly analyze the two variables: gender and credit
card balance.
Excel provides a means for analyzing two variables jointly. In Figure 4 we constructed
the frequency distribution for the 300 credit card balances using 10 classes. The class width
Example 6
Constructing Joint Frequency
Distributions
A joint frequency distribution is
constructed using the following
steps:
Obtain a set of data consisting
of paired responses for two
variables. The responses can be
qualitative or quantitative. If the
responses are quantitative they
can be discrete or continuous.
Construct a table with r rows
and c columns in which the
number of rows represents
the number of categories or
numeric classes of one vari
able and the number of columns
corresponds to the number of
categories or numeric classes
of the second variable.
Count the number of joint occur
rences at each row level and each
column level for all combina
tions of row and column values
and place these frequencies in
the appropriate cells.
Compute the row and column
totals which are called the
marginal frequencies.
If a joint relative frequency
distribution is desired divide
each cell frequency by the total
number of paired observations.
Excel Tutorial
Excel
tutorials
www.downloadslide.comslide 55:
Graphs Charts and Tables—Describing Your Data
was set at 150. Figure 5 shows a table that is called a joint frequency distribution. This type
of table is also called a crosstabulation table.
1
The Capital Credit Union managers can use a joint frequency table to analyze the credit
card balances for males versus females. For instance for the 42 customers with balances
of 390 to 539 Figure 5 shows that 33 were males and 9 were females. Previously we
discussed the concept of relative frequency proportions which Excel converts to percent
ages as a useful tool for making comparisons between two data sets. In this example
comparisons between males and females would be easier if the frequencies were converted
to proportions or percentages. The result is the joint relative frequency table shown in
Figure 6. Notice that the percentages in each cell are percentages of the total 300 people in
the survey. For example the 540to689 class had 20.33 61 of the 300 customers. The
male customers with balances in the 540to689 range constituted 15 45 of the 300
customers whereas females with that balance level made up 5.33 16 of all 300 custom
ers. On the surface this result seems to indicate a big difference between males and females
at this credit balance level.
Suppose we really wanted to focus on the male versus female issue and control for the
fact that there are far more male customers than female. We could compute the percentages
differently. Rather than using a base of 300 the entire sample size we might instead be inter
ested in the percentages of the males who have balances at each level and the same measure
for females.
2
There are many options for transferring data into useful information. Thus far we have
introduced frequency distributions joint frequency tables and histograms. In the next section
we discuss one of the most useful graphical tools: the bar chart.
1
In Excel the joint frequency distribution is developed using a tool called Pivot tables.
Minitab Instructions for similar results:
1. Open fle: Capital.MTW.
2. Click on Data Code Numeric to
Text.
3. Under Code data from columns
select data column.
4. Under Into columns specify
destination column: Classes.
5. In Original values defne each
data class range.
6. In New specify code for each class.
7. Click OK.
8. Click on Stat Tables Cross
Tabulation and ChiSquare.
9. Under Categorical Variables For rows
enter Classes column and For columns
enter Gender column.
10. Under Display check Counts.
11. Click OK.
Excel 2010 Instructions:
1. Open fle: Capital.xlsx.
2. Place cursor anywhere
in the data.
3. On the Insert tab click on
PivotTable and click OK.
4. On the Options tab
select Options in the
PivotTable group. Select
the Display tab and check
Classic PivotTable layout.
Click OK.
5. Drag Credit Card
Account Balance to “Drop
Row Fields Here” area.
6. Rightclick in Credit
Card Account Balance
numbers and click Group.
7. Change Start at to 90.
Change End to 1589.
Change By to 150.
7. Drag Gender to “Drop
Column Fields Here” area.
8. Drag Credit Card
Account Balance to “Drop
Value Fields Here” area.
9. Place cursor in the Data
Item area right click and
select Summarize Values
By and select Count.
FIGURE 5  Excel 2010 Joint Frequency Distribution for Capital Credit Union
2
Such distributions are known as marginal distributions.
www.downloadslide.comslide 56:
Graphs Charts and Tables—Describing Your Data
Excel 2010 Instructions:
1. Place cursor in the
Gender numbers of the
PivotTable.
2. Rightclick and select
Value Field Settings.
3. On the Show values as
tab click on the down
arrow and select of
Grand Total.
4. Click OK.
Minitab Instructions for similar results:
1. Open fle: Capital.MTW.
2. Steps 2–7 as in Figure 6.
3. Click on Stat Tables Cross
Tabulation and Chisquare.
4. Under Categorical variables For
rows enter Classes column and For
columns enter Gender column.
5. Under Display check Total Percents.
6. Click OK.
In Figure 6 we have used the
Data Field Settings in the
Excel PivotTable to represent
the data as percentages.
FIGURE 6 
Excel 2010 Joint Relative
Frequencies for Capital
Credit Union
Skill Development
21. Given the following data develop a frequency
distribution:
5 326 6
7 336 7
7 975 3
12 6 10 7 2
6 807 4
22. Assuming you have data for a variable with
2000 values using the 2
k
Ú n guideline what
is the smallest number of groups that should be
used in developing a grouped data frequency
distribution
23. A study is being conducted in which a variable of
interest has 1000 observations. The minimum value
in the data set is 300 points and the maximum is
2900 points.
a. Use the 2
k
Ú n guideline to determine the minimum
number of classes to use in developing a grouped
data frequency distribution.
b. Based on the answer to part a determine the class
width that should be used round up to the nearest
100 points.
24. Produce the relative frequency distribution from
a sample size of 50 that gave rise to the following
ogive:
0.0
0 100 200 300
Sales
400 500 600
0.2
0.4
0.6
Cumulative Relative Frequency
0.8
1.0
Ogive
25. You have the following data:
8 6 11 14 10
11972 8
9555 12
784 17 8
12788 7
10869 9
11 16 2 7 4
8445 5
9966 7
7954 5
14290 6
1 1 12 11 4
2 MyStatLab
www.downloadslide.comslide 57:
Graphs Charts and Tables—Describing Your Data
a. Construct a frequency distribution for these data.
Use the 2
k
Ú n guideline to determine the number
of classes to use.
b. Develop a relative frequency distribution using the
classes you constructed in part a.
c. Develop a cumulative frequency distribution and a
cumulative relative frequency distribution using the
classes you constructed in part a.
d. Develop a histogram based on the frequency
distribution you constructed in part a.
26. Fill in the missing components of the following frequency
distribution constructed for a sample size of 50:
Class Frequency
Relative
Frequency
Cumulative
Relative
Frequency
7.85 6 0.12
6 8.05 0.48
8.05 6 4
6 8.25 0.10
8.25 6
27. The following cumulative relative frequency
distribution summarizes data obtained in a study of
the ending overages in dollars for the cash register
balance at a business:
Class Frequency
Relative
Frequency
Cumulative
Relative
Frequency
–60.00––40.00 2 0.04 0.04
–40.00––20.00 2 0.04 0.08
–20.00––00.00 8 0.16 0.24
00.00– 20.00 16 0.32 0.56
20.00– 40.00 20 0.40 0.96
40.00– 60.00 2 0.04 1.00
a. Determine the proportion of the days in which there
were no shortages.
b. Determine the proportion of the days the cash
register was less than 20 off.
c. Determine the proportion of the days the cash
register was less than 40 over or at most 20 short.
28. You are given the following data:
6 10 6495
5 5 5762
5 5 5457
6 7 8684
7 5 5557
8 7 6754
6 4 4746
6 7 8676
7 8 5657
3 6 4744
a. Construct a frequency distribution for these data.
b. Based on the frequency distribution develop a
histogram.
c. Construct a relative frequency distribution.
d. Develop a relative frequency histogram.
e. Compare the two histograms. Why do they look alike
29. Using the data from Problem 28
a. Construct a grouped data relative frequency
distribution of the data. Use the 2
k
Ú n guideline to
determine the number of classes.
b. Construct a cumulative frequency distribution of the
data.
c. Construct a relative frequency histogram.
d. Construct an ogive.
Business Applications
210. Burger King is one of the largest fastfood franchise
operations in the world. Recently the district manager
for Burger King in Las Vegas conducted a study in
which she selected a random sample of sales receipts.
She was interested in the number of line items on the
receipts. For instance if a customer ordered two
1/4pound hamburgers one side of fries and two soft
drinks the number of line items would be five. The
following data were observed:
7 5 76 5548 65
8 7 65 6294 45
8 4 96 6589 91
6 5 106 7655 56
8 7 68 6696 12 7
5 6 7 11 4434 14
11 2 5 5 8234 96
6 5 86 3645 8 10
a. Develop a frequency distribution for these data.
Discuss briefly what the frequency distribution tells
you about these sample data.
b. Based on the results in part a construct a frequency
histogram for these sample data.
211. In a survey conducted by AIG investors were asked
to rate how knowledgeable they felt they were as
investors. Both online and traditional investors were
included in the survey. The survey resulted in the
following data:
Of the online investors 8 55 and 37 responded
th ey were “savvy” “experienced” and “novice”
res pecti v ely .
Of the traditional investors the percentages were 4
29 and 67 respectively.
Of the 600 investors surveyed 200 were traditional
investors.
a. Use the information to construct a joint frequency
distribution.
b. Use the information to construct a joint relative
frequency distribution.
c. Determine the proportion of investors who
were both online investors and rated themselves
experienced.
d. Calculate the proportion of investors who were
online investors.
www.downloadslide.comslide 58:
Graphs Charts and Tables—Describing Your Data
214. The bubble in U.S. housing prices burst in 2008
causing sales of houses to decline in almost every part
of the country. Many homes were foreclosed because
the owners could not make the payments. Below is
a sample of 100 residential properties and the total
balance on the mortgage at the time of foreclosure.
172229 211021 159205 247697 247469
176736 240815 195056 315097 257150
129779 207451 165225 178970 319101
87429 219808 242761 277389 213803
153468 205696 210447 179029 241331
117808 188909 376644 185523 168145
158094 135461 131457 263232 256262
240034 289973 302341 178684 226998
176440 268106 181507 118752 251009
196457 195249 195986 201680 233182
271552 123262 212411 246462 177673
103699 252375 192335 265992 232247
320004 213020 192546 295660 211876
265787 207443 203043 133014 289645
251560 302054 185381 284345 184869
237485 282506 278783 335920 199630
248272 232234 188833 168905 357612
241894 186956 114601 301728 251865
207040 221614 318154 156611 219730
201473 174840 196622 263686 159029
a. Using the 2
k
Ú n guideline what is the minimum
number of classes that should be used to display
these data in a grouped data frequency distribution
b. Referring to part a what should the class width be
assuming you round the width up to nearest 1000
c. Referring to parts a and b develop a grouped data
frequency distribution for these mortgage balance data.
d. Based on your answer to part c construct and
interpret a frequency histogram for the mortgage
balance data.
215. Wageweb exhibits salary data obtained from surveys. It
provides compensation information on more than 170
benchmark positions including finance positions. It
recently reported that salaries of chief finance officers
CFOs ranged from 127735 to 209981 before
bonuses. Suppose the following data represent a
sample of the annual salaries for 25 CFOs. Assume
that data are in thousands of dollars.
173.1 171.2 141.9 112.6 211.1 156.5 145.4 134.0
192.0 185.8 168.3 131.0 214.4 155.2 164.9 123.9
161.9 162.7 178.8 161.3 182.0 165.8 213.1 177.4
159.3
a. Using 11 classes construct a cumulative frequency
distribution.
b. Determine the proportion of CFO salaries that are at
least 175000.
c. Determine the proportion of CFO salaries that are
less than 205000 and at least 135000.
212. The sales manager for the Fox News TV station
affiliate in a southern Florida city recently surveyed 20
advertisers and asked each one to rate the service of the
station on the following scale:
Very Good Good Fair Poor Very Poor
12 3 4 5
He also tracked the general time slot when the
advertiser’s commercials were shown on the station.
The following codes were used:
1 morning 2 afternoon
3 evening 4 various times
The following sample data were observed:
Rating Time Slot Rating Time Slot
2143
1122
3333
2133
1121
4411
2211
1153
2124
2234
a. Construct separate relative frequency distributions
for each of the two variables.
b. Construct a joint frequency distribution for these
two variables.
c. Construct a joint relative frequency distribution
for these two variables. Write a short paragraph
describing what the data imply.
213. A St. Louis–based shipping company recently selected
a random sample of 49 airplane weight slips for
crates shipped from an automobile parts supplier. The
weights measured in pounds for the sampled crates
are as follows:
89 83 97 101 86 89 86
91 84 89 87 93 86 90
86 92 92 88 88 92 86
93 80 93 77 98 94 95
94 88 95 87 99 98 90
91 87 89 89 96 88 94
95 79 94 86 92 94 85
a. Create a data array of the weights.
b. Develop a frequency distribution using five classes
having equal widths.
c. Develop a histogram from the frequency distribution
you created in part b.
d. Develop a relative frequency and a cumulative
relative frequency distribution for the weights using
the same five classes created in part b. What percent
of the sampled crates have weights greater than 96
pounds
www.downloadslide.comslide 59:
Graphs Charts and Tables—Describing Your Data
216. One effect of the great recession was to lower the
interest rate on fixedrate mortgages. A sample of
30year fixedrate mortgage rates taken from financial
institutions in the Pacific Northwest resulted in the
following:
3.79 4.03 3.92 3.87 3.86
3.93 3.87 3.69 3.99 3.88
3.91 3.81 3.85 3.81 3.65
4.15 3.98 3.82 4.08 3.84
3.95 4.03 3.96 3.69 3.97
4.08 3.86 3.82 3.83 4.19
3.49 3.49 3.75 3.46 3.92
3.67 3.63 3.81 3.86 4.15
3.85 4.01 3.93 4.02 4.09
a. Construct a histogram with eight classes beginning
at 3.46.
b. Determine the proportion of 30year fixed mortgage
rates that are at least 3.76.
c. Produce an ogive for the data.
Computer Database Exercises
217. J.D. Power and Associates’ annual customer
satisfaction survey the Automotive Performance
Execution and Layout APEAL Study
SM
in its 13th
year was released on September 22 2008. The study
measures owners’ satisfaction with the design content
layout and performance of their new vehicles. A file
titled APEAL2 contains the satisfaction ratings for
2008 for each make of car.
a. Construct a histogram that starts at 710 and has
class widths of 20 for the APEAL ratings.
b. The past industry average APEAL rating was 866
for 2005. What does the 2008 data suggest in terms
of the relative satisfaction with the 2008 models
218. The Franklin Tire Company is interested in
demonstrating the durability of its steelbelted radial
tires. To do this the managers have decided to put
four tires on 100 different sport utility vehicles and
drive them throughout Alaska. The data collected
indicate the number of miles rounded to the nearest
1000 miles that each of the SUVs traveled before
one of the tires on the vehicle did not meet minimum
federal standards for tread thickness. The data file is
called Franklin.
a. Construct a frequency distribution and histogram
using eight classes. Use 51 as the lower limit of the
first class.
b. The marketing department wishes to know the tread
life of at least 50 of the tires the 10 that had the
longest tread life and the longest tread life of these
tires. Provide this information to the marketing
department. Also provide any other significant items
that point out the desirability of this line of steel
belted tires.
c. Construct a frequency distribution and histogram
using 12 classes using 51 as the lower limit of
the first class. Compare your results with those
in parts a and b. Which distribution gives the best
information about the desirability of this line of
steelbelted tires Discuss.
219. The California Golf Association recently conducted
a survey of its members. Among other questions the
members were asked to indicate the number of 18hole
rounds that they played last year. Data for a sample
of 294 members are provided in the data file called
Golf Survey.
a. Using the 2
k
Ú n guideline what is the minimum
number of classes that should be used to
display these data in a grouped data frequency
distribution
b. Referring to part a what should the class width be
assuming you round the width up to the nearest
integer
c. Referring to parts a and b develop a grouped data
frequency distribution for these golf data.
d. Based on your answer to part c construct and
interpret a frequency histogram for the data.
220. Ars Technia LLD published a news release Eric
Bangeman “Dell Still King of Market Share” that
presented the results of a study concerning the world
market share for the major manufacturers of personal
computers. It indicated that Dell held 17.9 of this
market. The file titled PCMarket contains a sample of
the market shares alluded to in the article.
a. Construct a histogram from this set of data and
identify the sample shares for each of the listed
manufacturers.
b. Excluding the data referred to as “other” determine
the total share of the sample for manufacturers that
have headquarters in the United States.
221. Orlando Florida is a wellknown popular vacation
destination visited by tourists from around the world.
Consequently the Orlando International Airport
is busy throughout the year. Among the variety
of data collected by the Greater Orlando Airport
Authority is the number of passengers by airline.
The file Orlando Airport contains passenger data
for December 2011. Suppose the airport manager is
interested in analyzing the column labeled “Total”
for this data.
a. Using the 2
k
Ú n guideline what is the minimum
number of classes that should be used to display
the data in the “Total” column in a grouped data
frequency distribution
b. Referring to part a what should the class width be
assuming you round the width up to the nearest
1000 passengers
c. Referring to parts a and b develop a grouped data
frequency distribution for these airport data.
d. Based on your answer to part c construct and
interpret a frequency histogram for the data.
222. The manager of AJ’s Fitness Center a fullservice
health and exercise club recently conducted a survey
of 1214 members. The objective of the survey was to
www.downloadslide.comslide 60:
Graphs Charts and Tables—Describing Your Data
223. The file Danish Coffee contains data on individual
coffee consumption in kg for 144 randomly selected
Danish coffee drinkers.
a. Construct a data array of the coffee consumption data.
b. Construct a frequency distribution of the coffee
consumption data. Within what class do more of the
observations fall
c. Construct a histogram of the coffee consumption
data. Briefly comment on what the histogram
reveals concerning the data.
d. Develop a relative frequency distribution and a
cumulative relative frequency distribution of the
coffee data. What percentage of the coffee drinkers
sampled consume 8.0 kg or more annually
determine the satisfaction level of his club’s customers.
In addition the survey asked for several demographic
factors such as age and gender. The data from the
survey are in a file called AJFitness.
a. One of the key variables is “Overall Customer
Satisfaction.” This variable is measured on an
ordinal scale as follows:
5 very satisfied 4 satisfied 3 neutral
2 dissatisfied 1 very dissatisfied
Develop a frequency distribution for this variable
and discuss the results.
b. Develop a joint relative frequency distribution for
the variables “Overall Customer Satisfaction” and
“Typical Visits Per Week.” Discuss the results.
END EXERCISES 21
Bar Charts Pie Charts and
Stem and Leaf Diagrams
Bar Charts
Section 1 introduced some of the basic tools for describing numerical variables both discrete
and continuous when the data are in their raw form. However in many instances you will be
working with categorical data or data that have already been summarized to some extent. In
these cases an effective presentation tool is often a bar chart.
BUSINESS APPLICATION DEVELOPING BAR CHARTS
NEW CAR SALES The automobile industry is a significant part of the U.S economy. When
car sales are up the economy is up and vice versa. Table 8 displays data showing the total
number of cars sold in January 2012 by the eight largest automobile companies in the world.
Although the table format is informative a graphical presentation is often desirable. Because
the car sales data are characterized by car company a bar chart would work well in this
instance. The bars on a bar chart can be vertical called a column bar chart or horizontal
called a horizontal bar chart. Figure 7 illustrates an example of a column bar chart. The
height of the bars corresponds to the number of cars sold by each company. This gives you an
idea of the sales advantage held by General Motors in January 2012.
One strength of the bar chart is its capability of displaying multiple variables on the same
chart. For instance a bar chart can conveniently compare new car sales data for January 2012 and
Bar Chart
A graphical representation of a categorical
data set in which a rectangle or bar is drawn
over each category or class. The length or
height of each bar represents the frequency
or percentage of observations or some other
measure associated with the category. The bars
may be vertical or horizontal. The bars may
all be the same color or they may be different
colors depicting different categories. Additionally
multiple variables can be graphed on the same
bar chart.
Chapter Outcome 4.
TABLE 8  January 2012 New Car Sales for the
Top Eight Automobile Companies United States
Car Company January 2012 Sales
General Motors 167900
Ford 136300
Chrysler 101150
Toyota 125500
Honda 83000
Nissan 79300
Hyundai 42700
Mazda 24000
Source: Wall Street Journal Online February 1 2012.
www.downloadslide.comslide 61:
Graphs Charts and Tables—Describing Your Data
180000
160000
140000
120000
100000
80000
60000
40000
20000
0
General
Motors
Ford Chrysler Toyota Honda Nissan Hyundai Mazda
Automobile Company
January 2012 New Car Sales
FIGURE 7 
Bar Chart Showing January
2012 New Car Sales
sales for the same month the previous year. Figure 8 is a horizontal bar chart that does just that.
Notice that only General Motors had lower sales in January 2012 than in the previous January.
People sometimes confuse histograms and bar charts. Although there are some similari
ties they are two very different graphical tools. Histograms are used to represent a frequency
distribution associated with a single quantitative ratio or intervallevel variable. Refer to
the histogram illustrations in Section 1. In every case the variable on the horizontal axis was
numerical with values moving from low to high. The vertical axis shows the frequency count
0 100000 150000 200000
Automobile Company
New Car Sales in U.S.
50000
Ford
General
Motors
Toyota
Chrysler
Honda
Nissan
Hyundai
Mazda
2012 Sales up 68
2012 Sales up 15
January 2011 Sales
2012 Sales up 10
2012 Sales up 9
2012 Sales up 8
2012 Sales up 44
2012 Sales up 7
2012 Sales down 6
January 2012 Sales
FIGURE 8 
Bar Chart Comparing January
2011 and January 2012
New Cars Sold
www.downloadslide.comslide 62:
Graphs Charts and Tables—Describing Your Data
or relative frequency for each numerical value or range of values. There are no gaps between
the histogram bars. On the other hand bar charts are used when one or more variables of
interest are categorical as in this case in which the category is “car company.”
EXAMPLE 7 BAR CHARTS
Investment Recommendations Fortune contained an article by David Stires called
“The Best Stocks to Buy Now.” The article identified 40 companies as good investment
opportunities. These companies were divided into five categories: Growth and Income Bar
gain Growth Deep V alue Small Wonders and Foreign V alue. For each company data for sev
eral key variables were reported including the price/earnings PE ratio based on the previous
12 months’ reported earnings. We are interested in constructing a bar chart of the PE ratios for
the eight companies classified as Growth and Income.
Step 1 Define the categories.
Data for stock price and PE ratio for each of eight companies is shown as follows:
Company Ticker Symbol PE Ratio Stock Price
Abbott Labs ABT 21 49
Altria Group MO 14 65
CocaCola KO 21 42
ColgatePalmolive CL 20 51
General Mills GIS 17 51
Pfizer PFE 13 29
Procter Gamble PG 21 53
Wyeth WYE 15 43
The category to be displayed is the company.
Step 2 Determine the appropriate measure to be displayed.
The measure of interest is the PE ratio.
Step 3 Develop the bar chart.
A column bar chart is developed by placing the eight companies on the
horizontal axis and constructing bars whose heights correspond to the value
of the company’s PE ratio. Each company is assigned a differentcolored bar.
The resulting bar chart is
Abbott
Labs
ABT
0
5
10
15
Price/Earnings Ratio
20
21
14
21
20
17
13
21
15
25
Price/Earnings Ratio
Altria
Group
MO
CocaCola
KO
Colgate
Palmolive
CL
Company
General
Mills
GIS
Pfizer
PFE
Procter
Gamble
PG
Wyeth
WYE
Example 7
Constructing Bar Charts
A bar chart is constructed using the
following steps:
Define the categories for the
variable of interest.
For each category determine the
appropriate measure or value.
For a column bar chart locate
the categories on the horizontal
axis. The vertical axis is set to a
scale corresponding to the values
in the categories. For a horizon
tal bar chart place the categories
on the vertical axis and set the
scale of the horizontal axis in
accordance with the values in
the categories. Then construct
bars either vertical or horizontal
for each category such that the
length or height corresponds to
the value for the category.
www.downloadslide.comslide 63:
Graphs Charts and Tables—Describing Your Data
Step 4 Interpret the results.
The bar chart shows three companies with especially low PE ratios. These are
Altria Group Pfizer and Wyeth. Thus of the eight recommended companies in
the Growth and Income group these three have the lowest PE ratios.
END EXAMPLE
TRY PROBLEM 27
BUSINESS APPLICATION CONSTRUCTING BAR CHARTS
BACH LOMBARD WILSON One of the most useful features of bar charts is that
they can display multiple issues. Consider Bach Lombard Wilson a New England law
firm. Recently the firm handled a case in which a woman was suing her employer a major
electronics firm claiming the company gave higher starting salaries to men than to women.
Consequently she stated even though the company tended to give equalpercentage raises to
women and men the gap between the two groups widened.
Attorneys at Bach Lombard Wilson had their staff assemble massive amounts of data.
Table 9 provides an example of the type of data they collected. A bar chart is a more effective
way to convey this information as Figure 9 shows. From this graph we can quickly see that in
all years except 2009 the starting salaries for males did exceed those for females. The bar chart
also illustrates that the general trend in starting salaries for both groups has been increasing
though with a slight downturn in 2011. Do you think the information in Figure 9 alone is suf
ficient to rule in favor of the claimant in this lawsuit Bar charts like the one in Figure 9 that
display two or more variables are referred to as cluster bar charts.
Average Starting Salaries
80000
60000
40000
20000
70000
50000
30000
10000
0
2006 2007 2008 2009 2010 2011 2012
Female
Male
Males tend to have higher starting salaries.
General upward trend in salaries.
FIGURE 9 
Bar Chart of Starting Salaries
TABLE 9  Salary Data for Bach Lombard Wilson
Year
Males: Average
Starting Salaries
Females: Average
Starting Salaries
2006 44456 41789
2007 47286 46478
2008 56234 53854
2009 57890 58600
2010 63467 59070
2011 61090 55321
2012 67543 64506
Excel Tutorial
Excel
tutorials
www.downloadslide.comslide 64:
Graphs Charts and Tables—Describing Your Data
Pie Charts
Another graphical tool that can be used to transform data into information is the
pie chart.
EXAMPLE 8 PIE CHARTS
Golf Equipment A survey was recently conducted of 300 golfers that asked questions
about the impact of new technology on the game. One question asked the golfers to indicate
which area of golf equipment is most responsible for improving an amateur golfer’s game.
The following data were obtained:
Equipment Frequency
Golf ball 81
Club head material 66
Shaft material 63
Club head size 63
Shaft length 3
Don’t know 24
To display these data in pie chart form use the following steps:
Step 1 Define the categories.
The categories are the six equipmentresponse categories.
Step 2 Determine the appropriate measure.
The appropriate measure is the proportion of the golfers surveyed. The
proportion for each category is determined by dividing the number of golfers
in a category by the total sample size. For example for the category “golf ball”
the percentage is 81300 0.27 27.
Step 3 Construct the pie chart.
The pie chart is constructed by dividing a circle into six slices one for each
category such that each slice is proportional to the percentage of golfers in
the category.
Club Head Material
22
Shaf Material
21
Club Head Size
21
Golf Ball
27
Don’t Know
8
Shaf Length
1
Golf Equipment Impact
END EXAMPLE
TRY PROBLEM 28
Example 8
Constructing Pie Charts
A pie chart is constructed using the
following steps:
Define the categories for the
variable of interest.
For each category determine
the appropriate measure or
value. The value assigned to
each category is the proportion
the category is to the total for all
categories.
Construct the pie chart by
displaying one slice for each
category that is proportional in
size to the proportion the cat
egory value is to the total of all
categories.
Pie Chart
A graph in the shape of a circle. The circle is divided
into “slices” corresponding to the categories or
classes to be displayed. The size of each slice is
proportional to the magnitude of the displayed
variable associated with each category or class.
www.downloadslide.comslide 65:
Graphs Charts and Tables—Describing Your Data
Pie charts are sometimes mistakenly used when a bar chart would be more appropri
ate. For example a few years ago the student leaders at Boise State University wanted to
draw attention to the funding inequities among the four public universities in Idaho. To do
so they rented a large billboard adjacent to a major thoroughfare through downtown Boise.
The billboard contained a large pie chart like the one shown in Figure 10 where each slice
indicated the funding per student at a given university. However for a pie chart to be appro
priate the slices of the pie should represent parts of a total. But in the case of the billboard
that was not the case. The amounts merely represented the dollars of state money spent
per student at each university. The sum of the four dollar amounts on the pie chart was a
meaningless number. In this case a bar chart like that shown in Figure 11 would have been
more appropriate.
Lewis and Clark College
5410
Idaho State University
6320
University of Idaho
7 143
Boise State University
5900
FIGURE 10 
Pie Chart: PerStudent
Funding for Universities
Boise State
University
0
1000
2000
3000
4000
5000
6000
7 000
8000
Boise State
University
5900
University of Idaho
7 143
Idaho State University
6320
Lewis and Clark
College 5410
University of
Idaho
Idaho State
University
Lewis and
Clark College
FIGURE 11 
Bar Chart: PerStudent
Funding for Universities
www.downloadslide.comslide 66:
Graphs Charts and Tables—Describing Your Data
Stem and Leaf Diagrams
Another graphical technique useful for doing an exploratory analysis of quantitative data
is called the stem and leaf diagram. The stem and leaf diagram is similar to the histogram
introduced in Section 1 in that it displays the distribution for the quantitative variable. How
ever unlike the histogram in which the individual values of the data are lost if the vari
able of interest is broken into classes the stem and leaf diagram shows the individual data
values.
Although Excel does not have a stem and leaf procedure the PHStat addins to Excel do
have a stem and leaf procedure.
EXAMPLE 9 STEM AND LEAF DIAGRAMS
WalkIn Health Clinic The administrator for the WalkIn Health Clinic in Sydney
Australia is interested in performing an analysis of the number of patients who enter
the clinic daily. One method for analyzing the data for a sample of 200 days is the stem
and leaf diagram. The following data represent the number of patients on each of the
200 days:
113 112 63 127 110 129 142 115 192 94
165 121 105 140 85 93 105 140 93 126
183 118 67 104 162 110 76 109 91 132
88 96 132 80 144 112 57 139 123 124
172 149 198 114 88 111 133 117 138 134
53 147 108 109 153 89 159 99 130 93
161 118 115 117 128 98 125 184 134 132
117 127 166 72 122 109 124 92 82 69
110 128 151 67 142 177 135 121 143 89
160 115 138 79 104 76 89 110 44 140
117 103 59 109 145 117 162 108 141 139
148 175 107 117 87 87 150 152 80 168
88 127 131 85 143 101 137 111 128 147
110 81 111 149 154 90 150 117 101 116
153 176 112 147 87 177 190 66 62 154
143 122 176 153 97 106 86 62 146 98
134 135 127 118 109 143 146 152 140 95
102 137 158 69 122 135 136 129 91 136
135 86 131 154 132 59 136 85 142 137
155 190 120 154 102 109 97 157 144 149
The stem and leaf diagram is constructed using the following steps:
Step 1 Sort the data from low to high.
The lowest value is 44 patients and the highest value is 198 patients.
Step 2 Split the values into a stem and leaf.
Stem tens place leaf units place
For example for the value 113 the stem is 11 and the leaf is 3. We are keeping
one digit for the leaf.
Step 3 List all possible stems from lowest to highest.
Step 4 Itemize the leaves from lowest to highest and place next to the appropriate
stems.
Example 9
Constructing Stem and Leaf
Diagrams
To construct the stem and leaf
diagram for a quantitative variable
use the following steps:
Sort the data from low to high.
Analyze the data for the vari
able of interest to determine
how you wish to split the values
into a stem and a leaf.
List all possible stems in a sin
gle column between the lowest
and highest values in the data.
For each stem list all leaves
associated with the stem.
Chapter Outcome 5.
www.downloadslide.comslide 67:
Graphs Charts and Tables—Describing Your Data
44
5 3 7 9 9
6 2 2 3 6 7 7 9 9
7 2 6 6 9
8 0 0 1 2 5 5 5 6 6 7 7 7 8 8 8 9 9 9
9 0 1 1 2 3 3 3 4 5 6 7 7 8 8 9
10 1 1 2 2 3 4 4 5 5 6 7 8 8 9 9 9 9 9 9
11 0 0 0 0 0 1 1 1 2 2 2 3 4 5 5 5 6 7 7 7 7 7 7 7 8 8 8
12 0 1 1 2 2 2 3 4 4 5 6 7 7 7 7 8 8 8 9 9
13 0 1 1 2 2 2 2 3 4 4 4 5 5 5 5 6 6 6 7 7 7 8 8 9 9
14 0 0 0 0 1 2 2 2 3 3 3 3 4 4 5 6 6 7 7 7 8 9 9 9
15 0 0 1 2 2 3 3 3 4 4 4 4 5 7 8 9
16 0 1 2 2 5 6 8
17 2 5 6 6 7 7
18 3 4
19 0 0 2 8
The stem and leaf diagram shows that most days have between 80 and
160 patients with the most frequent value in the 110 to 120patient range.
END EXAMPLE
TRY PROBLEM 25
MyStatLab
Skill Development
224. The following data reflect the percentages of
employees with different levels of education:
Education Level Percentage
Less than high school graduate 18
High school graduate 34
Some college 14
College graduate 30
Graduate degree 4
Total 100
a. Develop a pie chart to illustrate these data.
b. Develop a horizontal bar chart to illustrate these data.
225. Given the following data construct a stem and leaf
diagram:
0.7 1.7 2.8 3.8
0.8 1.8 3.3 4.3
1.0 2.0 4.4 5.4
1.1 2.1 5.3 6.3
1.4 2.4 5.4 6.4
2.0 3.0
226. A university has the following number of students at
each grade level.
Freshman 3450
Sophomore 3190
Junior 2780
Senior 1980
Graduate 750
a. Construct a bar chart that effectively displays these
data.
b. Construct a pie chart to display these data.
c. Referring to the graphs constructed in parts a
and b indicate which you would favor as the most
effective way of presenting these data. Discuss.
227. Given the following sales data for product category and
sales region construct at least two different bar charts
that display the data effectively:
Region
East West North South
Product
Type
XJ6 Model 200 300 50 170
X15Y Model 100 200 20 100
Craftsman 80 400 60 200
Generic 100 150 40 50
228. The 2010 Annual Report of Murphy Oil Corporation
reports the following refinery yields in barrels per
day by product category for the United States and the
United Kingdom.
United States 2010
Product Category
Refinery Yields—
barrels per day
Gasoline 61128
Kerosene 11068
Diesel and Home Heating Oils 41305
Residuals 18082
Asphalt LPG and other 14802
Fuel and Loss 834
2 www.downloadslide.comslide 68:
Graphs Charts and Tables—Describing Your Data
United Kingdom 2010
Product Category
Refinery Yields—
barrels per day
Gasoline 20889
Kerosene 11374
Diesel and Home Heating Oils 25995
Residuals 8296
Asphalt LPG and other 14799
Fuel and Loss 2810
a. Construct a pie chart to display United States
refinery yields by product per day. Display the
refinery yields data for each product category as a
percentage of total refinery yields for all product
categories.
b. Construct a pie chart to display United Kingdom
refinery yields by product per day. Display the
refinery yields data for each product category as a
percentage of total refinery yields for all product
categories.
c. Construct a bar chart that effectively compares
United States and United Kingdom refinery yields
by product category.
229. Boston Properties is a real estate investment trust
REIT that owns firstclass office properties in
selected markets. According to its 2010 annual report
its percentage of net operating income distribution by
region as a percent of total income for the year ended
December 31 2010 was
Region 2010 Income Distribution
New Y ork 41
Washington D.C. 23
Boston 21
San Francisco 12
Princeton 3
a. Construct a pie chart to display the percentage of net
operating income distribution by region for 2010.
b. Construct a bar chart to display the net operating
income distribution by region for 2010.
c. Briefly comment on which of the two charts you
believe better summarizes and displays the data.
230. Hileman Services Company recently released the
following data concerning its operating profits in
billions for the five years:
Year 2004 2005 2006 2007 2008
Profit 0.5 0.1 0.7 0.5 0.2
a. Construct a bar chart to graphically display these
data.
b. Construct a pie chart to graphically display these
data.
c. Select the display that most effectively displays the
data and provide reasons for your choice.
231. DaimlerChrysler recently sold its Chrysler division to a
private equity firm. Before the sale it reported its first
half revenues in billions as follows:
Division Mercedes Chrysler
Commercial
Vehicles
Financial
Services Total
Revenues 27.7 30.5 23.2 8.9 90.3
a. Produce a bar chart for these data.
b. Determine the proportion of firsthalf revenues
accounted for by its vehicle divisions.
Business Applications
232. At the March meeting of the board of directors for
the Graystone Services Company one of the regional
managers put the following data on the overhead
projector to illustrate the ratio of the number of units
manufactured to the number of employees at each of
Graystone’s five manufacturing plants:
Plant Location Units Manufactured/Employees
Bismarck ND 14.5
Boulder CO 9.8
Omaha NE 13.0
Harrisburg PA 17.6
Portland ME 5.9
a. Discuss whether a pie chart or a bar chart would be
most appropriate to present these data graphically.
b. Construct the chart you recommended in part a.
233. The first few years after the turn of the century saw a
rapid increase in housing values followed by a rapid
decline due in part to the subprime crisis. The following
table indicates the increase in the number of homes
valued at more than one million dollars before 2005.
Year Number of 1 Million Homes
2000 394878
2001 495600
2002 595441
2003 714467
2004 1034386
Develop a horizontal bar chart to represent these data
in graphical form.
234. The pharmaceutical industry is a very fastgrowing
segment of the U.S. and international economies.
Recently there has been controversy over how studies are
done to show that drugs are both safe and effective. One
drug product Cymbalta which is an antidepressant was
purported in a published abstract of an article in a medical
journal to be superior to other competing products. Y et
the article itself stated that no studies had actually been
done to show such comparisons between Cymbalta and
other competing products. In an August 2005 report in an
article titled “Reading Fine Print Insurers Question Drug
Studies” in The Wall Street Journal the following data
www.downloadslide.comslide 69:
Graphs Charts and Tables—Describing Your Data
were presented showing the U.S. sales of antidepressant
drugs by major brand. The sales data for the first half of
2005 are shown in the following table.
Antidepressant Drug
Sales First Half 2005
in Billions
Effexor XR 1.29
Lexapro 1.03
Zoloft 1.55
Cymbalta 0.27
Other 0.97
Construct an appropriate graph to display these data.
235. The number of branded retail outlets in the United
States and the United Kingdom for Murphy Oil
Corporation as of December 31 of each year from 2001
to 2010 is shown below:
Branded Retail
Outlets 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010
United States
total
815 914 994 1127 1201 1164 1126 1154 1169 1215
United
Kingdom
411 416 384 358 412 402 389 454 453 451
Develop a chart that effectively compares the number
of branded retail outlets in North America with the
number in the United Kingdom.
236. The 2011 Annual Report of the Procter Gamble
Company reported the following net sales information
for its six global business segments:
Global Segment 2011 Net Sales Millions
Beauty 20157
Grooming 8025
Health Care 12033
Snacks and Pet Care 3156
Fabric Care and Home Care 24837
Baby Care and Family Care 15606
a. Construct a bar chart that displays this information
by global business segment for 2011.
b. Construct a pie chart that displays each global
business segment’s net sales as a percentage of total
global segment net sales for 2011.
237. A fastfood restaurant monitors its drivethru service
times electronically to ensure that its speed of service
is meeting the company’s goals. A sample of 28 drive
thru times was recently taken and is shown here.
Speed of Service Time in Seconds
83 138 145 147
130 79 156 156
90 85 68 93
178 76 73 119
92 146 88 103
116 134 162 71
181 110 105 74
a. Construct a stem and leaf diagram of the speed of
service times.
b. What range of time might the restaurant say is the
most frequent speed of service
238. A random sample of 30 customer records for a
physician’s office showed the following time in days
to collect insurance payments:
Number of Days to Collect Payment
34 55 36 39 36
32 35 30 47 31
60 66 48 43 33
24 37 38 65 35
22 45 33 29 41
38 35 28 56 56
a. Construct a stem and leaf diagram of these data.
b. Within what range of days are most payments
collected
239. USA Today presented data to show that major airlines
accounting for more than half of capacity were
expected to be in bankruptcy court. The total seat
capacity of major airlines was 858 billion at the time.
For airlines expected to be in bankruptcy court the
following data were presented:
Airline Seat Capacity in Billions
Airline United Delta Northwest U.S. Airways A TA
Capacity 145 130 92 54 21
a. Construct a bar graph representing the contribution
to the total seat capacity of the major airlines for the
five airlines indicated.
b. Produce a pie chart exhibiting the percentage of
the total seat capacity for the five major airlines
expected to be in bankruptcy court and the
combined capacity of all others.
c. Calculate the percentage of the total capacity of the
airlines expected to be in bankruptcy court. Was
USA Today correct in the percentage stated
240. Many of the world’s most successful companies
rely on The NPD Group to provide global sales and
marketing information that helps clients make more
informed factbased decisions to optimize their
businesses. These customers need NPD help for
insight on what is selling where and why so that they
can understand and leverage the latest trends. They
recently July 2009 released the following results
of a survey intended to determine the market share
distribution for the major corporations that make
digital music devices:
Corporation Apple SanDisk
Creative
Technology iRiver Samsung
Market Share 74 6.4 3.9 3.6 2.6
www.downloadslide.comslide 70:
Graphs Charts and Tables—Describing Your Data
a. Generate a bar chart to display these data.
b. Generate a pie chart to display these data.
c. Which of the two displays most effectively presents
the data Explain your answer.
Computer Database Exercises
241. The Honda Ridgeline was among the highestranked
compact pickups in J.D. Power and Associates’ annual
customersatisfaction survey. The study also found that
models with high ratings have a tendency to stay on
dealers’ lots a shorter period of time. As an example
the Honda Ridgeline had stayed on dealers’ lots an
average of 24 days. The file titled Honda contains 50
lengths of stay on dealers’ lots for Ridgeline trucks.
a. Construct a stem and leaf display for these data.
b. Determine the average length of stay on dealers’
lots for the Honda Ridgeline. Does this agree with
the average obtained by J.D. Power and Associates
Explain the difference.
242. The manager for Capital Educators Federal Credit
Union has selected a random sample of 300 of the
credit union’s credit card customers. The data are in
a file called Capital. The manager is interested in
graphically displaying the percentage of card holders
of each gender.
a. Determine the appropriate type of graph to use in
this application.
b. Construct the graph and interpret it.
243. Recently a study was conducted in which a random
sample of hospitals was selected from each of four
categories of hospitals: university related religious related
community owned and privately owned. At issue is the
hospital charges associated with outpatient gall bladder
surgery. The following data are in the file called Hospitals:
University
Related
Religious
Affiliated
Municipally
Owned
Privately
Held
6120 4010 4320 5100
5960 3770 4650 4920
6300 3960 4575 5200
6500 3620 4440 5345
6250 3280 4900 4875
6695 3680 4560 5330
University
Related
Religious
Affiliated
Municipally
Owned
Privately
Held
6475 3350 4610 5415
6250 3250 4850 5150
6880 3400 5380
6550
a. Compute the average charge for each hospital
category.
b. Construct a bar chart showing the averages by
hospital category.
c. Discuss why a pie chart would not in this case be an
appropriate graphical tool.
244. Amazon.com has become one of the success stories of
the Internet age. Its growth can be seen by examining
its increasing sales volume in billions and the net
income/loss during Amazon’s operations. A file titled
Amazon contains these data for its first 13 years.
a. Construct one bar graph illustrating the relationship
between sales and income for each separate year of
Amazon’s existence.
b. Describe the type of relationship that exists between
the years in business and Amazon’s sales volume.
c. Amazon’s sales rose sharply. However its net
income yielded losses which increased during
the first few years. In which year did this situation
reverse itself and show improvement in the net
income balance sheet
245. In your capacity as assistant to the administrator at
Freedom Hospital you have been asked to develop a
graphical presentation that focuses on the insurance
carried by the geriatric patients at the hospital. The
data file Patients contains data for a sample of
geriatric patients. In developing your presentation
please do the following:
a. Construct a pie chart that shows the percentage of
patients with each health insurance payer.
b. Develop a bar chart that shows total charges for
patients by insurance payer.
c. Develop a stem and leaf diagram for the lengthof
stay variable.
d. Develop a bar chart that shows the number of males
and females by insurance carrier.
END EXERCISES 22
Line Charts and Scatter Diagrams
Line Charts
Most of the examples that have been presented thus far have involved crosssectional data or
data gathered from many observations all taken at the same time. However if you have time
series data that are measured over time e.g. monthly quarterly annually an effective tool
for presenting such data is a line chart.
BUSINESS APPLICATION CONSTRUCTING LINE CHARTS
MCGREGOR VINEYARDS McGregor Vineyards owns and operates a winery in the
Sonoma Valley in northern California. At a recent company meeting the financial manager
Chapter Outcome 6.
Line Chart
A twodimensional chart showing time on the
horizontal axis and the variable of interest on the
vertical axis.
Excel Tutorial
Excel
tutorials
www.downloadslide.comslide 71:
Graphs Charts and Tables—Describing Your Data
5. Use the Layout tab in the Chart Tools
to remove the Legend change the
Chart Title add the Axis Titles and
remove the grid lines.
6. Repeat Steps 2–5 for the Proft data.
Excel 2010 Instructions:
1. Open fle: McGregor.xlsx.
2. Select the Sales dollars data to be
graphed.
3. On the Insert tab click the Line chart.
4. Click the Line with Markers option.
Minitab Instructions
for similar results:
1. Open fle:
McGregor.MTW.
2. Choose Graph Times
Series Plot.
3. Select Simple.
4. Click OK.
5. In Series enter Sales and
Proft columns.
6. Select Multiple Graphs.
7. Under Show Graph
Variables select In
separate panels of
the same graph.
8. Click OK. OK.
Sales Increasing but
Profts Decreasing
FIGURE 12 
Excel 2010 Output Showing
McGregor Line Charts for
Sales and Profits
5. Use the Layout tab in the Chart Tools
to change the Chart Title add the Axis
Titles remove the border and remove
the grid lines.
Excel 2010 Instructions:
1. Open fle: McGregor.xlsx.
2. Select the two variables Sales dollars
and Proft to be graphed.
3. On the Insert tab click the Line chart.
4. Click the Line with Markers option.
Minitab Instructions
for similar results:
1. Open fle:
McGregor.MTW.
2. Choose Graph Times
Series Plot.
3. Select Multiple.
4. Click OK.
5. In Series enter Sales and
Proft columns.
6. Select Multiple Graphs.
7. Under Show Graph
Variables select Overlaid
on the same graph.
8. Click OK. OK.
FIGURE 13 
Excel 2010 Line Chart of
McGregor Profit and Sales
Using a Single Vertical Axis
www.downloadslide.comslide 72:
Graphs Charts and Tables—Describing Your Data
expressed concern about the company’s profit trend over the past 20 weeks. He presented
weekly profit and sales data to McGregor management personnel. The data are in the file
McGregor.
Initially the financial manager developed two separate line charts for this data: one for
sales the other for profits. These are displayed in Figure 12. These line charts provide an indi
cation that although sales have been increasing the profit trend is downward. But to fit both
Excel graphs on one page he had to compress the size of the graphs. This “flattened” the lines
somewhat masking the magnitude of the problem.
What the financial manager needed is one graph with both profits and sales. Figure 13
shows his first attempt. This is better but there still is a problem: The sales and profit vari
ables are of different magnitudes. This results in the profit line being flattened out to almost a
straight line. The profit trend is hidden.
To overcome this problem the financial manager needed to construct his graph using two
scales one for each variable. Figure 14 shows the improved graph. We can now clearly see
that although sales are moving steadily higher profits are headed downhill. For some reason
costs are rising faster than revenues and this graph should motivate McGregor Vineyards to
look into the problem.
EXAMPLE 10 LINE CHARTS
Grogan Builders Grogan Builders produces mobile homes in Alberta Canada. The owners
are planning to expand the manufacturing facilities. To do so requires additional financing. In
preparation for the meeting with the bankers the owners have assembled data on total annual
sales for the past 10 years. These data are shown as follows:
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
1426 1678 2591 2105 2744 3068 2755 3689 4003 3997
Example 10
Constructing Line Charts
A line chart also commonly called
a trend chart is developed using the
following steps:
Identify the timeseries vari
able of interest and determine
the maximum value and the
range of time periods covered
in the data.
Construct the horizontal axis
for the time periods using
equal spacing between time
periods. Construct the vertical
axis with a scale appropriate
for the range of values of the
timeseries variable.
Plot the points on the graph and
connect the points with straight
lines.
5. Move graph to separate page.
6. Select Proft Line on graph and rick click.
7. Click on the Format Data Series.
8. Click on Secondary Axis.
9. Click on Layout and add titles as desired.
Excel 2010 Instructions:
1. Open fle: McGregor.xlsx.
2. Select data from the proft and
sales column.
3. Click on Insert.
4. Click on Line Chart.
Minitab Instructions
for similar results:
1. Open fle:
McGregor.MTW.
2. Choose Graph Times
Series Plot.
3. Select Multiple.
4. Click OK.
5. In Series enter Sales and
Proft columns.
6. Select Multiple Graphs.
8. Click OK.
Two vertical axes:
Left Sale Right Profts
Profts and Sales moving
in different directions.
FIGURE 14 
Excel 2010 Line Chart for
Sales and Profits Using Two
Vertical Axes and Different
Scales
www.downloadslide.comslide 73:
Graphs Charts and Tables—Describing Your Data
The owners wish to present these data in a line chart to effectively show the company’s sales
growth over the 10year period. To construct the line chart the following steps are used:
Step 1 Identify the timeseries variable.
The timeseries variable is units sold measured over 10 years with a maximum
value of 4003.
Step 2 Construct the horizontal and vertical axes.
The horizontal axis will have the 10 time periods equally spaced. The vertical
axis will start at zero and go to a value exceeding 4003. We will use 4500.
The vertical axis will also be divided into 500unit increments.
Step 3 Plot the data values on the graph and connect the points with straight lines.
Scatter Diagrams
In Section 1 we introduced a set of statistical procedures known as joint frequency distribu
tions that allow the decision maker to examine two variables at the same time. Another pro
cedure used to study two quantitative variables simultaneously is the scatter diagram or the
scatter plot.
There are many situations in which we are interested in understanding the bivariate rela
tionship between two quantitative variables. For example a company would like to know
the relationship between sales and advertising. A bank might be interested in the relationship
between savings account balances and credit card balances for its customers. A real estate
agent might wish to know the relationship between the selling price of houses and the number
of days that the houses have been on the market. The list of possibilities is almost limitless.
Regardless of the variables involved there are several key relationships we are looking
for when we develop a scatter diagram. Figure 15 shows scatter diagrams representing some
key bivariate relationships that might exist between two quantitative variables.
Elsewhere the text introduces a statistical technique called regression analysis that
focuses on the relationship between two variables. These variables are known as dependent
and independent variables.
BUSINESS APPLICATION CREATING SCATTER DIAGRAMS
PERSONAL COMPUTERS Can you think of any product that has increased in quality and
capability as rapidly as personal computers Not that many years ago an 8MB RAM system
with a 486 processor and a 640KB hard drive sold in the mid2500 range. Now the same money
would buy a 3.0 GHz or faster machine with a 100+ GB hard drive and 512MB RAM or more
A few years ago we examined various Web sites looking for the best prices on personal
computers. The data file called Personal Computers contains data on several characteristics
Chapter Outcome 7.
Scatter Diagram or Scatter Plot
A twodimensional graph of plotted points in
which the vertical axis represents values of
one quantitative variable and the horizontal
axis represents values of the other quantitative
variable. Each plotted point has coordinates
whose values are obtained from the respective
variables.
Independent Variable
A variable whose values are thought to impact
the values of the dependent variable. The
independent variable or explanatory variable
is often within the direct control of the decision
maker. On a scatter plot the independent
variable or explanatory variable is graphed on
the x axis.
Dependent Variable
A variable whose values are thought to be a
function of or dependent on the values of
another variable called the independent variable.
On a scatter plot the dependent variable is
placed on the y axis and is often called the
response variable.
END EXAMPLE
TRY PROBLEM 47
0
500
4000
4500
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
1000
1500
2000
2500
3000
3500
Mobile Homes Sold
Year
Grogan Builders
Annual Sales
www.downloadslide.comslide 74:
Graphs Charts and Tables—Describing Your Data
including processor speed hard drive capacity RAM whether a monitor is included and price
for 13 personal computers. Of particular interest is the relationship between the computer price
and processing speed. Our objective is to develop a scatter diagram to graphically depict what
if any relationship exists between these two variables. The dependent variable is price and the
independent variable is processor speed. Figure 16 shows the Excel scatter diagram output.
The relationship between processor speed and price is somewhat curvilinear and positive.
y
x
a Linear
y
x
d Curvilinear
y
x
b Linear
y
x
e No Relationship
y
x
c Curvilinear
y
x
f No Relationship
FIGURE 15 
Scatter Diagrams Showing
Relationships Between x and y
Excel Tutorial
Excel
tutorials
3. On Insert tab click Scatter and then
click Scatter with only Markers option.
4. Move chart to separate page.
5. Use the Layout tab of the Chart Tools
to add titles and remove grid lines.
Excel 2010 Instructions:
1. Open fle: Personal Computers.xlsx.
2. Select data for chart Processor GHz
and Price. Hint: Use Crtl key to select
just the two desired columns.
Minitab Instructions
for similar results:
1. Open fle: Personal
Computers.MTW.
2. Choose Graph
Scatterplot.
3. Select Simple.
4. Click OK.
5. In Y enter Price column.
In X enter Processor
Speed column.
6. Click OK.
FIGURE 16 
Excel 2010 Scatter Diagrams
for Personal Computer Data
www.downloadslide.comslide 75:
Graphs Charts and Tables—Describing Your Data
EXAMPLE 11 SCATTER DIAGRAMS
F ortune’s Best Eight Companies Each year F ortune magazine surveys employees regarding
job satisfaction to try to determine which companies are the “best” companies to work for in the
United States. F ortune also collects a variety of data associated with these companies. For example
the table here shows data for the top eight companies on three variables: number of U.S. employees
number of training hours per year per employee and total revenue in millions of dollars.
Company U.S. Employees Training Hr/Yr Revenues Millions
Southwest Airlines 24757 15 3400
Kingston Technology 552 100 1300
SAS Institute 3154 32 653
FelPro 2577 60 450
TD Industries 976 40 127
MBNA 18050 48 3300
W.L. Gore 4118 27 1200
Microsoft 14936 8 8700
To better understand these companies we might be interested in the relationship between
number of U.S. employees and revenue and between training hours and U.S. employees. To
construct these scatter diagrams we can use the following steps:
Step 1 Identify the two variables of interest.
In the first case one variable is U.S. employees and the second is revenue. In
the second case one variable is training hours and the other is U.S. employees.
Step 2 Identify the dependent and independent variables.
In each case think of U.S. employees as the independent x variable. Thus
Case 1: y revenue vertical axis x U.S. employees horizontal axis
Case 2: y training hours vertical axis x U.S. employees horizontal axis
Step 3 Establish the scales for the vertical and horizontal axes.
The maximum value for each variable is
revenue +8700 U.S.employees 24757 training hours 100
Step 4 Plot the joint values for the two variables by placing a point in the x y space.
Example 11
Constructing Scatter Diagrams
A scatter diagram is a twodimen
sional graph showing the joint val
ues for two quantitative variables.
It is constructed using the following
steps:
Identify the two quantitative
variables and collect paired
responses for the two variables.
Determine which variable will
be placed on the vertical axis
and which variable will be
placed on the horizontal axis.
Often the vertical axis can be
considered the dependent vari
able y and the horizontal axis
can be considered the independ
ent variable x.
Define the range of values for
each variable and define the
appropriate scale for the x and y
axes.
Plot the joint values for the two
variables by placing a point in
the xy space. Do not connect
the points.
0
30000 25000 20000
1000
2000
3000
4000
5000
6000
7000
8000
9000
10000
Revenue Millions
U.S. Employees
Scatter Diagram
15000 10000 5000 0
General Positive Relationship
www.downloadslide.comslide 76:
Graphs Charts and Tables—Describing Your Data
Descriptive Statistics and Data Mining
The purpose of data mining and the descriptive statistics techniques discussed in the chapter
are essentially the same—that is to take large amounts of data and arrange it so some under
lying pattern is more easily identifiable to the decision maker. In fact several of the tools are
the same only under different names. A few examples should suffice.
Pareto Charts We once had the CEO of a Fortune 500 company tell us most companies
don’t get serious about quality until they face a crisis. The crisis his company faced was run
ning out of money. Pareto charts are a commonly used tool of quality but they are just histo
grams with a different name.
END EXAMPLE
TRY PROBLEM 46
Number 48 32 23 8 4
Percent 41.7
41.7
27.8 20.0 7.0 3.5
Cum 69.6 89.6 96.5 100.0
Misshapen Causes All Others Cracks Surface Scars Incomplete
50
40
30
20
10
0
Number
Pareto Chart of Causes
0
30000 25000 20000
20
40
60
80
100
120
Training Hours
U.S. Employees
Scatter Diagram
15000 10000 5000 0
General Negative Relationship
www.downloadslide.comslide 77:
Graphs Charts and Tables—Describing Your Data
2 Skill Development
246. The following data represent 11 observations of two
quantitative variables:
x contact hours with client
y profit generated from client.
x yxyx y x y
45 2345 54 3811 34 700 24 1975
56 4200 24 2406 45 3457 32 206
26 278 23 3250 47 2478
a. Construct a scatter plot of the data. Indicate whether
the plot suggests a linear or nonlinear relationship
between the dependent and independent variables.
b. Determine how much influence one data point
will have on your perception of the relationship
between the independent and dependent variables
by deleting the data point with the smallest x value.
What appears to be the relationship between the
dependent and independent variables
247. You have the following sales data for the past 12
months. Develop a line graph for these data.
Month Jan Feb Mar Apr May Jun
Sales 200 230 210 300 320 290
Month Jul Aug Sep Oct Nov Dec
Sales 300 360 400 410 390 450
MyStatLab
248. The following data have been selected for two
variables y and x. Construct a scatter plot for these two
variables and indicate what type of relationship if any
appears to be present.
yx
100 23.5
250 17.8
70 28.6
130 19.3
190 15.9
250 19.1
40 35.3
249. The yearend dollar value in millions of deposits for
Bank of the Ozarks Inc. for the years 1997–2010 are
shown below
Year 1997 1998 1999 2000 2001 2002 2003
Deposits 296 529 596 678 678 790 1062
Year 2004 2005 2006 2007 2008 2009 2010
Deposits 1380 1592 2045 2057 2341 2029 2541
Develop a chart that effectively displays the deposit
data over time.
The Pareto Principle named after Alfredo Pareto but identified in quality circum
stances by Joseph Juran has led business decision makers to realize the majority of sick
days taken by employees are taken by a minority of employees or the majority of customer
complaints are filed by a minority of customers or a majority of quality problems arise
from a minority of possible causes. Consider the previous Pareto chart taken from a popular
qualitycontrol text. The example involves a company trying to determine the reason for
complaints about its shipments to customers. The company was able to identify 20 possible
types of complaints.
This should look familiar. It is a histogram although it is called a Pareto Chart and shows
3 out of 20 possible causes account for almost 90 of the complaints. This form of histogram
will show a company where to concentrate its efforts to reduce customer complaints. Charts
like this are also easier to use for employees who are not comfortable using more complicated
statistical tools.
Scatter Diagrams Numerous police departments use a version of scatter diagrams to
determine where to concentrate their efforts. The area of the diagram is a map of the city and
the locations of crimes are plotted often using a color code to identify the type of crime. By
identifying clusters of crimes rather than spreading their efforts equally throughout the city
police are able to concentrate patrol and reduction efforts where the plots indicate clusters are
happening. Numerous cities such as San Francisco have found such scatter diagrams to be
effective tools in combating crime.
www.downloadslide.comslide 78:
Graphs Charts and Tables—Describing Your Data
250. VanAuker Properties’ controller collected the
following data on annual sales and the years of
experience of members of his sales staff:
Sales K: 200 191 135 236 305 183 50 192 184 73
Years: 10 459 12 6 2 76 2
a. Construct a scatter plot representing these data.
b. Determine the kind of relationship that exists if
any between years of experience and sales.
c. Approximate the increase in sales that accrues with
each additional year of experience for a member of
the sales force.
Business Applications
251. Amazon.com celebrated its 13th anniversary in
July 2007. Its growth can be seen by examining its
increasing sales volume in billions as reported by
Hoovers Inc.
Sales 0.0005 0.0157 0.1477 0.6098 1.6398
Year 1995 1996 1997 1998 1999
Sales 2.7619 3.1229 3.9329 5.2637 6.9211
Year 2000 2001 2002 2003 2004
Sales 8.490 10.711 14.835
Year 2005 2006 2007
a. Construct a line plot for Amazon’s sales.
b. Describe the type of relationship that exists between
the years in business and Amazon’s sales volume.
c. In which year does it appear that Amazon had the
sharpest increase in sales
252. In July 2005 Greg Sandoval of the Associated Press
authored a study of the video game industry that focused
on the efforts of the industry to interest women in the
games. In that study he cited another report by the
Entertainments Software Association that indicated that
the percentage of women who played video games in 2004
was 43 whereas only 12.5 of the software developers
were female. Sandoval also presented the following data
showing the U.S. computer/video game sales:
Year Sales Billions
1996 3.80
1997 4.30
1998 5.70
1999 6.10
2000 6.00
2001 6.30
2002 6.95
2003 7.00
2004 7.30
Construct a line chart showing these computer/video
game sales data. Write a short statement describing the
graph.
253. The recent performance of U.S. equity markets has
increased the popularity of dividendpaying stocks
for some investors. Shown below are the diluted net
earnings per common share and the dividends per
common share for the Procter Gamble Company
PG for the years 1996–2011.
Year
Diluted Net Earnings
per Common Share
Dividends per
Common Share
1996 1.00 0.40
1997 1.14 0.45
1998 1.28 0.51
1999 1.29 0.57
2000 1.23 0.64
2001 1.03 0.70
2002 1.54 0.76
2003 1.85 0.82
2004 2.32 0.93
2005 2.66 1.03
2006 2.64 1.15
2007 3.04 1.28
2008 3.64 1.45
2009 4.26 1.64
2010 4.11 1.80
2011 3.93 1.97
a. Construct a line chart of diluted net earnings per
common share for the years shown.
b. Construct a line chart of dividends per common
share for the years shown.
c. Construct the appropriate chart for determining
whether there is a relationship between diluted
net earnings per common share and dividends
per common share for the years shown. Briefly
comment on the nature of any relationship you
believe your chart reveals.
254. Business Week Reed Stanley et al. “Open Season
on Big Oil” September 26 2005 reported on data
provided by A. G. Edwards Sons concerning the
profits billions for 10 of the largest integrated
oil and gas companies over the period from 1999 to
2005.
Year 1999 2000 2001 2002 2003 2004 2005
Profit Billions 33.3 62.5 58.3 41.7 66.7 91.7 118.0
a. Produce a line plot of the profit versus the year.
b. Describe the types of relationships that exist
between years and profits during the specified time
period.
c. Which of the relationships would you use to project
the companies’ profits in the year 2006 Explain
your answer.
www.downloadslide.comslide 79:
Graphs Charts and Tables—Describing Your Data
Computer Database Exercises
255. Major League Baseball MLB is played in 30 North
American cities including Toronto Canada. Having
a team in a city is generally considered to provide an
economic boost to the community. Although winning
is the stated goal for all teams the business side of
baseball has to do with attendance. The data file MLB
Attendance2008 contains data for both home and
road game attendance for all 30 MLB teams for 2008.
Of interest is the relationship between average home
attendance and average road attendance. Using the
2008 attendance data construct the appropriate graph
to help determine the relationship between these two
variables and discuss the implications of the graph.
256. In the October 17 2005 issue of Fortune a special
advertising section focused on private jets. Included
in the section was an article about “fractional”
jet ownership in which wealthy individuals and
companies share ownership in private jets. The idea
is that the expensive airplanes can be better utilized if
more than one individual or company has an ownership
stake. AvData Inc. provided data showing the number
of fractional ownerships since 1986. These data are
in the file called JetOwnership. Using these data
develop a line chart that displays the trend in fractional
ownership between 1986 and 2004. Discuss.
257. Starting in 2005 a chain of events including the
war in Iraq Hurricane Katrina and the expanding
economies in India and China lead to a sharp increase
in fuel costs. As a result the U.S. airline industry has
been hit hard financially with many airlines declaring
bankruptcy. Some airlines are substituting smaller
planes on certain routes in an attempt to reduce fuel
costs. As an analyst for one of the major airlines you
have been asked to analyze the relationship between
passenger capacity and fuel consumption per hour.
Data for 19 commonly flown planes is presented in the
file called Airplanes. Develop the appropriate graph
to illustrate the relationship between fuel consumption
per hour and passenger capacity. Discuss.
258. Japolli Bakery tracks sales of its different bread
products on a daily basis. The data at the bottom of
this page show sales for 22 consecutive days at one
of its retail outlets in Nashville. Develop a line chart
that displays these data. The data are also located in a
data file called Japolli Bakery. Discuss what if any
conclusions you might be able to reach from the line
chart.
259. Energy prices have been a major source of economic
and political debate in the United States and around the
world. Consumers have recently seen gasoline prices
both rise and fall rapidly and the impact of fuel prices
has been blamed for economic problems in the United
States at different points in time. Although no longer
doing so for years the California Energy Commission
published yearly gasoline prices. The data found in the
file called Gasoline Prices reflect the average price of
regular unleaded gasoline in the state of California for
the years between 1970 and 2005. The first price column
is the actual average price of gasoline during each of
Japolli Bakery
Day of Week White Wheat Multigrain Black Cinnamon Raisin Sourdough French Light Oat
Friday 436 456 417 311 95 96 224
Saturday 653 571 557 416 129 140 224
Sunday 496 490 403 351 114 108 228
Monday 786 611 570 473 165 148 304
Tuesday 547 474 424 365 144 104 256
Wednesday 513 443 380 317 100 92 180
Thursday 817 669 622 518 181 152 308
Friday 375 390 299 256 124 88 172
Saturday 700 678 564 463 173 136 248
Sunday 597 502 457 383 140 144 312
Monday 536 530 428 360 135 112 356
Tuesday 875 703 605 549 201 188 356
Wednesday 421 433 336 312 100 104 224
Thursday 667 576 541 438 152 144 304
Friday 506 461 406 342 135 116 264
Saturday 470 352 377 266 84 92 172
Sunday 748 643 599 425 153 148 316
Monday 376 367 310 279 128 104 208
Tuesday 704 646 586 426 174 160 264
Wednesday 591 504 408 349 140 120 276
Thursday 564 497 415 348 107 120 212
Friday 817 673 644 492 200 180 348
www.downloadslide.comslide 80:
Graphs Charts and Tables—Describing Your Data
END EXERCISES 23
those years. The second column is the average price
adjusted for inflation with 2005 being the base year.
a. Construct an appropriate chart showing the actual
average price of gasoline in California over the
years between 1970 and 2005.
b. Add to the graph developed in part a the data for the
adjusted gasoline prices.
c. Based on the graph from part b what conclusions
might be reached about the price of gasoline over
the years between 1970 and 2005
260. Federal flood insurance underwritten by the federal
government was initiated in 1968. This federal flood
insurance coverage has according to USA Today
“How You Pay for People to Build in Flood Zones”
September 21 2005 more than tripled in the past
15 years. A file titled Flood contains the amount of
federal flood insurance coverage for each of the years
from 1990 to 2004.
a. Produce a line plot for these data.
b. Describe the type of relationship between the year
and the amount of federal flood insurance.
c. Determine the average increase per year in federal
flood insurance.
261. The Office of Management and Budget keeps data on
many facets of corporations. One item that has become
a matter of concern is the number of applications
for patents submitted compared to the backlog of
applications that have not been processed by the end
of the year. A file titled Patent provides data extracted
from a USA Today article that addresses the problem.
a. Construct the two line plots on the same axes.
b. Determine the types of relationships that exist
between the years and the two patentrelated
variables.
c. During which years did the backlog of
applications at the end of the year equal
approximately the same number of patent
applications
262. The subprime mortgage crisis that hit the world
economy also impacted the real estate market. Both
new and existing home sales were affected. A file titled
EHSales contains the number of existing homes sold
in millions from September of 2007 to September
2008.
a. Construct a line plot for these data.
b. The data file also contains data concerning the
median selling price thousands. Construct a
graph containing the line plot for both the number
of sales tens of thousands and the median
thousands price of these sales for the indicated
time period.
c. Describe the relationship between the two line plots
constructed in part b.
www.downloadslide.comslide 81:
Graphs Charts and Tables—Describing Your Data
1 Frequency Distributions and Histograms
2 Bar Charts Pie Charts and Stem and Leaf Diagrams
Outcome 4. Construct and interpret various types of bar charts.
Outcome 5. Build a stem and leaf diagram.
3 Line Charts and Scatter Diagrams
Outcome 6. Create a line chart and interpret the trend in the data.
Outcome 7. Create a scatter plot and interpret it.
Summary
When you are working with timeseries data and you are interested in displaying the
pattern in the data over time the chart that is used is called a line chart. The vertical
axis displays the value of the timeseries variable while the horizontal axis contains the
time increments. The points are plotted and are usually connected by straight lines. In
other cases you may be interested in the relationship between two quantitative
variables the graphical tool that is used is called a scatter diagram. The variable
judged to be the dependent variable is placed on the vertical axis and
independent variable goes on the horizontal axis. The joint values are plotted as
points in the twodimensional space. Do not connect the points with lines.
Conclusion
There are many types of charts graphs and tables that can be
used to display data. The technique that is used often depends
on the type and level of data you have. In cases where multiple
graphs or charts can apply you should select the one that most
effectively displays the data for your application. Figure 17
summarizes the different graphical options that are presented in
this chapter.
The old adage states that a picture is worth a thousand words. In many ways
this applies to descriptive statistics. The use of graphs charts and tables to
display data in a way that helps decisionmakers better understand the data is
one of the major applications of business statistics. This chapter has introduced
many of the most frequently used graphical techniques using examples and
business applications.
Summary
A frequency distribution is used to determine the number of occurrences in your data that fall at each
possible data value or within defned ranges of possible values. It represents a good summary of the data
and from a frequency distribution you can form a graph called a histogram. This histogram gives a visual
picture showing how the data are distributed. You can use the histogram to see where the data’s center is
and how spread out the data are. It is often helpful to convert the frequencies in a frequency distribution to
relative frequencies and to construct a relative frequency distribution and a relative frequency
histogram. Another option is to convert the frequency distribution to a cumulative frequency
distribution and then a graph called an ogive. Finally if you are analyzing two variables simultaneously
you may want to construct a joint frequency distribution.
Outcome 1. Construct frequency distributions both manually and with your computer.
Outcome 2. Construct and interpret a frequency histogram.
Outcome 3. Develop and interpret joint frequency distributions.
Summary
If your data are discrete or are nominal or ordinal level three charts introduced in this
section are often considered. These are bar charts pie charts and stem and leaf
diagrams. A bar chart can be arranged with the bars vertical or horizontal. A single bar
chart can be used to describe two or more variables. In situations where you wish to show
how the parts making up a total are distributed a pie chart is often used. The “slices” of the
pie are many times depicted as the percentage of the total. A lesser used graphical tool that
provides a quick view of how the data are distributed is the stem and leaf diagram.
www.downloadslide.comslide 82:
Graphs Charts and Tables—Describing Your Data
Equations
Cumulative
Relative
Frequency
Distribution
Quantitative Qualitative
Categorical
Nominal/Ordinal
Time
Series
CrossSectional
Grouped
or
Ungrouped
Discrete or
Continuous
Interval/Ratio
Line Chart
Bar Chart
Vertical
Frequency
Distribution
Bar Chart
Vertical or
Horizontal
Relative
Frequency
Distribution
Joint
Frequency
Distribution
Frequency
Distribution
Relative
Frequency
Distribution
Joint
Frequency
Distribution
Data
Type
Data
Class
Histogram
Pie Chart
Stem and Leaf
Diagram
Scatter
Diagram
Ogive
FIGURE 17 
Summary: Descriptive
Statistical Techniques
1 Relative Frequency
Relative frequency 5
f
n
i
2 Class Width
W
Largest value Smallest value
Number of clas sses
Key Terms
Allinclusive classes
Bar chart
Class boundaries
Class width
Continuous data
Cumulative frequency distribution
Cumulative relative frequency distribution
Chapter Exercises MyStatLab
Conceptual Questions
263. Discuss the advantages of constructing a relative frequency
distribution as opposed to a frequency distribution.
264. What are the characteristics of a data set that would lead
you to construct a bar chart
265. What are the characteristics of a data set that would
lead you to construct a pie chart
266. Discuss the differences in data that would lead you
to construct a line chart as opposed to a scatter
plot.
Dependent variable
Discrete data
Equalwidth classes
Frequency distribution
Frequency histogram
Independent variable
Line chart
Mutually exclusive classes
Ogive
Pie chart
Relative frequency
Scatter diagram or scatter plot
www.downloadslide.comslide 83:
Graphs Charts and Tables—Describing Your Data
Business Applications
267. USA T oday reported Anthony Breznican and Gary
Strauss “Where Have All the Moviegoers Gone” June
23 2005 that in the summer of 2005 ticket sales to movie
theaters had fallen for 17 straight weeks the industry’s
longest losing streak since 1985. To determine the long
term trends in ticket sales the following data representing
the number of admissions in billions were obtained from
the National Association of Theatre Owners:
Year 1987 1988 1989 1990 1991 1992
Admissions 1.09 1.08 1.26 1.19 1.14 1.17
Year 1993 1994 1995 1996 1997 1998
Admissions 1.24 1.28 1.26 1.34 1.39 1.48
Year 1999 2000 2001 2002 2003 2004
Admissions 1.47 1.42 1.49 1.63 1.57 1.53
Year 2005 2006 2007
Admissions 1.38 1.41 1.40
a. Produce a line plot of the data.
b. Describe any trends that you detect.
268. The following data represent the commuting distances
for employees of the PayandCarry Department store.
a. The personnel manager for PayandCarry would
like you to develop a frequency distribution and
histogram for these data.
b. Develop a stem and leaf diagram for these data.
c. Break the data into three groups under 3 miles 3 and
under 6 miles and 6 and over. Construct a pie chart to
illustrate the proportion of employees in each category.
Commuting Distance miles
3.5 2.0 4.0 2.5 0.3 1.0 12.0 17.5
3.0 3.5 6.5 9.0 3.0 4.0 9.0 16.0
3.5 0.5 2.5 1.0 0.7 1.5 1.4 12.0
9.2 8.3 1.0 3.0 7.5 3.2 2.0 1.0
3.5 3.6 1.9 2.0 3.0 1.5 0.4 6.4
11.0 2.5 2.4 2.7 4.0 2.0 2.0 3.0
d. Referring to part c construct a bar chart to depict
the proportion of employees in each category.
269. Anyone attending college realizes tuition costs have
increased rapidly. In fact tuition had risen at a faster
pace than inflation for more than two decades. Data
showing costs for both private and public colleges for
selected years are shown below.
Year 1984 1989 1994 1999 2004
Private College
Tuition
9202 12146 13844 16454 19710
Public College
Tuition
2074 2395 3188 3632 4694
a. Construct one bar graph illustrating the relationship
between private and public university tuition for the
displayed years.
b. Describe the tuition trend for both private and
public college tuition.
270. A recent article in USA Today used the following data
to illustrate the decline in the percentage of men who
receive college and advanced degrees:
Bachelor Doctorate
1989 2003 2014 1989 2003 2014
Men 47 43 40 64 57 49
Women 53 57 60 36 43 51
Education Department projection.
a. Use one graph that contains two bar charts each
of which represents the kind of degree received to
display the relationship between the percentages of
men and women receiving each type of degree.
b. Describe any trends that might be evident.
271. The Minnesota State Fishing Bureau has contracted with
a university biologist to study the length of walleyes
fish caught in Minnesota lakes. The biologist collected
data on a sample of 1000 fish caught and developed the
following relative frequency distribution:
Class Length inches Relative Frequency f
i
8 to less than 10 0.22
10 to less than 12 0.15
12 to less than 14 0.25
14 to less than 16 0.24
16 to less than 18 0.06
18 to less than 20 0.05
20 to less than 22 0.03
a. Construct a frequency distribution from this relative
frequency distribution and then produce a histogram
based on the frequency distribution.
b. Construct a pie chart from the relative frequency
distribution. Discuss which of the two graphs
the pie chart or the histogram you think is more
effective in presenting the fish length data.
272. A computer software company has been looking at the
amount of time customers spend on hold after their call
is answered by the central switchboard. The company
would like to have at most 2 of the callers wait two
minutes or more. The company’s calling service has
provided the following data showing how long each of
last month’s callers spent on hold:
Class Number
Less than 15 seconds 456
15 to less than 30 seconds 718
30 to less than 45 seconds 891
45 to less than 60 seconds 823
60 to less than 75 seconds 610
75 to less than 90 seconds 449
90 to less than 105 seconds 385
105 to less than 120 seconds 221
Continued www.downloadslide.comslide 84:
Graphs Charts and Tables—Describing Your Data
Class Number
120 to less than 150 seconds 158
150 to less than 180 seconds 124
180 to less than 240 seconds 87
More than 240 seconds 153
a. Develop a relative frequency distribution and ogive
for these data.
b. The company estimates it loses an average of 30
in business from callers who must wait two minutes
or more before receiving assistance. The company
thinks that last month’s distribution of waiting times
is typical. Estimate how much money the company
is losing in business per month because people have
to wait too long before receiving assistance.
273. The regional sales manager for American Toys Inc.
recently collected data on weekly sales in dollars for
the 15 stores in his region. He also collected data on the
number of salesclerk work hours during the week for
each of the stores. The data are as follows:
Store Sales Hours
1 23300 120
2 25600 135
3 19200 96
4 10211 102
5 19330 240
6 35789 190
7 12540 108
8 43150 234
9 27886 140
10 54156 300
11 34080 254
12 25900 180
13 36400 270
14 25760 175
15 31500 256
a. Develop a scatter plot of these data. Determine
which variable should be the dependent variable and
which should be the independent variable.
b. Based on the scatter plot what if any conclusions
might the sales manager reach with respect to the
relationship between sales and number of clerk
hours worked Do any stores stand out as being
different Discuss.
Computer Database Exercises
274. The Energy Information Administration published
a press release on September 26 2005 Paula Weir
and Pedro Saavedra “Two MultiPhase Surveys
That Combine Overlapping Sample Cycles at Phase
I”. The file titled Diesel contains the average
onhighway diesel prices for each of 53 weeks from
September 27 2004 to September 26 2005. It also
contains equivalent information for the state of
California recognized as having the highest national
prices.
a. Construct a chart containing line plots for both
the national average and California’s diesel prices.
Describe the relationship between the diesel prices
in California and the national average.
b. In what week did the California average diesel price
surpass 3.00 a gallon
c. Determine the smallest and largest price paid in
California for a gallon of diesel. At what weeks did
these occur Use this information to project when
California gas prices might exceed 4.00 assuming
a linear trend between California diesel prices and
the weeks in which they occurred.
275. A recent article in USA Today reported that Apple had
74 of the digital music device market according to
researcher The NPD Group. The NPD Group provides
global sales and marketing information that helps
clients make more informed factbased decisions to
optimize their businesses. The data in the file titled
Digital provide the brand of digital devices owned by
a sample of consumers that would produce the market
shares alluded to in the article. Produce a pie chart
that represents the market shares obtained from the
referenced sample. Indicate the market shares and the
identity of those manufacturers in the pie chart.
276. The file HomePrices contains information about
singlefamily housing prices in 100 metropolitan areas
in the United States.
a. Construct a frequency distribution and histogram
of 1997 median singlefamily home prices. Use
the 2
k
Ú n guideline to determine the appropriate
number of classes.
b. Construct a cumulative relative frequency
distribution and ogive for 1997 median single
family home prices.
c. Repeat parts a and b but this time use 1.5 times as
many class intervals as recommended by the 2
k
Ú n
guideline. What was the impact of using more class
intervals
277. Elliel’s Department Store tracks its inventory on a
monthly basis. Monthly data for the years 2008–2012
are in the file called Elliels.
a. Construct a line chart showing the monthly
inventory over the five years. Discuss what this
graph implies about inventory.
b. Sum the monthly inventory figures for each year.
Then present the sums in bar chart form. Discuss
whether you think this is an appropriate graph to
describe the inventory situation at Elliels.
278. The Energy Information Administration EIA surveys
the price of diesel fuel. The EIA888 is a survey of diesel
fuel outlet prices from truck stops and service stations
across the country. It produces estimates of national and
regional prices. The diesel fuel prices that are released are
used by the trucking industry to make rate adjustments
in hauling contracts. The file titled Diesel contains the
www.downloadslide.comslide 85:
Graphs Charts and Tables—Describing Your Data
average onhighway diesel prices for each of 53 weeks
from September 27 2004 to September 26 2005.
a. Construct a histogram with 11 classes beginning at
1.85.
b. Are there any data points that are unusually larger
than the rest of the data In which classes do these
points occur What is the interpretation of this
phenomenon
video
Video Case 2
DriveThru Service Times
McDonald’s
When you’re on the go and looking for a quick meal where do
you go If you’re like millions of people every day you make
a stop at McDonald’s. Known as “quick service restaurants”
in the industry not “fast food” companies such as McDon
ald’s invest heavily to determine the most efficient and effective
ways to provide fast highquality service in all phases of their
business.
Drivethru operations play a vital role. It’s not surprising that
attention is focused on the drivethru process. After all more than
60 of individual restaurant revenues in the United States come
from the drivethru experience. Yet understanding the process is
more complex than just counting cars. Marla King professor at
the company’s international training center Hamburger University
got her start 25 years ago working at a McDonald’s drivethru. She
now coaches new restaurant owners and managers. “Our stated
drivethru service time is 90 seconds or less. We train every man
ager and team member to understand that a quality customer expe
rience at the drivethru depends on them” says Marla. Some of the
factors that affect customers’ ability to complete their purchases
within 90 seconds include restaurant staffing equipment layout in
the restaurant training efficiency of the grill team and frequency
of customer arrivals to name a few. Also customer order patterns
also play a role. Some customers will just order drinks whereas
others seem to need enough food to feed an entire soccer team.
And then there are the special orders. Obviously there is plenty of
room for variability here.
Yet that doesn’t stop the company from using statistical tech
niques to better understand the drivethru action. In particular
McDonald’s uses graphical techniques to display data and to help
transform the data into useful information. For restaurant manag
ers to achieve the goal in their own restaurants they need training
in proper restaurant and drivethru operations. Hamburger Uni
versity McDonald’s training center located near Chicago Illi
nois satisfies that need. In the mockup restaurant service lab
managers go through a “before and after” training scenario. In
the “before” scenario they run the restaurant for 30 minutes as if
they were back in their home restaurants. Managers in the train
ing class are assigned to be crew customers drivethru cars spe
cialneeds guests such as hearing impaired indecisive clumsy
or observers. Statistical data about the operations revenues and
service times are collected and analyzed. Without the right train
ing the restaurant’s operations usually start breaking down after
1015 minutes. After debriefing and analyzing the data col
lected the managers make suggestions for adjustments and head
back to the service lab to try again. This time the results usually
come in well within standards. “When presented with the quanti
tative results managers are pretty quick to make the connections
between better operations higher revenues and happier custom
ers” Marla states.
When managers return to their respective restaurants the train
ing results and techniques are shared with staff who are charged
with implementing the ideas locally. The results of the training
eventually are measured when McDonald’s conducts a restaurant
operations improvement process study or ROIP. The goal is sim
ple: improved operations. When the ROIP review is completed
statistical analyses are performed and managers are given their
results. Depending on the results decisions might be made that
require additional financial resources building construction staff
training or reconfiguring layouts. Yet one thing is clear: Statis
tics drive the decisions behind McDonald’s drivethrough service
operations.
Discussion Questions:
1. After returning from the training session at Hamburger
University a McDonald’s store owner selected a random
sample of 362 drivethru customers and carefully
measured the time it took from when a customer entered
the McDonald’s property until the customer received the
order at the drivethru window. These data are in the file
called McDonald’s DriveThru Waiting Times. Note
the owner selected some customers during the breakfast
period others during lunch and others during dinner.
Construct any appropriate graphs and charts that will
effectively display these drivethru data. Prepare a short
discussion indicating the conclusions that this store owner
might reach after reviewing the graphs and charts you have
prepared.
2. Referring to question 1 suppose the manager comes
away with the conclusion that his store is not meeting the
90second customer service goal. As a result he plans to dig
deeper into the problem by collecting more data from the
drivethru process. Discuss what other measures you would
suggest the manager collect. Discuss how these data could be
of potential value in helping the store owner understand his
problem.
3. Visit a local McDonald’s that has a drivethru facility.
Randomly sample 20 drivethru customers and collect the
following data:
a. the total time from arrival on the property to departure
from the drivethru window
b. the time from when customers place the order until they
receive their order and exit the drivethru process
c. the number of cars in the line when the sampled vehicle
enters the drivethru process
d. Using the data that you have collected construct appropri
ate graphs and charts to describe these data. Write a short
report discussing the data
www.downloadslide.comslide 86:
Graphs Charts and Tables—Describing Your Data
Server Downtime
After getting outstanding grades in high school and scoring very
high on his ACT and SAT tests Clayton Haney had his choice of
colleges but wanted to follow his parents’ legacy and enrolled at
Northwestern University. Clayton soon learned that there is a big
difference between getting high grades in high school and being a
good student. Although he was recognized as being quite bright and
very quick to pick up on things he had never learned how to study.
As a result after slightly more than two years at Northwestern
Clayton was asked to try his luck at another university. To the cha
grin of his parents Clayton decided that college was not for him.
After short stints working for a computer manufacturer and
as a manager for a Blockbuster video store Clayton landed a job
working for EDS. EDS contracts to support information technol
ogy implementation and application for companies in the United
States and throughout the world. Clayton had to train himself in
virtually all aspects of personal computers and local area networks
and was assigned to work for a client in the Chicago area.
Clayton’s first assignment was to research the downtime on
one of the client’s primary network servers. He was asked to study
the downtime data for the month of April and to make a short pres
entation to the company’s management. The downtime data are in
a file called Server Downtime. These data are also shown in Table
C1A. Although Clayton is very good at solving computer prob
lems he has had no training or experience in analyzing data so he
is going to need some help.
Required Tasks:
1. Construct a frequency distribution showing the number of
times during the month that the server was down for each
downtime cause category.
2. Develop a bar chart that displays the data from the frequency
distribution in part a.
3. Develop a histogram that displays the downtime data.
TABLE C1A 
Date Problem Experienced Downtime Minutes
04/01 Lockups 25
04/02 Lockups 35
04/05 Memory Errors 10
04/07 Lockups 40
04/09 Weekly Virus Scan 60
04/09 Lockups 30
04/09 Memory Errors 35
04/09 Memory Errors 20
04/12 Slow Startup 45
04/12 Weekly Virus Scan 60
04/13 Memory Errors 30
04/14 Memory Errors 10
04/19 Manual Restart 20
04/20 Memory Errors 35
04/20 Weekly Virus Scan 60
04/20 Lockups 25
04/21 Memory Errors 35
04/22 Memory Errors 20
04/27 Memory Errors 40
04/28 Weekly Virus Scan 60
04/28 Memory Errors 15
04/28 Lockups 25
4. Develop a pie chart that breaks down the percentage of total
downtime that is attributed to each downtime cause during
the month.
5. Prepare a short written report that discusses the downtime
data. Make sure you include the graphs and charts in the
report.
Hudson Valley Apples Inc.
As a rule Stacey Fredrick preferred to work in the field rather
than do “office” work in her capacity as a midlevel manager with
Hudson Valley Apples Inc. a large grower and processor of apples
in the state of New Y ork. However after just leaving a staff meeting
at which she was asked to prepare a report of apple consumption in
the United States Stacey was actually looking forward to spend
ing some time at her computer “crunching some numbers.” Arden
Golchein senior marketing manager indicated that he would
email her a data file that contained apple consumption data from
1970 through 2009 and told her that he wanted a very nice report
using graphs charts and tables to describe apple consumption.
When she got to her desk the email was waiting and she
saved the file under the name Hudson Valley Apples. Stacey had
done quite a bit of descriptive analysis in her previous job with
the New York State Department of Agriculture so she had several
ideas for types of graphs and tables that she might construct. She
began by creating a list of the tasks that she thought would be
needed.
Required Tasks:
1. Construct a line chart showing the total annual availability of
apples.
2. Construct one line chart that shows two things: the annual
availability of fresh apples and the annual availability of
processed apples.
3. Construct a line chart that shows the annual availability for
each type of processed apples.
4. Construct a histogram for the total annual availability of
apples.
5. Write a short report that discusses the historical pattern of
apple availability. The report will include all pertinent charts
and graphs.
www.downloadslide.comslide 87:
Graphs Charts and Tables—Describing Your Data
Welco Lumber Company—Part A
Gene Denning wears several hats at the Welco Lumber Company
including process improvement team leader shipping manager
and assistant human resources manager. Welco Lumber makes
cedar fencing materials at its Naples Idaho facility employing
about 160 people.
More than 75 of the cost of the finished cedar fence boards
is in the cedar logs that the company buys on the open market.
Therefore it is very important that the company get as much fin
ished product as possible from each log. One of the most important
steps in the manufacturing process is referred to as the head rig.
The head rig is a large saw that breaks down the logs into slabs
and cants. Figure C3A shows the concept. From small logs with
diameters of 12 inches or less one cant and four or fewer usable
slabs are obtained. From larger logs multiple cants and four slabs
are obtained. Finished fence boards can be produced from both the
slabs and the cants.
At some companies the head rig cutting operation is
automated and the cuts are made based on a scanner system and
computer algorithms. However at Welco Lumber the head rig
is operated manually by operators who must look at a log as it
arrives and determine how best to break the log down to get the
most finished product. In addition the operators are responsible
for making sure that the cants are “centered” so that maximum
product can be gained from them.
Recently Gene Denning headed up a study in which he video
taped 365 logs being broken down by the head rig. All three opera
tors April Sid and Jim were involved. Each log was marked as to
its true diameter. Then Gene observed the way the log was broken
down and the degree to which the cants were properly centered. He
then determined the projected value of the finished product from
each log given the way it was actually cut. In addition he also
determined what the value would have been had the log been cut
in the optimal way. Data for this study is in a file called Welco
Lumber. A portion of the data is shown in Table C3A.
You have been asked to assist Gene by analyzing these data
using graphs charts and tables as appropriate. He wishes to focus
on the lost profit to the company and whether there are differences
among the operators. Also do the operators tend to do a better job
on small logs than on large logs In general he is hoping to learn
as much as possible from this study and needs your help with the
analysis.
TABLE C3A  Head Rig Data—Welco Lumber Company
5Nov06 Through
21Nov06
Head Rig Log Study Baseline
Log Operator Log Size
Large/Small
Log
Correct Cut
Yes or No Error Category Actual Value
Potential
Value Potential Gain
1 Sid 15 Large No Excessive Log Breakdown 59.00 65.97 6.97
2 Sid 17 Large No Excessive Log Breakdown 79.27 85.33 6.06
3 Sid 11 Small Yes No Error 35.40 35.40 0.00
4 Sid 11 Small No Off Center Cant 31.61 35.40 3.79
5 Sid 14 Large No Reduced V alue Cut 47.67 58.86 11.19
6 Sid 17 Large Yes No Error 85.33 85.33 0.00
7 Sid 8 Small Yes No Error 16.22 16.22 0.00
8 Sid 11 Small Yes No Error 35.40 35.40 0.00
9 Sid 9 Small Yes No Error 21.54 21.54 0.00
10 Sid 9 Small No Off Center Cant 18.92 21.54 2.62
11 Sid 10 Small Yes No Error 21.54 21.54 0.00
12 Sid 8 Small Yes No Error 16.22 16.22 0.00
13 Sid 10 Small No Off Center Cant 25.71 28.97 3.26
14 Sid 12 Small Yes No Error 41.79 41.79 0.00
15 Sid 11 Small Yes No Error 35.40 35.40 0.00
Slabs
Slabs
Cant
FIGURE C3A  Log Breakdown at the Head Rig
www.downloadslide.comslide 88:
Graphs Charts and Tables—Describing Your Data
3. a. 2
k
Ú n or 2
10
1024 Ú 1000. Thus use k 10 classes.
b. w
High  Low
Classes
2900  300
10
2600
10
260 round to 300
5. a. 2.833 which is rounded to 3.
b. Divide the number of occurrences frequency in each class
by the total number of occurrences.
c. Compute a running sum for each class by adding the
frequency for that class to the frequencies for all classes
above it.
d. Classes form the horizontal axis and the frequency forms
the vertical axis. Bars corresponding to the frequency of
each class are developed.
7. a. 1  0.24 0.76
b. 0.56  0.08 0.48
c. 0.96  0.08 0.86
9. a.
b. cumulative frequencies: 2 27 53 59 60
c. cumulative relative frequencies: 0.0333 0.4500 0.8833
0.9833 1.000
d. ogive
13. a. The weights are sorted from smallest to largest to create the
data array.
b.
c. The histogram can be created from the frequency
distrib ution.
d. 10.20
15. a. w
Largest  smallest
number of classes
214.4  105.0
11
9.945 S w 10.
b. 8 of the 25 or 0.32 of the salaries are at least 175000
c. 18 of the 25 or 0.72 having salaries that are at most
205000 but at least 135000
19. a. 9 classes
b. w
High  Low
Classes
32  10
9
22
9
2.44 round up to 3.0
c. The frequency distribution with nine classes and a class
width of 3.0 will depend on the starting point for the first
class. This starting value must be at or below the minimum
value of 10.
d. The distribution is mound shaped and fairly symmetrical. It
appears that the center is between 19 and 22 rounds
per year but the rounds played are quite spread out around
the center.
21. a. 2
5
32 and 2
6
64. Therefore 6 classes are chosen.
b. The maximum value is 602708 and the minimum
value is 160 from the Total column. The difference is
602708  160 602548. The class width would be
6025486 100424.67. Rounding up to the nearest
1000 produces a class width of 101000.
c.
d. The vast majority of airlines had fewer than 101000
monthly passengers.
23. a. Order the observations coffee consumption from smallest
to largest.
b. Using the 2
k
7 n guideline the number of classes k
would be 0.9 and w 110.1  3.528 0.825 which
is rounded to 0. 9. The class with the largest number is
6.2–7.0 kg of coffee.
c. The histogram can be created from the frequency distri
bution. The classes are shown on the horizontal axis and
the frequency on the vertical axis.
d. 8.33 1100  91.672 of the coffee drinkers sampled
consume 8.0 kg or more annually.
29. a. The pie chart categories are the regions and the measure is
the region’s percentage of total income.
b. The bar chart categories are the regions and the measure
for each category is the region’s percentage of total
income.
c. The bar chart however makes it easier to compare per
centages across regions.
31. b. 1  0.0985 0.9015
33. The bar chart is skewed indicating that the number of 1
million houses is growing rapidly. The growth is exponential
rather than linear.
35. A bar chart can be used to make the comparison.
37. a. The stem unit is 10 and the leaf unit is 1.
b. between 70 and 79 seconds
41. a. Leaf unit 1.0
b. Slightly skewed to the left. The center is between 24
and 26.
c. x
2428
50
48.56
Answers to Selected OddNumbered Problems
This section contains summary answers to most of the oddnumbered problems in the text. The Student Solutions Manual contains fully developed
solutions to all oddnumbered problems and shows clearly how each answer is determined.
Class Frequency Relative Frequency
2–3 2 0.0333
4–5 25 0.4167
6–7 26 0.4333
8–9 6 0.1000
10–11 1 0.0167
Weight Classes Frequency
77–81 3
82–86 9
87–91 16
92–96 16
97–101 5
Total 49
Class Frequency
0 6 101000 33
101000 6 202000 2
202000 6 303000 2
303000 6 404000 1
404000 6 505000 2
505000 6 606000 1
www.downloadslide.comslide 89:
Graphs Charts and Tables—Describing Your Data
51. b. curvilinear
c. The largest difference in sales occurred between 2010 and
2011. That difference was almost 14 billions.
55. A slight positive linear relationship
61. b. Both relationships seem to be linear in nature.
c. This occurred in 1998 1999 and 2001.
67. b. It appears that there is a positive linear relationship between
the attendance and the year.
73. a. The independent variable is hours and the dependent
variable is sales.
b. It appears that there is a positive linear relationship.
43. a.
b. The largest average charges occur at universityrelated
hospitals and the lowest average appears to be in
religiousaffiliated hospitals.
c. A pie chart showing how that total is divided among the
four hospital types would not be useful or appropriate.
47. The sales have trended upward over the past 12 months.
49. The line chart shows that yearend deposits have been increas
ing since 1997 but have increased more sharply since 2002 and
have leveled off between 2006 and 2007 and recently become
more volatile.
University
Related
Religious
Affiliated
Municipally
Owned
Privately
Held
6398 3591 4613 5191
Continuous Data Data whose possible values are uncountable
and that may assume any value in an interval.
Cumulative Frequency Distribution A summary of a set of
data that displays the number of observations with values
less than or equal to the upper limit of each of its classes.
Cumulative Relative Frequency Distribution A summary
of a set of data that displays the proportion of observations
with values less than or equal to the upper limit of each of its
classes.
Dependent Variable A variable whose values are thought to
be a function of or dependent on the values of another vari
able called the independent variable. On a scatter plot the
dependent variable is placed on the y axis and is often called
the response variable.
AllInclusive Classes A set of classes that contains all the pos
sible data values.
Bar Chart A graphical representation of a categorical data set
in which a rectangle or bar is drawn over each category or
class. The length or height of each bar represents the fre
quency or percentage of observations or some other meas
ure associated with the category. The bars may be vertical
or horizontal. The bars may all be the same color or they
may be different colors depicting different categories. Addi
tionally multiple variables can be graphed on the same bar
chart.
Class Boundaries The upper and lower values of each class.
Class Width The distance between the lowest possible value
and the highest possible value for a frequency class.
Glossary
Berenson Mark L. and David M. Levine Basic Business Sta
tistics: Concepts and Applications 12th ed. Upper Saddle
River NJ: Prentice Hall 2012.
Cleveland William S. “Graphs in Scientific Publications” The
American Statistician 38 November 1984 pp. 261–269.
Cleveland William S. and R. McGill “Graphical Perception:
Theory Experimentation and Application to the Develop
ment of Graphical Methods” Journal of the American Statis
tical Association 79 September 1984 pp. 531–554.
Cryer Jonathan D. and Robert B. Miller Statistics for Busi
ness: Data Analysis and Modeling 2nd ed. Belmont CA:
Duxbury Press 1994.
DeVeaux Richard D. Paul F. Velleman and David E. Bock
Stats Data and Models 3rd ed. New York: AddisonWesley
2012.
Microsoft Excel 2010 Redmond WA: Microsoft Corp. 2010.
Siegel Andrew F. Practical Business Statistics 5th ed. Burr
Ridge IL: Irwin 2002.
Tufte Edward R. Envisioning Information Cheshire CT:
Graphics Press 1990.
Tufte Edward R. The Visual Display of Quantitative Informa
tion 2nd ed. Cheshire CT: Graphics Press 2001.
Tukey John W. Exploratory Data Analysis Reading MA:
AddisonWesley 1977.
References
www.downloadslide.comslide 90:
Graphs Charts and Tables—Describing Your Data
Mutually Exclusive Classes Classes that do not overlap so
that a data value can be placed in only one class.
Ogive The graphical representation of the cumulative relative
frequency. A line is connected to points plotted above the
upper limit of each class at a height corresponding to the
cumulative relative frequency.
Pie Chart A graph in the shape of a circle. The circle is divided
into “slices” corresponding to the categories or classes to
be displayed. The size of each slice is proportional to the
magnitude of the displayed variable associated with each
category or class.
Relative Frequency The proportion of total observations that
are in a given category. Relative frequency is computed by
dividing the frequency in a category by the total number of
observations. The relative frequencies can be converted to
percentages by multiplying by 100.
Scatter Diagram or Scatter Plot A twodimensional graph
of plotted points in which the vertical axis represents values
of one quantitative variable and the horizontal axis repre
sents values of the other quantitative variable. Each plotted
point has coordinates whose values are obtained from the
respective variables.
Discrete Data Data that can take on a countable number of
possible values.
EqualWidth Classes The distance between the lowest pos
sible value and the highest possible value in each class is
equal for all classes.
Frequency Distribution A summary of a set of data that dis
plays the number of observations in each of the distribu
tion’s distinct categories or classes.
Frequency Histogram A graph of a frequency distribution
with the horizontal axis showing the classes the verti
cal axis showing the frequency count and for equal class
widths the rectangles having a height equal to the frequency
in each class.
Independent Variable A variable whose values are thought
to impact the values of the dependent variable. The inde
pendent variable or explanatory variable is often within the
direct control of the decision maker. On a scatter plot the
independent variable or explanatory variable is graphed on
the x axis.
Line Chart A twodimensional chart showing time on the hori
zontal axis and the variable of interest on the vertical axis.
www.downloadslide.comslide 91:
www.downloadslide.comslide 92:
1 Measures of Center and
Location
2 Measures of Variation
3 Using the Mean and
Standard Deviation Together
Outcome 3. Compute the range interquartile range
variance and standard deviation and know what these
values mean.
Outcome 4. Compute a z score and the coefficient of
variation and understand how they are applied in
decisionmaking situations.
Outcome 5. Understand the Empirical Rule and
Tchebysheff’s Theorem
Quick Prep Links
Review the definitions for nominal ordinal
interval and ratio data.
Examine the statistical software such as
Excel used during this course to identify
the tools for computing descriptive
measures. For instance in Excel look at
the function wizard and the descriptive
statistics tools on the Data tab under
Data Analysis.
Review the material on frequency
histograms paying special attention to how
histograms help determine where the data
are centered and how the data are spread
around the center.
Describing Data Using
Numerical Measures
Outcome 1. Compute the mean median mode and
weighted mean for a set of data and understand what
these values represent.
Outcome 2. Construct a box and whisker graph and
interpret it.
Why you need to know
Suppose you are the advertising manager for a major airline and you want to develop an ad campaign touting how
much cheaper your fares are than the competition’s. You must be careful that your claims are valid. First the Federal
Trade Commission FTC is charged with regulating advertising and requires that advertising be truthful. Second
customers who can show that they were misled by an incorrect claim about prices could sue your company. You
need to use statistical procedures to determine the validity of any claim you might want to make about your prices.
Graphs and charts provide effective tools for transforming data into information however they are only a starting
point. Graphs and charts do not reveal all the information contained in a set of data. To make your descriptive toolkit
complete you need to become familiar with key descriptive measures that are widely used in fully describing data.
You might start by sampling travel routes from your company and from the competition. You could then
determine the price of a roundtrip flight for each route for your airline and your competitors. You might graph the
data for each company as a histogram but a clear comparison using only this graph might be difficult. Instead
you could compute the summary flight price measures for the various airlines and show these values
side by side perhaps in a bar chart. Thus to effectively describe data you will need to combine the
graphical tools with the numerical measures introduced in this text.
1 Measures of Center and Location
Histograms are an effective way of converting quantitative data into use
ful information. The histogram provides a visual indication of where
data are centered and how much spread there is in the data around the center. Okea/Fotolia LLC
From Chapter 3 of Business Statistics A DecisionMaking Approach Ninth Edition. David F. Groebner
Patrick W. Shannon and Phillip C. Fry. Copyright © 2014 by Pearson Education Inc. All rights reserved.
www.downloadslide.comslide 93:
Describing Data Using Numerical Measures
However to fully describe a quantitative variable we also need to compute measures of its
center and spread. These measures can then be coupled with the histogram to give a clear
picture of the variable’s distribution. This section focuses on measures of the center of data.
Section 2 introduces measures of the spread of data.
Parameters and Statistics
Depending on whether we are working with a population or a sample a numerical measure is
known as either a parameter or a statistic.
Population Mean
There are three important measures of the center of a set of data. The first of these is the
mean or average of the data. To find the mean we sum the values and divide the sum by the
number of data values as shown in Equation 1.
Parameter
A measure computed from the entire population.
As long as the population does not change the
value of the parameter will not change.
Statistic
A measure computed from a sample that has
been selected from a population. The value of
the statistic will depend on which sample is
selected.
Mean
A numerical measure of the center of a set of
quantitative measures computed by dividing the
sum of the values by the number of values in
the data.
Population Mean
∑
x
N
i
i
N
1
1
where:
m Population mean 1mu2
N Population size
x
i
ith individual value of variable x
The population mean is represented by the Greek symbol m pronounced “mu.” The
formal notation in the numerator for the sum of the x values reads
xx i
i
i
N
i
1
Sum all values where goes from
∑
→ 1 1to N
In other words we are summing all N values in the population.
Because you almost always sum all the data values to simplify notation in this text we
generally will drop the subscripts after the first time we introduce a formula. Thus the for
mula for the population mean will be written as
∑ x
N
BUSINESS APPLICATION POPULATION MEAN
FOSTER CITY HOTEL The manager of a small hotel in Foster City California was asked
by the corporate vice president to analyze the Sunday night registration information for the
past eight weeks. Data on three variables were collected:
x
1
Total number of rooms rented
x
2
Total dollar revenue from the room rentals
x
3
Number of customer complaints that came from guests each Sunday
These data are shown in Table 1. They are a population because they include all data that
interest the vice president.
Figure 1 shows the frequency histogram for the number of rooms rented. If the manager
wants to describe the data further she can locate the center of the data by finding the bal
ance point for the histogram. Think of the horizontal axis as a plank and the histogram bars
as weights proportional to their area. The center of the data would be the point at which the
plank would balance. As shown in Figure 1 the balance point seems to be about 15 rooms.
Population Mean
The average for all values in the population
computed by dividing the sum of all values by
the population size.
Chapter Outcome 1.
www.downloadslide.comslide 94:
Describing Data Using Numerical Measures
Eyeing the histogram might yield a reasonable approximation of the center. However
computing a numerical measure of the center directly from the data is preferable. The most
frequently used measure of the center is the mean. The population mean for number of rooms
rented is computed using Equation 1 as follows:
∑
++ + + ++ +
x
N
22 13 10 16 23 13 11 13
8
121
8
15 125 .
Thus the average number of rooms rented on Sunday for the past eight weeks is 15.125.
This is the true balance point for the data. Take a look at Table 2 where we calculate what is
called a deviation 1x
i
m2 by subtracting the mean from each value x
i
.
TABLE 1  Foster City Hotel Data
Week Rooms Rented Revenue Complaints
1 22 1870 0
2 13 1590 2
3 10 1760 1
4 16 2345 0
5 23 4563 2
6 13 1630 1
7 11 2156 0
8 13 1756 0
TABLE 2  Centering Concept of the Mean Using Hotel Data
x x2μ 5 Deviation
22 22  15.125 6.875
13 13  15.1252.125
10 10  15.1255.125
16 16  15.125 0.875
23 23  15.125 7.875
13 13  15.1252.125
11 11  15.1254.125
13 13  15.125 2.125
Σ1xm2 0.000 d Sum of deviations equals zero.
Number of Occurrences
5 to 10 11 to 15 16 to 20 21 to 25
Approximate Balance Point
Rooms Rented
5
4
3
2
1
0
FIGURE 1 
Balance Point Rooms Rented
at Foster City Hotel
www.downloadslide.comslide 95:
Describing Data Using Numerical Measures
Note that the sum of the deviations of the data from the mean is zero. This is not a coinci
dence. For any set of data the sum of the deviations around the mean will be zero.
EXAMPLE 1 COMPUTING THE POPULATION MEAN
United Airlines As the airline industry becomes increas
ingly competitive in an effort to increase profits many airlines
are reducing flights. Therefore the supply of idled airplanes
has increased. United Airlines headquartered in Chicago has
decided to expand its fleet. Suppose United selects additional
planes from a list of 17 possible planes including such models
as the Boeing 747400 the Air Bus 300B4 and the DC 910.
At a recent meeting the chief operating officer asked a member of his staff to determine the
mean fuel consumption rate per hour of operation for the population of 17 planes.
Step 1 Collect data for the quantitative variable of interest.
The staff member was able to determine for each of the 17 planes the hourly
fuel consumption in gallons for a flight between Chicago and New York City.
These data are recorded as follows:
Airplane
Fuel Consumption
1galhr2
B747400 3529
L1011100/200 2215
DC1010 2174
A300B4 1482
A310300 1574
B767300 1503
B767200 1377
B757200 985
B727200 1249
MD80 882
B737300 732
DC950 848
B727100 806
B737100/200 1104
F100 631
DC93011 804
DC910 764
Step 2 Add the data values.
∑x 3529 2215 2174
...
764 22 659
Step 3 Divide the sum by the number of values in the population using Equation 1.
∑x
N
22 659
17
1 332 9
.
The mean number of gallons of fuel consumed per hour on these 17 planes is
1332.9.
END EXAMPLE
BUSINESS APPLICATION POPULATION MEAN
FOSTER CITY HOTEL CONTINUED In addition to collecting data on the number of
rooms rented on Sunday nights the Foster City Hotel manager also collected data on the room
rental revenue generated and the number of complaints on Sunday nights. Excel can quite
easily be used to calculate numerical measures such as the mean. Because these data are the
population of all nights of interest to the hotel manager she can compute the population mean
How to do it Example 1
Computing the Population Mean
When the available data consti
tute the population of interest the
population mean is computed using
the following steps:
1. Collect the data for the variable
of interest for all items in the
population. The data must be
quantitative.
2. Sum all values in the population
1Σx
i
2.
3. Divide the sum1Σx
i
2. by the
number of values N in the
population to get the popula
tion mean. The formula for the
population mean is
∑ x
N
m
Bruce Leibowitz/ Shutterstock
www.downloadslide.comslide 96:
Describing Data Using Numerical Measures
m revenue per night. The population mean is m +2208.75 as shown in the Excel output
in Figure 2. Likewise the mean number of complaints is m 0.75 per night. Note there are
other measures shown in the figures. We will discuss several of these later in the chapter.
Now for these eight Sunday nights the manager can report to the corporate vice presi
dent that the mean number of rooms rented is 15.13. This level of business generated a mean
nightly revenue of 2208.75. The number of complaints averaged 0.75 less than 1 per night.
These values are the true means for the population and are therefore called parameters.
Sample Mean
The data for the Foster City Hotel constituted the population of interest. Thus m 15.13
rooms rented is the parameter measure. However if the data constitute a sample rather than a
population the mean for the sample sample mean is computed using Equation 2. Sample Mean
The average for all values in the sample
computed by dividing the sum of all sample
values by the sample size. Sample Mean
x
x
n
i
i
n
5
51
∑
2
where:
x Sample mean 1pronounced xbar2
n Sample size
Excel Tutorial
Excel
tutorials
Notice Equation 2 is the same as Equation 1 except that we sum the sample values not
the population values and divide by the sample size not the population size.
FIGURE 2 
Excel 2010 Output Showing
Mean Revenue for the Foster
City Hotel
Mean Rooms Rented 15.13
Mean Revenue 2208.75
Mean Complaints 0.75
Minitab Instructions for similar results:
1. Open fle: Foster.MTW.
2. Choose Stat Basic Statistics Display
Descriptive Statistics.
3. In Variables enter columns Rooms
Rented Revenue and Complaints.
4. Click Statistics.
5. Check required statistics.
6. Click OK. OK.
Excel 2010 Instructions:
1. Open fle: Foster.xlsx.
2. Select the Data tab.
3. Click on Data Analysis
Descriptive Statistics.
4. Defne data range for
the desired variables.
5. Check Summary
Statistics.
6. Name new Output Sheet.
7. On Home tab adjust
decimal places as
desired.
Chapter Outcome 1.
www.downloadslide.comslide 97:
Describing Data Using Numerical Measures
The notation for the sample mean is x. Sample descriptors statistics are usually assigned
a Roman character. Recall that population values usually are assigned a Greek character.
EXAMPLE 2 COMPUTING A SAMPLE MEAN
Professor Salaries A newspaper reporter in Wisconsin collected a sample of seven univer
sity professors and determined their annual salaries. As part of her story she wished to be able
to report the mean salary. The following steps are used to calculate the sample mean salaries
for professors in Wisconsin.
Step 1 Collect the sample data.
5x
i
6 5Professor Salaries6 5+144000 +98000 +204000
+177000 +155000 +316000 +1000006
Step 2 Add the values in the sample.
Σx +144000 + +98000 + +204000 + +177000 + +155000
+ +316000 + +100000 +1194000
Step 3 Divide the sum by the sample size Equation 2.
x
x
n
55 5
∑
.
1 194 000
7
170 571 43
Therefore the mean salary for the sample of seven professors is 170571.43.
END EXAMPLE
The Impact of Extreme Values on the Mean
The mean population or sample is the balance point for data so using the mean as a measure
of the center generally makes sense. However the mean does have a potential disadvantage: The
mean can be affected by extreme values. There are many instances in business when this may
occur. For example in a population or sample of income data there likely will be extremes on the
high end that will pull the mean upward from the center. Example 3 illustrates how an extreme
value can affect the mean. In these situations a second measure called the median may be more
appropriate.
EXAMPLE 3 IMPACT OF EXTREME VALUES
Professor Salaries Suppose the sample of professor salaries see Example 2 had been
slightly different. If the salary recorded as 316000 had actually been 1000000 must also
be the college’s football coach how would the mean be affected We can see the impact as
follows:
Step 1 Collect the sample data.
x
i
Professor Salaries 144000 98000 204000
+177000 +155000 +1000000 +1000006
extreme value
Step 2 Add the values.
Σx +144000 + 98000 + 204000 + 177000 + 155000 + 1000000 + 100000
+1878000
Step 3 Divide the sum by the number of values in the sample.
x
x
n
55 5
∑
.
1 878 000
7
268 285 71
Recall in Example 2 the sample mean was 170571.43.
www.downloadslide.comslide 98:
Describing Data Using Numerical Measures
With only one value in the sample changed the mean is now substantially higher than
before. Because the mean is affected by extreme values it may be a misleading measure of
the data’s center. In this case the mean is larger than all but one of the starting salaries.
END EXAMPLE
TRY PROBLEM 15
Median
Another measure of the center is called the median.
The median is found by first arranging data in numerical order from smallest to largest.
Data that are sorted in order are referred to as a data array.
Equation 3 is used to find the index point corresponding to the median value for a set of
data placed in numerical order from low to high.
Median
The median is a center value that divides a data
array into two halves. We use u
∼
to denote the
population median and M
d
to denote the sample
median.
Data Array
Data that have been arranged in numerical
order. Median Index
in 5
1
2
3
where:
i The index of the point in the data array corresponding to the median value
n Sample size
If i is not an integer round its value up to the next highest integer. This next highest
integer then is the position of the median in the data array.
If i is an integer the median is the average of the values in position i and position
i + 1.
For instance suppose a personnel manager has hired 10 new employees. The ages of
each of these employees sorted from low to high is listed as follows:
23 25 25 34 35 45 46 47 52 54
Using Equation 3 to find the median index we get
in 55 5
1
2
1
2
10 5
Since the index is an integer the median value will be the average of the 5th and 6th
values in the data set. Thus the median is
M
d
55
35 45
2
40
+
Consider another case in which customers at a restaurant are asked to rate the service
they received on a scale of 1 to 100. A total of 15 customers were asked to provide the ratings.
The data sorted from low to high are presented as follows:
60 68 75 77 80 80 80 85 88 90 95 95 95 95 99
Using Equation 3 we get the median index:
in 55 5
1
2
1
2
15 7 5 .
Since the index is not an integer we round 7.5 up to 8. Thus the median 1M
d
2 is the 8th
data value from either end. In this case
M
d
85
Chapter Outcome 1.
www.downloadslide.comslide 99:
Describing Data Using Numerical Measures
EXAMPLE 4 COMPUTING THE MEDIAN
Professor Salaries Consider again the example involving the news reporter in Wisconsin
and the sample salary data in Example 2. The median for these data is computed using the
following steps:
Step 1 Collect the sample data.
5x
i
6 5Professor Salaries6 5+144000 +98000 +204000 +177000
+155000 +316000 +1000006
Step 2 Sort the data from smallest to largest forming a data array.
5x
i
6 5+98000 +100000 +144000 +155000 +177000 +204000 +3160006
Step 3 Calculate the median index.
Using Equation 3 we get i
1
2
73 5 .. Rounding up the median is the
fourth value from either end of the data array.
Step 4 Find the median.
5x
i
6 5+98000 +100000 +144000 +155000 +177000 +204000 +3160006
fourth value M
d
The median salary is 155000. The notation for the sample median is M
d
.
Note if the number of data values in a sample or population is an even number the median
is the average of the two middle values.
END EXAMPLE
TRY PROBLEM 2
Skewed and Symmetric Distributions
Data in a population or sample can be either symmetric or skewed depending on how the
data are distributed around the center.
In the original professor salary example Example 2 the mean for the sample of seven
managers was 170571.43. In Example 4 the median salary was 155000. Thus for these
data the mean and the median are not equal. This sample data set is right skewed because
xM
d
55 . . 170 571 43 155 000
Figure 3 illustrates examples of rightskewed leftskewed and symmetric distributions.
The greater the difference between the mean and the median the more skewed the distribu
tion. Example 5 shows that an advantage of the median over the mean is that the median is
not affected by extreme values. Thus the median is particularly useful as a measure of the
center when the data are highly skewed.
1
Symmetric Data
Data sets whose values are evenly spread
around the center. For symmetric data the mean
and median are equal.
Skewed Data
Data sets that are not symmetric. For skewed
data the mean will be larger or smaller than the
median.
RightSkewed Data
A data distribution is right skewed if the mean
for the data is larger than the median.
LeftSkewed Data
A data distribution is left skewed if the mean for
the data is smaller than the median.
1
Excel can be used to calculate a skewness statistic. The sign on the skewness statistic implies the direction of
skewness. The higher the absolute value the more the data are skewed.
FIGURE 3 
Skewed and Symmetric
Distributions
Frequency
x
a RightSkewed
Median Mean
Frequency
x
b LeftSkewed
Mean Median
Frequency
x
c Symmetric
Mean Median
www.downloadslide.comslide 100:
Describing Data Using Numerical Measures
EXAMPLE 5 IMPACT OF EXTREME VALUES ON THE MEDIAN
Professor Salaries Continued In Example 3 when we substituted a 1000000 sal
ary for the professor who had a salary of 316000 the sample mean salary increased from
170571.43 to 268285.71. What will happen to the median The median is determined
using the following steps:
Step 1 Collect the sample data.
The sample salary data including the extremely high salary are
5x
i
6 5Professor Salary6 5+144000 +98000 +204000 +177000
+155000 +1000000 +1000006
Step 2 Sort the data from smallest to largest forming a data array.
5x
i
6 5+98000 +100000 +144000 +155000 +177000 +204000 +10000006
Step 3 Calculate the median index.
Using Equation 3 we get i55
1
2
73 5 .. Rounding up the median is the
fourth value from either end of the data array.
Step 4 Find the median.
5x
i
6 5+98000 +100000 +144000 +155000 +177000 +204000 +10000006
fourth value M
d
The median professor salary is 155000 the same value as in Example 4 when the high sal
ary was not included in the data. Thus the median is not affected by the extreme values in the
data.
END EXAMPLE
TRY PROBLEM 2
Mode
The mean is the most commonly used measure of central location followed closely by the
median. However the mode is another measure that is occasionally used as a measure of
central location.
A data set may have more than one mode if two or more values tie for the most frequently
occurring value. Example 6 illustrates this concept and shows how the mode is determined.
EXAMPLE 6 DETERMINING THE MODE
Smoky Mountain Pizza The owners of Smoky Moun
tain Pizza are planning to expand their restaurant to include
an openair patio. Before finalizing the design the managers
want to know what the most frequently occurring group size
is so they can organize the seating arrangements to best meet
demand. They wish to know the mode which can be calcu
lated using the following steps:
Step 1 Collect the sample data.
A sample of 20 groups was selected at random. These data are
5x
i
6 5people6 52 4 1 2 3 2 4 2 3 6 8 4 2 1 7 4 2 4 4 36
Step 2 Organize the data into a frequency distribution.
x
i
Frequency
1 2
2 6
3 3
4 6
Mode
The mode is the value in a data set that occurs
most frequently.
Chapter Outcome 1.
Beth V an Trees/Shutterstock
www.downloadslide.comslide 101:
Describing Data Using Numerical Measures
x
i
Frequency
5 0
6 1
7 1
8 1
Total 20
Step 3 Determine the values that occurs occur most frequently.
In this case there are two modes because the values 2 and 4 each occurred six
times. Thus the modes are 2 and 4.
END EXAMPLE
TRY PROBLEM 2
A common mistake is to state the mode as being the frequency of the most frequently
occurring value. In Example 6 you might be tempted to say that the mode 6 because that
was the highest frequency. Instead there were two modes 2 and 4 each of which occurred
six times.
If no value occurs more frequently than any other the data set is said to not have a mode.
The mode might be particularly useful in describing the central location value for clothes
sizes. For example shoes come in full and half sizes. Consider the following sample data that
have been sorted from low to high:
5x6 57.5 8.0 8.5 9.0 9.0 10.0 10.0 10.0 10.5 10.5 11.0 11.56
The mean for these sample data is
x
x
n
55 5 5
∑+ ++ 75 80 115
12
115 50
12
963
.. . .
.
Although 9.63 is the numerical average the mode is 10 because more people wore that size
shoe than any other. In making purchasing decisions a shoe store manager would order more
shoes at the modal size than at any other size. The mean isn’t of any particular use in her pur
chasing decision.
Applying the Measures of Central Tendency
The cost of tuition is an important factor that most students and their families consider when
deciding where to attend college. The data file Colleges and Universities contains data for a
sample of 718 colleges and universities in the United States. The cost of outofstate tuition
is one of the variables in the data file. Suppose a guidance counselor who will be advising
students about college choices wishes to conduct a descriptive analysis for this quantitative
variable.
Figure 4 shows a frequency histogram generated using Excel. This histogram is a good
place to begin the descriptive analysis since it allows the analyst to get a good indication of
the center value the spread around the center and the general shape of the distribution of out
ofstate tuition for these colleges and universities. Given that the file contains 718 colleges
and universities using the 2
k
Ú n rule the guidance counselor used k 10 classes. The least
expensive school in the file is CUNY–Medgar Evers College in New York at 2600 and the
most expensive is Franklin and Marshall in Pennsylvania at 24940. Based on this histogram
in Figure 4 what would you conclude about the distribution of college tuition Is it skewed
right or left
The analysis can be extended by computing appropriate descriptive measures for the
outofstate tuition variable. Specifically we want to look at measures of central location.
Figure 5 shows the Excel output with descriptive measures for outofstate tuition. First
focus on the primary measures of central location: mean and median. These are
Mean +9933.38 Median +9433
These statistics provide measures of the center of the outofstate tuition variable. The mean
tuition value was 9933.38 whereas the median was 9433. Because the mean exceeds the
Excel
tutorials
Excel Tutorial
www.downloadslide.comslide 102:
Describing Data Using Numerical Measures
median we conclude that the data are right skewed—the same conclusion you should have
reached by looking at the histogram in Figure 4.
Issues with Excel In many instances data files will have “missing values.” That is the val
ues for one or more variables may not be available for some of the observations. The data may
have been lost or they were not measured when the data were collected. Many times when
you receive data like this the missing values will be coded in a special way. For example the
code NA might be used or a 99 might be entered to signify that the datum for that
observation is missing.
Statistical software packages typically have flexible procedures for dealing with missing
data. However Excel does not contain a missingvalue option. If you attempt to use certain
data analysis options in Excel such as Descriptive Statistics in the presence of nonnumeric
1NA2 data you will get an error message. When that happens you must clear the missing
values generally by deleting all rows with missing values. In some instances you can save
the good data in the row by using EditClearAll for the cell in question. However a bigger
problem exists when the missing value has been coded as an arbitrary numeric value 1992.
In this case unless you go into the data and clear these values Excel will use the 99 values
in the computations as if they are real values. The result will be incorrect calculations.
Also if a data set contains more than one mode Excel’s Descriptive Statistics tool will
only show the first mode in the list of modes and will not warn you that multiple modes
exist. For instance if you look at Figure 5 Excel has computed a mode +6550. If you
examine these data you will see that a tuition of 6550 occurred 14 times. This is the most
frequently occurring value. However had other tuition values occurred 14 times too Excel’s
Descriptive Statistics tool would not have so indicated. Excel 2010 does however have a
function MODE.MULT that will display multiple modes when they exist.
FIGURE 4 
Excel 2010 Frequency Histogram of College Tuition Prices
Minitab Instructions for similar results:
1. Open fle: Colleges and
Universities.MTW.
2. Choose Graph Histogram.
3. Click Simple.
4. Click OK.
5. In Graph variables enter
data column outofstate tuition.
6. Click OK.
Excel 2010 Instructions:
1. Open fle Colleges and
Universities.xlsx.
2. Set up an area in the
worksheet for the bins
upper limit of each class
as 4750 7000 etc. Be
sure to label the column
of upper limits as “Bins.”
3. On the Data tab click
Data Analysis
Histogram.
4. Input Range specifes the
actual data values as the
outofstate tuition.
5. Put on a new worksheet
and include Chart
Output.
6. Right click on the bars
and use the Format Data
Series Options to set gap
width to zero and add
lines to bars.
7. Convert the bins in
column A of the
histogram output sheet to
actual class labels. Note
the bin labeled 4750 is
changed to “under
4750.”
8. Click on Layout and set
titles as desired.
www.downloadslide.comslide 103:
Describing Data Using Numerical Measures
Other Measures of Location
Weighted Mean The arithmetic mean is the most frequently used measure of central loca
tion. Equations 1 and 2 are used when you have either a population or a sample. For instance
the sample mean is computed using
x
x
n
xx x x
n
n
55
∑
++ + +
12 3
In this case each x value is given an equal weight in the computation of the mean. However
in some applications there is reason to weight the data values differently. In those cases we
need to compute a weighted mean.
Equations 4 and 5 are used to find the weighted mean or weighted average for a popula
tion and for a sample respectively.
Weighted Mean
The mean value of data values that have been
weighted according to their relative importance.
Weighted Mean for a Population
w
ii
i
wx
w
∑
∑
4
FIGURE 5 
Excel 2010 Descriptive
Statistics for College Tuition
Data
Excel 2010 Instructions:
1. Open fle: Colleges and
Universities.xlsx.
2. Select the Data tab.
3. Click on Data Analysis
Descriptive Statistics.
4. Defne data range for the
desired variables.
5. Check Summary
Statistics.
6. Name new Output Sheet.
7. On Home tab adjust
decimal places as desired.
Note the Skewness statistic is a
small positive number indicating a
slight amount of positive right
skew to the tuition data. The higher
the absolute value of the Skewness
statistic the greater the skewness in
the data.
Minitab Instructions for similar results:
1. Open fle: Colleges and Universities.MTW.
2. Choose Stat Basic Statistics Display
Descriptive Statistics.
3. In Variables enter column outofstate
tuition.
4. Click Statistics.
5. Check required statistics.
6. Click OK. OK.
Mean Median Mode
Chapter Outcome 1.
www.downloadslide.comslide 104:
Describing Data Using Numerical Measures
EXAMPLE 7 CALCULATING A WEIGHTED POPULATION MEAN
Myers Associates Recently the law firm of Myers Associates was involved in litigating
a discrimination suit concerning ski instructors at a ski resort in Colorado. One ski instructor from
Germany had sued the operator of the ski resort claiming he had not received equitable pay com
pared with the other ski instructors from Norway and the United States. In preparing a defense the
Myers attorneys planned to compute the mean annual income for all seven Norwegian ski instruc
tors at the resort. However because these instructors worked different numbers of days during the
ski season a weighted mean needed to be computed. This was done using the following steps:
Step 1 Collect the desired data and determine the weight to be assigned to each
data value.
In this case the variable of interest was the income of the ski instructors. The
population consisted of seven Norwegian instructors. The weights were the
number of days that the instructors worked. The following data and weights
were determined:
x
i
Income: +7600 +3900 +5300 +4000 +7200 +2300 +5100
w
i
Days: 50 30 40 25 60 15 50
Step 2 Multiply each weight by the data value and sum these.
Σw
i
x
i
15021+76002 + 13021+39002+
c
+15021+51002 +1530500
Step 3 Sum the weights for all values the weights are the days.
Σw
i
50 + 30 + 40 + 25 + 60 + 15 + 50 270
Step 4 Compute the weighted mean.
Divide the weighted sum by the sum of the weights. Because we are
working with the population the result will be the population weighted
mean.
W
ii
i
wx
w
∑
∑
.
1 530 500
270
5 668 52
Thus taking into account the number of days worked the Norwegian ski
instructors had a mean income of 5668.52.
END EXAMPLE
TRY PROBLEM 8
One weightedmean example that you are probably very familiar with is your college
grade point average GPA. At most schools A 4 points B 3 points and so forth.
Each course has a certain number of credits usually 1 to 5. The credits are the weights.
Your GPA is computed by summing the product of points earned in a course and the
credits for the course and then dividing this sum by the total number of credits earned.
Percentiles In some applications we might wish to describe the location of the data in
terms other than the center of the data. For example prior to enrolling at your university you
took the SAT or ACT test and received a percentile score in math and verbal skills.
Percentiles
The pth percentile in a data array is a value that
divides the data set into two parts. The lower
segment contains at least p and the upper
segment contains at least 1 100  p2 of the
data. The 50th percentile is the median.
Weighted Mean for a Sample
x
wx
w
w
ii
i
5
∑
∑
5
where:
w
i
The weight of the ith data value
x
i
The ith data value
www.downloadslide.comslide 105:
Describing Data Using Numerical Measures
If you received word that your standardized exam score was at the 90th percentile it
means that you scored as high as or higher than 90 of the other students who took the exam.
The score at the 50th percentile would indicate that you were at the median where at least
50 scored at or below and at least 50 scored at or above your score.
2
To illustrate how to manually approximate a percentile value consider a situation in
which you have 309 customers enter a bank during the course of a day. The time rounded to
the nearest minute that each customer spends in the bank is recorded. If we wish to approxi
mate the 10th percentile we would begin by first sorting the data in order from low to high
then assign each data value a location index from 1 to 309 and next determine the location
index that corresponds to the 10th percentile using Equation 6.
2
More rigorously the percentile is that value or set of values such that at least p of the data is as small or smaller
than that value and at least 1100  p2 of the data is at least as large as that value. For introductory courses a
convention has been adopted to average the largest and smallest values that qualify as a certain percentile. This is
why the median was defined as it was earlier for data sets with an even number of data values.
Percentile Location Index
in 5
p
100
6
where:
p Desired percent
n Number of values in the data set
If i is not an integer round up to the next highest integer. The next integer greater than i
corresponds to the position of the pth percentile in the data set.
If i is an integer the pth percentile is the average of the values in position i and position
i + 1.
Thus the index value associated with the 10th percentile is
in 55 5
p
100
10
100
309 30 90 .
Because i 30.90 is not an integer we round to the next highest integer which is 31. The
10th percentile corresponds to the value in the 31st position from the low end of the sorted
data.
EXAMPLE 8 CALCULATING PERCENTILES
Henson Trucking The Henson Trucking Company is a
small company in the business of moving people from one
home to another within the Dallas Texas area. Historically
the owners have charged the customers on an hourly basis
regardless of the distance of the move within the Dallas city
limits. However they are now considering adding a surcharge
for moves over a certain distance. They have decided to base
this charge on the 80th percentile. They have a sample of traveldistance data for 30 moves.
These data are as follows:
13.5 8.6 16.2 21.4 21.0 23.7 4.1 13.8 20.5 9.6
11.5 6.5 5.8 10.1 11.1 4.4 12.2 13.0 15.7 13.2
13.4 13.1 21.7 14.6 14.1 12.4 24.9 19.3 26.9 11.7
The 80th percentile can be computed using these steps.
Step 1 Sort the data from lowest to highest
4.1 4.4 5.8 6.5 8.6 9.6 10.1 11.1 11.5 11.7
12.2 12.4 13.0 13.1 13.2 13.4 13.5 13.8 14.1 14.6
15.7 16.2 19.3 20.5 21.0 21.4 21.7 23.7 24.9 26.9
How to do it Example 8
Calculating Percentiles
To calculate a specific percentile
for a set of quantitative data you
can use the following steps:
1. Sort the data in order from the
lowest to highest value.
2. Determine the percentile loca
tion index i using Equation 6.
in
p
100
5
where
p Desired percent
n Number of values in the
data set
3. If i is not an integer then round
to next highest integer. The
pth percentile is located at the
rounded index position. If i is an
integer the pth percentile is the
average of the values at location
index positions i and i + 1.
Gilles Lougassi/Shutterstock
www.downloadslide.comslide 106:
Describing Data Using Numerical Measures
Step 2 Determine percentile location index i using Equation 6.
The 80th percentile location index is
in 55 5
p
100
80
100
30 24
Step 3 Locate the appropriate percentile.
Because i 24 is an integer value the 80th percentile is found by averaging
the values in the 24th and 25th positions. These are 20.5 and 21.0. Thus
the 80th percentile is 120.5 + 21.022 20.75 therefore any distance
exceeding 20.75 miles will be subject to a surcharge.
END EXAMPLE
TRY PROBLEM 7
Quartiles Another location measure that can be used to describe data is quartiles.
The first quartile corresponds to the 25th percentile. That is it is the value at or below
which there is at least 25 one quarter of the data and at or above which there is at least
75 of the data. The third quartile is also the 75th percentile. It is the value at or below which
there is at least 75 of the data and at or above which there is at least 25 of the data. The
second quartile is the 50th percentile and is also the median.
A quartile value can be approximated manually using the same method as for percentiles
using Equation 6. For the 309 bank customerservice times mentioned earlier the location of
the firstquartile 25th percentile index is found after sorting the data as
in 55 5
p
100
25
100
309 77 25 .
Because 77.25 is not an integer value we round up to 78. The first quartile is the 78th value
from the low end of the sorted data.
Issues with Excel The procedure that Excel uses to compute quartiles is not standard. There
fore the quartile and percentile values from Excel will be slightly different from those we find
manually using Equation 6. For example referring to Example 8 when Excel is used to compute
the 80th percentile for the moving distances the value returned is 20.60 miles. This is slightly dif
ferent from the 20.75 we found in Example 8. Equation 6 is generally accepted by statisticians to
be correct. However Excel will give reasonably close percentile and quartile values.
Box and Whisker Plots
A descriptive tool that many decision makers like to use is called a box and whisker plot or
a box plot. The box and whisker plot incorporates the median and the quartiles to graphically
display quantitative data. It is also used to identify outliers that are unusually small or large
data values that lie mostly by themselves.
EXAMPLE 9 CONSTRUCTING A BOX AND WHISKER PLOT
Chevron Corporation A demand analyst for Chevron
Corporation has recently performed a study at one of the com
pany’s stores in which he asked customers to set their trip
odometer to zero when they filled up. Then when the custom
ers returned for their next fillup he recorded the miles that
had been driven and wishes to construct a box and whisker
plot as part of a presentation to describe the driving patterns
of Chevron customers between fillups. The sorted sample data showing the miles between
fillups is as follows:
231 236 241 242 242 243 243 243 248
248 249 250 251 251 252 252 254 255
255 256 256 257 259 260 260 260 260
262 262 264 265 265 265 266 268 268
270 276 277 277 280 286 300 324 345
Quartiles
Quartiles in a data array are those values that
divide the data set into four equalsized groups.
The median corresponds to the second quartile.
Box and Whisker Plot
A graph that is composed of two parts: a box
and the whiskers. The box has a width that
ranges from the first quartile Q
3
to the third
quartile Q
3
. A vertical line through the box is
placed at the median. Limits are located at a
value that is 1.5 times the difference between
Q
1
and Q
3
below Q
1
and above Q
3
. The whiskers
extend to the left to the lowest value within the
limits and to the right to the highest value within
the limits.
Chapter Outcome 2.
Laura Gangi Pond/Shutterstock
www.downloadslide.comslide 107:
Describing Data Using Numerical Measures
The box and whisker plot is computed using the following steps:
Step 1 Sort the data from low to high.
Step 2 Calculate the 25th percentile Q
1
the 50th percentile median and the
75th percentile Q
3
.
The location index for Q
1
is
in 55 5
p
100
25
100
45 11 25 .
Thus Q
1
will be the 12th value which is 250 miles.
The median location is
in 55 5
p
100
50
100
45 22 5 .
In the sorted data the median is the 23rd value which is 259 miles. The
thirdquartile location is
in 55 5
p
100
75
100
45 33 75 .
Thus Q
3
is the 34th data value. This is 266 miles.
Step 3 Draw the box so the ends correspond to Q
1
and Q
3
.
230
Q
1
Q
3
240 250 260 270 280 290 300 310 320 330 340 350
Step 4 Draw a vertical line through the box at the median.
230
Q
1
Q
3
240 250 260 270
Median
280 290 300 310 320 330 340 350
Step 5 Compute the upper and lower limits.
The lower limit is computed as Q
1
 1.51Q
3
 Q
1
2. This is
Lower Limit 250  1.51266  2502 226
The upper limit is Q
3
+ 1.51Q
3
 Q
1
2. This is
Upper Limit 266 + 1.51266  2502 290
Any value outside these limits is identified as an outlier.
How to do it Example 9
Constructing a Box and
Whisker Plot
A box and whisker plot is a graphi
cal summary of a quantitative
variable. It is constructed using
the following steps:
1. Sort the data values from low to
high.
2. Use Equation 6 to find
the 25th percentile
1Q
1
first quartile2 the 50th
percentile 1Q
2
median2
and the 75th percentile
1Q
3
third quartile2.
3. Draw a box so that the ends of
the box are at Q
1
and Q
3
. This
box will contain the middle
50 of the data values in the
population or sample.
4. Draw a vertical line through
the box at the median. Half the
data values in the box will be on
either side of the median.
5. Calculate the interquartile
range 1IQR Q
3
 Q
1
2.
The interquartile range will
be discussed more fully in
Section 2. Compute the lower
limit for the box and whisker
plot as Q
1
 1.51Q
3
 Q
1
2.
The upper limit is
Q
3
+ 1.51Q
3
 Q
1
2. Any data
values outside these limits are
referred to as outliers.
6. Extend dashed lines called the
whiskers from each end of the
box to the lowest and highest
value within the limits.
7. Any value outside the limits
outlier found in step 5 is
marked with an asterisk .
www.downloadslide.comslide 108:
Describing Data Using Numerical Measures
Step 6 Draw the whiskers.
The whiskers are drawn to the smallest and largest values within the limits.
230
Q
1
Q
3
240 250 260 270
Median
280 290 300 310 320 330 340 350
Lower
Limit 226
Upper
Limit 290
Outliers
Step 7 Plot the outliers.
The outliers are plotted as values outside the limits.
END EXAMPLE
TRY PROBLEM 5
DataLevel Issues
You need to be very aware of the level of data you are working with before computing the
numerical measures introduced in this chapter. A common mistake is to compute means on
nominallevel data. For example a major electronics manufacturer recently surveyed a sam
ple of customers to determine whether they preferred black white or colored stereo cases.
The data were coded as follows:
1 black
2 white
3 colored
A few of the responses are
Color code 51 1 3 2 1 2 2 2 3 1 1 1 3 2 2 1 26
Using these codes the sample mean is
x
x
n
5
55
∑
30
17
1 765 .
As you can see reporting that customers prefer a color somewhere between black and
white but closer to white would be meaningless. The mean should not be used with nominal
data. This type of mistake tends to happen when people use computer software to perform
their calculations. Asking Excel or other statistical software to compute the mean median
and so on for all the variables in the data set is very easy. Then a table is created and before
long the meaningless measures creep into your report. Don’t let that happen.
There is also some disagreement about whether means should be computed on ordinal data.
For example in market research a 5 or 7point scale is often used to measure customers’ atti
tudes about products or TV commercials. For example we might set up the following scale:
1 Strongly agree
2 Agree
3 Neutral
4 Disagree
5 Strongly disagree
Customer responses to a particular question are obtained on this scale from 1 to 5. For a sam
ple of n 10 people we might get the following responses to a question:
Response 52 2 1 3 3 1 5 2 1 36
www.downloadslide.comslide 109:
Describing Data Using Numerical Measures
The mean rating is 2.3. We could then compute the mean for a second issue and compare the
means. However what exactly do we have First when we compute a mean for a scaled vari
able we are making two basic assumptions:
1. We are assuming the distance between a rating of 1 and 2 is the same as the distance
between 2 and 3. We are also saying these distances are exactly the same for the second
issue’s variable to which you wish to compare it. Although from a numerical standpoint
this is true in terms of what the scale is measuring is the difference between strongly
agree and agree the same as the difference between agree and neutral If not is the
mean really a meaningful measure
2. We are also assuming people who respond to the survey have the same definition of what
“strongly agree” means or what “disagree” means. When you mark a 4 disagree on your
survey are you applying the same criteria as someone else who also marks a 4 on the same
issue If not then the mean might be misleading.
Although these difficulties exist with ordinal data we see many examples in which means are
computed and used for decision purposes. In fact we once had a dean who focused on one particular
question on the course evaluation survey that was administered in every class each semester. This
question was “Considering all factors of importance to you how would you rate this instructor”
1 Excellent 2 Good 3 Average 4 Poor 5 Very poor
The dean then had his staff compute means for each class and for each professor. He then
listed classes and faculty in order based on the mean values and he based a major part of the
performance evaluation on where a faculty member stood with respect to mean score on this
one question. By the way he carried the calculations for the mean out to three decimal places
In general the median is the preferred measure of central location for ordinal data instead
of the mean.
Figure 6 summarizes the three measures of the center that have been discussed in this section.
Skill Development
31. A random sample of 15 articles in Fortune revealed the
following word counts per article:
5176 6005 5052 5310 4188
4132 5736 5381 4983 4423
5002 4573 4209 5611 4568
Compute the mean median first quartile and third
quartile for these sample data.
32. The following data reflect the number of defects
produced on an assembly line at the Dearfield
Electronics Company for the past 8 days.
30201352
51300133
43184240
a. Compute the mean number of defects for this
population of days.
31: Exercises
MyStatLab
FIGURE 6 
Descriptive Measures of the
Center
Descriptive
Measure
Computation
Method
Data
Level
Advantages/
Disadvantages
Mean Sum of values
divided by the
number of
values
Ratio
Interval
Numerical center of the data
Sum of deviations from the mean is zero
Sensitive to extreme values
Median Middle value
for data that
have been
sorted
Ratio
Interval
Ordinal
Not sensitive to extreme values
Computed only from the center values
t use information from all the data
Mode Values that
occur most
frequently in
the data
Ratio
Interval
Ordinal
Nominal
May not refect the center
May not exist
t have multiple modes
www.downloadslide.comslide 110:
Describing Data Using Numerical Measures
b. Compute the median number of defects produced
for this population of days.
c. Determine if there is a mode number of defects and
if so indicate the mode value.
33. A European cereal maker recently sampled 20 of its
mediumsize oat cereal packages to determine the
weights of the cereal in each package. These sample
data measured in ounces are as follows:
14.7 16.3 14.3 14.2 18.7 13.2 13.1 14.4 16.2 12.8
13.6 17.1 14.4 11.5 15.5 15.9 13.8 14.2 15.1 13.5
Calculate the first and third quartiles for these sample
data.
34. The time in seconds that it took for each of 16
vehicles to exit a parking lot in downtown Cincinnati is
106 153 169 116
135 78 51 129
100 141 72 101
130 125 128 139
Compute the mean median first quartile and third
quartile for the sample data.
35. A random sample of the miles driven by 20 rental car
customers is shown as follows:
90 85 100 150
125 75 50 100
75 60 35 90
100 125 75 85
50 100 50 80
Develop a box and whisker plot for the sample data.
36. Examine the following data:
23 65 45 19 35 28 39 100 50 26 25 27
24 17 12 106 23 19 39 70 20 18 44 31
a. Compute the quartiles.
b. Calculate the 90th percentile.
c. Develop a box and whisker plot.
d. Calculate the 20th and the 30th percentiles.
37. Consider the following data that represent the commute
distances for students who attend Emory University:
3.1 4.7 8.4 11.6 12.1 13.0 13.4 16.1 17.3 20.8
22.8 24.3 26.2 26.6 26.7 31.2 32.2 35.8 35.8 39.8
a. Determine the 80th percentile.
b. Determine numbers that are the 25th and 75th
percentiles.
c. Determine a number that qualifies as a median for
these data.
38. A professor wishes to develop a numerical method
for giving grades. He intends to base the grade on
homework two midterms a project and a final
examination. He wishes the final exam to have the
largest influence on the grade. He wants the project
to have 10 each midterm to have 20 and the
homework to have 10 of the influence of the
semester grade.
a. Determine the weights the professor should use to
produce a weighted average for grading purposes.
b. For a student with the following grades during the
quarter calculate a weighted average for the course:
Instrument Final Project Midterm 1 Midterm 2 Homework
Percentage
Grade
64 98 67 63 89
c. Calculate an unweighted average of these five
scores and discuss why the weighted average would
be preferable here.
Business Applications
39. The manager for the Jiffy Lube in Saratoga Florida
has collected data on the number of customers who
agreed to purchase an air filter when they were also
having their oil changed. The sample data are shown as
follows:
21 19 21 19 19 20 18 12 20 19 17 14
21 22 25 21 22 23 10 19 25 14 17 18
a. Compute the mean median and mode for these
data.
b. Indicate whether the data are skewed or
symmetrical.
c. Construct a box and whisker plot for these data.
Referring to your answer in part b does the box plot
support your conclusion about skewness Discuss.
310. During the past few years there has been a lot of
discussion about the price of university textbooks. The
complaints have come from many places including
students faculty parents and even government
officials. The publishing companies have been called
on to explain why textbooks cost so much. Recently
one of the major publishing companies was asked to
testify before a congressional panel in Washington
D.C. As part of the presentation the president of the
company organized his talk around four main areas:
production costs author royalties marketing costs
and bookstore markup. He used one of his company’s
business statistics texts as an example when he pointed
out the production costs—including editing proofing
printing binding inventory holding and distribution—
come to about 32 per book sold. Authors receive 12
per copy for the hundreds of hours of creative work
in writing the book and supplementary materials.
Marketing costs are pegged at about 5 per copy sold
and go to pay for the book sales force and examination
copies sent to professors. The book is then sold to
bookstores for 70 per copy a markup on costs of
about 40 to cover overhead and the publishing costs
associated with many upperdivision lowmarket texts
that lose money for the company. Once university
bookstores purchase the book they mark it up place
www.downloadslide.comslide 111:
Describing Data Using Numerical Measures
Speed of Service time in seconds
83 138 145 147
130 79 156 156
90 85 68 93
178 76 73 119
92 146 88 103
116 134 162 71
181 110 105 74
a. Compute the mean median and mode for these
sample data.
b. Indicate whether the data are symmetrical or
skewed.
c. Construct a box and whisker plot for the sample
data. Does the box and whisker plot support your
conclusions in part b concerning the symmetry or
skewness of these data
314. Todd Lindsey Associates a commercial real estate
company located in Boston owns six office buildings
in the Boston area that it leases to businesses. The lease
price per square foot differs by building due to location
and building amenities. Currently all six buildings are
fully leased at the prices shown here.
Price per
Square Foot
Number of
Square Feet
Building 1 75 125000
Building 2 85 37500
Building 3 90 77500
Building 4 45 35000
Building 5 55 60000
Building 6 110 130000
a. Compute the weighted average mean price per
square foot for these buildings.
b. Why is the weighted average price per square foot
preferred to a simple average price per square foot
in this case
315. Business Week recently reported that L. G. Philips
LCD Co. would complete a new factory in Paju
South Korea. It will be the world’s largest maker
of liquidcrystal display panels. The arrival of the
plant means that flatpanel LCD televisions would
become increasingly affordable. The average retail
cost of a 20″ LCD television in 2000 was 5139.
To obtain what the average retail cost of a 37″ LCD
was in 2008 a survey yielded the following data in
U.S.:
606.70 558.12 625.82 533.70 464.37
511.15 400.56 538.20 531.64 632.14
474.86 567.46 588.39 528.78 610.32
564.71 912.68 475.87 545.25 589.15
a. Calculate the mean cost for these data.
b. Examine the data presented. Choose an appropriate
measure of the center of the data justify the choice
and calculate the measure.
it on the shelf and sell it to the student. If books go
unsold they are returned to the publisher for a full
refund. The following data reflect the dollar markup on
the business statistics text for a sample of 20 college
bookstores:
33 32 42 31 31
37 37 34 47 31
42 29 36 32 25
29 47 26 32 40
a. Compute the mean markup on the business statistics
text by university bookstores in the sample.
b. Compute the median markup.
c. Determine the mode markup.
d. Write a short paragraph discussing the statistics
computed in parts a–c.
311. The Xang Corporation operates five clothing
suppliers in China to provide merchandise for Nike.
Nike recently sought information from the five
plants. One variable for which data were collected
was the total money in U.S. dollars the company
spent on medical support for its employees in the
first three months of the year. Data on number of
employees at the plants are also shown. These data
are as follows:
Medical 7400 14400 12300 6200 3100
Employees 123 402 256 109 67
a. Compute the weighted mean medical payments for
these five plants using number of employees as the
weights.
b. Explain why Nike would desire that a weighted
average be computed in this situation rather than a
simple numeric average.
312. The TruGreen Lawn Company provides yard care
services for customers throughout the Denver area.
The company owner recently tracked the time his field
employees spent at a sample of customer locations.
He was hoping to use these data to help him with his
scheduling and to establish billing rates. The following
sample data in minutes were recorded:
31 27 29 22 24 30 28 21 29 26
22 17 17 20 38 10 38 25 27 23
23 13 17 34 25 29 22 22 14 11
29 26 29 29 37 32 27 26 18 22
Describe the central tendency of these data by
computing the mean median and mode. Based on
these measures can you conclude that the distribution
of time spent at customer locations is skewed or
symmetric
313. Eastern States Bank and Trust monitors its drive
thru service times electronically to ensure that its
speed of service is meeting the company’s goals. A
sample of 28 drivethru times was recently taken and
is shown here.
www.downloadslide.comslide 112:
Describing Data Using Numerical Measures
c. The influence an observation has on a statistic
may be calculated by deleting the observation and
calculating the difference between the original
statistic and the statistic with the data point
removed. The larger the difference the more
influential the data point. Identify the data points
that have the most and the least influence in the
calculation of the sample mean.
316. The following table exhibits base salary data obtained
from a survey of over 170 benchmark positions
including finance positions. It reports the salaries of a
sample of 25 chief finance officers for midsized firms.
Assume the data are in thousands of dollars.
173.1 171.2 141.9 112.6 211.1 156.5 145.4 134.0 192.0
185.8 168.3 131.0 214.4 155.2 164.9 123.9 161.9 162.7
178.8 161.3 182.0 165.8 213.1 177.4 159.3
a. Calculate the mean salary of the CFOs.
b. Based on measures of the center of the data
determine if the CFO salary data are skewed.
c. Construct a box and whisker plot and summarize the
characteristics of the CFO salaries that it reveals.
317. The Federal Deposit Insurance Corporation FDIC
insures deposits in banks and thrift institutions for up
to 250000. Before the banking crisis of late 2008
there were 8885 FDIC–insured institutions with
deposits of 6826804000000 there were 7436 in
late 2011 with deposits of 7966700000000.
a. Calculate the average deposits per bank for FDIC–
insured institutions during both time periods.
b. Describe the relationship between the two averages
calculated in part a. Can you provide a reason for
the difference
c. Would the two averages be considered to be
parameters or statistics Explain.
Computer Database Exercises
318. Each year Business Week publishes information and
rankings of master of business administration MBA
programs. The data file MBA Analysis contains
data on several variables for eight reputable MBA
programs. The variables include pre– and post–MBA
salary percentage salary increase undergraduate
GPA average Graduate Management Admission Test
GMAT score annual tuition and expected annual
student cost. Compute the mean and median for each
of the variables in the database and write a short report
that summarizes the data. Include any appropriate
charts or graphs to assist in your report.
319. Dynamic randomaccess memory DRAM memory
chips are made from silicon wafers in manufacturing
facilities through a very complex process called wafer
fabs. The wafers are routed through the fab machines
in an order that is referred to as a recipe. The wafers
may go through the same machine several times as the
chip is created. The data file DRAM Chips contains
a sample of processing times measured in fractions
of hours at a particular machine center for one chip
recipe.
a. Compute the mean processing time.
b. Compute the median processing time.
c. Determine what the mode processing time is.
d. Calculate the 80th percentile for processing time.
320. Japolli Bakery tracks sales of its different bread
products on a daily basis. The data for 22 consecutive
days at one of its retail outlets in Nashville are in a file
called Japolli Bakery. Calculate the mean mode and
median sales for each of the bread categories and write
a short report that describes these data. Use any charts
or graphs that may be helpful in more fully describing
the data.
321. Before the subprime loan crisis and the end of the
“housing bubble” in 2008 the value of houses was
escalating rapidly as much as 40 a year in some
areas. In an effort to track housing prices the National
Association of Realtors developed the Pending
Home Sales Index PHSI a new leading indicator
for the housing market. An index of 100 is equal to
the average level of contract activity during 2001
the first year to be analyzed. The index is based on a
large national sample representing about 20 of home
sales. The file titled Pending contains the PHSI from
December 2010 to December 2011.
a. Determine the mean and median for the PHSI from
December 2010 through December 2011. Specify
the shape of the PHSI’s distribution.
b. The PHSI was at 91.5 in December 2010 and it
was at 100.1 in November of 2011. Determine
the average monthly increase in the PHSI for this
period.
c. Using your answer to part b suggest a weighting
scheme to calculate the weighted mean for the
months between December 2010 and November
2011. Use the scheme to produce the weighted
average of the PHSI in this time period.
d. Does the weighted average seem more appropriate
here Explain.
322. Homeowners and businesses pay taxes on the assessed
value of their property. As a result property taxes can
be a problem for elderly homeowners who are on a
fixed retirement income. Whereas these retirement
incomes remain basically constant because of rising
real estate prices the property taxes in many areas of
the country have risen dramatically. In some cases
homeowners are required to sell their homes because
they can’t afford the taxes. In Phoenix Arizona
government officials are considering giving certain
elderly homeowners a property tax reduction based
on income. One proposal calls for all homeowners
over the age of 65 with incomes at or below the 20th
percentile to get a reduction in property taxes. A
random sample of 50 people over the age of 65 was
selected and the household income as reported on the
most current federal tax return was recorded. These
data are also in the file called Property Tax Incomes.
www.downloadslide.comslide 113:
Describing Data Using Numerical Measures
b. Are the data skewed or symmetric
c. Approximately what percent of the data values are
between 2900 and 3250
324. USA Today reported a survey made by Nationwide
Mutual Insurance that indicated the average amount of
time spent to resolve identity theft cases was slightly
more than 88 hours. The file titled Theft contains data
that would produce this statistic.
a. Construct a stem and leaf display. Indicate the
shape of data displayed by the stem and leaf
display.
b. Use measures that indicate the shape of the
distribution. Do these measures give results that
agree with the shape shown in part a
c. Considering your answers to part a and b indicate
which measure you would recommend using to
indicate the center of the data.
Use these data to establish the income cutoff point to
qualify for the property tax cut.
35303 56855 7928 26006 28278
54215 38850 15733 29786 65878
46658 62874 49427 19017 46007
32367 31904 35534 66668 37986
10669 54337 8858 45263 37746
14550 8748 58075 23381 11725
45044 55807 54211 42961 62682
32939 38698 11632 66714 31869
57530 59233 14136 8824 42183
58443 34553 26805 16133 61785
323. Suppose a random sample of 137 households in Detroit
was taken as part of a study on annual household
spending for food at home. The sample data are
contained in the file Detroit Eats.
a. For the sample data compute the mean and the
median and construct a box and whisker plot.
2 Measures of Variation
BUSINESS APPLICATION MEASURING VARIATION USING THE RANGE
FLEETWOOD MOBILE HOMES Consider the situation involving two manufacturing
facilities for Fleetwood Mobile Homes. The division vice president asked the two plant
managers to record the number of mobile homes produced weekly over a fiveweek period.
The resulting sample data are shown in Table 3.
Instead of reporting these raw data the managers reported only the mean and median for
their data. The following are the computed statistics for the two plants:
Plant A Plant B
x 25 units
M
d
25 units
x 25 units
M
d
25 units
The division vice president looked at these statistics and concluded the following:
1. Average production is 25 units per week at both plants.
2. The median production is 25 units per week at both plants.
3. Because the mean and median are equal the distribution of production output at the two
plants is symmetrical.
4. Based on these statistics there is no reason to believe that the two plants are different in
terms of their production output.
However if he had taken a closer look at the raw data he would have seen there is a
very big difference between the two plants. The difference is the production variation from
week to week. Plant B is very stable producing almost the same number of units every week.
Plant A varies considerably with some highoutput weeks and some lowoutput weeks. Thus
looking at only measures of the data’s central location can be misleading. To fully describe a
set of data we need a measure of variation or spread.
There is variation in everything that is made by humans or that occurs in nature. The vari
ation may be small but it is there. Given a fine enough measuring instrument we can detect
the variation. Variation is either a natural part of a process or inherent to a product or can be
attributed to a special cause that is not considered random.
Several different measures of variation are used in business decision making. In this section
we introduce four of these measures: range interquartile range variance and standard deviation.
Variation
A set of data exhibits variation if all the data are
not the same value.
TABLE 3  Manufacturing
Output for Fleetwood
Mobile Homes
Plant A Plant B
15 units 23 units
25 units 26 units
35 units 25 units
20 units 24 units
30 units 27 units
END EXERCISES 31
www.downloadslide.comslide 114:
Describing Data Using Numerical Measures
Range
The simplest measure of variation is the range. It is both easy to compute and easy to understand.
The range is computed using Equation 7.
Range
The range is a measure of variation that is
computed by finding the difference between the
maximum and minimum values in a data set.
Range
R Maximum Value  MinimumValue
7
BUSINESS APPLICATION CALCULATING THE RANGE
FLEETWOOD MOBILE HOMES CONTINUED Table 3 showed the productionvolume
data for the two Fleetwood Mobile Home plants. The range in production for each plant is
determined using Equation 7 as follows:
Plant A Plant B
R Maximum  Minimum R Maximum  Minimum
R 35  15 R 27  23
R 20 R 4
We see Plant A has a range that is fve times as great as Plant B.
Although the range is quick and easy to compute it does have some limitations. First
because we use only the high and low values to compute the range it is very sensitive to
extreme values in the data. Second regardless of how many values are in the sample or popu
lation the range is computed from only two of these values. For these reasons it is considered
a weak measure of variation.
Interquartile Range
A measure of variation that tends to overcome the range’s susceptibility to extreme values is
called the interquartile range.
Equation 8 is used to compute the interquartile range.
Interquartile Range
The interquartile range is a measure of variation
that is determined by computing the difference
between the third and first quartiles.
Interquartile Range
Interquartile Range Third Quartile  First Quartile 8
EXAMPLE 10 COMPUTING THE INTERQUARTILE RANGE
Verizon Wireless A systems capacity manager for Verizon Wireless is interested in better
understanding Verizon customer text messaging use. To do this she has collected a random
sample of 100 customers under the age of 25 and recorded the number of text messages sent
in a oneweek period. She wishes to analyze the variation in these data by computing the
range and the interquartile range. She could use the following steps to do so:
Step 1 Sort the data into a data array from lowest to highest.
The 100 sorted values are as follows:
33 164 173 184 190 197 207 216 224 237
53 164 175 186 191 197 207 217 225 240
150 164 175 186 191 198 208 217 225 240
152 166 175 186 192 200 208 217 229 240
157 166 178 187 193 200 208 219 231 250
160 168 178 188 193 201 210 222 231 251
161 169 179 188 194 202 211 223 234 259
162 171 180 188 194 204 212 223 234 270
162 171 182 190 196 205 213 223 235 379
163 172 183 190 196 205 216 224 236 479
Chapter Outcome 3.
Chapter Outcome 3.
www.downloadslide.comslide 115:
Describing Data Using Numerical Measures
Step 2 Compute the range using Equation 7.
R Maximum value  Minimum value
R 479  33 446
Note the range is sensitive to extreme values. The small value of 33 and the
high value of 479 cause the range value to be very large.
Step 3 Compute the first and third quartiles.
Equation 6 can be used to find the location of the third quartile 75th
percentile and the first quartile 25th percentile.
For Q
3
the location i
75
100
100 75. Thus Q
3
is halfway between the
75th and 76th data values which is found as follows:
Q
3
1219 + 22222 220.50
For Q
1
the location is i
25
100
100 25. Then Q
1
is halfway between the
25th and 26th data values.
Q
1
1178 + 17822 178
Step 4 Compute the interquartile range.
The interquartile range overcomes the range’s problem of sensitivity to extreme
values. It is computed using Equation 8:
Interquartile range Q
3
 Q
1
220.50  178 42.50
Note the interquartile range would be unchanged even if the values on the high
or low end of the distribution were even more extreme than those shown in
these sample data.
END EXAMPLE
TRY PROBLEM 30
Population Variance and Standard Deviation
Although the range is easy to compute and understand and the interquartile range is designed
to overcome the range’s sensitivity to extreme values neither measure uses all the available
data in its computation. Thus both measures ignore potentially valuable information in data.
Two measures of variation that incorporate all the values in a data set are the variance
and the standard deviation.
These two measures are closely related. The standard deviation is the positive square root
of the variance. The standard deviation is in the original units dollars pounds etc. whereas
the units of measure in the variance are squared. Because dealing with original units is easier
than dealing with the square of the units we usually use the standard deviation to measure
variation in a population or sample.
BUSINESS APPLICATION CALCULATING THE VARIANCE AND STANDARD DEVIATION
FLEETWOOD MOBILE HOMES CONTINUED Recall the Fleetwood Mobile Home
application in which we compared the weekly production output for two of the company’s
plants. Table 3 showed the data which are considered a population for our purposes here.
Previously we examined the variability in the output from these two plants by com
puting the ranges. Although those results gave us some sense of how much more variable
Plant A is than Plant B we also pointed out some of the deficiencies of the range. The
variance and standard deviation offer alternatives to the range for measuring variation in
data.
Equation 9 is the formula for the population variance. Like the population mean the
population variance and standard deviation are assigned Greek symbols.
Variance
The population variance is the average of the
squared distances of the data values from the
mean.
Standard Deviation
The standard deviation is the positive square
root of the variance.
Chapter Outcome 3.
www.downloadslide.comslide 116:
Describing Data Using Numerical Measures
We begin by computing the variance for the output data from Plant A. The frst step in manu
ally calculating the variance is to fnd the mean using Equation 1.
∑ x
N
15 25 35 20 30
5
125
5
25
Next subtract the mean from each value as shown in Table 4. Notice the sum of the
deviations from the mean is 0. Recall from Section 1 that this will be true for any set of data.
The positive differences are cancelled out by the negative differences. To overcome this fact
when computing the variance we square each of the differences and then sum the squared
differences. These calculations are also shown in Table 4.
The final step in computing the population variance is to divide the sum of the squared
differences by the population size N 5.
2
2
250
5
50
∑− x
N
The population variance is 50 mobile homes squared.
Manual calculations for the population variance may be easier if you use an alternative
formula for s
2
that is the algebraic equivalent. This is shown as Equation 10.
Population Variance
2
x
N
i
i
N
2
1
∑
9
where:
m Population mean
N Population size
s
2
Population variance 1sigma squared2
TABLE 4  Computing the Population
Variance: Squaring the Deviations
x
i
1x
i
m2 1x
i
m2
2
15 15  25 10 100
25 25  25 0 0
35 35  25 10 100
20 20  255 25
30 30  25 5 25
Σ1x
i
m2 0 Σ1x
i
m2
2
250
Population Variance Shortcut
2
2
2
∑−
∑
x
x
N
N
5 s
10
Example 11 will illustrate using Equation 10 to fnd a population variance.
Because we squared the deviations to keep the positive values and negative values from
canceling the units of measure were also squared but the term mobile homes squared doesn’t
have a meaning. To get back to the original units of measure take the square root of the
www.downloadslide.comslide 117:
Describing Data Using Numerical Measures
variance. The result is the standard deviation. Equation 11 shows the formula for the popula
tion standard deviation.
Population Standard Deviation
2
2
1
x
N
i
i
N
−
∑
11
Therefore the population standard deviation of Plant A’s production output is
50
707 . mobile homes
The population standard deviation is a parameter and will not change unless the population
values change.
We could repeat this process using the data for Plant B which also had a mean output of
25 mobile homes. You should verify that the population variance is
2
2
10
5
2
∑− x
N
mobile homes squared s
m
555
The standard deviation is found by taking the square root of the variance.
2
1 414 . mobile homes
5 s
5 s
Thus Plant A has an output standard deviation that is five times larger than Plant B’s. The
fact that Plant A’s range was also five times larger than the range for Plant B is merely a
coincidence.
EXAMPLE 11 COMPUTING A POPULATION VARIANCE AND STANDARD DEVIATION
Boydson Shipping Company Boydson Shipping
Company owns and operates a fleet of tanker ships that
carry commodities between the countries of the world. In
the past six months the company has had seven contracts
that called for shipments between Vancouver Canada and
London England. For many reasons the travel time varies
between these two locations. The scheduling manager is
interested in knowing the variance and standard deviation in shipping times for these seven
shipments. To find these values he can follow these steps:
Step 1 Collect the data for the population.
The shipping times are shown as follows:
x shipping weeks
55 7 5 9 7 4 66
Step 2 Select Equation 10 to find the population variance.
2
2
2
∑−
∑
x
x
N
N
Step 3 Add the x values and square the sum.
Σx 5 + 7 + 5 + 9 + 7 + 4 + 6 43
1Σx
2
2 1432
2
1849
How to do it Example 11
Computing the Population
Variance and Standard Deviation
The population variance and stand
ard deviation are computed using
the following steps:
1. Collect quantitative data for the
variable of interest for the entire
population.
2. Use either Equation 9 or Equa
tion 10 to compute the variance.
3. If Equation 10 is used find the
sum of the xvalues 1Σx
2
2 and
then square this sum 1Σx2
2
.
4. Square each x value and sum
these squared values 1Σx
2
2.
5. Compute the variance using
2
2
2
∑−
∑
x
x
N
N
6. Compute the standard deviation
by taking the positive square
root of the variance:
2
Svitlana Kataieva/Shutterstock
www.downloadslide.comslide 118:
Describing Data Using Numerical Measures
Step 4 Square each of the x values and sum these squares.
Σx
2
5
2
+ 7
2
+ 5
2
+ 9
2
+ 7
2
+ 4
2
+ 6
2
281
Step 5 Compute the population variance.
2
2
2
281
1 849
7
7
2 4082
∑−
∑
− x
x
N
N
.
The variance is in units squared so in this example the population variance is
2.4082 weeks squared.
Step 6 Calculate the standard deviation as the square root of the variance.
2
2 4082 1 5518 . . weeks
Thus the standard deviation for the number of shipping weeks between Vancouver and
London for the seven shipments is 1.5518 weeks.
END EXAMPLE
TRY PROBLEM 27
Sample Variance and Standard Deviation
Equations 9 10 and 11 are the equations for the population variance and standard deviation. Any
time you are working with a population these are the equations that are used. However in most
instances you will be describing sample data that have been selected from the population. In
addition to using different notations for the sample variance and sample standard deviation the
equations are also slightly different. Equations 12 and 13 can be used to find the sample variance.
Note that Equation 13 is considered the shortcut formula for manual computations.
Sample Variance
s
xx
n
i
i
n
2
2
1
1
5
5
–
–
∑
12
The sample standard deviation is found by taking the square root of the sample variance as
shown in Equation 14.
Sample Variance Shortcut
s
x
x
n
n
2
2
2
1
5
∑
∑
–
–
13
where:
n Sample size
x Sample mean
s
2
Sample variance
Sample Standard Deviation
ss
xx
n
i
i
n
55
5 2
2
1
1
–
–
∑
14
www.downloadslide.comslide 119:
Describing Data Using Numerical Measures
Take note in Equations 12 13 and 14 that the denominator is n  1 sample size
minus 1. This may seem strange given that the denominator for the population vari
ance and the standard deviation is simply N the population size. The mathematical
justification for the n  1 divisor is outside the scope of this text. However the gen
eral reason for this is that we want the average sample variance to equal the population
variance. If we were to select all possible samples of size n from a given population
and for each sample we computed the sample variance using Equation 12 or Equation
13 the average of all the sample variances would equal s
2
the population variance
provided we used n  1 as the divisor. Using n instead of n  1 in the denominator
would produce an average sample variance that would be smaller than s
2
the popula
tion variance. Because we want an estimator on average to equal the population vari
ance we use n  1 in the denominator of s
2
.
EXAMPLE 12 COMPUTING A SAMPLE VARIANCE AND STANDARD DEVIATION
Balco Industries The internal auditor for Balco Industries a manufacturer of equip
ment used in the food processing business recently examined the transaction records
related to 10 of the company’s customers. For each client the auditor counted the number
of incorrectly recorded entries i.e “defects”. The ten customers’ accounts can be consid
ered to be samples of all possible Balco customers that could be analyzed. To fully ana
lyze the data the auditor can calculate the sample variance and sample standard deviation
using the following steps:
Step 1 Select the sample and record the data for the variable of interest.
Client Defects x Client Defects x
14 6 0
27 7 3
31 8 2
40 9 6
5 5 10 2
Step 2 Select either Equation 12 or Equation 13 to compute the sample variance.
If we use Equation 12
s
xx
n
2
2
1
5
∑ –
–
Step 3 Compute x.
The sample mean number of defectives is
x
x
n
55 5
∑ 30
10
30 .
Step 4 Determine the sum of the squared deviations of each x value from x.
Client Defectives x 1x  x2 1x  x2
2
14 1 1
27 4 16
31 2 4
40 3 9
55 2 4
60 3 9
73 0 0
82 1 1
www.downloadslide.comslide 120:
Describing Data Using Numerical Measures
Client Defectives x 1x  x2 1x  x2
2
96 3 9
10 2 1 1
Σ 30 Σ 0 Σ 54
Step 5 Compute the sample variance using Equation 12.
s
xx
n
2
2
1
54
9
6 555
∑−
−
The sample variance is measured in squared units. Thus the variance in this
example is 6 defectives squared.
Step 6 Compute the sample standard deviation by taking the square root of the
variance see Equation 14.
s
xx
n
s
555
5
∑−
−
.
2
1
54
9
6
2 4495 defects
This sample standard deviation measures the variation in the sample data for the number of
incorrectly recorded entries.
END EXAMPLE
TRY PROBLEM 25
BUSINESS APPLICATION CALCULATING MEASURES OF VARIATION USING EXCEL
COLLEGES AND UNIVERSITIES CONTINUED In Section 1 the guidance counselor
was interested in describing the data representing the cost of outofstate tuition for a large
number of colleges and universities in the United States. The data for 718 schools are in the
file called Colleges and Universities. Previously we determined the following descriptive
measures of the center for the variable outofstate tuition:
Mean +9933.38
Median +9433.00
Mode +6550
Next the analyst will turn her attention to measures of variability. The range maximum –
minimum is one measure of variability. Excel can be used to compute the range and the
standard deviation of tuition which is a more powerful measure of variation than the range.
Figure 7 shows the Excel descriptive statistics results. We find the following measures of vari
ation:
Range +22340.00
Standard Deviation +3920.07
These values are measures of the spread in the data. You should know that outlier values in
a data set will increase both the range and standard deviation. One guideline for identifying
outliers is the 3 standard deviation rule. That is if a value falls outside 3 standard devia
tions from the mean it is considered an outlier. Also as shown in Section 1 outliers can be
identified using box and whisker plots.
Excel Tutorial
Excel
tutorials
www.downloadslide.comslide 121:
Describing Data Using Numerical Measures
32: Exercises
FIGURE 7 
Excel 2010 Descriptive
Statistics for Colleges and
Universities Data
Excel 2010 Instructions:
1. Open fle: Colleges and
Universities.xlsx.
2. Select the Data tab.
3. Click on Data Analysis
Descriptive Statistics.
4. Defne data range for the
desired variables.
5. Check Summary
Statistics.
6. Name new Output Sheet.
7. On Home tab adjust
decimal places as desired.
Standard Deviation
Variance
Range
Minitab Instructions for similar results:
1. Open fle: Colleges and Universities.MTW.
2. Choose Stat Basic Statistics Display
Descriptive Statistics.
3. In Variables enter column outofstate
tuition.
4. Click Statistics.
5. Check required statistics.
6. Click OK. OK.
MyStatLab
a. Compute the range for these data.
b. Compute the variance and standard deviation.
c. Assuming that these data represent a sample
rather than a population compute the variance and
standard deviation. Discuss the difference between
the values computed here and in part b.
327. The following data are the population of ages of students
who have recently purchased a sports video game:
16 15 17 15 15 15
14 916151310
81820171717
18 23 7152010
14 14 12 12 24 21
a. Compute the population variance.
b. Compute the population standard deviation.
Skill Development
325. Google is noted for its generous employee benefits.
The following data reflect the number of vacation days
that a sample of employees at Google have left to take
before the end of the year:
30201352
51300133
43184240
a. Compute the range for these sample data.
b. Compute the variance for these sample data.
c. Compute the standard deviation for these sample data.
326. The following data reflect the number of times a
population of business executives flew on business
during the previous month:
469457
www.downloadslide.comslide 122:
Describing Data Using Numerical Measures
328. A county library in Minnesota reported the following
number of books checked out in 15 randomly selected
months:
5176 6005 5052 5310 4188
4132 5736 5381 4983 4423
5002 4573 4209 5611 4568
Determine the range variance and standard deviation
for the sample data.
329. The following data show the number of hours spent
watching television for 12 randomly selected freshmen
attending a liberal arts college in the Midwest:
Hours of Television Viewed Weekly
7.5 11.5 14.4 7.8
13.0 10.3 5.4 12.0
12.2 8.9 8.5 6.6
Calculate the range variance standard deviation and
interquartile range for the sample data.
330. Consider the following two separate samples:
27 27 25 12 15 10 20 37 31 35
and
1 3 216181616 4 16 118
a. Calculate the range variance standard deviation
and interquartile range for each data set.
b. Which data set is most spread out based on these
statistics
c. Now remove the largest number from each data set
and repeat the calculations called for in part a.
d. Compare the results of parts a and c. Which statistic
seems to be most affected by outliers
331. The following set of data shows the number of
alcoholic drinks that students at a Kansas university
reported they had consumed in the past month:
24 16 23 26 30 21 15 9
18 27 14 6 14 10 12
a. Assume the data set is a sample. Calculate the
range variance standard deviation and interquartile
range for the data set.
b. Assume the data set is a population. Calculate the
range variance standard deviation and interquartile
range for the data set.
c. Indicate the relationship between the statistics and
the respective parameters calculated in parts a and b.
Business Applications
332. Easy Connect Inc. provides access to computers for
business uses. The manager monitors computer use to
make sure that the number of computers is sufficient
to meet the needs of the customers. Recently the
manager collected data on a sample of customers and
tracked the time the customers started working at a
computer until they were finished. The elapsed times
in minutes are shown as follows:
40 42 18 32 43 35 11 39 36 37
8343450203931753317
Compute appropriate measures of the center and
variation to describe the time customers spend on the
computer.
333. A random sample of 20 pledges to a public radio fund
raiser revealed the following dollar pledges:
90 85 100 150
125 75 50 100
75 60 35 90
100 125 75 85
50 100 50 80
a. Compute the range variance standard deviation
and interquartile range for these sample data.
b. Briefly explain the difference between the range and
the interquartile range as a measure of dispersion.
334. Gold’s Gym selected a random sample of 10 customers
and monitored the number of times each customer
used the workout facility in a onemonth period. The
following data were collected:
10 19 17 19 12 20 20 15 16 13
Gold’s managers are considering a promotion in which
they reward frequent users with a small gift. They
have decided that they will only give gifts to those
customers whose number of visits in a onemonth
period is 1 standard deviation above the mean. Find the
minimum number of visits required to receive a gift.
335. The registrar at Whitworth College has been asked
to prepare a report about the graduate students.
Among other things she wants to analyze the ages of
the students. She has taken a sample of 10 graduate
students and has found the following ages:
32 22 24 27 27 33 28 23 24 21
a. Compute the range interquartile range and
standard deviation for these data.
b. An earlier study showed that the mean age of
graduate students in U.S. colleges and universities
is 37.8 years. Based on your calculations in part a
what might you conclude about the age of students
in Whitworth’s programs
336. The branch manager for the D. L. Evens Bank has been
asked to prepare a presentation for next week’s board
meeting. At the presentation she will discuss the status of
her branch’s loans issued for recreation vehicles RVs. In
particular she will analyze the loan balances for a sample
of 10 RV loans. The following data were collected:
11509 8088 13415 17028 16754
18626 4917 11740 16393 8757
www.downloadslide.comslide 123:
Describing Data Using Numerical Measures
a. Compute the mean loan balance.
b. Compute the loan balance standard deviation.
c. Write a oneparagraph statement that uses the
statistics computed in parts a and b to describe the
RV loan data at the branch.
337. A parking garage in Memphis monitors the time it
takes customers to exit the parking structure from the
time they get in their car until they are on the streets.
A sample of 28 exits was recently taken and is shown
here.
Garage Exit time in seconds
83 138 145 147
130 79 156 156
90 85 68 93
178 76 73 119
92 146 88 103
116 134 162 71
181 110 105 74
a. Calculate the range interquartile range
variance and standard deviation for these
sample data.
b. If the minimum time and the maximum time in
the sample data are both increased by 10 seconds
would this affect the value for the interquartile
range that you calculated in part a Why or why
not
c. Suppose the clock that electronically recorded
the times was not working properly when the
sample was taken and each of the sampled times
needs to be increased by 10 seconds. How would
adding 10 seconds to each of the sampled speed
of service times change the sample variance of
the data
338. Nielsen MonitorPlus a service of Nielsen Media
Research is one of the leaders in advertising
information services in the United States providing
advertising activity for 16 media including television
tracking in all 210 Designated Market Areas
DMAs. One of the issues it has researched is the
increasing amount of “clutter”—nonprogramming
minutes in an hour of prime time—including
network and local commercials and advertisements
for other shows. Recently it found the average
nonprogramming minutes in an hour of prime
time broadcasting for network television was 15:48
minutes. For cable television the average was 14:55
minutes.
a. Calculate the difference in the average clutter
between network and cable television.
b. Suppose the standard deviation in the amount of
clutter for both the network and cable television
was either 5 minutes or 15 seconds. Which standard
deviation would lead you to conclude that there
was a major difference in the two clutter averages
Comment.
339. The Bureau of Labor Statistics in its Monthly Labor
Review published the “overthemonth” percent change
in the price index for imports from December 2010 to
December 2011. These data are reproduced next.
Month Change in Index
Jan 0.1
Feb 0.2
Mar 1.3
Apr 0.4
May 2.5
Jun 0.7
Jul 0.7
Aug 0.5
Sep 0.3
Oct 0.3
Nov 1.2
Dec 0.9
a. Calculate the mean standard deviation and
interquartile range for these data.
b. Consider the mean calculated in part a. What does
this value indicate about the price index
c. What does the standard deviation indicate about the
price index
340. The U.S. Government Accountability Office recently
indicated the price of college textbooks has been rising
an average of 6 annually since the late 1980s. The
report estimated that the average cost of books and
supplies for firsttime fulltime students at fouryear
public universities for the academic year had reached
898. A data set that would produce this average
follows:
537.51 1032.52 1119.17 877.27 856.87 739.91 963.79
847.92 1393.81 524.68 1012.91 1176.46 944.60 708.26
1074.35 778.87 967.91 562.55 789.50 1051.65
a. Calculate the mean and standard deviation.
b. Determine the number of standard deviations the most
extreme cost is away from the mean. If you were to
advise a prospective student concerning the money
the student should save to afford the cost of books and
supplies for at least 90 of the colleges determine
the amount you would suggest. Hint: Don’t forget the
yearly inflation of the cost of books and supplies.
Computer Database Exercises
341. The manager of a phone kiosk in the Valley Mall
recently collected data on a sample of 50 customers
who purchased a cell phone and a monthly call plan.
The data she recorded are in the data file called Phone
Survey.
a. The manager is interested in describing the
difference between male and female customers with
respect to the price of the phone purchased. She
wants to compute mean and standard deviation of
phone purchase price for each group of customers.
www.downloadslide.comslide 124:
Describing Data Using Numerical Measures
and their 2005 outofpocket medical expenses for
prescription drugs were collected. The data are in the
file Drug Expenses.
a. Calculate the mean and median for the sample data.
b. Calculate the range variance standard deviation
and interquartile range for the sample data.
c. Construct a box and whisker plot for the sample
data.
d. Write a short report that describes outofpocket
drug expenses for privately insured adults whose
incomes are greater than 200 of the poverty level.
345. Executive MBA programs have become increasingly
popular. In an article titled “The Best Executive
MBAs” Business Week provided data concerning the
top 25 executive MBA programs for one specific year.
The tuition for each of the schools selected was given.
A file titled EMBA contains this data.
a. Calculate the 20th 40th 60th and 80th percentile
among the ranks.
b. Calculate the mean and standard deviation of the
tuition for the five subgroups defined by the rank
percentiles in part a. Hint: For this purpose are the
data subgroups samples or populations
c. Do the various subgroups’ descriptive statistics
echo their standing among the listed programs
Comment.
346. When PricewaterhouseCoopers Saratoga released
its Human Capital Index Report it indicated that the
average hiring cost for an American company to fill
a job vacancy was 3270. Sample data for recent job
hires is in a file titled Hired.
a. Calculate the variance and standard deviation for
the sample data.
b. Construct a box and whisker plot. Does this
plot indicate that extreme values outliers may
be inflating the measures of spread calculated
in part a
c. Suggest and calculate a measure of spread that is
not affected by outliers.
b. The manager is also interested in an analysis of
the phone purchase price based on whether the use
will be for home or business. Again she wants to
compute mean and standard deviation of phone
purchase price for each group of customers.
342. Each year Business Week publishes information
and rankings of MBA programs. The data file MBA
Analysis contains data on several variables for eight
reputable MBA programs. The variables include pre–
and post–MBA salary percentage salary increase
undergraduate GPA average GMAT score annual
tuition and expected annual student cost. Compute the
mean median range variance and standard deviation
for each of the variables in the database and write
a short report that summarizes the data using these
measures. Include any appropriate charts or graphs to
assist in your report.
343. The First City Real Estate Company lists and sells
residential real estate property in and around Yuma
Arizona. At a recent company meeting the managing
partner asked the office administrator to provide a
descriptive analysis of the asking prices of the homes
the company currently has listed. This list includes
319 homes the price data along with other home
characteristics are included in the data file called First
City Real Estate. These data constitute a population.
a. Compute the mean listing price.
b. Compute the median listing price.
c. Compute the range in listing prices.
d. Compute the standard deviation in listing prices.
e. Write a short report using the statistics computed
in parts a–d to describe the prices of the homes
currently listed by First City Real Estate.
344. Suppose there is an investigation to determine whether
the increased availability of generic drugs Internet
drug purchases and cost controls have reduced outof
pocket drug expenses. As a part of the investigation
a random sample of 196 privately insured adults with
incomes above 200 of the poverty level was taken
Chapter Outcome 4.
END EXERCISES 32
3 Using the Mean and Standard
Deviation Together
In the previous sections we introduced several important descriptive measures that are useful
for transforming data into meaningful information. Two of the most important of these meas
ures are the mean and the standard deviation. In this section we discuss several statistical
tools that combine these two.
Coefficient of Variation
The standard deviation measures the variation in a set of data. For decision makers the stan
dard deviation indicates how spread out a distribution is. For distributions having the same
mean the distribution with the largest standard deviation has the greatest relative spread.
When two or more distributions have different means the relative spread cannot be deter
mined by merely comparing standard deviations.
The coefficient of variation CV is used to measure the relative variation for distribu
tions with different means.
Coefficient of Variation
The ratio of the standard deviation to the mean
expressed as a percentage. The coefficient of
variation is used to measure variation relative to
the mean.
www.downloadslide.comslide 125:
Describing Data Using Numerical Measures
The coefficient of variation for a population is computed using Equation 15 whereas
Equation 16 is used for sample data.
Population Coefficient of Variation
CV 100 15
Sample Coefficient of Variation
CV
s
x
5 100 16
When the coefficients of variation for two or more distributions are compared the distribution
with the largest CV is said to have the greatest relative spread.
In finance the CV measures the relative risk of a stock portfolio. Assume portfolio A
has a collection of stocks that average a 12 return with a standard deviation of 3 and
portfolio B has an average return of 6 with a standard deviation of 2. We can compute the
CV values for each as follows:
CV A55
3
12
100 25
and
CV B55
2
6
100 33
Even though portfolio B has a lower standard deviation it would be considered more risky
than portfolio A because B’s CV is 33 and A’s CV is 25.
EXAMPLE 13 COMPUTING THE COEFFICIENT OF VARIATION
AgraTech Industries AgraTech Industries has recently introduced feed supplements for
both cattle and hogs that will increase the rate at which the animals gain weight. Three years
of feedlot tests indicate that steers fed the supplement will weigh an average of 125 pounds
more than those not fed the supplement. However not every steer on the supplement has the
same weight gain results vary. The standard deviation in weightgain advantage for the steers
in the threeyear study has been 10 pounds.
Similar tests with hogs indicate those fed the supplement average 40 additional pounds
compared with hogs not given the supplement. The standard deviation for the hogs was also
10 pounds. Even though the standard deviation is the same for both cattle and hogs the
mean weight gains differ. Therefore the coefficient of variation is needed to compare rela
tive variability. The coefficient of variation for each is computed using the following steps:
Step 1 Collect the sample or population data for the variable of interest.
In this case we have two samples: weight gain for cattle and weight gain for
hogs.
Step 2 Compute the mean and the standard deviation.
For the two samples in this example we get
Cattle: lb and lb
Hogs: lb and
xs
x
55
5
125 10
40 lb s 510
Step 3 Compute the coefficient of variation using Equation 15 for populations
or Equation 16 for samples.
Because the data in this example are from samples the CV is computed using
CV
s
x
5 100
www.downloadslide.comslide 126:
Describing Data Using Numerical Measures
For each data set we get
CV
CV
cattle
hogs
55
5
10
125
100 8
10
40
100
5 25
These results indicate that hogs exhibit much greater relative variability in weight gain com
pared with cattle.
END EXAMPLE
TRY PROBLEM 50
The Empirical Rule A tool that is helpful in describing data in certain circumstances is
called the Empirical Rule. For the Empirical Rule to be used the frequency distribution must
be bellshaped such as the one shown in Figure 8.
BUSINESS APPLICATION EMPIRICAL RULE
BURGER N’ BREW The standard deviation can be thought
of as a measure of distance from the mean. Consider the
Phoenix Burger n’ Brew restaurant chain which records the
number of each hamburger option it sells each day at each
location. The numbers of chili burgers sold each day for the
past 365 days are in the file called BurgerNBrew. Figure 9
shows the frequency histogram for those data. The distribution
is nearly symmetrical and is approximately bellshaped. The mean number of chili burgers
sold was 15.1 with a standard deviation of 3.1.
The Empirical Rule is a very useful statistical concept for helping us understand the data
in a bellshaped distribution. In the Burger N’ Brew example with x 15.1 and s 3.1
if we move 1 standard deviation in each direction from the mean approximately 68 of the
data should lie within the following range:
15.1 113.12
12.0 18.2
Empirical Rule
If the data distribution is bell shaped then the
interval
m 1s contains approximately 68 of
the values
m 2s contains approximately 95 of
the values
m 3s contains virtually all of the data
values
FIGURE 8 
Illustrating the Empirical
Rule for the BellShaped
Distribution
x
± 1
± 2
68
95
Excel
tutorials
Excel Tutorial
Chapter Outcome 5.
Dan Peretz/Shutterstock
www.downloadslide.comslide 127:
Describing Data Using Numerical Measures
FIGURE 9 
Excel 2010 Histogram for
Burger n’ Brew Data
Minitab Instructions for similar results:
1. Open fle: BurgerNBrew.MTW.
2. Choose Graph Histogram.
3. Click Simple.
4. Click OK.
5. In Graph variables enter
data column ChiliBurgers Sold.
6. Click OK.
Excel 2010 Instructions:
1. Open fle
BurgerNBrew.xlsx.
2. Set up an area in the
worksheet for the bins
upper limit of each class
as 7 9 etc. Be sure to
label the column of upper
limits as “Bins.”
3. On the Data tab click
Data Analysis
Histogram.
4. Input Range specifes the
actual data values.
5. Put on a new worksheet
and include Chart
Output.
6. Right click on the bars
and use the Format Data
Series Options to set gap
width to zero and add
lines to bars.
7. Convert the bins in
column A of the histogram
output sheet to actual
class labels. Note the bin
labeled 7 is changed to
“6  7”.
8. Click on Layout and set
titles as desired.
Mean 15.1
Standard Deviation 3.1
The actual number of days Burger n’ Brew sold between 12 and 18 chili burgers is 263. Thus out
of 365 days on 72 of the days Burger n’ Brew sold between 12 and 18 chili burgers. The rea
son that we didn’t get exactly 68 is that the distribution in Figure 9 is not perfectly bellshaped.
If we look at the interval 2 standard deviations from either side of the mean we would
expect approximately 95 of the data. The interval is
15.1 213.12
15.1 6.2
8.9 21.30
Counting the values between these limits we find 353 of the 365 values or 97. Again this
is close to what the Empirical Rule predicted. Finally according to the Empirical Rule we
would expect almost all of the data to fall within 3 standard deviations. The interval is
15.1 313.12
15.1 9.3
5.80 24.40
Looking at the data in Figure 9 we fnd that in fact all the data do fall within this interval.
Therefore if we know only the mean and the standard deviation for a set of data the Empiri
cal Rule gives us a tool for describing how the data are distributed if the distribution is
bellshaped.
Tchebysheff’s Theorem
The Empirical Rule applies when a distribution is bellshaped. But what about the many sit
uations in which a distribution is skewed and not bellshaped In these cases we can use
Tchebysheff’s theorem.
Tchebysheff’s Theorem
Regardless of how data are distributed at least
11  1k
2
2 of the values will fall within k
standard deviations of the mean. For example:
At least a1 
1
1
2
b 0 0 of the
values will
fall within k 1 standard deviation
of the mean.
At least a1 
1
2
2
b
3
4
75
of the values will lie within k 2 standard
deviations of the mean.
At least a1 
1
3
2
b
8
9
89 of
the values will lie within k 3 standard
deviations of the mean.
Chapter Outcome 5.
www.downloadslide.comslide 128:
Describing Data Using Numerical Measures
Tchebysheff’s theorem is conservative. It tells us nothing about the data within 1 stan
dard deviation of the mean. Tchebysheff indicates that at least 75 of the data will fall within
2 standard deviations—it could be more. If we applied Tchebysheff’s theorem to bellshaped
distributions the percentage estimates are very low. The thing to remember is that Tcheby
sheff’s theorem applies to any distribution. This gives it great flexibility.
Standardized Data Values
When you are dealing with quantitative data you will sometimes want to convert the meas
ures to a form called standardized data values. This is especially useful when we wish to
compare data from two or more distributions when the data scales for the two distributions are
substantially different.
BUSINESS APPLICATION STANDARDIZING DATA
HUMAN RESOURCES Consider a company that uses placement exams as part of its hiring
process. The company currently will accept scores from either of two tests: AIMS Hiring
and BHSScreen. The problem is that the AIMS Hiring test has an average score of 2000
and a standard deviation of 200 whereas the BHSScreen test has an average score of 80
with a standard deviation of 12. These means and standard deviations were developed from
a large number of people who have taken the two tests. How can the company compare
applicants when the average scores and measures of spread are so different for the two tests
One approach is to standardize the test scores.
Suppose the company is considering two applicants John and Mary. John took the AIMS
Hiring test and scored 2344 whereas Mary took the BHSScreen and scored 95. Their scores
can be standardized using Equation 17.
Standardized Data Values
The number of standard deviations a value is
from the mean. Standardized data values are
sometimes referred to as z scores.
Standardized Population Data
z
x −
17
where:
x Original data value
m Population mean
s Population standard deviation
z Standard score 1number of standard deviation x is from m2
Standardized Sample Data
z
xx
s
5
−
18
where:
x Original data value
x Sample mean
s Sample standard deviation
z The standard score
If you are working with sample data rather than a population Equation 18 can be used to
standardize the values.
We can standardize the test scores for John and Mary using
z
x −
For the AIMS Hiring test the mean m is 2000 and the standard deviation s equals 200.
John’s score of 2344 converts to
Chapter Outcome 4.
www.downloadslide.comslide 129:
Describing Data Using Numerical Measures
z
z
5
5
2 344 2 000
200
172
.
−
The BHSScreen’s m 80 and s 12. Mary’s score of 95 converts to
z
z
5
5
95 80
12
125
−
.
Compared to the average score on the AIMS Hiring test John’s score is 1.72 standard devia
tions higher. Mary’s score is only 1.25 standard deviations higher than the average score on the
BHSScreen test. Therefore even though the two tests used different scales standardizing the
data allows us to conclude John scored relatively better on his test than Mary did on her test.
EXAMPLE 14 CONVERTING DATA TO STANDARDIZED VALUES
SAT and ACT Exams Many colleges and universities require students to submit either SAT
or ACT scores or both. One eastern university requires both exam scores. However in assess
ing whether to admit a student the university uses whichever exam score favors the student
among all the applicants. Suppose the school receives 4000 applications for admission. To
determine which exam will be used for each student the school will standardize the exam
scores from both tests. To do this it can use the following steps:
Step 1 Collect data.
The university will collect the data for the 4000 SAT scores and the 4000
ACT scores for those students who applied for admission.
Step 2 Compute the mean and standard deviation.
Assuming that these data reflect the population of interest for the university
the population mean is computed using
SAT: ACT: 5
∑
∑ x
N
x
N
1 255 28 3 .
The standard deviation is computed using
SAT: ACT:
∑−
∑−
.
x
N
x
N
22
72 2 4
Step 3 Standardize the data.
Convert the x values to z values using
z
x
−
Suppose a particular applicant has an SAT score of 1228 and an ACT score of
27. These test scores can be converted to standardized scores.
SAT:
ACT:
z
x
z
x
−
−
−
−
1 228 1 255
72
0 375
2
.
77283
24
0 542
−
−
.
.
.
The negative z values indicate that this student is below the mean on both the
SAT and ACT exams. Because the university wishes to use the score that most
favors the student it will use the SAT score. The student is only 0.375 standard
deviations below the SAT mean compared with 0.542 standard deviations
below the ACT mean.
END EXAMPLE
TRY PROBLEM 52
How to do it Example 14
Converting Data to Standardized
Values
For a set of quantitative data each
data value can be converted to a
corresponding standardized value
by determining how many standard
deviations the value is from the
mean. Here are the steps to do this.
1. Collect the population or sam
ple values for the quantitative
variable of interest.
2. Compute the population mean
and standard deviation or the
sample mean and standard
deviation.
3. Convert the values to standard
ized zvalues using Equation 17
or Equation 18. For populations
z
x −
For samples
z
xx
s
−
www.downloadslide.comslide 130:
Describing Data Using Numerical Measures
MyStatLab
352. Given two distributions with the following
characteristics:
Distribution A Distribution B
m 45600 m 33.40
s 6333 s 4.05
If a value from distribution A is 50000 and a value
from distribution B is 40.0 convert each value to
a standardized z value and indicate which one is
relatively closer to its respective mean.
353. If a sample mean is 1000 and the sample standard
deviation is 250 determine the standardized value for
a. x 800
b. x 1200
c. x 1000
354. The following data represent random samples taken
from two different populations A and B:
A 31 10 69 25 62 61 46 74 57
B 1030 1111 1155 978 943 983 932 1067 1013
a. Compute the mean and standard deviation for the
sample data randomly selected from population A.
b. Compute the mean and standard deviation for the
sample data randomly selected from population B.
c. Which sample has the greater spread when
measured by the standard deviation
d. Compute the coefficient of variation for the
sample data selected from population A and from
population B. Which sample exhibits the greater
relative variation
355. Consider the following sample:
22 46 25 37 35 84 33 54 80 37
76 34 48 86 41 13 49 45 62 47
72 70 91 51 91 43 56 25 12 65
a. Calculate the mean and standard deviation for this
data.
b. Determine the percentage of data values
that fall in each of the following intervals:
x s x 2s x 3s.
c. Compare these with the percentages that should
be expected from a bellshaped distribution. Does
it seem plausible that these data came from a bell
shaped population Explain.
356. Consider the following population:
71 89 65 97 46 52 99 41 62 88
73 50 91 71 52 86 92 60 70 91
73 98 56 80 70 63 55 61 40 95
Skill Development
347. A population of unknown shape has a mean of 3000
and a standard deviation of 200.
a. Find the minimum proportion of observations in the
population that are in the range 2600 to 3400.
b. Determine the maximum proportion of the
observations that are above 3600.
c. What statement could you make concerning the
proportion of observations that are smaller than
2400
348. The mean time that a certain model of light bulb will
last is 400 hours with a standard deviation equal to
50 hours.
a. Calculate the standardized value for a light bulb that
lasts 500 hours.
b. Assuming that the distribution of hours that light
bulbs last is bellshaped what percentage of bulbs
could be expected to last longer than 500 hours
349. Consider the following set of sample data:
78 121 143 88 110 107 62 122 130 95 78 139 89 125
a. Compute the mean and standard deviation for these
sample data.
b. Calculate the coefficient of variation for these
sample data and interpret its meaning.
c. Using Tchebysheff’s theorem determine the range
of values that should include at least 89 of the
data. Count the number of data values that fall into
this range and comment on whether your interval
range was conservative.
350. You are given the following parameters for two
populations:
Population 1 Population 2
m 700 m 29000
s 50 s 5000
a. Compute the coefficient of variation for each
population.
b. Based on the answers to part a which population
has data values that are more variable relative to the
size of the population mean
351. Two distributions of data are being analyzed.
Distribution A has a mean of 500 and a standard
deviation equal to 100. Distribution B has a mean
of 10 and a standard deviation equal to 4.0. Based
on this information use the coefficient of variation
to determine which distribution has greater relative
variation.
33: Exercises
www.downloadslide.comslide 131:
Describing Data Using Numerical Measures
359. Lockheed Martin is a supplier for the aerospace
industry. Recently the company was considering
switching to Cirus Systems Inc. a new supplier for one
of the component parts it needs for an assembly. At issue
is the variability of the components supplied by Cirus
Systems Inc. compared to that of the existing supplier.
The existing supplier makes the desired part with a mean
diameter of 3.75 inches and a standard deviation of
0.078 inches. Unfortunately Lockheed Martin does not
have any of the exact same parts from the new supplier.
Instead the new supplier has sent a sample of 20 parts
of a different size that it claims are representative of the
type of work it can do. These sample data are shown
here and in the data file called Cirus.
Diameters in inches
18.018 17.856 18.095 17.992 18.086 17.812
17.988 17.996 18.129 18.003 18.214 18.313
17.983 18.153 17.996 17.908
17.948 18.219 18.079 17.799
Prepare a short letter to LockheedMartin indicating
which supplier you would recommend based on
relative variability.
360. A recent article in The Washington Post Weekly Edition
indicated that about 80 of the estimated 200 billion
of federal housing subsidies consists of tax breaks
mainly deductions for mortgage interest payments
and preferential treatment for profits on home sales.
Federal housing benefits average 8268 for those with
incomes between 50000 and 200000 and 365 for
those with income of 40000 to 50000. Suppose the
standard deviations of the housing benefits in these two
categories were equal to 2750 and 120 respectively.
a. Examine the two standard deviations. What do these
indicate about the range of benefits enjoyed by the
two groups
b. Repeat part a using the coefficient of variation as
the measure of relative variation.
361. Anaheim Human Resources Inc. performs employment
screening for large companies in southern California.
It usually follows a twostep process. First potential
applicants are given a test that covers basic knowledge
and intelligence. If applicants score within a certain
range they are called in for an interview. If they score
below a certain point they are sent a rejection letter.
If applicants score above a certain point they are sent
directly to the client’s human resources office without
the interview. Recently Anaheim Human Resources
began working with a new client and formulated a new
test just for this company. Thirty people were given
the test which is supposed to produce scores that are
distributed according to a bellshaped distribution. The
following data reflect the scores of those 30 people:
76 75 74 56 61 76
62 96 68 62 78 76
a. Determine the mean and variance.
b. Determine the percentage of data values
that fall in each of the following intervals:
x 2s x 3s x 4s.
c. Compare these with the percentages specified by
Tchebysheff’s theorem.
Business Applications
357. Pfizer Inc. a major U.S. pharmaceutical company
is developing a new drug aimed at reducing the pain
associated with migraine headaches. Two drugs are
currently under development. One consideration in
the evaluation of the medication is how long the pain
killing effects of the drugs last. A random sample of
12 tests for each drug revealed the following times in
minutes until the effects of the drug were neutralized.
The random samples are as follows:
Drug A 258 214 243 227 235 222 240 245 245 234 243 211
Drug B 219 283 291 277 258 273 289 260 286 265 284 266
a. Calculate the mean and standard deviation for each
of the two drugs.
b. Based on the sample means calculated in part a
which drug appears to be effective longer
c. Based on the sample standard deviations calculated
in part a which drug appears to have the greater
variability in effect time
d. Calculate the sample coefficient of variation for the
two drugs. Based on the coefficient of variation
which drug has the greater variability in its time
until the effect is neutralized
358. Wells Fargo Bank’s call center has representatives that
speak both English and Spanish. A random sample of
11 calls to Englishspeaking service representatives
and a random sample of 14 calls to Spanishspeaking
service representatives was taken and the time to
complete the calls was measured. The results in
seconds are as follows:
Time to Complete the Call in seconds
EnglishSpeaking 131 80 140 118 79 94 103 145 113 100 122
SpanishSpeaking 170 177 150 208 151 127 147 140 109 184 119
149 129 152
a. Compute the mean and standard deviation for the
time to complete calls to Englishspeaking service
representatives.
b. Compute the mean and standard deviation for the
time to complete calls to Spanishspeaking service
representatives.
c. Compute the coefficient of variation for the time to
complete calls to Englishspeaking and Spanish
speaking service representatives. Which group
has the greater relative variability in the time to
complete calls
d. Construct box and whisker plots for the time
required to complete the two types of calls and
briefly discuss.
www.downloadslide.comslide 132:
Describing Data Using Numerical Measures
Include in your article such descriptive statistics as
the mean median and standard deviation. You might
consider using percentiles the coefficient of variation
and Tchebysheff’s theorem to help describe the data.
364. Nike ONE Black is one of the golf balls Nike Inc.
produces. It must meet the specifications of the United
States Golf Association USGA. The USGA mandates
that the diameter of the ball shall not be less than 1.682
inches 42.67 mm. To verify that this specification is
met sample golf balls are taken from the production
line and measured. These data are found in the file
titled Diameter.
a. Calculate the mean and standard deviation of this
sample.
b. Examine the specification for the diameter of the
golf ball again. Does it seem that the data could
possibly be bell shaped Explain.
c. Determine the proportion of diameters in the
following intervals: x 2s x 3s x 4s.
Compare these with the percentages specified by
Tchebysheff’s theorem.
365. The Centers for Disease Control and Prevention CDC
started the Vessel Sanitation Program VSP in the
early 1970s because of several disease outbreaks on
cruise ships. The VSP was established to protect the
health of passengers and crew by minimizing the risk
of gastrointestinal illness on cruise ships. Inspections
are scored on a point system of maximum 100 and
cruise ships earn a score based on the criteria. Ships
that score an 86 or higher have a satisfactory sanitation
level. Data from a recent inspection are contained in a
file titled Cruiscore.
a. Calculate the mean standard deviation median and
interquartile range. Which of these measures would
seem most appropriate to characterize this data set
b. Produce a box and whisker plot of the data. Would
the Empirical Rule or Tchebysheff’s theorem be
appropriate for describing this data set Explain.
c. If you wished to travel only on those ships that
are at the 90th percentile or above in terms of
sanitation what would be the lowest sanitation
score you would find acceptable
366. Airfare prices were collected for a round trip from
Los Angeles LAX to Salt Lake City SLC. Airfare
prices were also collected for a round trip from Los
Angeles LAX to Barcelona Spain BCN. Airfares
were obtained for the designated and nearby airports
during high travel months. The passenger was to fly
coach class roundtrip staying seven days. The data are
contained in a file titled Airfare.
a. Calculate the mean and standard deviation for each
of the flights.
b. Calculate an appropriate measure of the relative
variability of these two flights.
c. A British friend of yours is currently in Barcelona
and wishes to fly to Los Angeles. If the flight
84 67 60 96 77 59
67 81 66 71 69 65
58 77 82 75 76 67
Anaheim Human Resources has in the past issued
a rejection letter with no interview to the lower 16
taking the test. They also send the upper 2.5 directly
to the company without an interview. Everyone else is
interviewed. Based on the data and the assumption of a
bellshaped distribution what score should be used for
the two cutoffs
362. The College Board’s Annual Survey of Colleges provides
uptodate information on tuition and other expenses
associated with attending public and private nonprofit
institutions of postsecondary education in the United
States. Each fall the College Board releases the survey
results on how much colleges and universities are
charging undergraduate students in the new academic
year. The survey indicated that the average published
tuition and fees for 2005–2006 were 8244 at public
fouryear colleges and universities and 28500 at
private nonprofit fouryear colleges and universities. The
standard deviation was approximately 4500 at public
fouryear colleges and universities and approximately
12000 for private colleges and universities.
a. Do the private nonprofit fouryear colleges and
universities have the larger relative variability
Provide statistical evidence to support your answer.
b. If the data on published tuition and fees were bell
shaped determine the largest and smallest amount
paid at the fouryear private nonprofit colleges and
universities.
c. Based on your answer to part b do you believe that
the data are bell shaped Support your answer using
statistical reasoning.
Computer Database Exercises
363. April 15 of every year is a day that most adults in the
United States can relate to—the day that federal and
state income taxes are due. Although there have been
several attempts by Congress and the Internal Revenue
Service over the past few years to simplify the income
tax process many people still have a difficult time
completing their tax returns properly. To draw attention
to this problem a West Coast newspaper has asked 50
certified public accountant CPA firms to complete the
same tax return for a hypothetical head of household.
The CPA firms have their tax experts complete the
return with the objective of determining the total
federal income tax liability. The data in the file Taxes
show the taxes owed as figured by each of the 50 CPA
firms. Theoretically they should all come up with the
same taxes owed.
Based on these data write a short article for the
paper that describes the results of this experiment.
www.downloadslide.comslide 133:
Describing Data Using Numerical Measures
the world’s countries. The data file Country Growth
contains the most recent United Nations data on the
population and the growth rate for the last decade for
231 countries throughout the world. Based on these
data which countries had growth rates that exceeded 2
standard deviations higher than the mean growth rate
Which countries had growth rates more than 2 standard
deviations below the mean growth rate
fares are the same but priced in English pounds
determine his mean standard deviation and
measure of relative dispersion for that data. Note:
+1 0.566 GBP.
367. Doing business internationally is no longer something
reserved for the largest companies. In fact medium
size and in some cases even small companies are
finding themselves with the opportunity to do business
internationally. One factor that will be important for
world trade is the growth rate of the population of
END EXERCISES 33
www.downloadslide.comslide 134:
Describing Data Using Numerical Measures
Visual Summary
1 Measures of Center and Location
2 Measures of Variation
Outcome 3. Compute the range variance and standard deviation and know what these values mean.
3 Using the Mean and Standard Deviation Together
Outcome 4. Compute a z score and the coeffcient of variation and understand
how they are applied in decisionmaking situations.
Outcome 5. Understand the Empirical Rule and Tchebysheff’s Theorem
Summary
The real power of statistical measures of the center and variation come when they are used
together to fully describe the data. One particular measure that is used a great deal in
business especially in fnancial analysis is the coeffcient of variation . When comparing
two or more data sets the larger the coeffcient of variation the greater the relative variation
of the data. Another very important way in which the mean and standard deviation are
used together is evident in the empirical rule which allows decision makers to better
understand the data from a bellshaped distribution. In cases where the data are not
bellshaped the data can be described using Tchebysheff’s Theorem. The fnal way
discussed in this chapter in which the mean and standard deviation are used together is the
zvalue. Zvalues for each individual data point measure the number of standard deviations
a data value is from the mean.
Conclusion
A very important part of the descriptive tools in
statistics is the collection of numerical measures that
can be computed. When these measures of the
center and variation in the data are combined with
charts and graphs you can fully describe the data.
Figure 10 presents a summary of the key numerical
measures that are discussed in this chapter.
Remember measures computed from a population
are called parameters while measures computed from
a sample are called statistics.
To fully describe your data not only do you need to use the graphs
charts and tables you need to provide the measures of the center and
measures of variations in the data that are presented in this chapter.
Together the numeric measures and the graphs and charts can paint a
complete picture of the data that transform it from just data to useful
information for decisionmaking purposes.
Summary
The three numerical measures of the center for a set of data are the mean median and the mode. The mean is the
arithmetic average and is the most frequently used measure. However if the data are skewed or are ordinal
level the median is suggested. Unlike the mean which is sensitive to extreme values in the data the median is
unaffected by extremes. The mode is less frequently used as a measure of the center since it is simply the value in the
data that occurs most frequently. When one of these measures is computed from a population the measure is said
to be a parameter but if the measure is computed from sample data the measure is called a statistic. Other
measures of location that are commonly used are percentiles and quartiles. Finally many decision makers prefer
to construct a box and whisker plot which uses a box to display the range of the middle 50 percent of the data.
The limits of whiskers are calculated based on the numerical distance between the frst and third quartiles.
Outcome 1. Compute the mean median mode and weighted average for a set of data and
understand what these values represent.
Outcome 2. Construct a box and whisker graph and interpret it.
Summary
One of the major issues that business decision makers face every day is the variation that exists in their operations
processes and people. Because virtually all data exhibit variation it is important to measure it. The simplest measure of
variation is the range which is the difference between the highest value and the lowest value in the data. An alternative to the
range that ignores the extremes in the data is the interquartile range which measures the numerical distance between the
3rd and 1st quartiles. But the two most frequently used measures of variation are the variance and the standard
deviation. The equations for these two measures differ slightly depending on whether you are working with a population or
a sample. The standard deviation is measured in the same units as the variable of interest and is a measure of the average
deviation of the individual data items around the mean.
www.downloadslide.comslide 135:
Describing Data Using Numerical Measures
FIGURE 10 
Summary of Numerical
Statistical Measures
Mode
Median
Mean
Mode
Median
Mode
Percentiles/
Quartiles
Range
Interquartile
Range
Variance and
Standard
Deviation
Percentiles/
Quartiles
Box and
Whisker
Coefcient
of Variation
Standardized
zvalues
Ordinal Nominal
Ratio/Interval
Location Variation
Location Location
Descriptive
Analysis
Comparisons
Data
Level
Type of
Measures
Equations
1 Population Mean
∑
x
N
i
i
N
1
2 Sample Mean
x
x
n
i
i
n
5
51
∑
3 Median Index
in 5
1
2
4 Weighted Mean for a Population
w
ii
i
wx
w
∑
∑
5 Weighted Mean for a Sample
x
wx
w
w
ii
i
5
∑
∑
6 Percentile Location Index
in 5
p
100
7 Range
R Maximum value  Minimum value
8 Interquartile Range
Interquartile range Third quartile  First quartile
9 Population Variance
2
x
N
i
i
N
2
1
∑
10 Population Variance Shortcut
2
2
2
∑−
∑
x
x
N
N
5 s
11 Population Standard Deviation
2
2
1
x
N
i
i
N
−
∑
www.downloadslide.comslide 136:
Describing Data Using Numerical Measures
15 Population Coefficient of Variation
CV 100
16 Sample Coefficient of Variation
CV
s
x
5 100
17 Standardized Population Data
z
x −
18 Standardized Sample Data
z
xx
s
5
−
12 Sample Variance
s
xx
n
i
i
n
2
2
1
1
5
5
–
–
∑
13 Sample Variance Shortcut
s
x
x
n
n
2
2
2
1
5
∑
∑
–
–
14 Sample Standard Deviation
ss
xx
n
i
i
n
55
5 2
2
1
1
–
–
∑
Key Terms
Box and whisker plot
Coefficient of variation
Data array
Empirical Rule
Interquartile range
Leftskewed data
Mean
Median
Mode
Parameter
Percentiles
Population mean
Quartiles
Range
Rightskewed data
Sample mean
Skewed data
Standard deviation
Standardized data values
Statistic
Symmetric data
Tchebysheff’s theorem
Variance
Variation
Weighted mean
Chapter Exercises
Conceptual Questions
368. Consider the following questions concerning the sample
variance:
a. Is it possible for a variance to be negative Explain.
b. What is the smallest value a variance can be
Under what conditions does the variance equal this
smallest value
c. Under what conditions is the sample variance smaller
than the corresponding sample standard deviation
369. For a continuous variable that has a bellshaped
distribution determine the percentiles associated with the
endpoints of the intervals specified in the Empirical Rule.
370. Consider that the Empirical Rule stipulates that virtually
all of the data values are within the interval m 3s.
Use this stipulation to determine an approximation for
the standard deviation involving the range.
371. At almost every university in the United States the
university computes student grade point averages
GPAs. The following scale is typically used by
universities:
A 4 points B 3 points C 2 points
D 1 point F 0 points
Discuss what if any problems might exist when GPAs
for two students are compared. What about comparing
GPAs for students from two different universities
372. Since the standard deviation of a set of data requires
more effort to compute than the range does what
advantages does the standard deviation have when
discussing the spread in a set of data
373. The mode seems like a very simple measure of the
location of a distribution. When would the mode be
preferred over the median or the mean
Business Applications
374. Home Pros sells supplies to “doityourselfers.” One of
the things the company prides itself on is fast service.
It uses a number system and takes customers in the
order they arrive at the store. Recently the assistant
manager tracked the time customers spent in the store
from the time they took a number until they left. A
sample of 16 customers was selected and the following
data measured in minutes were recorded:
15 14 16 14 14 14 13 8
12 9 7 17 10 15 16 16
MyStatLab
www.downloadslide.comslide 137:
Describing Data Using Numerical Measures
229.00 345.00 599.00 229.00 429.00 605.00
339.00 339.00 229.00 279.00 344.00 407.00
a. Compute the mean quoted airfare.
b. Compute the variance and standard deviation in
airfares quoted. Treat the data as a sample.
378. The manager of the Cottonwood Grille recently selected
a random sample of 18 customers and kept track of
how long the customers were required to wait from
the time they arrived at the restaurant until they were
actually served dinner. This study resulted from several
complaints the manager had received from customers
saying that their wait time was unduly long and that it
appeared that the objective was to keep people waiting
in the lounge for as long as possible to increase the
lounge business. The following data were recorded
with time measured in minutes:
34 24 43 56 74 20 19 33 55
43 54 34 27 34 36 24 54 39
a. Compute the mean waiting time for this sample of
customers.
b. Compute the median waiting time for this sample of
customers.
c. Compute the variance and standard deviation of
waiting time for this sample of customers.
d. Develop a frequency distribution using six classes
each with a class width of 10. Make the lower limit
of the first class 15.
e. Develop a frequency histogram for the frequency
distribution.
f. Construct a box and whisker plot of these data.
g. The manager is considering giving a complementary
drink to customers whose waiting time is longer
than the third quartile. Determine the minimum
number of minutes a customer would have to wait
in order to receive a complementary drink.
379. Simplot AgriChemical has decided to implement
a new incentive system for the managers of its
three plants. The plan calls for a bonus to be paid
next month to the manager whose plant has the
greatest relative improvement over the average
monthly production volume. The following data
reflect the historical production volumes at the
three plants:
Plant 1 Plant 2 Plant 3
m 700 m 2300 m 1200
s 200 s 350 s 30
At the close of next month the monthly output for the
three plants was
Plant 1 810 Plant 2 2600 Plant 3 1320
a. Compute the mean median mode range
interquartile range and standard deviation.
b. Develop a box and whisker plot for these data.
375. More than 272 million computer and video games
were sold in 2010—more than two games for every
U.S. household. Gamers spend an average of 3 to 4
hours playing games online every day. The average
age of players is 28. Video games and gamers have
even created a new form of marketing—called
“advergaming.” “Advergaming is taking games—
something that people do for recreation—and inserting
a message” said Julie Roehm director of marketing
communications for the Chrysler Group which
sells Chrysler Jeep and Dodge brand vehicles. “It’s
important we go to all the places our consumers are.”
Suppose it is possible to assume the standard deviation
of the ages of video game users is 9 years and that
the distribution is bell shaped. To assist the marketing
department in obtaining demographics to increase sales
determine the proportion of players who are
a. between 19 and 28
b. between 28 and 37
c. older than 37
376. Travelers are facing increased costs for both driving
and flying to chosen destinations. With rising costs
for both modes of transportation what really weighs
on the decision to drive or to fly To gain a better
understanding of the “fly or drive” decision a recent
study compared the costs for trips between Los Angeles
and Denver 1016 miles one way. Los Angeles to
Denver roundtrip costs 218 by car and 159.00 by
plane. Cost flexibility is greater with the flying trips
because of greater airfare choices. The driving trip costs
except for the onroad lunches are pretty much set in
place. Assume the standard deviation for the cost of
flying trips is approximately 54.
a. If a flight to Denver from Los Angeles was chosen
at random determine the proportion of the time that
the cost would be smaller than 164. Assume the
flight costs are bellshaped.
b. Determine a flight cost that would qualify as the
25th percentile.
c. If nothing can be assumed about the distribution of
the flight costs determine the largest percentile that
could be attributed to an airfare of 128.
377. With the ups and downs in the economy since 2008
many discount airline fares are available if a customer
knows how to obtain the discount. Many travelers
complain that they get a different price every time they
call. The American Consumer Institute recently priced
tickets between Spokane Washington and St. Louis
Missouri. The passenger was to fly coach class round
trip staying seven days. Calls were made directly to
airlines and to travel agents with the following results.
Note that the data reflect roundtrip airfare.
www.downloadslide.comslide 138:
Describing Data Using Numerical Measures
reason for his interest in this data is that the club is
thinking of applying a discount to members who golf
more than a specified number of rounds per year. The
sample of eight people produced the following number
of rounds played:
13 32 12 9 16 17 16 12
a. Compute the mean for these sample data.
b. Compute the median for these sample data.
c. Compute the mode for these sample data.
d. Calculate the variance and standard deviation for
these sample data.
e. Note that one person in the sample played 32
rounds. What effect if any does this large value
have on each of the three measures of location
Discuss.
f. For these sample data which measure of location
provides the best measure of the center of the data
Discuss.
g. Given this sample data suppose the manager wishes
to give discounts to golfers in the top quartile. What
should the minimum number of rounds played be to
receive a discount
383. Stock investors often look to beat the performance of the
SP 500 Index which generally serves as a yardstick
for the market as a whole. The following table shows a
comparison of the fiveyear cumulative total shareholder
returns for IDACORP common stock the SP 500
Index and the Edison Electric Institute EEI Electric
Utilities Index. The data assume that 100 was invested
on December 31 2002 with beginningofperiod
weighting of the peer group indices based on market
capitalization and monthly compounding of returns
Source: IDACORP 2007 Annual Report.
Year IDACORP SP 500
EEI Electric
Utilities Inde x
2002 100.00 100.00 100.00
2003 128.86 128.67 123.48
2004 137.11 142.65 151.68
2005 136.92 149.66 176.02
2006 186.71 173.27 212.56
2007 176.26 182.78 247.76
Using the information provided construct appropriate
statistical measures that illustrate the performance
of the three investments. How well has IDACORP
performed over the time periods compared to the SP
500 How well has it performed relative to its industry
as measured by the returns of the EEI Electric Utilities
Index
384. The Zagat Survey
®
a leading provider of leisurebased
survey results released its San Francisco Restaurants
Survey involving participants who dined out an average
of 3.2 times per week. The report showed the average
Suppose the division manager has awarded the bonus
to the manager of Plant 2 since her plant increased its
production by 300 units over the mean more than that
for any of the other managers. Do you agree with the
award of the bonus for this month Explain using the
appropriate statistical measures to support your position.
380. According to the annual report issued by Wilson
Associates an investment firm in Bowling Green the
stocks in its Growth Fund have generated an average
return of 8 with a standard deviation of 2. The
stocks in the Specialized Fund have generated an
average return of 18 with a standard deviation of 6.
a. Based on the data provided which of these
funds has exhibited greater relative variability
Use the proper statistical measure to make your
determination.
b. Suppose an investor who is very risk averse is
interested in one of these two funds. Based strictly
on relative variability which fund would you
recommend Discuss.
c. Suppose the distributions for the two stock funds
had a bellshaped distribution with the means and
standard deviations previously indicated. Which
fund appears to be the best investment assuming
future returns will mimic past returns Explain.
381. The Dakota Farm Cooperative owns and leases prime
farmland in the upper Midwest. Most of its 34000
acres are planted in grain. The cooperative performs
a substantial amount of testing to determine what
seed types produce the greatest yields. Recently the
cooperative tested three types of corn seed on test plots.
The following values were observed after the first test
year:
Seed Type A Seed Type B Seed Type C
Mean Bushels/Acre 88 56 100
Standard Deviation 25 15 16
a. Based on the results of this testing which seed
seems to produce the greatest average yield per
acre Comment on the type of testing controls that
should have been used to make this study valid.
b. Suppose the company is interested in consistency.
Which seed type shows the least relative variability
c. Assuming the Empirical Rule applies describe the
production distribution for each of the three seed
types.
d. Suppose you were a farmer and had to obtain at
least 135 bushels per acre to escape bankruptcy.
Which seed type would you plant Explain your
choice.
e. Rework your answer to part d assuming the farmer
needed 115 bushels per acre instead.
382. The Hillcrest Golf and Country Club manager selected
a random sample of the members and recorded the
number of rounds of golf they played last season. The
www.downloadslide.comslide 139:
Describing Data Using Numerical Measures
at least as good mileage in city driving conditions as
the mean mileage for highway driving for all cars
387. American Express estimates current Halloween
spending to be about 53 per person. Much of the
spending was expected to come from young adults. A
file titled Halloween contains sample data on Halloween
spending.
a. Calculate the mean and standard deviation of these
data.
b. Determine the following intervals for this data set:
x 1s x 2s x 3s.
c. Suppose your responsibility as an assistant manager
was to determine the price of costumes to be sold.
The manager has informed you to set the price of
one costume so that it was beyond the budget of
only 2.5 of the customers. Assume that the data
set has a bellshaped distribution.
388. PayScale is a source of online compensation
information providing access to accurate compensation
data for both employees and employers. PayScale
allows users to obtain compensation information
providing a snapshot of the job market. Recently it
published statistics for the salaries of MBA graduates.
The file titled Payscale contains data with the same
characteristics as those obtained by PayScale for
California and Florida.
a. Calculate the standard deviations of the salaries for
both states’ MBA graduates. Which state seems
to have the widest spectrum of salaries for MBA
graduates
b. Calculate the average and median salary for each
state’s MBA graduates.
c. Examining the averages calculated in part b
determine which state’s MBA graduates have the
largest relative dispersion.
389. Yahoo Finance makes available historical stock prices.
It lists the opening high and low stock prices for each
stock available on NYSE and NASDAQ. A file titled
GEstock gives this data for General Electric GE for a
recent 99day period.
a. Calculate the difference between the opening and
closing stock prices for GE over this time period.
Then calculate the mean median and standard
deviation of these differences.
b. Indicate what the mean in part a indicates about
the relative prices of the opening and closing stock
prices for GE.
c. Compare the dispersion of the opening stock prices
with the difference between the opening and closing
stock prices.
390. Zepolle’s Bakery makes a variety of bread types
that it sells to supermarket chains in the area. One of
Zepolle’s problems is that the number of loaves of
each type of bread sold each day by the chain stores
varies considerably making it difficult to know how
price per meal falling from the previous year from
34.07 to 33.75.
a. If the standard deviation of the price of meals in San
Francisco was 10 determine the largest proportion
of meal prices that could be larger than 50.
b. If the checks were paid in Chinese currency
1 USD 8.0916 Chinese Yuan determine the
mean and standard deviation of meal prices in San
Francisco. How would this change of currency
affect your answer to part a
Computer Database Exercises
385. The data in the file named Fast100 were collected
by D. L. Green Associates a regional investment
management company that specializes in working with
clients who wish to invest in smaller companies with
high growth potential. To aid the investment firm in
locating appropriate investments for its clients Sandra
Williams an assistant client manager put together
a database on 100 fastgrowing companies. The
database consists of data on eight variables for each of
the 100 companies. Note that in some cases data are
not available. A code of –99 has been used to signify
missing data. These data will have to be omitted from
any calculations.
a. Select the variable Sales. Develop a frequency
distribution and histogram for Sales.
b. Compute the mean median and standard deviation
for the Sales variable.
c. Determine the interquartile range for the Sales
variable.
d. Construct a box and whisker plot for the Sales
variable. Identify any outliers. Discard the outliers
and recalculate the measures in part b.
e. Each year a goal is set for sales. Next year’s goal
will be to have average sales that are at this year’s
65th percentile. Identify next year’s sales goal.
386. The Environmental Protection Agency EPA tests all
new cars and provides a mileage rating for both city
and highway driving conditions. Thirty cars were tested
and are contained in the data file Automobiles. The file
contains data on several variables. In this problem focus
on the city and highway mileage data.
a. Calculate the sample mean miles per gallon mpg
for both city and highway driving for the 30 cars.
Also calculate the sample standard deviation for the
two mileage variables. Do the data tend to support
the premise that cars get better mileage on the
highway than around town Discuss.
b. Referring to part a what can the EPA conclude
about the relative variability between car models
for highway versus city driving Hint: Compute
the appropriate measure to compare relative
variability.
c. Assume that mileage ratings are approximately bell
shaped. Approximately what proportion of cars gets
www.downloadslide.comslide 140:
Describing Data Using Numerical Measures
b. For this new variable computed in part a develop a
frequency distribution.
c. For the new variable computed in part a determine
the mean median and standard deviation.
d. Determine the percentile that would correspond to
the “correct” tax figure if the IRS figure were one
of the CPA firms’ estimated tax figures. Describe
what this implies about the agreement between the
IRS and consultants’ calculated tax.
392. The Cozine Corporation operates a garbage hauling
business. Up to this point the company has been
charged a flat fee for each of the garbage trucks that
enters the county landfill. The flat fee is based on
the assumed truck weight of 45000 pounds. In two
weeks the company is required to appear before the
county commissioners to discuss a rate adjustment.
In preparation for this meeting Cozine has hired an
independent company to weigh a sample of Cozine’s
garbage trucks just prior to their entering the landfill.
The data file Cozine contains the data the company has
collected.
a. Based on the sample data what percentile does the
45000pound weight fall closest to
b. Compute appropriate measures of central location
for the data.
c. Construct a frequency histogram based on
the sample data. Use the 2
k
Ú n guideline to
determine the number of classes. Also construct
a box and whisker plot for these data. Discuss the
relative advantages of histograms and box and
whisker plots for presenting these data.
d. Use the information determined in parts a–c to
develop a presentation to the county commissioners.
Make sure the presentation attempts to answer
the question of whether Cozine deserves a rate
reduction.
many loaves to bake. A sample of daily demand data is
contained in the file Bakery.
a. Which bread type has the highest average daily
demand
b. Develop a frequency distribution for each bread
type.
c. Which bread type has the highest standard deviation
in demand
d. Which bread type has the greatest relative
variability Which type has the lowest relative
variability
e. Assuming that these sample data are representative
of demand during the year determine how many
loaves of each type of bread should be made such
that demand would be met on at least 75 of the
days during the year.
f. Create a new variable called Total Loaves Sold. On
which day of the week is the average for total loaves
sold the highest
391. The Internal Revenue Service IRS has come under a
great deal of criticism in recent years for various actions
it is purported to have taken against U.S. citizens related
to collecting federal income taxes. The IRS is also
criticized for the complexity of the tax code although
the tax laws are actually written by congressional
staff and passed by Congress. For the past few years
one of the country’s biggest taxpreparing companies
has sponsored an event in which 50 certified public
accountants from all sizes of CPA firms are asked to
determine the tax owed for a fictitious citizen. The IRS
is also asked to determine the “correct” tax owed. Last
year the “correct” figure stated by the IRS was 11560.
The file Taxes contains the data for the 50 accountants.
a. Compute a new variable that is the difference
between the IRS number and the number
determined by each accountant.
Video Case 3 video
DriveThru Service Times
McDonald’s
When you’re on the go and looking for a quick meal where do you
go If you’re like millions of people every day you make a stop at
McDonald’s. Known as “quick service restaurants” in the industry
not “fast food” companies such as McDonald’s invest heavily
to determine the most efficient and effective ways to provide fast
highquality service in all phases of their business.
Drivethru operations play a vital role. It’s not surprising that
attention is focused on the drivethru process. After all more than
60 of individual restaurant revenues in the United States come
from the drivethru experience. Yet understanding the process is
more complex than just counting cars. Marla King professor at
the company’s international training center Hamburger University
got her start 25 years ago working at a McDonald’s drivethru. She
now coaches new restaurant owners and managers. “Our stated
drivethru service time is 90 seconds or less. We train every man
ager and team member to understand that a quality customer expe
rience at the drivethru depends on them” says Marla. Some of the
factors that affect customers’ ability to complete their purchases
within 90 seconds include restaurant staffing equipment layout in
the restaurant training efficiency of the grill team and frequency
of customer arrivals to name a few. Also customer order patterns
play a role. Some customers will just order drinks whereas others
seem to need enough food to feed an entire soccer team. And then
there are the special orders. Obviously there is plenty of room for
variability here.
www.downloadslide.comslide 141:
Describing Data Using Numerical Measures
drive the decisions behind McDonald’s drivethru service opera
tions.
Discussion Questions:
1. After returning from the training session at Hamburger
University a McDonald’s store owner selected a random sample
of 362 drivethru customers and carefully measured the time it
took from when a customer entered the McDonald’s property
until the customer had received the order at the drivethru
window. These data are in the file called McDonald’s Drive
Thru Waiting Times. Note the owner selected some customers
during the breakfast period others during lunch and others
during dinner. For the overall sample compute the key measures
of central tendency. Based on these measures what conclusion
might the owner reach with respect to how well his store is
doing in meeting the 90second customer service goal Discuss.
2. Referring to question 1 compute the key measures of central
tendency for drivethru times broken down by breakfast
lunch and dinner time periods. Based on these calculations
does it appear that the store is doing better at one of these
time periods than the others in providing shorter drivethru
waiting times Discuss.
3. Referring to questions 1 and 2 compute the range and standard
deviation for drivethru times for the overall sample and for the
three different times of the day. Also calculate the appropriate
measure of relative variability for each time period. Discuss
these measures of variability and what they might imply about
what customers can expect at this McDonald’s drivethru.
4. Determine the 1st and 3rd quartiles for drivethru times and
develop a box and whisker diagram for the overall sample
data. Are there any outliers identified in these sample data
Discuss.
Yet that doesn’t stop the company from using statistical tech
niques to better understand the drivethru action. In particular
McDonald’s utilizes numerical measures of the center and spread in
the data to help transform the data into useful information. For res
taurant managers to achieve the goal in their own restaurants they
need training in proper restaurant and drivethru operations. Ham
burger University McDonald’s training center located near Chi
cago Illinois satisfies that need. In the mockup restaurant service
lab managers go through a “before and after” training scenario. In
the “before” scenario they run the restaurant for 30 minutes as if
they were back in their home restaurants. Managers in the training
class are assigned to be crew customers drivethru cars special
needs guests such as hearing impaired indecisive clumsy or
observers. Statistical data about the operations revenues and ser
vice times are collected and analyzed. Without the right training
the restaurant’s operations usually start breaking down after 10 to
15 minutes. After debriefing and analyzing the data collected the
managers make suggestions for adjustments and head back to the
service lab to try again. This time the results usually come in well
within standards. “When presented with the quantitative results
managers are pretty quick to make the connections between better
operations higher revenues and happier customers” Marla states.
When managers return to their respective restaurants the
training results and techniques are shared with staff charged with
implementing the ideas locally. The results of the training eventu
ally are measured when McDonald’s conducts a restaurant opera
tions improvement process study or ROIP. The goal is simple:
improved operations. When the ROIP review is completed statis
tical analyses are performed and managers are given their results.
Depending on the results decisions might be made that require
additional financial resources building construction staff train
ing or layout reconfiguration. Yet one thing is clear: Statistics
Case 1
WGI—Human Resources
WGI is a large international construction company with opera
tions in 43 countries. The company has been a major player in
the reconstruction efforts in Iraq with a number of subcontracts
under the major contractor Haliburton Inc. However the com
pany is also involved in many small projects both in the United
States and around the world. One of these is a sewer line installa
tion project in Madison Wisconsin. The contract is what is called
a “cost plus” contract meaning that the city of Madison will pay
for all direct costs including materials and labor of the project
plus an additional fee to WGI. Roberta Bernhart is the human
resources HR manager for the Madison project and is respon
sible for overseeing all aspects of employee compensation and
HR issues.
WGI is required to produce a variety of reports to the Madison
city council on an annual basis. Recently the council asked WGI to
prepare a report showing the current hourly rates for the nonsalaried
work crew on the project. Specifically the council is interested in
any proposed pay increases to the work crew that will ultimately be
passed along to the city of Madison. In response to the city’s request
Ro berta put together a data file for all 19 nonsalaried work crew mem
bers called WGI which shows their current hourly pay rate and the
proposed increase to take place the first of next month. These data are
as follows:
Name Current Rate New Rate
Jody 20.55 22.55
Tim 22.15 23.81
Thomas 14.18 15.60
Shari 14.18 15.60
John 18.80 20.20
Jared 18.98 20.20
Loren 25.24 26.42
Mike 18.36 19.28
Patrick 17.20 18.06
Sharon 16.99 17.84
Sam 16.45 17.27
Susan 18.90 19.66
Chris 18.30 19.02
Steve F 27.45 28.12
Kevin 16.00 16.64
Larry 17.47 18.00
MaryAnn 23.99 24.47
Mark 22.62 23.08
Aaron 15.00 15.40
www.downloadslide.comslide 142:
Describing Data Using Numerical Measures
3. Compute a new variable called Pay Increase that reflects
the difference between the proposed new pay rate and the
current rate. Develop a histogram for this variable and then
compute key measures of the center and variation for the
new variable.
4. Compute a new variable that is the percentage increase
in hourly pay rate. Prepare a graphical and numerical
description of this new variable.
5. Prepare a report to the city council that contains the results
from tasks 1–4.
The city council expects the report to contain both graphic and
numerical descriptive analyses. Roberta has outlined the following
tasks and has asked you to help her.
Required Tasks:
1. Develop and interpret histograms showing the distributions
of current hourly rates and proposed new hourly rates for the
crew members.
2. Compute and interpret key measures of central tendency and
variation for the current and new hourly rates. Determine the
coefficient of variation for each.
Case 2
National Call Center
Candice Worthy and Philip Hanson are day shift supervisors at
National Call Center’s Austin Texas facility. National provides
contract call center services for a number of companies includ
ing banks and major retail companies. Candice and Philip have
both been with the company for slightly more than five years
having joined National right after graduating with bachelor’s
degrees from the University of Texas. As they walked down
the hall together after the weekly staff meeting the two friends
were discussing the assignment they were just handed by Mark
Gonzales the division manager. The assignment came out of
a discussion at the meeting in which one of National’s clients
wanted a report describing the calls being handled for them
by National. Mark had asked Candice and Philip to describe
the data in a file called National Call Center and produce a
report that would both graphically and numerically analyze the
data. The data are for a sample of 57 calls and for the following
v ariables:
Account Number
Caller Gender
Account Holder Gender
Past Due Amount
Current Account Balance
Nature of Call Billing Question or Other
By the time they reached their office Candice and Philip had out
lined some of the key tasks that they needed to do.
Required Tasks:
1. Develop bar charts showing the mean and median current
account balance by gender of the caller.
2. Develop bar charts showing the mean and median current
account balance by gender of the account holder.
3. Construct a scatter diagram showing current balance on the
horizontal axis and past due amount on the vertical axis.
4. Compute the key descriptive statistics for the center and for
the variation in current account balance broken down by
gender of the caller gender of the account holder and by the
nature of the call.
5. Repeat task 4 but compute the statistics for the past due
balances.
6. Compute the coefficient of variation for current account
balances for male and female account holders.
7. Develop frequency and relative frequency distributions for
the gender of callers gender of account holders and nature
of the calls.
8. Develop joint frequency and joint relative frequency
distributions for the account holder gender by whether or not
the account has a past due balance.
9. Write a report to National’s client that contains the results
for tasks 1–8 along with a discussion of these statistics and
graphs.
Case 3
Welco Lumber Company—Part B
Welco Lumber makes cedar fencing materials at its Naples Idaho
facility employing about 160 people.
The head rig is a large saw that breaks down the logs into
slabs and cants. Gene recently conducted a study in which he
videotaped 365 logs being broken down by the head rig. All
three operators April Sid and Jim were involved. Each log
was marked as to its true diameter. Then Gene observed the way
the log was broken down and the degree to which the cants were
properly centered. He then determined the projected value of
the finished product from each log given the way it was actually
cut. In addition he also determined what the value would have
been had the log been cut in the optimal way. Data for this study
are in a file called Welco Lumber.
In addition to the graphical analysis that you helped Gene per
form you have been asked to assist Gene by analyzing these data
using appropriate measures of the center and variation. He wishes
to focus on the lost profit to the company and whether there are
differences among the operators. Also do the operators tend to
perform better on small logs than on large logs In general he is
hoping to learn as much as possible from this study and needs your
help with the analysis.
www.downloadslide.comslide 143:
Describing Data Using Numerical Measures
Case 4
AJ’s Fitness Center
When A. J. Reeser signed papers to take ownership of the fitness
center previously known as the Park Center Club he realized that
he had just taken the biggest financial step in his life. Every asset
he could pull together had been pledged against the mortgage. If
the new AJ’s Fitness Center didn’t succeed he would be in really
bad shape financially. But A. J. didn’t plan on failing. After all he
had never failed at anything.
As a high school football AllAmerican A. J. had been heav
ily recruited by major colleges around the country. Although he
loved football he and his family had always put academics ahead
of sports. Thus he surprised almost everyone other than those who
knew him best when he chose to attend an Ivy League university
not particularly noted for its football success. Although he excelled
at football and was a member of two winning teams he also suc
ceeded in the classroom and graduated in four years. He spent six
years working for McKinsey Company a major consulting firm
at which he gained significant experience in a broad range of busi
ness situations.
He was hired away from McKinsey Company by the
Dryden Group a management services company that specializes
in running health and fitness operations and recreational resorts
throughout the world. After eight years of leading the Fitness
Center section at Dryden A. J. found that earning a high salary
and the perks associated with corporate life were not satisfying
him. Besides the travel was getting old now that he had married
and had two young children. When the opportunity to purchase
the Park Center Club came he decided that the time was right to
control his own destiny.
A key aspect of the deal was that AJ’s Fitness Club would keep
its existing clientele consisting of 1833 memberships. One of the
things A. J. was very concerned about was whether these members
would stay with the club after the sale or move on to other fitness
clubs in the area. He knew that keeping existing customers is a lot
less expensive than attracting new customers.
Within days of assuming ownership A. J. developed a survey
that was mailed to all 1833 members. The letter that accompa
nied the survey discussed A. J.’s philosophy and asked several
key questions regarding the current level of satisfaction. Survey
respondents were eligible to win a free lifetime membership in a
drawing—an inducement that was no doubt responsible for the
1214 usable responses.
To get help with the analysis of the survey data A. J.
approached the college of business at a local university with the
idea of having a senior student serve as an intern at AJ’s Fitness
Center. In addition to an hourly wage the intern would get free use
of the fitness facilities for the rest of the academic year.
The intern’s first task was to key the data from the survey into
a file that could be analyzed using a spreadsheet or a statistical
software package. The survey contained eight questions that were
keyed into eight columns as follows:
Column 1: Satisfaction with the club’s weight and exer
ciseequipment facilities
Column 2: Satisfaction with the club’s staff
Column 3: Satisfaction with the club’s exercise programs
aerobics etc.
Column 4: Satisfaction with the club’s overall service
Note columns 1 through 4 were coded on an ordinal scale as
follows:
1 2345
Very
unsatisfied
Unsatisfied Neutral Satisfied Very
satisfied
Column 5: Number of years that the respondent had been
a member at this club
Column 6: Gender 11 Male 2 Female2
Column 7: Typical number of visits to the club per week
Column 8: Age
The data saved in the file AJFitness were clearly too much
for anyone to comprehend in raw form. At yesterday’s meeting
A. J. asked the intern to “make some sense of the data.” When
the intern asked for some direction A. J.’s response was “That’s
what I’m paying you the big bucks for. I just want you to develop
a descriptive analysis of these data. Use whatever charts graphs
and tables that will help us understand our customers. Also use
any pertinent numerical measures that will help in the analysis. For
right now give me a report that discusses the data. Why don’t we
set a time to get together next week to review your report”
www.downloadslide.comslide 144:
Describing Data Using Numerical Measures
Answers to Selected OddNumbered Problems
This section contains summary answers to most of the oddnumbered problems in the text. The Student Solutions Manual contains fully developed
solutions to all oddnumbered problems and shows clearly how each answer is determined.
x
a
n
i 1
x
i
n
26110 26.1
s
2
a
n
i 1
x  x
2
n  1
148.910  1 16.5444
s 2s
2
216.5444 4.0675
Interquartile range 28  23 5
b. Ages are lower at Whitworth than for the U.S. colleges and
universities as a group.
37. a. The range is 113.0 the IQR is 62.25 the variance is
1217.14 and the standard deviation is 34.89.
b. No
c. Adding a constant to all the data values leaves the variance
unchanged.
39. a. Mean.125 Standard Deviation 1.028. IQR 1.1
b. Prices have fallen slightly.
c. Variation in prices is greater than the average decrease.
41. a. Men spent an average of 117 whereas women spent an
average of 98 for their phones. The standard deviation for
men was nearly twice that for women.
b. Business users spent an average of 166.67 on their phones
whereas home users spent an average of 105.74. The vari
ation in phone costs for the two groups was about equal.
43. a. The population mean is
m
a
x
N
+178465
b. The population median is
m
∼
+173000
c. The range is:
R High  Low
R +361100  +54100
+307000
d. The population standard deviation is
s
Q
a
x m
2
N
+63172
47. a. at least 75 in the range 2600 to 3400 m2s.
b. The range 2400 to 3600 should contain at least 89 of
the data values.
c. less than 11
49. a. 25.008
b. CV 23.55
c. The range from 31.19 to 181.24 should contain at least
89 of the data values.
51. For Distribution A: CV
s
m
100
100
500
100 20
For Distribution B: CV
s
m
100
4.0
10.0
100 40
1. Q
1
4423 Median 5002 Q
3
5381
3. Q
1
13.5 + 13.6
2
13.55
Q
3
15.5 + 15.9
2
15.7
7. a. 31.2 + 32.22 31.7
b. 26.7 + 31.22 28.95
c. 20.8 + 22.82 21.8
9. a. Mean 19 Median 19 + 192 19 Mode 19
b. Symmetrical Mean Median Mode
11. a. 11213.48
b. Use weighted average.
13. a. Mean 114.21 Median 107.50 Mode 156
b. skewed right
15. a. 562.99
b. 551.685
c. 562.90
17. a. 2008 Average
Σx
i
n
6826804000000
8885
768351603.83
2011 Average
Σx
i
n
7966700000000
7436
10713690156
b. Deposits went up and the number of institutions went down.
c. If however these data were considered “typical” and were
to be used to describe past current averages they would be
considered statistics.
19. a. Mean .33
b. Median .31
c. Mode .24
d. The 800th percentile .40 minutes.
21. a. x
1177.1
13
90.55 rightskewed
b. 0.392
c. 91.55
d. weighted average
25. a. Range 8  0 8
b. 3.99
c. 1.998
27. a. 16.87
b. 4.11
29. Standard deviation 2.8
31. a. Standard deviation 7.21 IQR 24  12 12
b. Range Largest  Smallest 30  6 24
Standard deviation 6.96 IQR 12
c. s
2
is smaller than s
2
by a factor of N  12N. s is smaller
than s by a factor of 1N  1N. The range and IQR are
not affected.
33. a. The variance is 815.79 and the standard deviation is 28.56.
b. Interquartile range overcomes the susceptibility of the
range to being highly influenced by extreme values.
35. a. Range 33  21 12
www.downloadslide.comslide 145:
Describing Data Using Numerical Measures
53. a. z
800  x
s
800  1000
250
0.80
b. z 0.80
c. z 0.00
55. a. x
1530
30
51 Variance 510.76 Standard deviation 22.60
b. 5122.60 51222.60 51322.60 i.e.
28.4 73.6 5.8 96.2 and 16.8 118.8. There are
1930100 63.3 of the data within 28.4 73.6
3030100 100 of the data within 5.8 96.2
3030100 100 of the data within 16.8 118.8.
c. bellshaped population
57. a.
b. Based on the sample means of the time each drug is effec
tive Drug B appears to be effective longer than Drug A.
c. Based on the standard deviation of effect time Drug B
exhibits a higher variability in effect time than Drug A.
d. Drug A CV 5.93 Drug B CV 7.35. Drug B has a
higher coefficient of variation and the greater relative spread.
59. Existing supplier: CV
0.078
3.75
100 2.08
New supplier: CV
0.135
18.029
100 0.75
61. Anyone scoring below 61.86 rounded to 62 will be rejected
without an interview.
Anyone scoring higher than 91.98 rounded to 92 will be sent
directly to the company.
63. CV
3083.45
11144.48
100 27.67
At least 75 of CPA firms will compute a tax owed between
+4977.58+17311.38
65. a.
Variable Mean StDev Variance Q1 Median Q3 IQR
Scores 94.780 4.130 17.056 93.000 96.000 98.000 5.000
b. Tchebysheff’s Theorem would be preferable.
c. 99
Drug A Drug B
Mean 234.75 270.92
Standard Deviation 13.92 19.90
Variable Mean StDev Median
CloseOpen 0.0354 0.2615 0.0600
73. The mode is a useful measure of location of a set of data if the
data set is large and involves nominal or ordinal data.
75. a. 0.34
b. 0.34
c. 0.16
77. a. 364.42
b. Variance 16662.63 Standard deviation 129.08
81. a. Comparing only the mean bushels/acre you would say that
Seed Type C produces the greatest average yield per acre.
b. CV of Seed Type A 2588 0.2841 or 28.41
CV of Seed Type B 1556 0.2679 or 26.79
CV of Seed Type C 16/100 0.1600 or 16
Seed Type C shows the least relative variability.
c. Seed Type A: 68 between 63 and 113
95 between 38 and 138
approximately 100 between 13 and 163
Seed Type B: 68 between 41 and 71
95 between 26 and 86
approximately 100 between 11 and 101
Seed Type C: 68 between 84 and 116
95 between 68 and 132
approximately 100 between 52 and 148
d. Seed Type A
e. Seed Type C
87. a. Mean 54.00
Standard Deviation 3.813
b. x1s 543.813 150.187 57.813
x2s 46.374 61.6262
x3s 42.561 65.4392
c. The Empirical Rule indicates that 95 of the data is con
tained within x 2s. This would mean that each tail has
11  0.9522 0.025 of the data. Therefore the costume
should be priced at 46.37.
89. a.
b. It means that the closing price for GE stock is an average
of approximately four 0.0354 cents lower than the
opening price.
c.
Variable Mean StDev Median
Open 33.947 0.503 33.980
CloseOpen 0.0354 0.2615 0.0600
www.downloadslide.comslide 146:
Describing Data Using Numerical Measures
Range The range is a measure of variation that is computed by
finding the difference between the maximum and minimum
values in a data set.
RightSkewed Data A data distribution is right skewed if the
mean for the data is larger than the median.
Sample Mean The average for all values in the sample com
puted by dividing the sum of all sample values by the sam
ple size.
Skewed Data Data sets that are not symmetric. For skewed
data the mean will be larger or smaller than the median.
Standard Deviation The standard deviation is the positive
square root of the variance.
Standardized Data Values The number of standard devia
tions a value is from the mean. Standardized data values are
sometimes referred to as z scores.
Statistic A measure computed from a sample that has been
selected from a population. The value of the statistic will
depend on which sample is selected.
Symmetric Data Data sets whose values are evenly spread
around the center. For symmetric data the mean and median
are equal.
Tchebysheff’s Theorem Regardless of how data are distrib
uted at least 11  1k
2
2 of the values will fall within k
standard deviations of the mean. For example:
At least a1 
1
1
2
b 0 0 of the values will fall
within k 1 standard deviation of the mean.
At least a1 
1
2
2
b
3
4
75 of the values will lie
within k 2 standard deviations of the mean.
At least a1 
1
3
2
b
8
9
89 of the values will lie
within k 3 standard deviations of the mean.
Variance The population variance is the average of the squared
distances of the data values from the mean.
Variation A set of data exhibits variation if all the data are not
the same value.
Weighted Mean The mean value of data values that have been
weighted according to their relative importance.
Box and Whisker Plot A graph that is composed of two parts:
a box and the whiskers. The box has a width that ranges
from the first quartile 1Q
1
2 to the third quartile 1Q
3
2. A verti
cal line through the box is placed at the median. Limits are
located at a value that is 1.5 times the difference between Q
1
and Q
3
below Q
1
and above Q
3
. The whiskers extend to the
left to the lowest value within the limits and to the right to
the highest value within the limits.
Coefficient of Variation The ratio of the standard deviation to
the mean expressed as a percentage. The coefficient of varia
tion is used to measure variation relative to the mean.
Data Array Data that have been arranged in numerical order.
Empirical Rule If the data distribution is bell shaped then the
interval
m 1s contains approximately 68 of the values
m 2s contains approximately 95 of the values
m 3s contains virtually all of the data values
Interquartile Range The interquartile range is a measure of
variation that is determined by computing the difference
between the third and first quartiles.
LeftSkewed Data A data distribution is left skewed if the
mean for the data is smaller than the median.
Mean A numerical measure of the center of a set of quantita
tive measures computed by dividing the sum of the values
by the number of values in the data.
Median The median is a center value that divides a data array
into two halves. We use m
∼
to denote the population median
and M
d
to denote the sample median.
Mode The mode is the value in a data set that occurs most fre
quently.
Parameter A measure computed from the entire population.
As long as the population does not change the value of the
parameter will not change.
Percentiles The pth percentile in a data array is a value that
divides the data set into two parts. The lower segment con
tains at least p and the upper segment contains at least
1 100  p25 of the data. The 50th percentile is the median.
Population Mean The average for all values in the population
computed by dividing the sum of all values by the popula
tion size.
Quartiles Quartiles in a data array are those values that divide
the data set into four equalsized groups. The median cor
responds to the second quartile.
Glossary
www.downloadslide.comslide 147:
Describing Data Using Numerical Measures
Berenson Mark L. and David M. Levine Basic Business Sta
tistics: Concepts and Applications 12th ed. Upper Saddle
River NJ: Prentice Hall 2012.
DeVeaux Richard D. Paul F. Velleman and David E. Bock
Stats Data and Models 3rd ed. New York: AddisonWesley
2012.
Microsoft Excel 2010 Redmond WA: Microsoft Corp. 2010.
Siegel Andrew F. Practical Business Statistics 5th ed. Burr
Ridge IL: Irwin 2002.
Tukey John W. Exploratory Data Analysis Reading MA:
AddisonWesley 1977.
References
www.downloadslide.comslide 148:
Special Review Section
The Where Why and How of Data Collection
Graphs Charts and Tables—Describing Your Data
Describing Data Using Numerical Measures
This text introduces data data collection and statistical tools for describing data. The steps
needed to gather “good” statistical data transform it to usable information and present the
information in a manner that allows good decisions are outlined in the following figures.
Information Statistical Tools
Transforming Data into Information
Data
From Special Review Section 1–3 of Business Statistics A DecisionMaking Approach Ninth Edition. David F. Groebner
Patrick W. Shannon and Phillip C. Fry. Copyright © 2014 by Pearson Education Inc. All rights reserved.
www.downloadslide.comslide 149:
Special Review Section
A Typical Application Sequence
Determine a Need for Data
• Research the issue
• Analyze business alternatives
• Respond to request for information
Defne your
data requirements
Defne the Population
• All items of interest—Who What
Determine What Data Y ou Will Need
• Identify the key variables e.g. age income diameter processing time satisfaction rating
• What categorical breakdowns will be needed e.g. analyze by gender race region and class standing
Decide How the Data Will Be Collected
• Experiment
• Observation
• Automation
• Telephone Survey
• Written Survey
• Personal Interview
Decide on a Census or a Sample
• Census: All items in the population
Decide on Statistical or Nonstatistical Sampling
• Statistical Sampling:
• Nonstatistical Sampling:
Convenience Sample
Judgment Sample
Simple Random Sample
Stratifed Random Sample
Systematic Random Sample
Cluster Random Sample
Determine Data Types and Measurement Level
Te method of descriptive statistical analysis that can be performed depends on the type of data and
the level of data measurement for the variables in the data set. Typical studies will involve multiple types
of variables and data levels.
• Types of Data • Data Timing
Quantitative Qualitative CrossSectional TimeSeries
ta Level
Nominal
Ordinal
Interval/Ratio
Lowest Level
MidLevel
Highest Level
Categories—no ordering implied
Categories—defned ordering
Measurements
• Sample: A subset of the population
• Determine how to gain access to the population
www.downloadslide.comslide 150:
Special Review Section
Select Graphic Presentation Tools
Quantitative Qualitative
Data
Timing
Line Chart
Data
Type
Frequency
Distribution
Bar Chart
Vertical or
Horizontal
Bar Chart
Vertical
Relative
Frequency
Distribution
Pie Chart
Joint
Frequency
Histogram
Frequency
Distribution
CrossSectional
Grouped
or
Ungrouped
Time
Series
Relative
Frequency
Distribution
Stem and Leaf
Diagram
Scatter
Diagram Cumulative
Relative
Frequency
Distribution
Joint
Frequency
Distribution
Box and
Whisker Plot
Ogive
Discrete or
Continuous
Interval/Ratio
Categorical/
Nominal/Ordinal
Data
Level
Mode
Median
Mean
Mode
Median
Mode
Percentiles/
Quartiles
Range
Interquartile
Range
Variance and
Standard
Deviation
Percentiles/
Quartiles
Box and
Whisker
Coefcient
of Variation
Standardized
zvalues
Ordinal Nominal
Ratio/Interval
Central Location Variation
Central Location
Descriptive
Analysis
Comparisons
Compute Numerical Measures
Te choice of nu merical descriptive analysis depends on the level of data measurement. If the data
are ratio or interval you have the widest range of numerical tools available.
Type of
Measures
www.downloadslide.comslide 151:
Special Review Section
Write the Statistical Report
Tere is no one set format for writing a statistical report. However there are a few suggestions
you may fnd useful.
Describe the data collection methodology :
Explain how the data were gathered and the sampling techniques were used.
Lay the foundation :
Provide background and motivation for the analysis.
Use a logical sequence :
Follow a systematic plan for presenting your fndings and analysis.
Label fgures and tables by number :
Employ a consistent numbering and labeling format.
Exercises
Integrative Application Exercises
This text introduced you to the basics of descriptive statistics.
Many of the business application problems advanced business
application problems and cases will give you practice in
performing descriptive statistical analysis. However too
often you are told which procedure you should use or you
can surmise which to use by the location of the exercise. It is
important that you learn to identify the appropriate procedure
on your own in order to solve problems for test purposes. But
more important this ability is essential throughout your career
when you are required to select procedures for the tasks you
will undertake. The following exercises will provide you with
identification practice.
SR.1. Go to your university library and obtain the Statistical
Abstract of the United States.
a. Construct a frequency distribution for
unemployment rate by state for the most current
year available.
b. Justify your choice of class limits and number of
classes.
c. Locate the unemployment rate for the state in which
you are attending college. 1 What proportion of
the unemployment rates are below that of your
state 2 Describe the distribution’s shape with
respect to symmetry. 3 If you were planning
to build a new manufacturing plant what state
would you choose in which to build Justify your
answer. 4 Are there any unusual features of this
distribution Describe them.
SR.2. The State Industrial Development Council is currently
working on a financial services brochure to send to
outofstate companies. It is hoped that the brochure
will be helpful in attracting companies to relocate
to your state. You are given the following frequency
distribution on banks in your state:
Deposit Size
in millions
Number of
Banks
Total Deposits
in millions
Less than 5 2 7.2
5 to less than 10 7 52.1
10 to less than 25 6 111.5
25 to less than 50 3 95.4
50 to less than 100 2 166.6
100 to less than 500 2 529.8
Over 500 2 1663.0
a. Does this frequency distribution violate any of the
rules of construction for frequency distributions If
so reconstruct the frequency distribution to remedy
this violation.
b. The Council wishes to target companies that
would require financial support from banks
that have at least 25 million in deposits.
Reconstruct the frequency distribution to attract
such companies to relocate to your state. Do
this by considering different classes that would
accomplish such a goal.
c. Reconstruct the frequency distribution to attract
companies that require financial support from banks
that have between 5 million and 25 million in
deposits.
d. Present an eyecatching twoparagraph summary
of what the data would mean to a company that
is considering moving to the state. Your boss has
said you need to include relative frequencies in this
presentation.
MyStatLab
www.downloadslide.comslide 152:
Special Review Section
in part a to prepare a written report that describes
the results of the test. Be sure to include in your
report a conclusion regarding whether the scanner
outperforms the manual process.
c. Which process scanner or manual generated
the most values that were more than 2 standard
deviations from the mean
d. Which of the two processes has the least relative
variability
SR.5.
Excel
tutorials
The commercial banking industry is undergoing
rapid changes due to advances in technology and
competitive pressures in the financial services sector.
The data file Banks contains selected information
tabulated by Fortune concerning the revenues
profitability and number of employees for the 51 largest
U.S. commercial banks in terms of revenues. Use the
information in this file to complete the following:
a. Compute the mean median and standard deviation
for the three variables revenues profits and number
of employees.
b. Convert the data for each variable to a z value.
Consider Mellon Bank Corporation headquartered
in Pittsburgh. How does it compare to the average
bank in the study on the three variables Discuss.
c. As you can see by examining the data and by
looking at the statistics computed in part a not all
banks had the same revenue same profit or the
same number of employees. Which variable had the
greatest relative variation among the banks in the
study
d. Calculate a new variable: profits per employee.
Develop a frequency distribution and a histogram
for this new variable. Also compute the mean
median and standard deviation for the new variable.
Write a short report that describes the profits per
employee for the banks.
e. Referring to part d how many banks had a profit
peremployee ratio that exceeded 2 standard
deviations from the mean
SR.3.
Excel
tutorials
As an intern for Intel Corporation suppose you
have been asked to help the vice president prepare a
newsletter to the shareholders. Y ou have been given
access to the data in a file called Intel that contains Intel
Corporation financial data for the years 1987 through
1996. Go to the Internet or to Intel’s annual report and
update the file to include the same variables for the years
1997 to the present. Then use graphs to effectively
present the data in a format that would be usable for the
vice president’s newsletter. Write a short article that
discusses the information shown in your graphs.
SR.4.
Excel
tutorials
The Woodmill Company makes windows and
door trim products. The first step in the process is to
rip dimension 2 8 2 10 etc. lumber into
narrower pieces. Currently the company uses a manual
process in which an experienced operator quickly
looks at a board and determines what rip widths to use.
The decision is based on the knots and defects in the
wood.
A company in Oregon has developed an optical
scanner that can be used to determine the rip widths.
The scanner is programmed to recognize defects and
to determine rip widths that will optimize the value of
the board. A test run of 100 boards was put through the
scanner and the rip widths were identified. However
the boards were not actually ripped. A lumber grader
determined the resulting values for each of the 100
boards assuming that the rips determined by the
scanner had been made. Next the same 100 boards
were manually ripped using the normal process. The
grader then determined the value for each board after
the manual rip process was completed. The resulting
data in the file Woodmill consist of manual rip values
and scanner rip values for each of the 100 boards.
a. Develop a frequency distribution for the board
values for the scanner and the manual process.
b. Compute appropriate descriptive statistics for
both manual and scanner values. Use these data
along with the frequency distribution developed
END SRS 13
www.downloadslide.comslide 153:
Special Review Section
were selected in each county and the state police set up roadblocks
on a randomly selected day. Vehicles with instate license plates
were stopped at random until approximately 100 vehicles had been
stopped at each location. The target total was about 2400 vehicles
statewide.
The issue of primary interest was whether the vehicle was
insured. This was determined by observing whether the vehicle
was carrying the required certificate of insurance. If so the officer
took down the insurance company name and address and the policy
number. If the certificate was not in the car but the owner stated
that insurance was carried the owner was given a postcard to
return within five days supplying the required information. A vehi
cle was determined to be uninsured if no postcard was returned or
if subsequently the insurance company reported that the policy
was not valid on the day of the survey.
In addition to the issue of insurance coverage Herb Kriner
wanted to collect other information about the vehicle and the
owner. This was done using a personal interview during which
the police officer asked a series of questions and observed cer
tain things such as seat belt usage and driver’s and vehicle license
expiration status. Also the owners’ driving records were obtained
through the Transportation Department’s computer division and
added to the information gathered by the state police.
The Data
The data are contained in the file Liabins. The sheet titled
“Description” contains an explanation of the data set and the vari
ables.
Issues to Address
Herb Kriner has two weeks before making a presentation to the
legislative subcommittee that has been dealing with the liability
insurance issue. As Herb’s chief analyst your job is to perform a
comprehensive analysis of the data and to prepare the report that
Herb will deliver to the legislature. Remember this report will
go a long way in determining whether the state should spend the
1.5 million to implement a full liability insurance audit system.
State Department of Insurance
Excel
tutorials
This case study describes the efforts undertaken by the director of
the Insurance Division to assess the magnitude of the uninsured
motorist problem in a western state. The objective of the case study
is to introduce you to a data collection application and show how
one organization developed a database. The database Liabins con
tains a subset of the data actually collected by the state department.
The impetus for the case came from the legislative transpor
tation committee which heard much testimony during the recent
legislative session about the problems that occur when an unin
sured motorist is involved in a traffic accident in which damages to
individuals and property occur. The state’s law enforcement offic
ers also testified that a large number of vehicles are not covered by
liability insurance.
Because of both political pressure and a sense of duty to do
what is right the legislative committee spent many hours wres
tling with what to do about drivers who do not carry the mandatory
liability insurance. Because the actual magnitude of the problem
was unknown the committee finally arrived at a compromise plan
which required the state Insurance Division to perform random
audits of vehicles to determine whether the vehicle was covered
by liability insurance. The audits are to be performed on approxi
mately 1 of the state’s 1 million registered vehicles each month.
If a vehicle is found not to have liability insurance the vehicle
license and the owner’s driver’s license will be revoked for three
months and a 250 fine will be imposed.
However before actually implementing the audit process
which is projected to cost 1.5 million per year Herb Kriner
director of the Insurance Division was told to conduct a prelimi
nary study of the uninsured motorists’ problem in the state and to
report back to the legislative committee in six months.
The Study
A random sample of 12 counties in the state was selected in a man
ner that gave the counties with higher numbers of registered vehi
cles proportionally higher chances of being selected. Two locations
Here is an integrative case study designed to give you more experience. In addition we have included several term project assignments that
require you to collect and analyze data.
Review Case 1
Term Project Assignments
For the project selected you are to devise a sampling plan collect
appropriate data and carry out a full descriptive analysis aimed
at shedding light on the key issues for the project. The finished
project will include a written report of a length and format
specified by your professor.
Project A
Issue: Your College of Business and Economics seeks input
from business majors regarding class scheduling. Some potential
issues are
● Day or evening
● Morning or afternoon
● Oneday twoday or threeday schedules
● Weekend
● Location on or off campus
Project B
Issue: Intercollegiate athletics is a part of most major universi
ties. Revenue from attendance at major sporting events is one
key to financing the athletic program. Investigate the drivers of
www.downloadslide.comslide 154:
Special Review Section
many companies are listed in both files but some are just in one
or the other.
The two files have many of the same variables but the 2003
file has a larger range of financial variables than the 2005 file.
For some companies the data for certain variables are not avail
able and a code of NA is used to so indicate. The 2003 file has a
special worksheet that contains the description of each variable.
These descriptions apply to the 2005 data file as well.
You have been given access to these two data files for use in
preparing your reports. Your role will be to perform certain statis
tical analyses that can be used to help convert these data into use
ful information in order to respond to the clients’ questions.
This morning one of the partners of your company received
a call from a client who asked for a report that would compare
companies in the financial services industry SIC codes in the
6000s to companies in productionoriented businesses SIC
codes in the 2000s and 3000s. There are no firm guidelines on
what the report should entail but the partner has suggested the
following:
● Start with the 2005 data file. Pull the data for all companies
with the desired SIC codes into a new worksheet.
● Prepare a complete descriptive analysis of key financial vari
ables using appropriate charts and graphs to help compare
the two types of businesses.
● Determine whether there are differences between the two
classes of companies in terms of key financial measures.
● Using data from the 2003 file for companies that have these
SIC codes and that are also in the 2005 file develop a compari
son that shows the changes over the time span both within SIC
code grouping and between SIC code groupings.
Project Deliverables
To successfully complete this capstone project you are required
to deliver a management report that addresses the partner’s
requests listed above and also contains at least one other sub
stantial type of analysis not mentioned by the partner. This latter
work should be set off in a special section of the report.
The final report should be presented in a professional format
using the style or format suggested by your instructor.
attendance at your university’s men’s basketball and football
games. Some potential issues:
● Game times
● Game days basketball
● Ticket prices
● Athletic booster club memberships
● Competition for entertainment dollars
Project C
Issue: The department of your major is interested in surveying
department alumni. Some potential issues are
● Satisfaction with degree
● Employment status
● Job satisfaction
● Suggestions for improving course content
Capstone Project
Project Objective
The objective of this business statistics capstone project is to pro
vide you with an opportunity to integrate the statistical tools and
concepts that you have learned thus far in your business statistics
course. Like all realworld applications completing this project
will not require you to utilize every statistical technique. Rather
an objective of the assignment is for you to determine which of
the statistical tools and techniques are appropriate for the situation
you have selected.
Project Description
Assume that you are working as an intern for a financial manage
ment company. Your employer has a large number of clients who
trust the company managers to invest their funds. In your posi
tion you are responsible for producing reports for clients when
they request information. Your company has two large data files
with financial information for a large number of U.S. companies.
The first is called US Companies 2003 which contains financial
information for the companies’ 2001 or 2002 fiscal yearend. The
second file is called US Companies 2005 which has data for the
fiscal 2003 or 2004 yearend. The 2003 file has data for 7098
companies. The 2005 file has data for 6992 companies. Thus
www.downloadslide.comslide 155:
www.downloadslide.comslide 156:
Introduction to
Probability
From Chapter 4 of Business Statistics A DecisionMaking Approach Ninth Edition. David F. Groebner
Patrick W. Shannon and Phillip C. Fry. Copyright © 2014 by Pearson Education Inc. All rights reserved.
www.downloadslide.comslide 157:
Introduction to
Probability
Quick Prep Links
Review the discussion of statistical sampling
in.
Examine recent business periodicals and
newspapers looking for examples in which
probability concepts are discussed.
Think about how you determine what
decision to make in situations in which you
are uncertain about your choices.
1 The Basics of Probability
2 The Rules of Probability
Outcome 1. Understand the three approaches to assessing
probabilities.
Outcome 2. Be able to apply the Addition Rule.
Outcome 3. Know how to use the Multiplication Rule.
Outcome 4. Know how to use Bayes’ Theorem for
applications involving conditional probabilities.
Why you need to know
Recently the Powerball lottery raised the cost of buying a ticket from 1 to 2. With the higher ticket prices lottery offi
cials expect the jackpot prize value to increase more rapidly and thereby entice even greater ticket sales. Most people
recognize when buying a lottery ticket that there is a very small probability of winning and that whether they win or
lose is based on chance alone. In case you are not familiar with the Powerball lottery system a drum contains 59 balls
numbered 1 to 59. The player must choose or have a computer choose five numbers between 1 and 59. The player
also chooses a 6th number called the power ball. On the night of the drawing five balls are randomly selected and
then placed in numerical order. Lastly a sixth ball the power ball is randomly selected. To win the jackpot the player
must match all five numbers plus the power ball. The odds of winning are shown on the Powerball website to be 1 in
175223510 or about 0.00000000571. Later in the chapter you will learn the method for computing probabilities like
this. One analogy might put this in perspective. Suppose we take a college football field and cover it with 175223510
tiny red ants. One of these ants has a yellow dot on it. If you
were blindfolded your chances of picking the one ant with
the yellow dot from the millions of ants on the football field
would be the same as winning the Powerball jackpot We
suggest you come up with a different retirement strategy.
In business decision making in many instances chance
is involved in determining the outcome of a decision. For
instance when a TV manufacturer establishes a warranty on
its television sets there is a certain probability that any given
TV will last less than the warranty life and customers will have
to be compensated. Accountants perform audits on the finan
cial statements of a client and sign off on the statements as
accurate while realizing there is a chance that problems exist
that were not uncovered by the audit. A food processor manu
facturer recognizes that there is a chance that one or more of
its products will be substandard and dissatisfy the customer.
Airlines overbook flights to make sure that the seats on the
plane are as full as possible because they know there is a
certain probability that customers will not show for their flight.
Professional poker players base their decisions to fold or play
a hand based on their assessment of the chances that their
hand beats those of their opponents.
Photo_Ma/Fotolia
www.downloadslide.comslide 158:
Introduction to Probability
If we always knew what the result of our decisions would be our life as decision makers would be a lot less
stressful. However in most instances uncertainty exists. To deal with this uncertainty we need to know how to incor
porate probability concepts into the decision process. This text takes the first step in teaching you how to do this by
introducing the basic concepts and rules of probability. You need to have a solid understanding of these basics before
moving on to the more practical probability applications that you will encounter in business.
1 The Basics of Probability
The mathematical study of probability originated more than 300 years ago. The Chevalier de Méré
a French nobleman who today would probably own a gaming house in Monte Carlo began ask
ing questions about games of chance. He was mostly interested in the probability of observing
various outcomes when dice were repeatedly rolled. The French mathematician Blaise Pascal you
may remember studying Pascal’s triangle in a mathematics class with the help of his friend Pierre
de Fermat was able to answer de Méré’s questions. Of course Pascal began asking more and more
complicated questions of himself and his colleagues and the formal study of probability began.
Important Probability Terms
Several explanations of what probability is have come out of this mathematical study. How
ever the definition of probability is quite basic.
For instance if we look out the window and see rain we can say the probability of rain
today is 1 since we know for sure that it will rain. If an airplane has a top speed of 450 mph
and the distance between city A and city B is 900 miles we can say the probability the plane
will make the trip in 1.5 hours is zero—it can’t happen. These examples involve situations in
which we are certain of the outcome and our 1 and 0 probabilities reflect this.
However in most business situations we do not have certainty but instead are uncertain.
For instance if a real estate investor has the option to purchase a small shopping mall deter
mining rate of return on this investment involves uncertainty. The investor does not know with
certainty whether she will make a profit break even or lose money. After looking closely at
the situation she might say the chance of making a profit is 0.30. This value between 0 and 1
reflects her uncertainty about whether she will make a profit from purchasing the shopping mall.
Events and Sample Space Data come in many forms and are gathered in many ways. In
probability language the process that produces the outcomes is an experiment.
For instance a very simple experiment might involve flipping a coin one time. When this
experiment is performed two possible experimental outcomes can occur: head and tail. If the coin
tossing experiment is expanded to involve two flips of the coin the experimental outcomes are
Head on first flip and head on second flip denoted by HH
Head on first flip and tail on second flip denoted by HT
Tail on first flip and head on second flip denoted by TH
Tail on first flip and tail on second flip denoted by TT
In business situations the experiment can be things like an investment decision a person
nel decision or a choice of warehouse location.
The collection of possible experimental outcomes is called the sample space.
EXAMPLE 1 DEFINING THE SAMPLE SPACE
Five Guys Hamburgers The sales manager for the Five Guys
Hamburger chain is interested in analyzing the sales of its three
bestselling hamburgers. As part of this analysis he might be
interested in determining the sample space possible outcomes
for two randomly selected customers. To do this he can use the
following steps.
Step 1 Define the experiment.
The experiment is the sale. The item of interest is the product sold.
Probability
The chance that a particular event will occur.
The probability value will be in the range 0
to 1. A value of 0 means the event will not
occur. A probability of 1 means the event will
occur. Anything between 0 and 1 reflects the
uncertainty of the event occurring. The definition
given is for a countable number of events.
Experiment
A process that produces a single outcome
whose result cannot be predicted with certainty.
Sample Space
The collection of all outcomes that can result
from a selection decision or experiment.
ExQuisine/Fotolia
www.downloadslide.comslide 159:
Introduction to Probability
Step 2 Define the outcomes for one trial of the experiment.
The manager can define the outcomes to be
e
1
Hamburger
e
2
Cheeseburger
e
3
Bacon Burger
Step 3 Define the sample space.
The sample space SS for an experiment involving a single sale is
SS 5e
1
e
2
e
3
6
If the experiment is expanded to include two sales the sample space is
SS 5e
1
e
2
e
3
e
4
e
5
e
6
e
7
e
8
e
9
6
where the outcomes include what happens on both sales and are defined as
Outcome Sale 1 Sale 2
e
1
Hamburger Hamburger
e
2
Hamburger Cheeseburger
e
3
Hamburger Bacon Burger
e
4
Cheeseburger Hamburger
e
5
Cheeseburger Cheeseburger
e
6
Cheeseburger Bacon Burger
e
7
Bacon Burger Hamburger
e
8
Bacon Burger Cheeseburger
e
9
Bacon Burger Bacon Burger
END EXAMPLE
TRY PROBLEM 3
Using Tree Diagrams A tree diagram is often a useful way to define the sample space for
an experiment that helps ensure no outcomes are omitted or repeated. Example 2 illustrates
how a tree diagram is used.
EXAMPLE 2 USING A TREE DIAGRAM TO DEFINE THE SAMPLE SPACE
Clearwater Marketing Research Clearwater Marketing Research is involved in a project
in which television viewers were asked whether they objected to hardliquor advertisements
being shown on television. The analyst is interested in listing the sample space using a tree
diagram as an aid when three viewers are interviewed. The following steps can be used:
Step 1 Define the experiment.
Three people are interviewed and asked “Would you object to hardliquor
advertisements on television” Thus the experiment consists of three trials.
Step 2 Define the outcomes for a single trial of the experiment.
The possible outcomes when one person is interviewed are
no
yes
Step 3 Define the sample space for three trials using a tree diagram.
Begin by determining the outcomes for a single trial. Illustrate these with tree
branches beginning on the left side of the page:
No
Yes
www.downloadslide.comslide 160:
Introduction to Probability
For each of these branches add branches depicting the outcomes for a second
trial. Continue until the tree has the number of sets of branches corresponding
to the number of trials.
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
No No No
No No Y es
No Y es No
No Y es Y es
Y es No No
Y es No Y es
Y es Y es No
Y es Y es Y es
Trial 1 Trial 2 Trial 3 Experimental Outcomes
END EXAMPLE
TRY PROBLEM 4
A collection of possible outcomes is called an event. An example will help clarify these
terms.
EXAMPLE 3 DEFINING AN EVENT OF INTEREST
KPMG Accounting The KPMG Accounting firm is interested in the sample space for
an audit experiment in which the outcome of interest is the audit’s completion status. The
sample space is the list of all possible outcomes from the experiment. The accounting firm is
also interested in specifying the outcomes that make up an event of interest. This can be done
using the following steps:
Step 1 Define the experiment.
The experiment consists of two randomly chosen audits.
Step 2 List the outcomes associated with one trial of the experiment.
For a single audit the following completionstatus possibilities exist:
Audit done early
Audit done on time
Audit done late
Step 3 Define the sample space.
For two audits two trials we define the sample space as follows:
Experimental Outcome Audit 1 Audit 2
e
1
Early Early
e
2
Early On time
Event
A collection of experimental outcomes.
www.downloadslide.comslide 161:
Introduction to Probability
Experimental Outcome Audit 1 Audit 2
e
3
Early Late
e
4
On time Early
e
5
On time On time
e
6
On time Late
e
7
Late Early
e
8
Late On time
e
9
Late Late
Step 4 Define the event of interest.
The event of interest at least one audit is completed late is composed of all
the outcomes in which one or more audits are late. This event E is
E 5e
3
e
6
e
7
e
8
e
9
6
There are five ways in which one or more audits are completed late.
END EXAMPLE
TRY PROBLEM 11
Mutually Exclusive Events Keeping in mind the definitions for experiment sample space
and events we introduce two additional concepts. The first is mutually exclusive events.
BUSINESS APPLICATION MUTUALLY EXCLUSIVE EVENTS
KPMG ACCOUNTING Consider again the KPMG Accounting firm example. The possible
outcomes for two audits are
Experimental Outcomes Audit 1 Audit 2
e
1
Early Early
e
2
Early On time
e
3
Early Late
e
4
On time Early
e
5
On time On time
e
6
On time Late
e
7
Late Early
e
8
Late On time
e
9
Late Late
Suppose we define one event as consisting of the outcomes in which at least one of the
two audits is late.
E
1
5e
3
e
6
e
7
e
8
e
9
6
Further suppose we defne a second event as follows:
E
2
Neither audit is late 5e
1
e
2
e
4
e
5
6
Events E
1
and E
2
are mutually exclusive: If E
1
occurs E
2
cannot occur if E
2
occurs E
1
cannot
occur. That is if at least one audit is late then it is not possible for neither audit to be late. We
can verify this fact by observing that no outcomes in E
1
appear in E
2
. This observation pro
vides another way of defning mutually exclusive events: Two events are mutually exclusive if
they have no common outcomes.
EXAMPLE 4 MUTUALLY EXCLUSIVE EVENTS
TechWorks Inc. TechWorks Inc. located in Dublin Ireland does contract assembly
work for companies such as HewlettPackard. Each item produced on the assembly line can
be thought of as an experimental trial. The managers at this facility can analyze their process
to determine whether the events of interest are mutually exclusive using the following steps:
Step 1 Define the experiment.
The experiment is producing a part on an assembly line.
Mutually Exclusive Events
Two events are mutually exclusive if the
occurrence of one event precludes the
occurrence of the other event.
www.downloadslide.comslide 162:
Introduction to Probability
Step 2 Define the outcomes for a single trial of the experiment.
On each trial the outcome is either a good or a defective item.
Step 3 Define the sample space.
If two products are produced two trials the following sample space is
defined:
Experimental Outcomes
Product 1 Product 2
e
1
Good Good
e
2
Good Defective
e
3
Defective Good
e
4
Defective Defective
Step 4 Determine whether the events are mutually exclusive.
Let event E
1
be defined as both products produced are good and let event E
2
be defined as at least one product is defective:
E
1
Both good 5e
1
6
E
2
At least one defective 5e
2
e
3
e
4
6
Then events E
1
and E
2
are determined to be mutually exclusive because the
two events have no outcomes in common. Having two good items and at the
same time having at least one defective item is not possible.
END EXAMPLE
TRY PROBLEM 9
Independent and Dependent Events A second probability concept is that of
independent versus dependent events.
BUSINESS APPLICATION INDEPENDENT AND DEPENDENT EVENTS
PETROGL YPH Petroglyph is a subsidiary of the Intermountain Gas Company and is responsible
for natural gas exploration in the western U.S. and Canada. During the exploration phase seismic
surveys are conducted that provide information about the Earth’s underground formations. Based
on history the company knows that if the seismic readings are favorable gas will more likely be
discovered than if the seismic readings are not favorable. However the readings are not perfect
indicators. Suppose the company currently is exploring in eastern Colorado. The possible outcomes
for the seismic survey are defined as
e
1
Favorable
e
2
Unfavorable
If the company decides to drill the outcomes are defned as
e
3
Strike gas
e
4
Dry hole
If we let the event E
1
be that the seismic survey is favorable and event E
2
be that the hole is
dry we can say that the events A and B are not mutually exclusive because one event’s occur
rence does not preclude the other event from occurring. We can also say that the two events
are dependent because the probability of a dry hole depends on whether the seismic survey
is favorable or unfavorable. If the result of drilling was not related to the seismic survey the
events would be independent.
Methods of Assigning Probability
Part of the confusion surrounding probability may be due to the fact that probability can be
assigned to outcomes in more than one way. There are three common ways to assign prob
ability to outcomes: classical probability assessment relative frequency assessment and
Independent Events
Two events are independent if the occurrence of
one event in no way influences the probability of
the occurrence of the other event.
Dependent Events
Two events are dependent if the occurrence of
one event impacts the probability of the other
event occurring.
Chapter Outcome 1.
www.downloadslide.comslide 163:
Introduction to Probability
subjective probability assessment. The following notation is used when we refer to the prob
ability of an event:
P1E
i
2 Probability of event E
i
occurring
Classical Probability Assessment The first method of probability assessment involves
classical probability.
You are probably already familiar with classical probability. It had its beginning with
games of chance and is still most often discussed in those terms.
Consider again the experiment of flipping a coin one time. There are two possible out
comes: head and tail. Each of these is equally likely. Thus using the classical assessment
method the probability of a head is the ratio of the number of ways a head can occur 1 way
to the total number of ways any outcome can occur 2 ways. Thus we get
P. Head
way
ways
1
2
1
2
050
The chance of a head occurring is 1 out of 2 or 0.50.
In those situations in which all possible outcomes are equally likely the classical prob
ability measurement is defined in Equation 1.
Classical Probability
Assessment
The method of determining probability based
on the ratio of the number of ways an outcome
or event of interest can occur to the number of
ways any outcome or event can occur when the
individual outcomes are equally likely.
Classical Probability Assessment
PE
E
i
i
Number of ways can occur
Total numbe er of possible outcomes
1
EXAMPLE 5 CLASSICAL PROBABILITY ASSESSMENT
Active Sporting Goods Inc. The managers at Active Sport
ing Goods plan to hold a special promotion over Labor Day
Weekend. Each customer making a purchase exceeding 100 will
qualify to select an envelope from a large drum. Inside the enve
lope are coupons for percentage discounts off the purchase total.
At the beginning of the weekend there were 500 coupons. Four
hundred of these were for a 10 discount 50 were for 20 45
were for 30 and 5 were for 50. The probability of getting a particular discount amount can
be determined using classical assessment with the following steps:
Step 1 Define the experiment.
An envelope is selected from a large drum.
Step 2 Determine whether the possible outcomes are equally likely.
In this case the envelopes with the different discount amounts are unmarked
from the outside and are thoroughly mixed in the drum. Thus any one
envelope has the same probability of being selected as any other envelope.
The outcomes are equally likely.
Step 3 Determine the total number of outcomes.
There are 500 envelopes in the drum.
Step 4 Define the event of interest.
We might be interested in assessing the probability that the first customer will
get a 20 discount.
Step 5 Determine the number of outcomes associated with the event of interest.
There are 50 coupons with a discount of 20 marked on them.
Step 6 Compute the classical probability using Equation 1:
PE
E
i
i
Number of ways can occur
Total numbe er of possible outcomes
discount
Num
P 20
b ber of ways can occur
Total number of p
20
o ossible outcomes
50
500
0.10
Michael Flippo/Fotolia
www.downloadslide.comslide 164:
Introduction to Probability
Note: After the first customer selects an envelope from the drum the
probability that the next customer will get a particular discount will change
because the values in the denominator and possibly the numerator will change.
END EXAMPLE
TRY PROBLEM 10
As you can see the classical approach to probability measurement is fairly straightforward.
Many games of chance are based on classical probability assessment. However classical prob
ability assessment does not apply in many business situations. Rarely are the individual out
comes equally likely. For instance you might be thinking of starting a business. The sample
space is
SS 5Succeed Fail6
Would it be reasonable to use classical assessment to determine the probability that your busi
ness will succeed If so we would make the following assessment:
PSucceed
1
2
If this were true then the chance of any business succeeding would be 0.50. Of course this is
not true. Many factors go into determining the success or failure of a business. The possible
outcomes Succeed Fail are not equally likely. Instead we need another method of probabil
ity assessment in these situations.
Relative Frequency Assessment The relative frequency assessment approach is based
on actual observations.
Equation 2 shows how the relative frequency assessment method is used to assess
probabilities.
Relative Frequency Assessment
The method that defines probability as the
number of times an event occurs divided by
the total number of times an experiment is
performed in a large number of trials.
Relative Frequency Assessment
PE
E
N
i
i
Number of times occurs
2
where:
E
i
The event of interest
N Number of trials
BUSINESS APPLICATION RELATIVE FREQUENCY ASSESSMENT
HATHAWAY HEATING AIR CONDITIONING The sales manager at Hathaway
Heating Air Conditioning has recently developed the customer profile shown in Table 1.
The profile is based on a random sample of 500 customers. As a promotion for the company
the sales manager plans to randomly select a customer once a month and perform a free
service on the customer’s system. What is the probability that the first customer selected
is a residential customer What is the probability that the first customer has a Hathaway
heating system
TABLE 1  Hathaway Heating Air Conditioning Co.
Customer Category
E
1
E
2
Commercial Residential Total
E
3
Heating Systems 55 145 200
E
4
AirConditioning Systems 45 255 300
Total 100 400 500
www.downloadslide.comslide 165:
Introduction to Probability
To determine the probability that the customer selected is residential we determine from
Table 1 the number of residential customers and divide by the total number of customers both
residential and commercial. We then apply Equation 2:
PE P .
2
400
500
080 Residential
Thus there is an 80 chance the customer selected will be a residential customer.
The probability that the customer selected has a Hathaway heating system is determined
by the ratio of the number of customers with heating systems to the number of total customers.
PE P .
3
200
500
040 Heating
There is a 40 chance the randomly selected customer will have a Hathaway heating system.
The sales manager hopes the customer selected is a residential customer with a Hatha
way heating system. Because there are 145 customers in this category the relative frequency
method assesses the probability of this event occurring as follows:
PE E P
23
145
and Residential with heating
5 500
029 .
There is a 29 chance the customer selected will be a residential customer with a Hathaway
heating system.
EXAMPLE 6 RELATIVE FREQUENCY PROBABILITY ASSESSMENT
Starbucks Coffee The international coffee chain Starbucks has a store in a busy mall in
Pittsburgh Pennsylvania. Starbucks sells caffeinated and decaffeinated drinks. One of the
difficulties in this business is determining how much of a given product to prepare for the
day. The manager is interested in determining the probability that a customer will select a
decaf versus a caffeinated drink. She has maintained records of customer purchases for the
past three weeks. The probability can be assessed using relative frequency with the following
steps:
Step 1 Define the experiment.
A randomly chosen customer will select between decaf and caffeinated.
Step 2 Define the events of interest.
The manager is interested in the event E
1
customer selects caffeinated.
Step 3 Determine the total number of occurrences.
In this case the manager has observed 2250 sales of decaf and caffeinated in
the past week. Thus N 2250.
Step 4 For the event of interest determine the number of occurrences.
In the past week 1570 sales were for caffeinated drinks.
Step 5 Use Equation 2 to determine the probability assessment.
PE
E
N
Number of times occurs
1
1
1 570
2 250
0 0 6978 .
Thus based on past history the chance that a customer will purchase a
caffeinated drink is just under 0.70.
END EXAMPLE
TRY PROBLEM 9
POTENTIAL ISSUES WITH THE RELATIVE FREQUENCY ASSESSMENT
METHOD There are a couple of concerns that you should be aware of before applying
the relative frequency assessment method. First for this method to be useful all of the
observed frequencies must be comparable. For instance consider again the case in which
www.downloadslide.comslide 166:
Introduction to Probability
you are interested in starting a small business. Two outcomes can occur: business succeeds or
business fails. If we are interested in the probability that the business will succeed we might
be tempted to study a sample of say 200 small businesses that have been started in the past
and determine the number of those that have succeeded—say 50. Using Equation 2 for the
relative frequency method we get
P. Succeed
50
200
025
However before we can conclude the chance your small business will succeed is 0.25 you
must be sure that the conditions of each of the 200 businesses match your conditions that is
location type of business management expertise and experience financial standing and so
on. If not then the relative frequency method should not be used.
Another issue involves the size of the denominator in Equation 2. If the number of pos
sible occurrences is quite small the probability assessment may be unreliable. For instance
suppose a basketball player took five free throws during the first ten games of the season and
missed them all. The relative frequency method would determine the probability that he will
make the next free throw to be
P. Make
made
5 shots
00
5
00
But do you think that there is a zero chance that he will make his next free throw No
even the notoriously poor freethrow shooter Shaquille O’Neal former National Basketball
Association NBA star player made some of his free throws. The problem is that the base of
five free throws is too small to provide a reliable probability assessment.
Subjective Probability Assessment Unfortunately even though managers may have
some experience to guide their decision making new factors will always be affecting each
decision making that experience only an approximate guide to the future. In other cases
managers may have little or no experience and therefore may not be able to use a relative
frequency as even a starting point in assessing the desired probability. When experience is
not available decision makers must make a subjective probability assessment. A subjective
probability is a measure of a personal conviction that an outcome will occur. Therefore in this
instance probability represents a person’s belief that an event will occur.
BUSINESS APPLICATION SUBJECTIVE PROBABILITY ASSESSMENT
BECHTEL CORPORATION The Bechtel Corporation is preparing a bid for a major
infrastructure construction project. The company’s engineers are very good at defining all
the elements of the projects labor materials and so on and know the costs of these with a
great deal of certainty. In finalizing the bid amount the managers add a profit markup to the
projected costs. The problem is how much markup to add. If they add too much they won’t be
the low bidder and may lose the contract. If they don’t mark the bid up enough they may get
the project and make less profit than they might have made had they used a higher markup.
The managers are considering four possible markup values stated as percentages of base
costs:
10 12 15 20
To make their decision the managers need to assess the probability of winning the con
tract at each of these markup levels. Because they have never done a project exactly like this
one they can’t rely on relative frequency assessment. Instead they must subjectively assess
the probability based on whatever information they currently have available such as who the
other bidders are the rapport Bechtel has with the potential client and so forth.
After considering these values the Bechtel managers make the following assessments:
P1Win at 102 0.30
P1Win at 122 0.25
P1Win at 152 0.15
P1Win at 202 0.05
Subjective Probability
Assessment
The method that defines probability of an event
as reflecting a decision maker’s state of mind
regarding the chances that the particular event
will occur.
www.downloadslide.comslide 167:
Introduction to Probability
These assessments indicate the managers’ state of mind regarding the chances of winning
the contract. If new information for example a competitor drops out of the bidding be
comes available before the bid is submitted these assessments could change.
Each of the three methods by which probabilities are assessed has specific advantages
and specific applications. Regardless of how decision makers arrive at a probability assess
ment the rules by which people use these probabilities in decision making are the same.
These rules will be introduced in Section 2.
MyStatLab
Skill Development
41. A special roulette wheel which has an equal number
of red and black spots has come up red four times in
a row. Assuming that the roulette wheel is fair what
concept allows a player to know that the probability the
next spin of the wheel will come up black is 0.5
42. In a survey respondents were asked to indicate their
favorite brand of cereal Post or Kellogg’s. They
were allowed only one choice. What is the probability
concept that implies it is not possible for a single
respondent to state both Post and Kellogg’s to be the
favorite cereal
43. If two customers are asked to list their choice of ice
cream flavor from among vanilla chocolate and
strawberry list the sample space showing the possible
outcomes.
44. Use a tree diagram to list the sample space for the
number of movies rented by three customers at a video
store where customers are allowed to rent one two
or three movies assuming that each customer rents at
least one movie.
45. In each of the following indicate what method of
probability assessment would most likely be used to
assess the probability.
a. What is the probability that a major earthquake will
occur in California in the next three years
b. What is the probability that a customer will return a
purchase for a refund
c. An inventory of appliances contains four white
washers and one black washer. If a customer selects
one at random what is the probability that the black
washer will be selected
46. Longtime friends Pat and Tom agree on many things
but not the outcome of the American League pennant
race and the World Series. Pat is originally from
Boston and Tom is from New York. They have a steak
dinner bet on next year’s race with Pat betting on the
Red Sox and Tom on the Yankees. Both are convinced
they will win.
a. What probability assessment technique is being
used by the two friends
b. Why would the relative frequency technique not be
appropriate in this situation
47. Students who live on campus and purchase a meal
plan are randomly assigned to one of three dining
halls: the Commons Northeast and Frazier. What
is the probability that the next student to purchase a
meal plan will be assigned to the Commons
48. The results of a census of 2500 employees of a mid
sized company with 401k retirement accounts are
as follows:
Account Balance
to nearest Male Female
6+25000 635 495
+25000  +49999 185 210
+50000  +99999 515 260
7+100000 155 45
Suppose researchers are going to sample employees
from the company for further study.
a. Based on the relative frequency assessment
method what is the probability that a randomly
selected employee will be a female
b. Based on the relative frequency assessment
method what is the probability that a randomly
selected employee will have a 401k account
balance of between 25000 and 49999
c. Compute the probability that a randomly selected
employee will be a female with an account
balance between 50000 and 99999.
49. Cross County Bicycles makes two mountain bike
models the XB50 and the YZ99 in three distinct
colors. The following table shows the production
volumes for last week:
Color
Model Blue Brown White
XB50 302 105 200
YZ99 40 205 130
a. Based on the relative frequency assessment
method what is the probability that a mountain
bike is brown
b. What is the probability that the mountain bike is a
YZ99
41: Exercises
www.downloadslide.comslide 168:
Introduction to Probability
CEO stating that the probability the company will
earn a profit in excess of 20 million next year is 80.
Comment on this probability assessment.
413. Five doctors work at the Evergreen Medical Clinic.
The plan is to staff Saturdays with three doctors. The
office manager has decided to make up Saturday
schedules in such a way that no set of three doctors
will be in the office together more than once. How
many weeks can be covered by this schedule Hint:
Use a tree diagram to list the sample space.
414. Prince Windows Inc. makes highquality windows for
the residential home market. Recently three marketing
managers were asked to assess the probability that
sales for next year will be more than 15 higher
than the current year. One manager stated that the
probability of this happening was 0.40. The second
manager assessed the probability to be 0.60 and the
third manager stated the probability to be 0.90.
a. What method of probability assessment are the
three managers using
b. Which manager is expressing the least uncertainty
in the probability assessment
c. Why is it that the three managers did not provide
the same probability assessment
415. The marketing manager for the Charlotte Times
newspaper has commissioned a study of the
advertisements in the classified section. The results for
the Wednesday edition showed that 204 are helpwanted
ads 520 are real estate ads and 306 are other ads.
a. If the newspaper plans to select an ad at random
each week to be published free what is the
probability that the ad for a specific week will be a
helpwanted ad
b. What method of probability assessment is used to
determine the probability in part a
c. Are the events that a helpwanted ad is chosen and
that an ad for other types of products or services
is chosen for this promotion on a specific week
mutually exclusive Explain.
416. Before passing away in 2009 Larry Miller owned the
Utah Jazz basketball team of the NBA and several
automobile dealerships in Utah and Idaho. One of
the dealerships sells Buick Cadillac and Pontiac
automobiles. It also sells used cars that it gets as trade
ins on new car purchases. Supposing two cars are sold
on Tuesday by the dealership what is the sample space
for the type of cars that might be sold
417. The Pacific Northwest has a substantial volume
of cedar forests and cedar product manufacturing
companies. Welco Lumber manufactures cedar fencing
material in Marysville Washington. The company’s
quality manager inspected 5900 boards and found that
4100 could be rated as a 1 grade.
a. If the manager wanted to assess the probability that
a board being produced will be a 1 grade what
method of assessment would he likely use
b. Referring to your answer in part a what would you
assess the probability of a 1 grade board to be
c. What is the joint probability that a randomly
selected mountain bike is a YZ99 and brown
d. Suppose a mountain bike is chosen at random.
Consider the following two events: the event that
model YZ99 is chosen and the event that a white
product is chosen. Are these two events mutually
exclusive Explain.
410. CyberPlastics Inc. is in search of a CEO and a CFO.
The company has a short list of candidates for each
position. The CEO candidates graduated from Chicago
C and three Ivy League universities: Harvard H
Princeton P and Yale Y. The four CFO candidates
graduated from MIT M Northwestern N and two
Ivy League universities: Dartmouth D and Brown
B. One candidate from each of the respective lists
will be chosen randomly to fill the positions. The
event of interest is that both positions are filled with
candidates from the Ivy League.
a. Determine whether the outcomes are equally likely.
b. Determine the number of equally likely outcomes.
c. Define the event of interest.
d. Determine the number of outcomes associated with
the event of interest.
e. Compute the classical probability of the event of
interest using Equation 1.
411. Three consumers go to a Best Buy to shop for high
definition televisions HDTVs. Let B indicate that
one of the consumers buys an HDTV . Let D be that the
consumer doesn’t buy an HDTV . Assume these events
are equally likely. Consider the following: 1 only two
consumers buy an HDTV 2 at most two consumers buy
HDTVs and 3 at least two consumers buy HDTVs.
a. Determine whether the outcomes 1 2 and 3 are
equally likely.
b. Determine the total number of equally likely
outcomes for the three shoppers.
c. Define the events of interest in each of 1 2 and 3.
To define the events of interest list the possible
outcomes in each of the following events:
.
only two consumers buy an HDTV 1E
1
2
.
at most two consumers buy HDTVs 1E
2
2
.
at least two consumers buy HDTVs 1E
3
2
d. Determine the number of outcomes associated
with each of the events of interest. Use the
classical probability assessment approach to assign
probabilities to each of the possible outcomes and
calculate the probabilities of the events.
e. Compute the classical probabilities of each of the
events in part d by using Equation 1.
Business Applications
412. Cyber Communications Inc. has a new cell phone
product under development in the research and
development RD lab. It will increase the megapixel
capability of cell phone cameras to the 6+ range. The
head of RD made a presentation to the company
www.downloadslide.comslide 169:
Introduction to Probability
418. The results of Fortune Personnel Consultants’ survey
of 405 workers was reported in USA Today. One of the
questions in the survey asked “Do you feel it’s OK
for your company to monitor your Internet use” The
possible responses were: 1 Only after informing me
2 Does not need to inform me 3 Only when company
believes I am misusing 4 Company does not have right
and 5 Only if I have previously misused. The following
table contains the results for the 405 respondents:
Response 12 3 4 5
Number of Respondents 223 130 32 14 6
a. Calculate the probability that a randomly chosen
respondent would indicate that there should be
some restriction concerning the company’s right to
monitor Internet use.
b. Indicate the method of probability assessment used
to determine the probability in part a.
c. Are the events that a randomly selected respondent
chose response 1 and that another randomly selected
respondent chose response 2 independent mutually
exclusive or dependent events Explain.
419. Famous Dave’s is a successful barbeque chain and
sells its beef pork and chicken items to three kinds of
customers: dinein delivery and carryout. Last year’s
sales showed that 12753 orders were dinein 5893
were delivery orders and 3122 orders were carryout.
Suppose an audit of last year’s sales is being conducted.
a. If a customer order is selected at random what is
the probability it will be a carryout order
b. What method of probability assessment was used to
determine the probability in part a
c. If two customer orders are selected at random list
the sample space indicating the type of order for
both customers.
420. VERCOR provides merger and acquisition consultants
to assist corporations when an owner decides to offer the
business for sale. One of their news releases “Tax Audit
Frequency Is Rising” written by David L. Perkins Jr. a
VERCOR partner originally appeared in The Business
Owner . Perkins indicated that audits of the largest
businesses those corporations with assets of 10 million
and over climbed to 9560 in the previous year. That was
up from a low of 7125 a year earlier. He indicated one
in six large corporations was being audited.
a. Designate the type of probability assessment
method that Perkins used to assess the probability of
large corporations being audited.
b. Determine the number of large corporations that
filed tax returns for the previous fiscal year.
c. Determine the probability that a large corporation
was not audited using the relative frequency
probability assessment method.
Computer Database Exercises
421. According to a September 2005 article on the
Womensenews.org Web site “Caesarean sections in
which a baby is delivered by abdominal surgery have
increased fivefold in the past 30 years prompting
concern among health advocates . . .” The data in the
file called Babies indicate whether the past 50 babies
delivered at a local hospital were delivered using the
Caesarean method.
a. Based on these data what is the probability that a
baby born in this hospital will be born using the
Caesarean method
b. What concerns might you have about using these
data to assess the probability of a Caesarean birth
Discuss.
422. Recently a large state university conducted a survey
of undergraduate students regarding their use of
computers. The results of the survey are contained in
the data file ComputerUse.
a. Based on the data from the survey what is the
probability that undergraduate students at this
university will have a major that requires them to
use a computer on a daily basis
b. Based on the data from this survey if a student
is a business major what is the probability of the
student believing that the computer lab facilities are
very adequate
423. A company produces scooters used by small
businesses such as pizza parlors that find them
convenient for making short deliveries. The company
is notified whenever a scooter breaks down and the
problem is classified as being either mechanical or
electrical. The company then matches the scooter to
the plant where it was assembled. The file Scooters
contains a random sample of 200 breakdowns. Use the
data in the file and the relative frequency assessment
method to find the following probabilities:
a. What is the probability a scooter was assembled at
the Tyler plant
b. What is the probability that a scooter breakdown
was due to a mechanical problem
c. What is the probability that a scooter was assem bled
at the Lincoln plant and had an electrical problem
424. A Harris survey on cell phone use asked in part what
was the most important reason that people give for
not using a wireless phone exclusively. The responses
were: 1 Like the safety of traditional phone 2
Need line for Internet access 3 Pricing not attractive
enough 4 Weak or unreliable cell signal at home
5 Coverage not good enough and 6 Other. The file
titled Wireless contains the responses for the 1088
respondents.
a. Calculate the probability that a randomly chosen
respondent would not use a wireless phone
exclusively because of some type of difficulty in
placing and receiving calls with a wireless phone.
b. Calculate the probability that a randomly chosen
person would not use a wireless phone exclusively
because of some type of difficulty in placing and
receiving calls with a wireless phone and is over the
age of 55.
www.downloadslide.comslide 170:
Introduction to Probability
c. Determine the probability that a randomly chosen
person would not use a wireless phone exclusively
because of a perceived need for Internet access and
the safety of a traditional phone.
d. Of those respondents under 36 determine the
probability that an individual in this age group
would not use a wireless phone exclusively because
of some type of difficulty in placing and receiving
calls with a wireless phone.
425. CNN staff writer Pariia Bhatnagar reported “Coke
Pepsi Losing the Fizz” March 8 2005 that Atlanta
based Coke saw its domestic market share drop
to 43.1 in 2004. New Yorkbased PepsiCo had
used its “Pepsi Challenge” advertising approach to
increase its market share which stood at 31.7 in
2004. A selection of softdrink users is asked to taste
the two disguised soft drinks and indicate which
they prefer. The file titled Challenge contains the
results of a simulated Pepsi Challenge on a college
campus.
a. Determine the probability that a randomly chosen
student prefers Pepsi.
b. Determine the probability that one of the students
prefers Pepsi and is less than 20 years old.
c. Of those students who are less than 20 years old
calculate the probability that a randomly chosen
student prefers 1 Pepsi and 2 Coke.
d. Of those students who are at least 20 years old
calculate the probability that a randomly chosen
student prefers 1 Pepsi and 2 Coke.
END EXERCISES 41
2 The Rules of Probability
Measuring Probabilities
The probability attached to an event represents the likelihood the event will occur on a speci
fied trial of an experiment. This probability also measures the perceived uncertainty about
whether the event will occur.
Possible Values and the Summation of Possible Values If we are certain about
the outcome of an event we will assign the event a probability of 0 or 1 where P1E
i
20
indicates the e v ent E
i
will not occur and P1E
i
21 means that E
i
will definitely occur.
1
If we
are uncertain about the result of an experiment we measure this uncertainty by assigning a
probability between 0 and 1. Probability Rule 1 shows that the probability of an event occur
ring is always between 0 and 1.
1
These statements are true only if the number of outcomes of an experiment is countable. They do not apply when
the number of outcomes is infinitely uncountable. This will be discussed when continuous probability distributions
are discussed.
Probability Rule 2
Pe
i
i
k
1
1
S
4
where:
k Number of outcomes in the sample
e
i
ith outcome
Probability Rule 1
For any event E
i
0 … P1E
i
2 … 1 for all i 3
All possible outcomes associated with an experiment form the sample space. Therefore
the sum of the probabilities of all possible outcomes is 1 as shown by Probability Rule 2.
Addition Rule for Individual Outcomes If a single event is composed of two or more
individual outcomes then the probability of the event is found by summing the probabilities
of the individual outcomes. This is illustrated by Probability Rule 3.
Chapter Outcome 2.
www.downloadslide.comslide 171:
Introduction to Probability
Probability Rule 3: Addition Rule for Individual Outcomes
The probability of an event E
i
is equal to the sum of the probabilities of the individual
outcomes forming E
i
. For example if
E
i
5e
1
e
2
e
3
6
then
P1E
i
2 P1e
1
2 + P1e
2
2 + P1e
3
2 5
BUSINESS APPLICATION ADDITION RULE
GOOGLE Google has become synonymous with Web searches and is clearly the market
leader in the search engine marketplace. Officials at the northern California headquarters
have recently performed a survey of computer users to determine how many Internet
searches individuals do daily using Google. Table 2 shows the results of the survey of
Internet users.
The sample space for the experiment for each respondent is
SS 5e
1
e
2
e
3
e
4
6
where the possible outcomes are
e
1
at least 10 searches
e
2
3 to 9 searches
e
3
1 to 2 searches
e
4
0 searches
Using the relative frequency assessment approach we assign the following probabilities.
P1e
1
2 4005000 0.08
P1e
2
2 19005000 0.38
P1e
3
2 15005000 0.30
P1e
4
2 12005000 0.24
g 1.00
Assume we are interested in the event respondent performs 1 to 9 searches per month.
E Internet User Performs 1 to 9 searches per day
The outcomes that make up E are
E 5e
2
e
3
6
We can find the probability PE by using Probability Rule 3 Equation 5 as follows:
P1E2 P1e
2
2 + P1e
3
2
0.38 + 0.30
0.68
TABLE 2  Google’s Survey Results
Searches Per Day Frequency Relative Frequency
at least 10 400 0.08
3 to 9 1900 0.38
1 to 2 1500 0.30
0 1200 0.24
Total 5000 1.00
www.downloadslide.comslide 172:
Introduction to Probability
EXAMPLE 7 THE ADDITION RULE FOR INDIVIDUAL OUTCOMES
KFI 640 Radio The KFI 640 radio station in Los Angeles is
a combination news/talk and “oldies” station. During a 24hour
day a listener can tune in and hear any of the following four
programs being broadcast:
“Oldies” music
News stories
Talk programming
Commercials
Recently the station has been having trouble with its transmitter. Each day the station’s
signal goes dead for a few seconds it seems that these outages are equally likely to occur at
any time during the 24hour broadcast day. There seems to be no pattern regarding what is
playing at the time the transmitter problem occurs. The station manager is concerned about
the probability that these problems will occur during either a news story or a talk program.
Step 1 Define the experiment.
The station conducts its broadcast starting at 12:00 midnight extending until a
transmitter outage is observed.
Step 2 Define the possible outcomes.
The possible outcomes are the type of programming that is playing when the
transmitter outage occurs. There are four possible outcomes:
e
1
Oldies
e
2
News
e
3
Talk programs
e
4
Commercials
Step 3 Determine the probability of each possible outcome.
The station manager has determined that out of the 1440 minutes per day
540 minutes are oldies 240 minutes are news 540 minutes are talk programs
and 120 minutes are commercials. Therefore the probability of each type of
programming being on at the moment the outage occurs is assessed as follows:
Outcome e
i
P1e
i
2
e
1
Oldies
Pe
.
1
540
1 440
0375
e
2
News
Pe
.
2
240
1 440
0 167
e
3
Talk programs
Pe
.
3
540
1 440
0 375
e
4
Commercials
Pe
.
4
120
1 440
0 083
a
1.000
Note based on Equation 4 Probability Rule 2 the sum of the probabilities of
the individual possible outcomes is 1.0.
Step 4 Define the event of interest.
The event of interest is a transmitter problem occurring during a news or talk
program. This is
E 5e
2
e
3
6
Step 5 Use Probability Rule 3 Equation 5 to compute the desired probability.
P1E2 P1e
2
2 + P1e
3
2
P1E2 0.167 + 0.375
P1E2 0.542
Tsian/Fotolia
www.downloadslide.comslide 173:
Introduction to Probability
Thus there is slightly higher than a 0.5 probability that when a transmitter problem
occurs it will happen during either a news or talk program.
END EXAMPLE
TRY PROBLEM 26
Complement Rule Closely connected with Probability Rules 1 and 2 is the complement
of an event. The complement of event E is represented by E. The Complement Rule is a corol
lary to Probability Rules 1 and 2.
Complement
The complement of an event E is the collection of
all possible outcomes not contained in event E.
Complement Rule
PE PE 1 6
That is the probability of the complement of event E is 1 minus the probability of
event E.
EXAMPLE 8 THE COMPLEMENT RULE
Capital Consulting The managing partner for Capital Consulting is working on a proposal
for a consulting project with a client in Sydney Australia. The manager lists four possible net
profits from the consulting engagement and his subjectively assessed probabilities related to
each profit level.
Outcome POutcome
0 0.70
2000 0.20
15000 0.07
50000 0.03
g 1.00
Note that each probability is between 0 and 1 and that the sum of the probabilities is 1 as
required by Rules 1 and 2.
The manager plans to submit the proposal if the consulting engagement will have a posi
tive profit so he is interested in knowing the probability of an outcome greater than 0. This
probability can be found using the Complement Rule with the following steps:
Step 1 Determine the probabilities for the outcomes.
P1+02 0.70
P1+20002 0.20
P1+150002 0.07
P1+500002 0.03
Step 2 Find the desired probability.
Let E be the consulting outcome event +0. The probability of the zero
outcome is
P1E2 0.70
The complement E is all investment outcomes greater than 0. Using the
Complement Rule the probability of profit greater than 0 is
P1Profit 7 +02 1  P1Profit +02
P1Profit 7 +02 1  0.70
P1Profit 7 +02 0.30
Based on his subjective probability assessment there is a 30 chance the
consulting project will have a positive profit.
END EXAMPLE
TRY PROBLEM 32
www.downloadslide.comslide 174:
Introduction to Probability
Addition Rule for Any Two Events
BUSINESS APPLICATION ADDITION RULE
GOOGLE CONTINUED Suppose the staff who conducted the survey for Google discussed
earlier also asked questions about the computer users’ ages. The Google managers consider
age important in designing their search engine methodologies. Table 3 shows the breakdown
of the sample by age group and by the number of times a user performs a search each day.
Table 3 shows that there are seven events defined. For instance E
1
is the event that a com
puter user performs 10 or more searches per day. This event is composed of three individual
outcomes associated with the three age categories. These are
E
1
5e
1
e
2
e
3
6
In another case event E
5
corresponds to a survey respondent being younger than 30 years
of age. It is composed of four individual outcomes associated with the four levels of search
activity. These are
E
5
5e
1
e
4
e
7
e
10
6
Table 3 illustrates two important concepts in data analysis: joint frequencies and mar
ginal frequencies. Joint frequencies are the values inside the table. They provide information
on age group and search activity jointly. Marginal frequencies are the row and column totals.
These values give information on only the age group or only Google search activity.
For example 2100 people in the survey are in the 30 to 50year age group. This column
total is a marginal frequency for the age group 30 to 50 years which is represented by E
6
.
Now notice that 600 respondents are younger than 30 years old and perform three to nine
searches a day. The 600 is a joint frequency whose outcome is represented by e
4
. The joint
frequencies are the number of times their associated outcomes occur.
Table 4 shows the relative frequencies for the data in Table 3. These values are the prob
abilities of the events and outcomes.
Suppose we wish to find the probability of E
4
0 searches or E
6
being in the 30to50
age group. That is
P1E
4
or E
6
2
Chapter Outcome 2.
TABLE 3  Google Search Study
Age Group
Searches Per Day
E
5
Less than 30
E
6
30 to 50
E
7
Over 50 Total
E
1
Ú 10 Searches e
1
200 e
2
100 e
3
100 400
E
2
3 to 9 Searches e
4
600 e
5
900 e
6
400 1900
E
3
1 to 2 Searches e
7
400 e
8
600 e
9
500 1500
E
4
0 Searches e
10
700 e
11
500 e
12
0 1200
Total 1900 2100 1000 5000
TABLE 4  Google—Joint Probability Table
Age Group
Searches per Day E
5
Less than 30 E
6
30 to 50 E
7
Over 50 Total
E
1
Ú 10 Searches e
1
2005000 0.04 e
2
1005000 0.02 e
3
1005000 0.02 4005000 0.08
E
2
3 to 9 Searches e
4
6005000 0.12 e
5
9005000 0.18 e
6
4005000 0.08 19005000 0.38
E
3
1 to 2 Searches e
7
4005000 0.08 e
8
6005000 0.12 e
9
5005000 0.10 15005000 0.30
E
4
0 Searches e
10
7005000 0.14 e
11
5005000 0.10 e
12
05000 0.00 12005000 0.24
Total 19005000 0.38 21005000 0.42 10005000 0.20 50005000 1.00
www.downloadslide.comslide 175:
Introduction to Probability
FIGURE 1 
Venn Diagram—Addition Rule
for Any Two Events
E
1
E
2
E
1
and E
2
PE
1
or E
2
PE
1
+ PE
2
– PE
1
and E
2
To fnd this probability we must use Probability Rule 4.
Probability Rule 4: Addition Rule for Any Two Events E
1
and E
2
P1E
1
or E
2
2 P1E
1
2 + P1E
2
2  P1E
1
and E
2
2 7
The key word in knowing when to use Rule 4 is or. The word or indicates addition.
You may have covered this concept as a union in a math class. P1E
1
or E
2
2 P1E
1
h E
2
2.
Figure 1 is a Venn diagram that illustrates the application of the Addition Rule for Any Two
Events. Notice that the probabilities of the outcomes in the overlap between the two events
E
1
and E
2
are doublecounted when the probabilities of the outcomes in E
1
are added to those
of E
2
. Thus the probabilities of the outcomes in the overlap which is E
1
and E
2
need to be
subtracted to avoid the double counting.
Referring to the Google situation the probability of E
4
0 searches or E
6
being in the
30to50 age group is
P1E
4
or E
6
2
Table 5 shows the relative frequencies with the events of interest shaded. The overlap corre
sponds to the joint occurrence intersection of conducting 0 searches and being in the 30to
50 age group. The probability of the outcomes in the overlap is represented by P1E
4
and E
6
2
and must be subtracted. This is done to avoid doublecounting the probabilities of the out
comes that are in both E
4
and E
6
when calculating the P1E
4
or E
6
2. Thus
P1E
4
or E
6
2 P1E
4
2 + P1E
6
2P1E
4
and E
6
2
0.24 + 0.42  0.10
0.56
Therefore the probability that a respondent will either be in the 30to50 age group or per
form zero searches on a given day is 0.56.
TABLE 5  Google Searches—Addition Rule Example
Age Group
Searches Per Day E
5
Less than 30 E
6
30 to 50 E
7
Over 50 Total
E
1
Ú 10 Searches e
1
2005000 0.04 e
2
1005000 0.02 e
3
1005000 0.02 4005000 0.08
E
2
3 to 9 Searches e
4
6005000 0.12 e
5
9005000 0.18 e
6
4005000 0.08 19005000 0.38
E
3
1 to 2 Searches e
7
4005000 0.08 e
8
6005000 0.12 e
9
5005000 0.10 15005000 0.30
E
4
0 Searches e
10
7005000 0.14 e
11
5005000 0.10 e
12
05000 0.00 12005000 0.24
Total 19005000 0.38 21005000 0.42 10005000 0.20 50005000 1.00
www.downloadslide.comslide 176:
Introduction to Probability
What is the probability a respondent will perform 1 to 2 searches or be in the over50 age
group Again we can use Rule 4:
P1E
3
or E
7
2 P1E
3
2 + P1E
7
2  P1E
3
and E
7
2
Table 6 shows the relative frequencies for these events. We have
P1E
3
or E
7
2 0.30 + 0.20  0.10 0.40
Thus there is a 0.40 chance that a respondent will perform 1 2 searches or be in the “over
50” age group.
EXAMPLE 9 ADDITION RULE FOR ANY TWO EVENTS
British Columbia Forest Products British Columbia Forest Products manufactures lum
ber for large material supply centers like Home Depot and Lowe’s in the U.S. and Canada.
A representative from Home Depot is due to arrive at the BC plant for a meeting to discuss
lumber quality. When the Home Depot representative arrives he will ask BC’s managers to
randomly select one board from the finished goods inventory for a quality check. Boards of
three dimensions and three lengths are in the inventory. The following chart shows the num
ber of boards of each size and length.
Dimension
Length E
4
2″ 4″ E
5
2″ 6″ E
6
2″ 8″ Total
E
1
8 feet 1400 1500 1100 4000
E
2
10 feet 2000 3500 2500 8000
E
3
12 feet 1600 2000 2400 6000
Total 5000 7000 6000 18000
The BC manager will be selecting one board at random from the inventory to show the Home
Depot representative. Suppose he is interested in the probability that the board selected will
be 8 feet long or a 2″ 6″. To find this probability he can use the following steps:
Step 1 Define the experiment.
One board is selected from the inventory and its dimensions are obtained.
Step 2 Define the events of interest.
The manager is interested in boards that are 8 feet long.
E
1
8 foot boards
He is also interested in the 2″ 6″ dimension so
E
5
2″ 6″ boards
Step 3 Determine the probability for each event.
There are 18000 boards in inventory and 4000 of these are 8 feet long so
PE
.
1
4 000
18 000
0 2222
TABLE 6  Google—Addition Rule Example
Age Group
Searches Per Day E
5
Less than 30 E
6
30 to 50 E
7
Over 50 Total
E
1
Ú 10 Searches e
1
2005000 0.04 e
2
1005000 0.02 e
3
1005000 0.02 4005000 0.08
E
2
3 to 9 Searches e
4
6005000 0.12 e
5
1005000 0.18 e
6
4005000 0.08 19005000 0.38
E
3
1 to 2 Searches e
7
4005000 0.08 e
8
600/5000 0.12 e
9
5005000 0.10 15005000 0.30
E
4
0 Searches e
10
7005000 0.14 e
11
5005000 0.10 e
12
05000 0.00 12005000 0.24
Total 19005000 0.38 21005000 0.42 10005000 0.20 50005000 1.00
www.downloadslide.comslide 177:
Introduction to Probability
Of the 18000 boards 7000 are 2″ 6″ so the probability is
PE
.
5
7 000
18 000
0 3889
Step 4 Determine whether the two events overlap and if so compute the joint
probability.
Of the 18000 total boards 1500 are 8 feet long and 2″ 6″. Thus the joint
probability is
PE E
.
15
1 500
18 000
0 0833 and
Step 5 Compute the desired probability using Probability Rule 4.
P1E
1
or E
5
2 P1E
1
2 + P1E
5
2P1E
1
and E
5
2
P1E
1
or E
5
2 0.2222 + 0.3889  0.0833
0.5278
The chance of selecting an 8foot board or a 2″ 6″ board is just under 0.53.
END EXAMPLE
TRY PROBLEM 31
Addition Rule for Mutually Exclusive Events We indicated previously that when two
events are mutually exclusive both events cannot occur at the same time. Thus for mutually
exclusive events
P1E
1
and E
2
2 0
Therefore when you are dealing with mutually exclusive events the Addition Rule assumes a
different form shown as Rule 5.
Probability Rule 5: Addition Rule for Mutually Exclusive Events
For two mutually exclusive events E
1
and E
2
P1E
1
or E
2
2 P1E
1
2 + P1E
2
2 8
Figure 2 is a Venn diagram illustrating the application of the Addition Rule for Mutually
Exclusive Events.
Conditional Probability
In dealing with probabilities you will often need to determine the chances of two or more
events occurring either at the same time or in succession. For example a quality control man
ager for a manufacturing company may be interested in the probability of selecting two suc
cessive defective products from an assembly line. If the probability of this event is low the
quality control manager will be surprised when it occurs and might readjust the production
process. In other instances the decision maker might know that an event has occurred and
may then want to know the probability of a second event occurring. For instance suppose
that an oil company geologist who believes oil will be found at a certain drilling site makes a
favorable report. Because oil is not always found at locations with a favorable report the oil
FIGURE 2 
Venn Diagram—Addition Rule
for Two Mutually Exclusive
Events
E
1
E
2
PE
1
or E
2
PE
1
+ PE
2
www.downloadslide.comslide 178:
Introduction to Probability
company’s exploration vice president might well be interested in the probability of finding
oil given the favorable report.
Situations such as this refer to a probability concept known as conditional probability.
Probability Rule 6 offers a general rule for conditional probability. The notation
P1E
1
E
2
2 reads “probability of event E
1
given event E
2
has occurred.” Thus the probability
of one event is conditional upon a second event having occurred.
Conditional Probability
The probability that an event will occur given
that some other event has already happened.
Probability Rule 6: Conditional Probability for Any Two Events
For any two events E
1
E
2
PE E
PE E
PE

12
12
2
and
9
where:
P1E
2
2 7 0
Rule 6 uses a joint probability P1E
1
and E
2
2 and a marginal probability P1E
2
2 to
calculate the conditional probability P1E
1
E
2
2. Note that to find a conditional probability
we find the ratio of how frequently E
1
occurs to the total number of observations given that
we restrict our observations to only those cases in which E
2
has occurred.
BUSINESS APPLICATION CONDITIONAL PROBABILITY
SYRINGA NETWORKS Syringa Networks is an Internet service provider to rural areas in
the western United States. The company has studied its customers’ Internet habits. Among the
information collected are the data shown in Table 7.
The company is focusing on highvolume users and one of the factors that will influence
Syringa Networks’ marketing strategy is whether time spent using the Internet is related to a
customer’s gender. For example suppose the company knows a user is female and wants to
know the chances this woman will spend between 20 and 40 hours a month on the Internet. Let
E
2
5e
3
e
4
6 Event: Person uses services 20 to 40 hours per month
E
4
5e
1
e
3
e
5
6 Event: User is female
A marketing analyst needs to know the probability of E
2
given E
4
.
One way to find the desired probability is as follows:
1. We know E
4
has occurred customer is female. There are 850 females in the survey.
2. Of the 850 females 300 use Internet services 20 to 40 hours per month.
3. Then
PE E 
.
24
300
850
035
However we can also apply Rule 6 as follows:
PE E
PE E
PE

24
24
4
and
TABLE 7  Joint Frequency Distribution for Syringa Network
Gender
Hours per Month E
4
Female E
5
Male Total
E
1
6 20 e
1
450 e
2
500 950
E
2
20 to 40 e
3
300 e
4
800 1100
E
3
7 40 e
5
100 e
6
350 450
Total 850 1650 2500
www.downloadslide.comslide 179:
Introduction to Probability
Table 8 shows the relative frequencies of interest. From Table 8 we get the joint
probability P1E
2
and E
4
2 0.12
and
P1E
4
2 0.34
Then applying Rule 6
PE E 
.
.
.
24
012
034
035
EXAMPLE 10 COMPUTING CONDITIONAL PROBABILITIES
Retirement Planning Most financial publications suggest that the older the investor is the
more conservative his or her investment strategy should be. For example younger investors
might hold more in stocks while older investors might hold more in bonds. A recent survey
conducted by a major financial publication yielded the following table which shows the num
ber of people in the study by age group and percentage of retirement funds in the stock market.
Percentage of Retirement Investments in the Stock Market
Age of Investor
E
5
6 5
E
6
5 6 10
E
7
10 6 30
E
8
30 6 50
E
9
50 or more Total
E
1
6 30 years 70 240 270 80 55 715
E
2
30 6 50 years 90 300 630 1120 1420 3560
E
3
50 6 65 years 110 305 780 530 480 2205
E
4
65+ years 200 170 370 260 65 1065
Total 470 1015 2050 1990 2020 7545
The publication’s editors are interested in knowing the probability that someone 65 or older
will have 50 or more of retirement funds invested in the stock market. Assuming the data
collected in this study reflect the population of investors the editors can find this conditional
probability using the following steps:
Step 1 Define the experiment.
A randomly selected person age 65 or older has his or her portfolio analyzed
for percentage of retirement funds in the stock market.
Step 2 Define the events of interest.
In this case we are interested in two events:
E
4
At least 65 years old
E
9
50 or more in stocks
Step 3 Define the probability statement of interest.
The editors are interested in
P1E
9
E
4
2 Probability of 50 or more stocks given at least 65 years
Step 4 Convert the data to probabilities using the relative frequency assessment
method.
We begin with the event that is given to have occurred 1E
4
2. A total of 1065
people in the study were at least 65 years of age. Of the 1065 people 65 had
50 or more of their retirement funds in the stock market.
TABLE 8  Joint Relative Frequency Distribution for Syringa Networks
Gender
Hours per Month E
4
Female E
5
Male Total
E
1
6 20 e
1
4502500 0.18 e
2
5002500 0.20 9502500 0.38
E
2
20 to 40 e
3
3002500 0.12 e
4
8002500 0.14 11002500 0.44
E
3
7 40 e
5
1002500 0.14 e
6
3502500 0.14 4502500 0.18
Total 8502500 0.34 16502500 0.66 25002500 1.00
www.downloadslide.comslide 180:
Introduction to Probability
PE E 
.
94
65
1 065
0 061
Thus the conditional probability that someone at least 65 will have 50 or
more of retirement assets in the stock market is 0.061. This value can be found
using Step 5 as well.
Step 5 Use Probability Rule 6 to find the conditional probability.
PE E
PE E
PE

94
94
4
and
The necessary probabilities are found using the relative frequency assessment method:
PE
.
4
1 065
7 545
0 1412
and the joint probability is
PE E
.
94
65
7 545
0 0086 and
Then using Probability Rule 6 we get
PE E
PE E
PE

.
.
.
94
94
4
0 0086
0 1412
00
and
6 61
END EXAMPLE
TRY PROBLEM 34
Tree Diagrams Another way of organizing the events of an experiment that aids in the
calculation of probabilities is the tree diagram.
BUSINESS APPLICATION USING TREE DIAGRAMS
SYRINGA NETWORKS CONTINUED Figure 3 illustrates the tree diagram for Syringa
Networks the Internet service provider discussed earlier. Note that the branches at each node
in the tree diagram represent mutually exclusive events. Moving from left to right the first
two branches indicate the two customer types male and female—mutually exclusive events.
Three branches grow from each of these original branches representing the three possible
categories for Internet use. The probabilities for the events male and female are shown on the
first two branches. The probabilities shown on the right of the tree are the joint probabilities
for each combination of gender and hours of use. These figures are found using Table 8
which was shown earlier. The probabilities on the branches following the male and female
branches showing hours of use are conditional probabilities. For example we can find the
probability that a male customer 1E
5
2 will spend more than 40 hours on the Internet 1E
3
2 by
PE E
PE E
PE
.
.
.
35
35
5
014
066
0 2121
and
Conditional Probability for Independent Events We previously discussed that two
events are independent if the occurrence of one event has no bearing on the probability that
the second event occurs. Therefore when two events are independent the rule for conditional
probability takes a different form as indicated in Probability Rule 7.
Probability Rule 7: Conditional Probability for Independent Events
For independent events E
1
E
2
PE E P E P E 
12 1 2
0
10
and
PE E P E P E 
21 2 1
0
www.downloadslide.comslide 181:
Introduction to Probability
As Rule 7 shows the conditional probability of one event occurring given a second independ
ent event has already occurred is simply the probability of the event occurring.
EXAMPLE 11 CHECKING FOR INDEPENDENCE
British Columbia Forest Products In Example 9 the manager at the British Columbia
Forest Products Company reported the following data on the boards in inventory:
Dimension
Length
E
4
2″ 4″
E
5
2″ 6″
E
6
2″ 8″ Total
E
1
8 feet 1400 1500 1100 4000
E
2
10 feet 2000 3500 2500 8000
E
3
12 feet 1600 2000 2400 6000
Total 5000 7000 6000 18000
He will be selecting one board at random from the inventory to show a visiting customer.
Of interest is whether the length of the board is independent of the dimension. This can be
determined using the following steps:
Step 1 Define the experiment.
A board is randomly selected and its dimensions determined.
Step 2 Define one event for length and one event for dimension.
Let E
2
Event that the board is 10 feet long and E
5
Event that the board is
a 2″ 6″ dimension.
Step 3 Determine the probability for each event.
PE PE
.
25
8 000
18 000
0 4444
7 000
18 00
and
0 0
0 3889 .
Step 4 Assess the joint probability of the two events occurring.
PE E
.
25
3 500
18 000
0 1944 and
Step 5 Compute the conditional probability of one event given the other using
Probability Rule 6.
PE E
PE E
PE

.
.
.
25
25
5
0 1944
0 3889
05
and
0 0
FIGURE 3 
Tree Diagram for Syringa
Networks
PE
1
and E
5
0.20
PE
1
E
5
20 hours
PE
3
E
5
40 hours
0.4848 PE
2
E
5
20 to 40 hours
Male
PE
5
0.66
Female
PE
4
0.34
PE
2
and E
5
0.32
PE
3
and E
5
0.14
PE
1
and E
4
0.18
PE
2
and E
4
0.12
PE
3
and E
4
0.04
0.3529 PE
2
E
4
20 to 40 hours
PE
1
E
4
20 hours
0.1176 PE
3
E
4
40 hours
0.3030
0.2121
0.5294
www.downloadslide.comslide 182:
Introduction to Probability
Step 6 Check for independence using Probability Rule 7.
Because P1E
2
E
5
2 0.50 7 P1E
2
2 0.4444 the two events board length
and board dimension are not independent.
END EXAMPLE
TRY PROBLEM 42
Multiplication Rule
We needed to find the joint probability of two events in the discussion on addition of two
events and in the discussion on conditional probability. We were able to find P1E
1
and E
2
2
simply by examining the joint relative frequency tables. However we often need to find
P1E
1
and E
2
2 when we do not know the joint relative frequencies. When this is the case we
can use the multiplication rule for two events.
Multiplication Rule for Two Events
Chapter Outcome 3.
Probability Rule 8: Multiplication Rule for Any Two Events
For two events E
1
and E
2
P1E
1
and E
2
2 P1E
1
2P1E
2
E
1
2 11
BUSINESS APPLICATION MULTIPLICATION RULE
HONG KONG FIREWORKS To illustrate how to find a joint probability consider an
example involving the Hong Kong Fireworks company a manufacturer of fireworks used
by cities fairs and other commercial establishments for largescale fireworks displays. The
company uses two suppliers of material used in making a particular product. The materials
from the two suppliers are intermingled on the manufacturing process. When a case of
fireworks is being made the material is pulled randomly from inventory without regard to
which company made it. Recently a customer ordered two products. At the time of assembly
the material inventory contained 30 units of MATX and 50 units of Quinex. What is the
probability that both fireworks products ordered by this customer will have MATX material
To answer this question we must recognize that two events are required to form the
desired outcome. Therefore let
E
1
Event: MATX Material in first product
E
2
Event: MATX Material in second product
The probability that both fireworks products contain MATX material is written as
P1E
1
and E
2
2. The key word here is and as contrasted with the Addition Rule in which the
key word is or. The and signifies that we are interested in the joint probability of two events
as noted by P1E
1
and E
2
2. To find this probability we employ Probability Rule 8.
P1E
1
and E
2
2 P1E
1
2P1E
2
E
1
2
We start by assuming that each unit of material in the inventory has the same chance of
being selected for assembly. For the frst freworks product
PE
1
5
5 5
Number of MATX units
Number of Firework Materials in inventory
30
80
0 375 .
Then because we are not replacing the frst frework material we fnd P1E
2
E
1
2 by
PE E 
21
Number of remaining MATX units
Number of remaining Firework Materials units
29
79
0 3671 .
5
5 5
www.downloadslide.comslide 183:
Introduction to Probability
Now by Rule 8
P1E
1
and E
2
2 P1E
1
2P1E
2
E
1
2 10.375210.36712
0.1377
Therefore there is a 13.77 chance the two freworks products will contain the MA TX material
Using a Tree Diagram
BUSINESS APPLICATION MULTIPLICATION RULE
HONG KONG FIREWORKS CONTINUED A tree diagram can be used to display the
situation facing Hong Kong Fireworks. The company uses material from two suppliers
which is intermingled in the inventory. Recently a customer ordered two products and
found that both contained the MATX material. Assuming that the inventory contains 30
MATX and 50 Quinex units to determine the probability of both products containing
the MATX material you can use a tree diagram. The two branches on the left side of
the tree in Figure 4 show the possible material options for the first product. The two
branches coming from each of the first branches show the possible material options for
the second product. The probabilities at the far right are the joint probabilities for the
material options for the two products. As we determined previously the probability that
both products will contain a MATX unit is 0.1377 as shown on the top right on the tree
diagram.
We can use the Multiplication Rule and the Addition Rule in one application when we
determine the probability that two products will have different materials. Looking at Figure 4
we see there are two ways this can happen.
P31MATX and Quinex2 or 1Quinex and MATX24
If the first product is a MATX and the second one is a Quinex then the first cannot be
a Quinex and the second a MATX. These two events are mutually exclusive and therefore
Rule 5 can be used to calculate the required probability. The joint probabilities generated
from the Multiplication Rule are shown on the right side of the tree. To find the desired prob
ability using Rule 5 we can add the two joint probabilities:
P31MATX and Quinex2 or Quinex and MATX4
0.2373 + 0.2373 0.4746
The chance that a customer buying two products will get two different materials is 47.46.
Multiplication Rule for Independent Events When we determined the probability that
two products would have MATX material we used the general multiplication rule Rule 8.
The general multiplication rule requires that conditional probability be used because the
FIGURE 4 
Tree Diagram for the
Fireworks Product Example
PMATX and MATX 0.375 3 0.3671 0.1377
MATX
P 29/79 0.3671
Quinex
P 50/79 0.6329
MATX
P 30/80 0.375
Quinex
P 50/80 0.625
Product 1 Product 2
PMATX and Quinex 0.375 3 0.6329 0.2373
PQuinex and MATX 0.625 3 0.3797 0.2373
PQuinex and Quinex 0.625 3 0.6203 0.3877
MATX
P 30/79 0.3797
Quinex
P 49/79 0.6203
www.downloadslide.comslide 184:
Introduction to Probability
probability associated with the second product depends on the material selected for the first
product. The chance of obtaining a MATX was lowered from 30/80 to 29/79 given the first
material was a MATX.
However if the two events of interest are independent the imposed condition does
not alter the probability and the Multiplication Rule takes the form shown in Probability
Rule 9.
Probability Rule 9: Multiplication Rule for Independent Events
For independent events E
1
E
2
P1E
1
and E
2
2 P1E
1
2P1E
2
2 12
The joint probability of two independent events is simply the product of the probabilities of
the two events. Rule 9 is the primary way that you can determine whether any two events are
independent. If the product of the probabilities of the two events equals the joint probability
then the events are independent.
EXAMPLE 12 USING THE MULTIPLICATION RULE AND THE ADDITION RULE
Christiansen Accounting Christiansen Accounting prepares tax returns for individuals
and companies. Over the years the firm has tracked its clients and has discovered that 12
of the individual returns have been selected for audit by the Internal Revenue Service. On
one particular day the firm signed two new individual tax clients. The firm is interested in
the probability that at least one of these clients will be audited. This probability can be found
using the following steps:
Step 1 Define the experiment.
The IRS randomly selects a tax return to audit.
Step 2 Define the possible outcomes.
For a single client the following outcomes are defined:
A Audit
N No audit
For each of the clients we define the outcomes as
Client 1: A
1
N
1
Client 2: A
2
N
2
Step 3 Define the overall event of interest.
The event that Christiansen Accounting is interested in is
E At least one client is audited
Step 4 List the outcomes for the events of interest.
The possible outcomes for which at least one client will be audited are as
follows:
E
1
: A
1
A
2
both are audited
E
2
: A
1
N
2
only one client is audited
E
3
: N
1
A
2
Step 5 Compute the probabilities for the events of interest.
Assuming the chances of the clients being audited are independent of each
other probabilities for the events are determined using Probability Rule 9 for
independent events:
P1E
1
2 P1A
1
and A
2
2 0.12 0.12 0.0144
P1E
2
2 P1A
1
and N
2
2 0.12 0.88 0.1056
P1E
3
2 P1N
1
and A
2
2 0.88 0.12 0.1056
www.downloadslide.comslide 185:
Introduction to Probability
Step 6 Determine the probability for the overall event of interest.
Because events E
1
E
2
and E
3
are mutually exclusive compute the probability
of at least one client being audited using Rule 5 the Addition Rule for
Mutually Exclusive Events:
P1E
1
or E
2
or E
3
2 P1E
1
2 + P1E
2
2 + P1E
3
2
0.0144 + 0.1056 + 0.1056
0.2256
The chance of one or both of the clients being audited is 0.2256.
END EXAMPLE
TRY PROBLEM 30
Bayes’ Theorem
As decision makers you will often encounter situations that require you to assess probabili
ties for events of interest. Your assessment may be based on relative frequency or subjectivity.
However you may then come across new information that causes you to revise the prob
ability assessment. For example a human resources manager who has interviewed a person
for a sales job might assess a low probability that the person will succeed in sales. However
after seeing the person’s very high score on the company’s sales aptitude test the manager
might revise her assessment upward. A medical doctor might assign an 80 chance that a
patient has a particular disease. However after seeing positive results from a lab test he might
increase his assessment to 95.
In these situations you will need a way to formally incorporate the new informa
tion. One very useful tool for doing this is called Bayes’ Theorem which is named
for the Reverend Thomas Bayes who developed the special application of conditional
probability in the 1700s. Letting event B be an event that is given to have occurred the
conditional probability of event E
i
occurring can be computed as shown earlier using
Equation 9:
PE B
PE B
PB
i
i

and
The numerator can be reformulated using the Multiplication Rule Equation 11 as
P1E
i
and B2 P1E
i
2P1B E
i
2
The conditional probability is then
PE B
PE P B E
PB
i
ii


The denominator P1B2 can be found by adding the probability of the k possible ways that
event B can occur. This is
P1B2 P1E
1
2P1B E
1
2 + P1E
2
2P1B E
2
2 + g+ P1E
k
2P1B E
k
2
Then Bayes’ Theorem is formulated as Equation 13.
Chapter Outcome 4.
Bayes’ Theorem
PE B
PE P B E
PE P B E P E P B E
i
ii


 
11 2 2 2

...
PE P B E
kk
13
where:
E
i
ith event of interest of the k possible events
B Event that has occurred that might impact P1E
i
2
Events E
1
to E
k
are mutually exclusive and collectively exhaustive.
www.downloadslide.comslide 186:
Introduction to Probability
BUSINESS APPLICATION BAYES’ THEOREM
TILE PRODUCTION The Glerum Tile and Flooring Company
has two production facilities one in Ohio and one in Virginia. The
company makes the same type of tile at both facilities. The Ohio
plant makes 60 of the company’s total tile output and the Virginia
plant 40. All tiles from the two facilities are sent to regional
warehouses where they are intermingled. After extensive study
the quality assurance manager has determined that 5 of the tiles
produced in Ohio and 10 of the tiles produced in Virginia are unusable due to quality problems.
When the company sells a defective tile it incurs not only the cost of replacing the item but
also the loss of goodwill. The vice president for production would like to allocate these costs
fairly between the two plants. To do so he knows he must first determine the probability that a
defective tile was produced by a particular production line. Specifically he needs to answer these
questions:
1. What is the probability that the tile was produced at the Ohio plant given that the tile is
defective
2. What is the probability that the tile was produced at the Virginia plant given that the tile is
defective
In notation form with D representing the event that an item is defective what the manager
wants to know is
P1Ohio D2
P1Virginia D2
We can use Bayes’ Theorem Equation 13 to determine these probabilities as follows:
PD
PPD
PD


Ohio
Ohio Ohio
We know that event DDefective tile can happen if it is made in either Ohio or Virginia.
Thus
P1D2 P1Ohio and Defective2 + P1Virginia and Defective2
P1D2 P1Ohio2P1D Ohio2 + P1Virginia2P1D Virginia2
We already know that 60 of the tiles come from Ohio and 40 from Virginia. So
P1Ohio2 0.60 and P1Virginia2 0.40. These are called the prior probabilities. Without
Bayes’ Theorem we would likely allocate the total cost of defects in a 60/40 split between
Ohio and Virginia based on total production. However the new information about the
quality from each line is
P1D Ohio2 0.05 and P1D Virginia2 0.10
which can be used to properly allocate the cost of defects. This is done using Bayes’ Theorem.
PD
PPD
PPD



Ohio
Ohio Ohio
Ohio OhioPPD  Virginia Virginia
then
PD 
. .
. . .
Ohio
060 0 05
060 005 040 0 ..
.
10
0 4286
and
PD
PPD
P


Virginia
Virginia Virginia
Virg ginia Virginia Ohio Ohio
Virg
 
PD P P D
P i inia 
. .
. . . .
D
040 010
040 010 060 0 0 05
0 5714
.
These probabilities are revised probabilities. The prior probabilities have been revised given
the new quality information. We now see that 42.86 of the cost of defects should be al
located to the Ohio plant and 57.14 should be allocated to the Virginia plant.
severija kirilovaite/Fotolia
www.downloadslide.comslide 187:
Introduction to Probability
TABLE 9  Bayes’ Theorem Calculations for Glerum Tile and Flooring
Events
Prior
Probabilities
Conditional
Probabilities Joint Probabilities Revised Probabilities
Ohio 0.60 0.05 10.60210.052 0.03 0.030.07 0.4286
Virginia 0.40 0.10 10.40210.102 0.04 0.040.07 0.5714
0.07 1.0000
Note the denominator P1D2 is the overall probability of a defective tile. This pro
bability is
P1D2 P1Ohio2P1D Ohio2 + P1Virginia2P1D Virginia2
10.60210.052 + 10.40210.102
0.03 + 0.04
0.07
Thus 7 of all the tiles made by Glerum are defective.
You might prefer to use a tabular approach like that shown in Table 9 when you apply
Bayes’ Theorem.
EXAMPLE 13 BAYES’ THEOREM
Techtronics Equipment Corporation The Techtronics Equipment Corporation has
developed a new electronic device that it would like to sell to the U.S. military for use
in fighter aircraft. The sales manager believes there is a 0.60 chance that the military
will place an order. However after making an initial sales presentation military officials
will often ask for a second presentation to other military decision makers. Historically
70 of successful companies are asked to make a second presentation whereas 50 of
unsuccessful companies are asked back a second time. Suppose Techtronics Equipment
has just been asked to make a second presentation what is the revised probability that
the company will make the sale This probability can be determined using the following
steps:
Step 1 Define the events.
In this case there are two events:
S Sale N No sale
Step 2 Determine the prior probabilities for the events.
The probability of the events prior to knowing whether a second presentation
will be requested are
P1S2 0.60 P1N2 0.40
Step 3 Define an event that if it occurs could alter the prior probabilities.
In this case the altering event is the invitation to make a second presentation.
We label this event as SP.
Step 4 Determine the conditional probabilities.
The conditional probabilities are associated with being invited to make a
second presentation:
P1SP S2 0.70 P1SP N2 0.50
Step 5 Use the tabular approach for Bayes’ Theorem to determine the revised
probabilities.
These correspond to
P1S SP2 and P1N SP2
www.downloadslide.comslide 188:
Introduction to Probability
Skill Development
426. Based on weather data collected in Racine Wisconsin
on Christmas Day the weather had the following
distribution:
Event Relative Frequency
Clear dry 0.20
Cloudy dry 0.30
Rain 0.40
Snow 0.10
a. Based on these data what is the probability that
next Christmas will be dry
b. Based on the data what is the probability that next
Christmas will be rainy or cloudy and dry
c. Supposing next Christmas is dry determine the
probability that it will also be cloudy.
427. The Jack In The Box franchise in Bangor Maine
has determined that the chance a customer will order
a soft drink is 0.90. The probability that a customer
will order a hamburger is 0.60. The probability that a
customer will order french fries is 0.50.
a. If a customer places an order what is the probability
that the order will include a soft drink and no fries if
these two events are independent
b. The restaurant has also determined that if a
customer orders a hamburger the probability the
customer will also order fries is 0.80. Determine the
probability that the order will include a hamburger
and fries.
428. Ponderosa Paint and Glass carries three brands of
paint. A customer wants to buy another gallon of paint
to match paint she purchased at the store previously.
She can’t recall the brand name and does not wish to
return home to find the old can of paint. So she selects
two of the three brands of paint at random and buys
them.
Event
Prior
Probabilities
Conditional
Probabilities
Joint
Probabilities
Revised
Probabilities
S Sale 0.60
P1SP S2 0.70 P1S2P1SP S210.60210.702 0.42 0.420.62 0.6774
N No sale 0.40
P1SP N2 0.50 P1N2P1SP N210.40210.502 0.20 0.200.62 0.3226
0.62 1.0000
Thus using Bayes’ Theorem if Techtronics Equipment gets a second presentation
opportunity the probability of making the sale is revised upward from 0.60 to 0.6774.
END EXAMPLE
TRY PROBLEM 33
MyStatLab
42: Exercises
a. What is the probability that she matched the paint
brand
b. Her husband also goes to the paint store and fails to
remember what brand to buy. So he also purchases
two of the three brands of paint at random.
Determine the probability that both the woman and
her husband fail to get the correct brand of paint.
Hint: Are the husband’s selections independent of
his wife’s selections
429. The college basketball team at West Texas State
University has 10 players 5 are seniors 2 are juniors
and 3 are sophomores. Two players are randomly
selected to serve as captains for the next game. What is
the probability that both players selected are seniors
430. Micron Technology has sales offices located in four
cities: Dallas Seattle Boston and Los Angeles. An
analysis of the company’s accounts receivables reveals
the number of overdue invoices by days as shown
here.
Days Overdue Dallas Seattle Boston Los Angeles
Under 30 days 137 122 198 287
30–60 days 85 46 76 109
61–90 days 33 27 55 48
Over 90 days 18 32 45 66
Assume the invoices are stored and managed from a
central database.
a. What is the probability that a randomly selected
invoice from the database is from the Boston sales
office
b. What is the probability that a randomly selected
invoice from the database is between 30 and 90 days
overdue
c. What is the probability that a randomly selected
invoice from the database is over 90 days old and
from the Seattle office
www.downloadslide.comslide 189:
Introduction to Probability
d. If a randomly selected invoice is from the Los
Angeles office what is the probability that it is 60
or fewer days overdue
431. Three events occur with probabilities
P1E
1
2 0.35 P1E
2
2 0.15 P1E
3
2 0.40.
If the event B occurs the probability becomes
P1E
1
B2 0.25 P1B2 0.30.
a. Calculate P1E
1
and B2
b. Compute P1E
1
or B2
c. Assume that E
1
E
2
and E
3
are independent events.
Calculate P1E
1
and E
2
and E
3
2.
432. The URS construction company has submitted two
bids one to build a large hotel in London and the other
to build a commercial office building in New York
City. The company believes it has a 40 chance of
winning the hotel bid and a 25 chance of winning
the office building bid. The company also believes that
winning the hotel bid is independent of winning the
office building bid.
a. What is the probability the company will win both
contracts
b. What is the probability the company will win at
least one contract
c. What is the probability the company will lose both
contracts
433. Suppose a quality manager for Dell Computers has
collected the following data on the quality status of
disk drives by supplier. She inspected a total of 700
disk drives.
Drive Status
Supplier Working Defective
Company A 120 10
Company B 180 15
Company C 50 5
Company D 300 20
a. Based on these inspection data what is the
probability of randomly selecting a disk drive from
company B
b. What is the probability of a defective disk drive
being received by the computer company
c. What is the probability of a defect given that
company B supplied the disk drive
434. Three events occur with probabilities of P1E
1
2 0.35
P1E
2
2 0.25 P1E
3
2 0.40. Other probabilities
are: P1B E
1
2 0.25 P1B E
2
2 0.15 P1B E
3
2
0.60.
a. Compute P1E
1
B2.
b. Compute P1E
2
B2.
c. Compute P1E
3
B2.
435. Men have a reputation for not wanting to ask for
directions. A Harris study conducted for Lincoln
Mercury indicated that 42 of men and 61 of
women would stop and ask for directions. The U.S.
Census Bureau’s 2007 population estimate was that
for individuals 18 or over 48.2 were men and 51.8
were women. This exercise addresses this age group.
a. A randomly chosen driver gets lost on a road trip.
Determine the probability that the driver is a woman
and stops to ask for directions.
b. Calculate the probability that the driver stops to ask
for directions.
c. Given that a driver stops to ask for directions
determine the probability that the driver was a man.
Business Applications
436. A local FedEx/Kinkos has three blackandwhite copy
machines and two color copiers. Based on historical
data the chance that each blackandwhite copier will
be down for repairs is 0.10. The color copiers are more
of a problem and are down 20 of the time each.
a. Based on this information what is the probability
that if a customer needs a color copy both color
machines will be down for repairs
b. If a customer wants both a color copy and a black
andwhite copy what is the probability that the
necessary machines will be available Assume that
the color copier can also be used to make a black
andwhite copy if needed.
c. If the manager wants to have at least a 99 chance
of being able to furnish a blackandwhite copy on
demand is the present configuration sufficient
Assume that the color copier can also be used to
make a blackandwhite copy if needed. Back
up your answer with appropriate probability
computations.
d. What is the probability that all five copiers will
be up and running at the same time Suppose the
manager added a fourth blackandwhite copier
how would the probability of all copiers being ready
at any one time be affected
437. Suppose the managers at FedEx/Kinkos wish to meet
the increasing demand for color photocopies and to
have more reliable service. Refer to Problem 36. As a
goal they would like to have at least a 99.9 chance of
being able to furnish a blackandwhite copy or a color
copy on demand. They also wish to purchase only four
copiers. They have asked for your advice regarding the
mix of blackandwhite and color copiers. Supply them
with your advice. Provide calculations and reasons to
support your advice.
438. The Snappy Service gas station manager is thinking
about a promotion that she hopes will bring in more
business to the fullservice island. She is considering
the option that when a customer requests a fillup if
the pump stops with the dollar amount at 19.99 the
customer will get the gasoline free. Previous studies
show that 70 of the customers require more than
20.00 when they fill up so would not be eligible for
the free gas. What is the probability that a customer
will get free gas at this station if the promotion is
implemented
www.downloadslide.comslide 190:
Introduction to Probability
439. Suppose the manager in Problem 38 is concerned about
alienating customers who buy more than 20.00 since
they would not be eligible to win the free gas under
the original concept. To overcome this she is thinking
about changing the contest. The customer will get free
gas if any of the following happens:
21.11 22.22 23.33 24.44 25.55 26.66
27.77 28.88 29.99
Past data show that only 5 of all customers require
30.00 or more. If one of these bigvolume customers
arrives he will get to blindly draw a ball from a box
containing 100 balls 99 red 1 white. If the white ball
is picked the customer gets his gas free. Considering
this new promotion what is the probability that a
customer will get free gas
440. Hubble Construction Company has submitted a bid
on a state government project that is to be funded by
the federal government’s stimulus money in Arizona.
The price of the bid was predetermined in the bid
specifications. The contract is to be awarded on the
basis of a blind drawing from those who have bid. Five
other companies have also submitted bids.
a. What is the probability of the Hubble Construction
Company winning the bid
b. Suppose that there are two contracts to be awarded
by a blind draw. What is the probability of Hubble
winning both contracts Assume sampling with
replacement.
c. Referring to part b what is the probability of
Hubble not winning either contract
d. Referring to part b what is the probability of
Hubble winning exactly one contract
441. Drake Marketing and Promotions has randomly
surveyed 200 men who watch professional sports. The
men were separated according to their educational
level college degree or not and whether they preferred
the NBA or the National Football League NFL. The
results of the survey are shown:
Sports
Pr efer ence
College
Degree
No College
Degree
NBA 40 55
NFL 10 95
a. What is the probability that a randomly selected
survey participant prefers the NFL
b. What is the probability that a randomly selected
survey participant has a college degree and prefers
the NBA
c. Suppose a survey participant is randomly selected
and you are told that he has a college degree.
What is the probability that this man prefers the
NFL
d. Is a survey participant’s preference for the NBA
independent of having a college degree
442. Until the summer of 2008 the real estate market in
Fresno California had been booming with prices
skyrocketing. Recently a study showed the sales
patterns in Fresno for singlefamily homes. One chart
presented in the commission’s report is reproduced
here. It shows the number of homes sold by price range
and number of days the home was on the market.
Days on the Market
Price Range 000 178 30 Over 30
Under 200 125 15 30
200–500 200 150 100
501–1000 400 525 175
Over 1000 125 140 35
a. Using the relative frequency approach to probability
assessment what is the probability that a house will
be on the market more than 7 days
b. Is the event 17 days on the market independent of
the price 200–500
c. Suppose a home has just sold in Fresno and was on
the market less than 8 days what is the most likely
price range for that home
443. Vegetables from the summer harvest are currently
being processed at Skone and Conners Foods Inc. The
manager has found a case of cans that have not been
properly sealed. There are three lines that processed
cans of this type and the manager wants to know
which line is most likely to be responsible for this
mistake. Provide the manager this information.
Line
Contribution
to Total
Proportion
Defective
1 0.40 0.05
2 0.35 0.10
3 0.25 0.07
444. A corporation has 11 manufacturing plants. Of
these seven are domestic and four are outside the
United States. Each year a performance evaluation is
conducted for four randomly selected plants. What
is the probability that a performance evaluation will
include at least one plant outside the United States
Hint: Begin by finding the probability that only
domestic plants are selected.
445. Parts and materials for the skis made by the Downhill
Adventures Company are supplied by two suppliers.
Supplier A’s materials make up 30 of what is used
with Supplier B providing the rest. Past records
indicate that 15 of Supplier A’s materials are
defective and 10 of B’s are defective. Since it is
impossible to tell which supplier the materials came
from once they are in inventory the manager wants to
know which supplier more likely supplied the defective
materials the foreman has brought to his attention.
Provide the manager this information.
www.downloadslide.comslide 191:
Introduction to Probability
446. A major electronics manufacturer has determined that
when one of its televisions is sold there is 0.08 chance
that the set will need service before the warranty
period expires. It has also assessed a 0.05 chance that a
DVD player will need service prior to the expiration of
the warranty.
a. Suppose a customer purchases one of the company’s
televisions and one of the DVD players. What is
the probability that at least one of the products will
require service prior to the warranty expiring
b. Suppose a retailer sells four televisions on a
particular Saturday. What is the probability that
none of the four will need service prior to the
warranty expiring
c. Suppose a retailer sells four televisions on a
particular Saturday. What is the probability that at
least one will need repair
447. The Committee for the Study of the American
Electorate indicated that 60.7 of the votingage
voters cast ballots in the 2004 presidential election. It
also indicated that 85.3 of registered voters voted in
the election. The percentage of those who voted for
President Bush was 50.8.
a. Determine the proportion of votingage voters who
voted for President Bush.
b. Determine the proportion of votingage voters who
were registered to vote.
448. A distributor of outdoor yard lights has four suppliers.
This past season she purchased 40 of the lights from
Franklin Lighting 30 from Wilson Sons 20
from Evergreen Supply and the rest from A. L. Scott.
In prior years 3 of Franklin’s lights were defective
6 of the Wilson lights were defective 2 of
Evergreen’s were defective and 8 of the Scott lights
were defective. When the lights arrive at the distributor
she puts them in inventory without identifying the
supplier. Suppose that a defective light string has been
pulled from inventory what is the probability that it
was supplied by Franklin Lighting
449. USA Today reported “Study Finds Better Survival
Rates at ‘HighV olume’ Hospitals” that “high
volume” hospitals performed at least 77 of bladder
removal surgeries “lowvolume” hospitals performed
at most 23. Assume the percentages are 77 and
23. In the first two weeks after surgery 3.1 of
patients at lowvolume centers died compared to 0.7
at the highvolume hospitals.
a. Calculate the probability that a randomly chosen
bladdercancer patient had surgery at a highvolume
hospital and survived the first two weeks after
surgery.
b. Calculate the probability that a randomly chosen
bladdercancer patient survived the first two weeks
after surgery.
c. If two bladdercancer patients were chosen
randomly determine the probability that only one
would survive the first two weeks after surgery.
d. If two bladdercancer patients were chosen
randomly determine the probability that at least one
would survive the first two weeks after surgery.
450. Suppose an auditor has 18 tax returns 12 of which
are for physicians. If three of the 18 tax returns are
randomly selected then what is the probability that at
least one of the three selected will be a physician’s tax
return
451. A box of 50 remote control devices contains 3 that
have a defective power button. If devices are randomly
sampled from the box and inspected one at a time
determine
a. The probability that the first control device is
defective.
b. The probability that the first control device is good
and the second control device is defective.
c. The probability that the first three sampled devices
are all good.
Computer Database Exercises
452. The data file Colleges contains data on more than
1300 colleges and universities in the United States.
Suppose a company is planning to award a significant
grant to a randomly selected college or university.
Using the relative frequency method for assessing
probabilities and the rules of probability respond to the
following questions. If data are missing for a needed
variable reduce the number of colleges in the study
appropriately.
a. What is the probability that the grant will go to a
private college or university
b. What is the probability that the grant will go to a
college or university that has a student/faculty ratio
over 20
c. What is the probability that the grant will go to a
college or university that is both private and has a
student/faculty ratio over 20
d. If the company decides to split the grant into two
grants what is the probability that both grants will
go to California colleges and universities What
might you conclude if this did happen
453. A Courtyard Hotel by Marriott conducted a survey
of its guests. Sixtytwo surveys were completed.
Based on the data from the survey found in the
file CourtyardSurvey determine the following
probabilities using the relative frequency assessment
method.
a. Of two customers selected what is the probability
that both will be on a business trip
b. What is the probability that a customer will be on
a business trip or will experience a hotel problem
during a stay at the Courtyard
c. What is the probability that a customer on business
has an instate area code phone number
d. Based on the data in the survey can the Courtyard
manager conclude that a customer’s rating regarding
staff attentiveness is independent of whether he or
www.downloadslide.comslide 192:
Introduction to Probability
she is traveling on business pleasure or both Use
the rules of probability to make this determination.
454. Continuing with the Marriott survey done by
the managers of a Marriott Courtyard Hotel
based on the data from the survey found in the
file CourtyardSurvey determine the following
probabilities using the relative frequency assessment
method.
a. Of two customers selected what is the probability
that neither will be on a business trip
b. What is the probability that a customer will be on a
business trip or will not experience a hotel problem
during a stay at the Courtyard
c. What is the probability that a customer on a
pleasure trip has an instate area code phone
number
455. A Harris survey asked in part what the most important
reason was that people give for not using a wireless
phone exclusively. The responses were: 1 Like the
safety of traditional phone 2 Need line for Internet
access 3 Pricing not attractive enough 4 Weak or
unreliable cell signal at home 5 Coverage not good
enough and 6 Other. The file titled Wireless contains
the responses for the 1088 respondents.
a. Of those respondents 36 or older determine the
probability that an individual in this age group
would not use a wireless phone exclusively because
of some type of difficulty in placing and receiving
calls with a wireless phone.
b. If three respondents were selected at random from
those respondents younger than 36 calculate the
probability that at least one of the respondents
stated the most important reason for not using a
wireless exclusively was that they need a line for
Internet access.
456. A recent news release published by Ars Technica LLD
presented the results of a study concerning the world
and domestic market share for the major manufacturers
of personal computers PCs. The file titled PCMarket
contains a sample that would produce the market
shares alluded to in the article and the highest
academic degrees achieved by the owners of those PCs.
a. Determine the probability that the person had
achieved at least a bachelor’s degree and owns a
Dell PC.
b. If a randomly selected person owned a Dell PC
determine the probability that the person had
achieved at least a bachelor’s degree.
c. Consider these two events: 1 At least a bachelor’s
degree and 2 Owns a Dell PC. Are these events
independent dependent or mutually exclusive
Explain.
457. PricewaterhouseCoopers Saratoga in its 2005/2006
Human Capital Index Report indicated the average
number of days it took for an American company
to fill a job vacancy in 2004 was 48 days. Sample
data similar to those used in the study are in a file
titled Hired. Categories for the days and hire cost
are provided under the headings “Time” and “Cost”
respectively.
a. Calculate the probability that a company vacancy
took at most 100 days or cost at most 4000 to fill.
b. Of the vacancies that took at most 100 days to fill
calculate the probability that the cost was at most
4000.
c. If three of the vacancies were chosen at random
calculate the probability that two of the vacancies
cost at most 4000 to fill.
458. A company produces scooters used by small
businesses such as pizza parlors that find them
convenient for making short deliveries. The
company is notified whenever a scooter breaks
down and the problem is classified as being either
mechanical or electrical. The company then matches
the scooter to the plant where it was assembled.
The file Scooters contains a random sample of
200 breakdowns. Use the data in the file to find the
following probabilities.
a. If a scooter was assembled in the Tyler plant what
is the probability its breakdown was due to an
electrical problem
b. Is the probability of a scooter having a mechanical
problem independent of the scooter being
assembled at the Lincoln plant
c. If mechanical problems are assigned a cost of
75 and electrical problems are assigned a cost
of 100 how much cost would be budgeted for
the Lincoln and Tyler plants next year if a total
of 500 scooters were expected to be returned for
repair
END EXERCISES 42
www.downloadslide.comslide 193:
Introduction to Probability
Visual Summary
1 The Basics of Probability
2 The Rules of Probability
Outcome 2. Be able to apply the Addition Rule.
Outcome 3. Know how to use the Multiplication Rule.
Outcome 4. Know how to use Bayes’ Theorem for applications involving
conditional probabilities
Conclusion
Probability is how we measure our uncertainty about whether an outcome will occur.
The closer the probability assessment is to 1.0 or 0.0 the more certain we are that
event will or will not occur. Assessing probabilities and then using those probabilities
to help make decisions is a central part of what business decisionmakers do on a
regular basis. This chapter has introduced the fundamentals of probability and the rules
that are used when working with probability.
Probability is used in our everyday lives and in business decisionmaking all the time. You
might base your decision to call ahead for dinner reservations based on your assessment of
the probability of having to wait for seating. A company may decide to switch suppliers
based on their assessment of the probability that the new supplier will provide higher
quality products or services. Probability is the way we measure our uncertainty about
events. However in order to properly use probability you need to know the probability
rules and the terms associated with probability.
Summary
In order to effectively use probability it is important to understand key concepts and terminology.
Some of the most important of these are discussed in section 1 including sample space
dependent and independent events and mutually exclusive events. Probabilities are
assessed in three main ways classical assessment relative frequency assessment
and subjective assessment.
Outcome 1. Understand the three approaches to assessing probabilities.
Summary
To effectively work with probability it is important to know the probability rules. Section 2
introduces nine rules including three addition rules and two multiplication rules. Rules
for conditional probability and the complement rule are also very useful. Bayes’
Theorem is used to calculate conditional probabilities in situations where the probability of the
given event is not provided and must be calculated.
www.downloadslide.comslide 194:
Introduction to Probability
8 Probability Rule 5
Addition rule for mutually exclusive events E
1
E
2
:
P1E
1
or E
2
2 P1E
1
2 + P1E
2
2
9 Probability Rule 6
Conditional probability for any two events E
1
E
2
:
PE E
PE E
PE

12
12
2
and
10 Probability Rule 7
Conditional probability for independent events E
1
E
2
:
PE E P E P E 
12 1 2
0
and
PE E P E P E 
21 2 1
0
11 Probability Rule 8
Multiplication rule for any two events E
1
and E
2
:
P1E
1
and E
2
2 P1E
1
2P1E
2
E
1
2
12 Probability Rule 9
Multiplication rule for independent events E
1
E
2
:
P1E
1
and E
2
2 P1E
1
2P1E
2
2
13 Bayes’ Theorem
PE B
PE P B E
PE P B E P E P B E
i
ii


 
11 2 2 2

...
PE P B E
kk
1 Classical Probability Assessment
PE
E
i
i
Number of ways can occur
Total numbe er of possible outcomes
2 Relative Frequency Assessment
PE
E
N
i
i
Number of times occurs
3 Probability Rule 1
0 … P1E
i
2 … 1 for all i
4 Probability Rule 2
Pe
i
i
k
1
1
S
5 Probability Rule 3
Addition rule for individual outcomes:
The probability of an event E
i
is equal to the sum of the
proba bilities of the possible outcomes forming E
i
. For example if
E
i
5e
1
e
2
e
3
6
then
P1E
i
2 P1e
1
2 + P1e
2
2 + P1e
3
2
6 Complement Rule
PE PE 1
7 Probability Rule 4
Addition rule for any two events E
1
and E
2
:
P1E
1
or E
2
2 P1E
1
2 + P1E
2
2P1E
1
and E
2
2
Equations
Key Terms
Classical probability assessment
Complement
Conditional probability
Dependent events
Event
Experiment
Independent events
Mutually exclusive events
Probability
Relative frequency assessment
Sample space
Subjective probability assessment
Chapter Exercises
Conceptual Questions
459. Discuss what is meant by classical probability
assessment and indicate why classical assessment is not
often used in business applications.
460. Discuss what is meant by the relative frequency
assessment approach to probability assessment. Provide
a businessrelated example other than the one given in
the text in which this method of probability assessment
might be used.
461. Discuss what is meant by subjective probability.
Provide a businessrelated example in which subjective
probability assessment would likely be used. Also
provide an example of when you have personally used
subjective probability assessment.
462. Examine the relationship between independent
dependent and mutually exclusive events. Consider two
events A and B that are mutually exclusive such that
P1A2 ≠ 0.
a. Calculate P1A B2.
b. What does your answer to part a say about whether
two mutually exclusive events are dependent or
independent
c. Consider two events C and D such that
P1C2 0.4 and P1C D2 0.15. 1 Are events
MyStatLab
www.downloadslide.comslide 195:
Introduction to Probability
b. Calculate the probability that a randomly chosen
U.S. adult does not take a multiple formula.
c. If three U.S. adults were chosen at random compute
the probability that only two of them take a multiple
formula as their primary source.
467. USA Today reported the IRS audited 1 in 63 wealthy
individuals and families about 1 of every 107
individuals 20 of corporations in general and 44
of the largest corporations with assets of at least 250
million.
a. Calculate the probability that at least 1 wealthy
individual in a sample of 10 would be audited.
b. Compute the probability that at least 1 from a
sample of 10 corporations with assets of at least
250 million would be audited.
c. Calculate the probability that a randomly chosen
wealthy CEO of a corporation with assets of 300
million would be audited or that the corporation
would be audited.
468. Simmons Furniture Company is considering changing
its starting hour from 8:00 a.m. to 7:30 a.m. A census
of the company’s 1200 office and production workers
shows 370 of its 750 production workers favor the
change and a total of 715 workers favor the change.
To further assess worker opinion the region manager
decides to talk with random workers.
a. What is the probability a randomly selected worker
will be in favor of the change
b. What is the probability a randomly selected worker
will be against the change and be an office worker
c. Are the events job type and opinion independent
Explain.
469. A survey released by the National Association of
Convenience Stores NACS indicated that 70 of gas
purchases paid for at the pump were made with a credit
or debit card.
a. Indicate the type of probability assessment method
that NACS would use to assess this probability.
b. In one local store 10 randomly chosen customers
were observed. All 10 of these customers used a
credit or a debit card. If the NACS statistic applies
to this area determine the probability that 10 out of
10 customers would use a credit or debit card.
c. If 90 of gas purchases paid for at the pump were
made with a credit or debit card determine the
probability that 10 out of 10 customers would use a
credit or debit card.
d. Based on your answers to parts b and c does it
appear that a larger percentage of local individuals
use credit or debit cards than is true for the nation as
a whole Explain.
470. Ponderosa Paint and Glass makes paint at three plants.
It then ships the unmarked paint cans to a central
warehouse. Plant A supplies 50 of the paint and past
records indicate that the paint is incorrectly mixed 10
of the time. Plant B contributes 30 with paint mixed
C and D mutually exclusive 2 Are events C and
D independent or dependent Are dependent events
necessarily mutually exclusive events
463. Consider the following table:
A A Totals
B 800 200 1000
B
600 400 1000
Totals 1400 600 2000
Explore the complements of conditional events:
a. Calculate the following probabilities: P1A B2
P1A B2 P1A B2 P1A B2.
b. Now determine which pair of events are
complements of each other. Hint: Use the
probabilities calculated in part a and the
Complement Rule.
464. Examine the following table:
A A Totals
B 200 800 1000
B
300 700 1000
Totals 500 1500 2000
a. Calculate the following probabilities: P1A2 P1A2
P1A B2 P1A B2 P1A B2 and P1A B2.
b. Show that 1 A and B 2 A and B 3 A and B
4 A and B are dependent events.
Business Applications
465. An accounting professor at a state university in V ermont
recently gave a threequestion multiplechoice quiz.
Each question had four optional answers.
a. What is the probability of getting a perfect score if
you were forced to guess at each question
b. Suppose it takes at least two correct answers out of
three to pass the test. What is the probability of passing
if you are forced to guess at each question What does
this indicate about studying for such an exam
c. Suppose through some latenight studying you are
able to correctly eliminate two answers on each
question. Now answer parts a and b.
466. Simmons Market Research conducted a national
consumer study of 13787 respondents. A subset of the
respondents was asked to indicate the primary source
of the vitamins or mineral supplements they consume.
Six out of 10 U.S. adults take vitamins or mineral
supplements. Of those who do 58 indicated a multiple
formula was their choice.
a. Calculate the probability that a randomly chosen
U.S. adult takes a multiple formula as her or his
primary source.
www.downloadslide.comslide 196:
Introduction to Probability
b. What is the probability that a randomly selected
executive from this group is a male whose
functional background is marketing
c. Assume that an executive is selected and you are
told that the executive’s functional background
was in operations. What is the probability that this
executive is a female
d. Assume that an executive is selected and you
are told that the executive is a female. What is
the probability the executive’s functional area is
marketing
e. Are gender and functional background independent
for this set of executives
474. A manufacturing firm has two suppliers for an electrical
component used in its process: one in Mexico and one
in China. The supplier in Mexico ships 82 of all the
electrical components used by the firm and has a defect
rate of 4. The Chinese supplier ships 18 of the
electrical components used by the firm and has a defect
rate of 6.
a. Calculate the probability that an electrical
component is defective.
b. Suppose an electrical component is defective. What
is the probability that component was shipped from
Mexico Hint: Use Bayes’ theorem.
Computer Database Exercises
475. A survey of 300 CEOs was conducted in which the
CEOs were to list their corporation’s geographical
location: Northeast NE Southeast SE Midwest
MW Southwest SW and West W. They were also
requested to indicate their company’s industrial type:
Communication C Electronics E Finance F and
Manufacturing M. The file titled CEOInfo contains
sample data similar to that used in this study.
a. Determine the probability that a randomly chosen
CEO would have a corporation in the West.
b. Compute the probability that a randomly chosen
CEO would have a corporation in the West and head
an electronics corporation.
c. Calculate the probability that a randomly chosen
CEO would have a corporation in the East or head a
communications corporation.
d. Of the corporations located in the East calculate
the probability that a randomly selected CEO would
head a communications corporation.
476. The ECCO company makes backup alarms for
machinery like forklifts and commercial trucks. When a
customer returns one of the alarms under warranty the
quality manager logs data on the product. From the data
available in the file named Ecco use relative frequency
to find the following probabilities.
a. What is the probability the product was made at the
Salt Lake City plant
b. What is the probability the reason for the return was
a wiring problem
incorrectly 5 of the time. Plant C supplies 20 with
paint mixed incorrectly 20 of the time. If Ponderosa
guarantees its product and spent 10000 replacing
improperly mixed paint last year how should the cost be
distributed among the three plants
471. Recently several longtime customers at the Sweet Haven
Chocolate Company have complained about the quality of
the chocolates. It seems there are several partially covered
chocolates being found in boxes. The defective chocolates
should have been caught when the boxes were packed. The
manager is wondering which of the three packers is not
doing the job properly. Clerk 1 packs 40 of the boxes and
usually has a 2 defective rate. Clerk 2 packs 30 with a
2.5 defective rate. Clerk 3 boxes 30 of the chocolates
and her defective rate is 1.5. Which clerk is most likely
responsible for the boxes that raised the complaints
472. Tamarack Resorts and Properties is considering opening
a skiing area near McCall Idaho. It is trying to decide
whether to open an area catering to family skiers or to
some other group. To help make its decision it gathers
the following information. Let
A
1
Family will ski
A
2
Family will not ski
B
1
Family has children but none in the 8–16 age group
B
2
Family has children in the 8–16 age group
B
3
Family has no children
then for this location
P1A
1
2 0.40
P1B
2
2 0.35
P1B
1
2 0.25
P1A
1
B
2
2 0.70
P1A
1
B
1
2 0.30
a. Use the probabilities given to construct a joint
probability distribution table.
b. What is the probability a family will ski and have
children who are not in the 8–16 age group How
do you write this probability
c. What is the probability a family with children in the
8–16 age group will not ski
d. Are the categories skiing and family composition
independent
473. Fifty chief executive officers of small to mediumsized
companies were classified according to their gender and
functional background as shown in the table below:
Functional Background Male Female Total
Marketing 4 10 14
Finance 11 5 16
Operations 17 3 20
Total 32 18 50
a. If a chief executive is randomly selected from this
group what is the probability that the executive is a
female
www.downloadslide.comslide 197:
Introduction to Probability
b. Determine the probability that an individual
who has saved for retirement has saved less than
50000. Use relative frequencies.
c. Determine the probability that a randomly chosen
individual has saved less than 50000 toward
retirement.
d. Calculate the probability that at least two of four
individuals have saved less than 50000 toward
retirement.
479. USA Today reported on the impact of Generation
Y on the workforce. The workforce is comprised
of 1 Silent generation born before 1946
7.5 2 Baby boomers 1946–1964 42
3 Generation X 1196519762 29.5 and
4 Generation Y 1197719892 21. Ways of
communication are changing. Randstad Holding
an international supplier of services to businesses
and institutions examined the different methods of
communication preferred by the different elements
of the workforce. The file titled Communication
contains sample data comparable to those found in
this study.
a. Construct a frequency distribution for each of the
generations. Use the communication categories 1
Gp Meeting 2 FacetoFace 3 Email and 4
Other.
b. Calculate the probability that a randomly chosen
member of the workforce prefers communicating
face to face.
c. Given that an individual in the workforce prefers
to communicate face to face determine the
generation of which the individual is most likely a
member.
c. What is the joint probability the returned item was
from the Salt Lake City plant and had a wiring
related problem
d. What is the probability that a returned item was
made on the day shift at the Salt Lake plant and had
a cracked lens problem
e. If an item was returned what is the most likely
profile for the item including plant location shift
and cause of problem
477. Continuing with the ECCO company from Problem
76 when a customer returns one of the alarms under
warranty the quality manager logs data on the product.
From the data available in the Ecco file use relative
frequency to find the following probabilities.
a. If a part was made in the Salt Lake plant what is
the probability the cause of the returned part was
wiring
b. If the company incurs a 30 cost for each returned
alarm what percentage of the cost should be
assigned to each plant if it is known that 70 of all
production is done in Boise 20 in Salt Lake and
the rest in Toronto
478. The Employee Benefit Research Institute EBRI issued
a news release “Saving in America: Three Key Sets of
Figures” on October 25 2005. In 2005 about 69 of
workers said they have saved for retirement. The file
titled Retirement contains sample data similar to those
used in this study.
a. Construct a frequency distribution of the total
savings and investments using the intervals 1 Less
than 25000 2 25000–49999 3 50000–
99999 4 100000–249999 and 5 250000
or more.
Case 1
Great Air Commuter Service
The Great Air Commuter Service Company started in 1984 to pro
vide efficient and inexpensive commuter travel between Boston
and New York City. People in the airline industry know Peter Wil
son the principal owner and operating manager of the company
as “a real promoter.” Before founding Great Air Peter operated a
small regional airline in the Rocky Mountains with varying suc
cess. When Cascade Airlines offered to buy his company Peter
decided to sell and return to the East.
Peter arrived at his office near Fenway Park in Boston a little
later than usual this morning. He had stopped to have a business
breakfast with Aaron Little his longtime friend and sometime
partner in various business deals. Peter needed some advice and
through the years has learned to rely on Aaron as a ready source
no matter what the subject.
Peter explained to Aaron that his commuter service needed a
promotional gimmick to improve its visibility in the business com
munities in Boston and New York. Peter was thinking of running
a contest on each flight and awarding the winner a prize. The idea
would be that travelers who commute between Boston and New
York might just as well have fun on the way and have a chance to
win a nice prize.
As Aaron listened to Peter outlining his contest plans his mind
raced through contest ideas. Aaron thought that a large variety of
contests would be needed because many of the passengers would
likely be repeat customers and might tire of the same old thing.
In addition some of the contests should be chancetype contests
whereas others should be skill based.
“Well what do you think” asked Peter. Aaron finished his
scrambled eggs before responding. When he did it was completely
in character. “I think it will fly” Aaron said and proceeded to offer
a variety of suggestions.
Peter felt good about the enthusiastic response Aaron had
given to the idea and thought that the ideas discussed at breakfast
presented a good basis for the promotional effort. Now back at the
office Peter does have some concerns with one part of the plan.
Aaron thought that in addition to the regular inflight contests for
www.downloadslide.comslide 198:
Introduction to Probability
match and therefore the chance of giving away a grand prize on
any one flight is small. Peter likes the idea but when he asks Mar
garet what the probability is that a match will occur her response
does not sound quite right. She believes the probability for a match
will be 40/365 for a full plane and less than that when there are
fewer than 40 passengers aboard.
After Margaret leaves Peter decides that it would be useful to
know the probability of one or more birthday matches on flights
with 20 30 and 40 passengers. He realizes that he will need some
help from someone with knowledge of statistics.
Required Tasks:
1. Assume that there are 365 days in a year in other words
there is no leap year. Also assume there is an equal
probability of a passenger’s birthday falling on any one of
the 365 days. Calculate the probability that there will be at
least one birthday match for a flight containing exactly 20
passengers. Hint: This calculation is made easier if you
will first calculate the probability that there are no birthday
matches for a flight containing 20 passengers.
2. Repeat 1 above for a flight containing 30 passengers and
a flight containing 40 passengers. Again it will be easier to
compute the probabilities of one or more matches if you first
compute the probability of no birthday matches.
3. Assuming that each of the six daily flights carries 20
passengers calculate the probability that the airline will
have to award two or more major prizes that month. Hint: It
will be easier to calculate the probability of interest by first
calculating the probability that the airline will award one or
fewer prizes in a month.
prizes such as free flights dictation equipment and business peri
odical subscriptions each month on a randomly selected day a
major prize should be offered on all Great Air flights. This would
encourage regular business fliers to fly Great Air all the time.
Aaron proposed that the prize could be a trip to the Virgin Islands
or somewhere similar or the cash equivalent.
Great Air has three flights daily to New York and three flights
returning to Boston for a total of six flights. Peter is concerned
that the cost of funding six prizes of this size each month plus six
daily smaller prizes might be excessive. He also believes that it
might be better to increase the size of the large prize to something
such as a new car but use a contest that will not guarantee a winner.
But what kind of a contest can be used Just as he is about
to dial Aaron’s number Margaret Runyon Great Air’s market
ing manager enters Peter’s office. He has been waiting for her to
return from a meeting so he can run the contest idea past her and
get her input.
Margaret’s response is not as upbeat as Aaron’s but she does
think the idea is worth exploring. She offers an idea for the large
prize contest that she thinks might be workable. She outlines the
contest as follows.
On the first of each month she and Peter will randomly select
a day for that month on which the major contest will be run. That
date will not be disclosed to the public. Then on each flight that
day the flight attendant will have passengers write down their
birthdays month and day. If any two people on the plane have the
same birthday they will place their names in a hat and one name
will be selected to receive the grand prize.
Margaret explains that because the capacity of each flight is 40
passengers plus the crew there is a very low chance of a birthday
Case 2
Let’s Make a Deal
Quite a few years ago a popular show called Let’s Make a Deal
appeared on network television. Contestants were selected from the
audience. Each contestant would bring some silly item that he or she
would trade for a cash prize or a prize behind one of three doors.
Suppose that you have been selected as a contestant on the show.
Y ou are given a choice of three doors. Behind one door is a new sports
car. Behind the other doors are a pig and a chicken—booby prizes to
be sure Let’s suppose that you pick door number one. Before open
ing that door the host who knows what is behind each door opens
door two to show you the chicken. He then asks you “Would you be
willing to trade door one for door three” What should you do
Required Tasks:
1. Given that there are three doors one of which hides a sports
car calculate the probability that your initial choice is the
door that hides the sports car. What is the probability that you
have not selected the correct door
2. Given that the host knows where the sports car is and has
opened door 2 which revealed a booby prize does this
affect the probability that your initial choice is the correct
one
3. Given that there are now only two doors remaining
and that the sports car is behind one of them is it to
your advantage to switch your choice to door 3 Hint:
Eliminate door 2 from consideration. The probability
that door 1 is the correct door has not changed from your
initial choice. Calculate the probability that the prize
must be behind door 3. This problem was discussed in the
movie 21 starring Jim Sturgess Kate Bosworth and Kevin
Spacey.
www.downloadslide.comslide 199:
Introduction to Probability
Answers to Selected OddNumbered Problems
This section contains summary answers to most of the oddnumbered problems in the text. The Student Solutions Manual contains fully developed
solutions to all oddnumbered problems and shows clearly how each answer is determined.
c. For Pepsi Probability
5+6+6
12+12+11
17
35
0.486
For Coke Probability
6+6+6
12+12+11
18
35
0.514
d. For Pepsi Probability
7+6+8+5
19+16+14+16
26
65
0.4
For Coke Probability
12+10+6+11
19+16+14+16
39
65
0.6
27. a. 0.9 1  0.5 0.45
b. 0.60.8 0.48
29. P1senior 1 and senior 22 a
5
10
ba
4
9
b
20
90
0.22
31. a. PE
1
and B PE
1
BPB 0.250.30 0.075
b. PE
1
or B PE
1
+ PB  PE
1
and B
0.35 + 0.30  0.075 0.575
c. PE
1
and E
2
and E
3
PE
1
PE
2
PE
3
0.350.150.40 0.021
33. a. PB
Number of drives from B
Total drives
195
700
0.2786
b. PDefect
Number of defective drives
Total drives
50
700
0.0714
c. PDefect B
PDefect and B
PB
0.0214
0.2786
0.0769
PDefect B
Number of defective drives from B
Number of drives from B
15
195
0.0769
35. a. 0.61 0.316
b. 0.42 0.202 0.518
c. 0.39
37. They cannot get to 99.9 on color copies.
39. PFree gas 0.00015 + 0.00585 + 0.0005 0.0065
41. a. PNFL 105200 0.5250
b. PCollege degree and NBA 40200 0.20
c. 1050 0.20
d. The two events are not independent.
43. PLine 1 Defective 0.050.40.0725 0.2759
PLine 2 Defective 0.100.350.0725 0.4828
PLine 3 Defective 0.070.250.0725 0.2413
The unsealed cans probably came from Line 2.
45. PSupplier A Defective 0.150.30.115 0.3913
PSupplier B Defective 0.100.70.115 0.6087
Supplier B is the most likely to have supplied the defective
parts.
47. a. PE
1
and E
2
PE
1
E
2
PE
2
0.5080.607
0.308
b. PE
1
and E
3
PE
1
E
3
0.6070.853 0.712
1. independent events
3. V V V C V S C V C C C S
S V S C S S
5. a. subjective probability based on expert opinion
b. relative frequency based on previous customer return
history
c. 15 0.20
7. 13 0.333333
9. a. PBrown BrownTotal 310982 0.3157
b. PYZ99 YZ99Total 375982 0.3819
c. PYZ99 and Brown 205982 0.2088
d. not mutually exclusive since their joint probability is
0.1324
11. 0.375
15.
a. 0.1981
b. relative frequency
c. yes
17. a. relative frequency assessment method
b. P1
4000
5900
0.69
19. a. 312221768 0.1434
b. relative frequency assessment method
21. a. PCaesarean
22
50
0.44
b. New births may not exactly match the 50 in this study.
23. The following joint frequency table developed using Excel’s
pivot table feature summarizes the data.
Electrical Mechanical Total
Lincoln 28 39 67
Tyler 64 69 133
Total 92 108 200
a. 133200 0.665
b. 108200 0.54
c. 28200 0.14
25. a.
43
100
0.43
b.
5 + 6 + 6
100
0.17
Type of Ad Occurrences
HelpWanted Ad 204
Real Estate Ad 520
Other Ad 306
Total 1030
www.downloadslide.comslide 200:
Introduction to Probability
71. Clerk 1 is most likely responsible for the boxes that raised the
complaints.
73. a. 1850 0.36
b. 450 0.08
c. 320 0.15
d. 1018 0.556
e. There are a higher proportion of females whose functional
background is marketing and a higher proportion of males
whose functional background is operations.
75. a.
100
300
0.33
b.
30
300
0.10
c. PEast or C PEast + PC  PEast and C
0.25 + 0.333  0.103 0.48
d. PC East PC and EastPEast 0.1030.25
0.41
77. a. 0.3333
b. Boise will get 70.91 of the cost Salt Lake will get
21.82 and Toronto will get 7.27 regardless of
production volume.
49. a. 0.76
b. 0.988
c. 0.024
d. 0.9999
51. a. Probability first sampled device is defective 350 0.06
b. 4750349 0.0576
c. 475046494548 0.8273
53. a. 0.1856
b. 0.50
c. 0.0323
d. 0.3653
55. a. 0.119
b. 0.4727
57. a. 0.50
b. 0.755
c. 0.269
63. a. 0.80 0.40 0.20 0.60
b. A B and A B are complements.
65. a. 0.0156
b. 0.1563
c. 0.50
67. a. 0.149
b. 0.997
c. 0.449
69. a. the relative frequency assessment approach
b. 0.028
c. 0.349
d. yes
Blyth C. R. “Subjective vs. Objective Methods in Statistics”
American Statistician 26 June 1972 pp. 20–22.
DeV eaux Richard D. Paul F. V elleman and David E. Bock Stats
Data and Models 3rd ed. New Y ork: AddisonWesley 2012.
Hogg R. V . and Elliot A. Tanis Probability and Statistical Infer
ence 8th ed. Upper Saddle River NJ: Prentice Hall 2010.
Larsen Richard J. and Morris L. Marx An Introduction to
Mathematical Statistics and Its Applications 5th ed. Upper
Saddle River NJ: Prentice Hall 2012.
Microsoft Excel 2010 Redmond WA: Microsoft Corp. 2010.
Mlodinow Leonard The Drunkard’s Walk: How Randomness
Rules Our Lives New York: Pantheon Books 2008.
Raiffa H. Decision Analysis: Introductory Lectures on Choices
Under Uncertainty Reading MA: AddisonWesley 1968.
Siegel Andrew F. Practical Business Statistics 5th ed. Burr
Ridge IL: Irwin 2002.
References
www.downloadslide.comslide 201:
Introduction to Probability
Mutually Exclusive Events Two events are mutually exclusive
if the occurrence of one event precludes the occurrence of
the other event.
Probability The chance that a particular event will occur. The
probability value will be in the range 0 to 1. A value of 0
means the event will not occur. A probability of 1 means
the event will occur. Anything between 0 and 1 reflects the
uncertainty of the event occurring. The definition given is
for a countable number of events.
Relative Frequency Assessment The method that defines
probability as the number of times an event occurs divided
by the total number of times an experiment is performed in a
large number of trials.
Sample Space The collection of all outcomes that can result
from a selection decision or experiment.
Subjective Probability Assessment The method that defines
probability of an event as reflecting a decision maker’s state
of mind regarding the chances that the particular event will
occur.
Classical Probability Assessment The method of determin
ing probability based on the ratio of the number of ways
an outcome or event of interest can occur to the number of
ways any outcome or event can occur when the individual
outcomes are equally likely.
Complement The complement of an event E is the collection
of all possible outcomes not contained in event E.
Conditional Probability The probability that an event will
occur given that some other event has already happened.
Dependent Events Two events are dependent if the occurrence
of one event impacts the probability of the other event occur
ring.
Event A collection of experimental outcomes.
Experiment A process that produces a single outcome whose
result cannot be predicted with certainty.
Independent Events Two events are independent if the occur
rence of one event in no way influences the probability of
the occurrence of the other event.
Glossary
www.downloadslide.comslide 202:
Discrete Probability
Distributions
From Chapter 5 of Business Statistics A DecisionMaking Approach Ninth Edition. David F. Groebner
Patrick W. Shannon and Phillip C. Fry. Copyright © 2014 by Pearson Education Inc. All rights reserved.
www.downloadslide.comslide 203:
Why you need to know
PakSense a manufacturer of temperature sensor equipment for the food industry receives component parts
for its sensors weekly from suppliers. When a batch of parts arrives the qualityassurance section randomly
samples a fixed number of parts and tests them to see if any are defective. Sup
pose in one such test a sample of 50 parts is selected from a supplier whose
contract calls for at most 2 defective parts. How many defective parts in the
sample of 50 should PakSense expect if the contract is being satisfied What
should be concluded if the sample contains three defects Answers to these ques
tions require calculations based on a probability distribution known as the binomial
distribution.
How many toll stations should be constructed when a new toll bridge is built in
California If there are four toll stations will drivers have to wait too long or will there
be too many toll stations and excess employees To help answer these questions
decision makers use a probability distribution known as the Poisson distribution.
A personnel manager has a chance to promote 3 people from 10 equally quali
fied candidates. Suppose none of six women are selected by the manager. Is this
evidence of gender bias or would we expect to see this type of result A distribution
known as the hypergeometric distribution would be very helpful in addressing this
issue.
The binomial Poisson and hypergeometric distributions are three discrete
probability distributions used in business decision making. This chapter introduces
discrete probability distributions and shows how they are used in business settings.
Probability is the way decision makers express their uncertainty about outcomes and
events. Through the use of wellestablished discrete probability distributions you will
be better prepared for making decisions in an uncertain environment.
including the Addition and Multiplication
Rules.
Review the discussion of weighted averages.
Review the basic rules of probability
Quick Prep Links
Review the concepts of simple random
sampling.
Outcome 3. Be able to compute probabilities for the
Poisson and hypergeometric distributions and apply these
distributions to decisionmaking situations.
Outcome 2. Be able to apply the binomial distribution to
business decisionmaking situations.
Outcome 1. Be able to calculate and interpret the expected
value of a discrete random variable.
Introduction to Discrete
Probability Distributions
The Binomial Probability
Distribution
Other Discrete Probability
Distributions
Discrete Probability
Distributions
Cobalt Creative/Shutterstock
www.downloadslide.comslide 204:
Discrete Probability Distributions
Introduction to Discrete Probability
Distributions
Random Variables
When a random experiment is performed some outcome must occur. When the experiment
has a quantitative characteristic we can associate a number with each outcome. For example
an inspector who examines three plasma flatpanel televisions can judge each television as
“acceptable” or “unacceptable.” The outcome of the experiment defines the specific number
of acceptable televisions. The possible outcomes are
x 50 1 2 36
The value x is called a random variable since the numerical values it takes on are random
and vary from trial to trial. Although the inspector knows these are the possible values for
the variable before she samples she does not know which value will occur in any given trial.
Further the value of the random variable may be different each time three plasma televisions
are inspected.
Two classes of random variables exist: discrete random variables and continuous
random variables. For instance if a bank auditor randomly examines 15 accounts to verify
the accuracy of the balances the number of inaccurate account balances can be represented by
a discrete random variable with the following values:
x 50 1c 156
In another situation 10 employees were recently hired by a major electronics company.
The number of females in that group can be described as a discrete random variable with pos
sible values equal to
x 50 1 2 3c 106
Notice that the value for a discrete random variable is often determined by counting. In the
bank auditing example the value of variable x is determined by counting the number of
accounts with errors. In the hiring example the value of variable x is determined by counting
the number of females hired.
In other situations the random variable is said to be continuous. For example the exact
time it takes a city bus to complete its route may be any value between two points say 30
minutes to 35 minutes. If x is the time required then x is continuous because if measured
precisely enough the possible values x can be any value in the interval 30 to 35 minutes.
Other examples of continuous variables include measures of distance and measures of weight
when measured precisely. A continuous random variable is generally defined by measuring
which is contrasted with a discrete random variable whose value is typically determined by
counting.
Displaying Discrete Probability Distributions Graphically The probability distribu
tion for a discrete random variable is composed of the values the variable can assume and the
probabilities for each of the possible values. For example if three parts are tested to determine
if they are defective the probability distribution for the number of defectives might be
x Number of Defectives Px
0 0.10
1 0.30
2 0.40
3 0.20
g 1.00
Graphically the discrete probability distribution associated with these defectives can be
represented by the areas of rectangles in which the base of each rectangle is one unit wide and
the height corresponds to the probability. The areas of the rectangles sum to 1. Figure 1 illus
trates two examples of discrete probability distributions. Figure 1a shows a discrete random
Random Variable
A variable that takes on different numerical
values based on chance.
Discrete Random Variable
A random variable that can only assume a finite
number of values or an infinite sequence of
values such as 0 1 2.…
Continuous Random Variables
Random variables that can assume an
uncountably infinite number of values.
www.downloadslide.comslide 205:
Discrete Probability Distributions
variable with only three possible outcomes. Figure 1b shows the probability distribution for
a discrete variable that has 21 possible outcomes. Note as the number of possible outcomes
increases the distribution becomes smoother and the individual probability of any particular
value tends to be reduced. In all cases the sum of the probabilities is 1.
Discrete probability distributions have many applications in business decisionmaking
situations. In the remainder of this section we discuss several important issues that are of
particular importance to discrete probability distributions.
Mean and Standard Deviation of
Discrete Distributions
A probability distribution like a frequency distribution can be only partially described by a
graph. To aid in a decision situation you may need to calculate the distribution’s mean and
standard deviation. These values measure the central location and spread respectively of the
probability distribution.
Calculating the Mean The mean of a discrete probability distribution is also called the
expected value of the random variable from an experiment. The expected value is actually
a weighted average of the random variable values in which the weights are the probabilities
assigned to the values. The expected value is given in Equation 1.
Expected Value
The mean of a probability distribution. The
average value when the experiment that
generates values for the random variable is
repeated over the long run.
Probability
10 20 30
Possible Values of x
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0
a Discrete Probability Distribution 3 possible outcomes
x
Px
Probability
123456789 101112131415161718192021
Possible Values of x
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0
b Discrete Probability Distribution 21 possible outcomes
x
Px
FIGURE 1 
Discrete Probability
Distributions
Chapter Outcome 1.
www.downloadslide.comslide 206:
Discrete Probability Distributions
Calculating the Standard Deviation The standard deviation measures the spread or
dispersion in a set of data. The standard deviation also measures the spread in the values of
a random variable. To calculate the standard deviation for a discrete probability distribution
use Equation 2.
Expected Value of a Discrete Probability Distribution
E1x2
a
xP1x2 1
where:
E1x2 Expected value of x
x Values of the random variable
P1x2 probability of the random variable taking on the value x
Standard Deviation of a Discrete Probability Distribution
x
xEx Px ∑−
2
2
where:
x Values of the random variable
E1x2 Expected value of x
P1x2 Probability of the random variable taking on the value x
Equation 2 is different in form than other equations given for standard deviation. This is
because we are now dealing with a discrete probability distribution rather than population or
sample values.
EXAMPLE 1 COMPUTING THE MEAN AND STANDARD DEVIATION
OF A DISCRETE RANDOM VARIABLE
Swenson Security Sales Swenson Security Sales is a com
pany that sells and installs home security systems throughout the
eastern United States. Each week the company’s quality manag
ers examine one randomly selected security installation in all 21
states where the company operates to see whether the installers
made errors. The discrete random variable x is the number of
errors discovered on each installation examined ranging from 0
to 3. The following frequency distribution was developed:
x Frequency
0 150
1 110
250
3 90
g 400
Assuming that these data reflect typical performance by the installers the company leadership
wishes to develop a discrete probability distribution for the number of install errors and compute
the mean and standard deviation for the distribution. This can be done using the following steps:
Step 1 Convert the frequency distribution into a probability distribution using
the relative frequency assessment method.
x Frequency
0 150400 0.375
1 110400 0.275
Scanrail/Fotolia
www.downloadslide.comslide 207:
Discrete Probability Distributions
x Frequency
2 50400 0.125
3 90400 0.225
g 1.000
Step 2 Compute the expected value using Equation 1.
Ex xP x
Ex
. .
∑
0 0 375 1 0 275 2 0. . .
.
125 3 0 225
120 Ex
The expected value is 1.20 errors per install. Thus assuming the distribution of
the number of errors is representative of that of each individual installation the
longrun average number of errors per installation will be 1.20.
Step 3 Compute the standard deviation using Equation 2.
x
xEx Px ∑
2
xP1x23x  E1x24 3x  E1x24
2
3x  E1x24
2
P1x2
0 0.375 01.21.20 1.44 0.540
1 0.275 11.2 0.20 0.04 0.011
2 0.125 21.2 0.80 0.64 0.080
3 0.225 31.2 1.80 3.24 0.729
g 1.360
x
136 117 ..
The standard deviation of the discrete probability distribution is 1.17 errors per
security system installed.
END EXAMPLE
TRY PROBLEM 4
BUSINESS APPLICATION EXPECTED VALUES
BENSON LIGHT FIXTURES Benson Light Fixtures imports fixtures from Taiwan
and other Far East countries for distribution in France and Spain. For one particular
product line Benson currently has two suppliers. Both suppliers have poor records
when it comes to quality. Benson is planning to purchase 100000 of a particular light
fixture and wants to use the leastcost supplier for the entire purchase. Supplier A is less
expensive by 1.20 per fixture and has an ongoing record of supplying 10 defects.
Supplier B is more expensive but may be a higherquality supplier. Benson records
indicate that the rate of defects from supplier B varies. Table 1 shows the probability
distributions for the defect percentages for supplier B. Each defect is thought to cost
the company 9.50.
Looking first at supplier A at a defect rate of 0.10 out of 100000 units the number of
defects is expected to be 10000. The cost of these is +9.50 10000 +95000. For sup
plier B the expected defect rate is found using Equation 1 as follows:
E1Defect rate2
a
x P1x2
E1Defect rate2 10.01210.32 + 10.05210.42 + 10.10210.22 + 10.15210.12
E1Defect rate2 0.058
Thus supplier B is expected to supply 5.8 defects or 5800 out of the 100000 units
ordered for an expected cost of +9.50 5800 +55100. Based on defect cost alone
supplier B is less expensive 55100 versus 95000. However recall that supplier B’s
TABLE 1  Probability
Distribution—Defect Rate
for Supplier B
Defect Rate x Probability Px
0.01 0.3
0.05 0.4
0.10 0.2
0.15 0.1
www.downloadslide.comslide 208:
Discrete Probability Distributions
product sells for 1.20 per unit more. Thus on a 100000unit order supplier B costs an extra
+1.20 100000 +120000 more than supplier A. The relative costs are
Supplier A +95000 Supplier B +55100 + +120000 +175100
Therefore based on expected costs supplier A should be selected to supply the 100000
light fixtures.
Skill Development
51. An economics quiz contains six multiplechoice
questions. Let x represent the number of questions a
student answers correctly.
a. Is x a continuous or discrete random variable
b. What are the possible values of x
52. Two numbers are randomly drawn without replacement
from a list of five. If the five numbers are 2 2 4
6 8 what is the probability distribution of the sum
of the two numbers selected Show the probability
distribution graphically.
53. If the Prudential Insurance Company surveys its
customers to determine the number of children under
age 22 living in each household
a. What is the random variable for this survey
b. Is the random variable discrete or continuous
54. Given the following discrete probability distribution
xPx
50 0.375
65 0.15
70 0.225
75 0.05
90 0.20
a. Calculate the expected value of x.
b. Calculate the variance of x.
c. Calculate the standard deviation of x.
55. Because of bad weather the number of days next week
that the captain of a charter fishing boat can leave port
is uncertain. Let x number of days that the boat is
able to leave port per week. The following probability
distribution for the variable x was determined based
on historical data when the weather was poor:
xPx
0 0.05
1 0.10
2 0.10
3 0.20
4 0.20
5 0.15
6 0.15
7 0.05
Based on the probability distribution what is the expected
number of days per week the captain can leave port
5 Exercises
MyStatLab
56. Consider the following discrete probability distribution:
xP x
3 0.13
6 0.12
9 0.15
12 0.60
a. Calculate the variance and standard deviation of the
random variable.
b. Let y x + 7. Calculate the variance and standard
deviation of the random variable y.
c. Let z 7x. Calculate the variance and standard
deviation of the random variable z.
d. From your calculations in part a and part b indicate
the effect that adding a constant to a random
variable has on its variance and standard deviation.
e. From your calculations in part a and part c indicate
the effect that multiplying a random variable with
a constant has on the variance and the standard
deviation of the random variable.
57. Given the following discrete probability distribution
xP x
100 0.25
125 0.30
150 0.45
a. Calculate the expected value of x.
b. Calculate the variance of x.
c. Calculate the standard deviation of x.
58. The roll of a pair of dice has the following probability
distribution where the random variable x is the sum of
the values produced by each die:
xP x
2 1/36
3 2/36
4 3/36
5 4/36
6 5/36
7 6/36
8 5/36
9 4/36
10 3/36
11 2/36
12 1/36
www.downloadslide.comslide 209:
Discrete Probability Distributions
512. Jennings Assembly in Hartford Connecticut uses a com
ponent supplied by a company in Brazil. The component
is expensive to carry in inventory and consequently is not
always available in stock when requested. Furthermore
shipping schedules are such that the lead time for trans
portation of the component is not a constant. Using his
torical records the manufacturing firm has developed the
following probability distribution for the product’s lead
time. The distribution is shown here where the random
variable x is the number of days between the placement
of the replenishment order and the receipt of the item.
xP x
2 0.15
3 0.45
4 0.30
5 0.075
6 0.025
a. What is the average lead time for the component
b. What is the coefficient of variation for delivery lead
time
c. How might the manufacturing firm in the United
States use this information
513. Marque Electronics is a familyowned electronics
repair business in Kansas City. The owner has read an
advertisement from a local competitor that guarantees
all highdefinition television HDTV repairs within
four days. Based on his company’s experience he
wants to know if he can offer a similar guarantee.
His past service records are used to determine the
following probability distribution:
Number of Days Probability
1 0.15
2 0.25
3 0.30
4 0.18
5 0.12
a. Calculate the mean number of days his customers
wait for an HDTV repair.
b. Also calculate the variance and standard deviation.
c. Based on the calculations in parts a and b what
conclusion should the manager reach regarding his
company’s repair times
514. Cramer’s Bar and Grille in Dallas can seat 130 people at a
time. The manager has been gathering data on the number
of minutes a party of four spends in the restaurant from
the moment they are seated to when they pay the check.
What is the mean number of minutes for a dinner party of
four What is the variance and standard deviation
Number of Minutes Probability
60 0.05
70 0.15
a. Calculate the expected value of x.
b. Calculate the variance of x.
c. Calculate the standard deviation of x.
59. Consider the following discrete probability
distribution:
xP x
5 0.10
10 0.15
15 0.25
20 0.50
a. Calculate the expected value of the random variable.
b. Let y x + 5. Calculate the expected value of the
new random variable y.
c. Let z 5x. Calculate the expected value of the new
random variable z.
d. From your calculations in part a and part b indicate
the effect that adding a constant to a random
variable has on the expected value of the random
variable.
e. From your calculations in part a and part c indicate
the effect that multiplying a random variable by a
constant has on the expected value of the random
variable.
510. Examine the following probability distribution:
x 5 101520253035404550
Px 0.01 0.05 0.14 0.20 0.30 0.15 0.05 0.04 0.01 0.05
a. Calculate the expected value and standard deviation
for this random variable.
b. Denote the expected value as m. Calculate m s
and m +s.
c. Determine the proportion of the distribution that is
contained within the interval m s.
d. Repeat part c for 112m 2s and 122m 3s.
Business Applications
511. The U.S. Census Bureau Annual Social Economic
Supplement collects demographics concerning the
number of people in families per household. Assume
the distribution of the number of people per household
is shown in the following table:
xP x
2 0.27
3 0.25
4 0.28
5 0.13
6 0.04
7 0.03
a. Calculate the expected number of people in families
per household in the United States.
b. Compute the variance and standard deviation of the
number of people in families per household.
www.downloadslide.comslide 210:
Discrete Probability Distributions
x Weekly Demand Px
0 0.05
1 0.05
2 0.10
3 0.20
4 0.40
5 0.10
6 0.05
7 0.05
a. What is the expected weekly demand for the alarm
clock radio
b. What is the probability that weekly demand will be
greater than the number of available radios
c. What is the expected weekly profit from the sale of
the alarm clock radio Remember: There are only
four clock radios available in any week to meet
demand.
d. On average how much profit is lost each week
because the radio is not available when demanded
517. Fiero Products LTD of Bologna Italy makes a variety
of footwear including indoor slippers children’s
shoes and flipflops. To keep up with increasing
demand it is considering three expansion plans: 1 a
small factory with yearly costs of 150000 that will
increase the production of flipflops to 400000 2 a
midsized factory with yearly costs of 250000 that
will increase the production of flipflops by 600000
and 3 a large factory with yearly costs of 350000
that will increase the production of flipflops by
900000. The profit per flipflop is projected to be
0.75. The probability distribution of demand for flip
flops is considered to be
Demand 300000 700000 900000
Probability 0.2 0.5 0.3
a. Compute the expected profit for each of the
expansion plans.
b. Calculate the standard deviation for each of the
expansion plans.
c. Which expansion plan would you suggest Provide
the statistical reasoning behind your selection.
518. A large corporation in search of a CEO and a CFO has
narrowed the fields for each position to a short list.
The CEO candidates graduated from Chicago C and
three Ivy League universities: Harvard H Princeton
P and Yale Y. The four CFO candidates graduated
from MIT M Northwestern N and two Ivy League
universities Dartmouth D and Brown B. The
personnel director wishes to determine the distribution
of the number of Ivy League graduates who could fill
these positions.
a. Assume the selections were made randomly.
Construct the probability distribution of the number
of Ivy League graduates who could fill these
positions.
Number of Minutes Probability
80 0.20
90 0.45
100 0.10
110 0.05
515. Rossmore Brothers Inc. sells plumbing supplies for
commercial and residential applications. The company
currently has only one supplier for a particular type of
faucet. Based on historical data that the company has
maintained the company has assessed the following
probability distribution for the proportion of defective
faucets that it receives from this supplier:
Proportion Defective Probability
0.01 0.4
0.02 0.3
0.05 0.2
0.10 0.1
This supplier charges Rossmore Brothers Inc. 29.00
per unit for this faucet. Although the supplier will
replace any defects free of charge Rossmore managers
figure the cost of dealing with the defects is about
5.00 each.
a. Assuming that Rossmore Brothers is planning to
purchase 2000 of these faucets from the supplier
what is the total expected cost to Rossmore Brothers
for the deal
b. Suppose that Rossmore Brothers has an opportunity
to buy the same faucets from another supplier at
a cost of 28.50 per unit. However based on its
investigations Rossmore Brothers has assessed the
following probability distribution for the proportion
of defective faucets that will be delivered by the
new supplier:
Proportion Defective x Probability Px
0.01 0.1
0.02 0.1
0.05 0.7
0.10 0.1
Assuming that the defect cost is still 5.00 each and
based on total expected cost for an order of 2000
faucets should Rossmore buy from the new supplier
or stick with its original supplier
516. Radio Shack stocks four alarm clock radios. If it
has fewer than four clock radios available at the end
of a week the store restocks the item to bring the
instock level up to four. If weekly demand is greater
than the four units in stock the store loses the sale.
The radio sells for 25 and costs the store 15. The
Radio Shack manager estimates that the probability
distribution of weekly demand for the radio is as
follows:
www.downloadslide.comslide 211:
Discrete Probability Distributions
It presents the number of days Revolution could remain
effective when applied to mature cats.
a. Produce a frequency distribution for these data. Convert
the frequency distribution into a probability distribution
using the relative frequency assessment method.
b. Calculate the expected value and standard deviation
for the number of days Revolution could remain
effective.
c. If the marketing department wished to advertise
the number of days that 90 of the cats remain
protected while using Revolution what would this
number of days be
521. Fiber Systems makes boat tops for a number of boat
manufacturers. Its fabric has a limited twoyear
warranty. Periodic testing is done to determine if
the warranty policy should be changed. One such
study may have examined those covers that became
unserviceable while still under warranty. Data that
could be produced by such a study are contained in the
file titled Covers. The data represent assessment of the
number of months a cover was used until it became
unserviceable.
a. Produce a frequency distribution for these data.
Convert the frequency distribution into a probability
distribution using the relative frequency assessment
method.
b. Calculate the expected value and standard deviation
for the time until the covers became unserviceable.
c. The qualitycontrol department thinks that among those
covers that do become unserviceable while still under
warranty the majority last longer than 19 months.
Produce the relevant statistic to verify this assumption.
b. Would it be surprising if both positions were filled
with Ivy League graduates
c. Calculate the expected value and standard deviation
of the number of Ivy League graduates who could
fill these positions.
Computer Database Exercises
519. Starbucks has entered into an agreement with a
publisher to begin selling a food and beverage
magazine on a trial basis. The magazine retails for
3.95 in other stores. Starbucks bought it for 1.95
and sold it for 3.49. During the trial period Starbucks
placed 10 copies of the magazine in each of 150 stores
throughout the country. The file titled Sold contains the
number of magazines sold in each of the stores.
a. Produce a frequency distribution for these data. Convert
the frequency distribution into a probability distribution
using the relative frequency assessment method.
b. Calculate the expected profit from the sale of these
10 magazines.
c. Starbucks is negotiating returning all unsold magazines
for a salvage price. Determine the salvage price
Starbucks will need to obtain to yield a positive
expected profit from selling 10 magazines.
520. Pfizer Inc. is the manufacturer of Revolution
Selamectin a topical parasiticide used for the
treatment control and prevention of flea infestation
heartworm and ear mites for dogs and cats. One of its
selling points is that it provides protection for an entire
month. Such claims are made on the basis of research
and statistical studies. The file titled Fleafree contains
data similar to those obtained in Pfizer’s research.
END EXERCISES 51
The Binomial Probability Distribution
In Section 1 you learned that random variables can be classified as either discrete or continu
ous. In most instances the value of a discrete random variable is determined by counting. For
instance the number of customers who arrive at a store each day is a discrete variable. Its
value is determined by counting the customers.
Several theoretical discrete distributions have extensive application in business decision
making. A probability distribution is called theoretical when the mathematical properties of
its random variable are used to produce its probabilities. Such distributions are different from
the distributions that are obtained subjectively or from observation. Sections 2 and 3 focus on
theoretical discrete probability distributions.
The Binomial Distribution
The first theoretical probability distribution we will consider is the binomial distribution
that describes processes whose trials have only two possible outcomes. The physical events
described by this type of process are widespread. For instance a qualitycontrol system in a
manufacturing plant labels each tested item as either defective or acceptable. A firm bidding
for a contract either will or will not get the contract. A marketing research firm may receive
responses to a questionnaire in the form of “Yes I will buy” or “No I will not buy.” The
personnel manager in an organization is faced with two possible outcomes each time he offers
a job—the applicant either accepts the offer or rejects it.
Binomial Probability Distribution
Characteristics
A distribution that gives the probability of x
successes in n trials in a process that meets the
following conditions:
1. A trial has only two possible outcomes: a
success or a failure.
2. There is a fixed number n of identical tri
als.
3. The trials of the experiment are independent
of each other. This means that if one out
come is a success this does not influence
the chance of another outcome being a
success.
4. The process must be consistent in gener
ating successes and failures. That is the
probability p associated with a success
remains constant from trial to trial.
5. If p represents the probability of a success
then11  p2 q is the probability of a
failure.
Chapter Outcome 2.
www.downloadslide.comslide 212:
Discrete Probability Distributions
Characteristics of the Binomial Distribution
The binomial distribution requires that the experiment’s trials be independent. This can
be assured if the sampling is performed with replacement from a finite population. This
means that an item is sampled from a population and returned to the population after its
characteristics have been recorded before the next item is sampled. However sampling
with replacement is the exception rather than the rule in business applications. Most often
the sampling is performed without replacement. Strictly speaking when sampling is per
formed without replacement the conditions for the binomial distribution cannot be satisfied.
However the conditions are approximately satisfied if the sample selected is quite small
relative to the size of the population from which the sample is selected. A commonly used
rule of thumb is that the binomial distribution can be applied if the sample size is at most 5
of the population size.
BUSINESS APPLICATION USING THE BINOMIAL DISTRIBUTION
LAKE CITY AUTOMOTIVE Lake City Automotive performs
300 automobile transmission repairs every week. Transmission
repair work is not an exact science and a percentage of the
work done must be reworked. Lake City shop managers have
determined that even when the mechanics perform the work in a
proper manner 10 of the time the car will have to be worked on
again to fix a problem. The binomial distribution applies to this
situation because the following conditions exist:
1. There are only two possible outcomes when a car is repaired: It either needs rework or it
doesn’t. We are interested in the cars that need rework so we will consider a reworked
car to be a “success.” A success occurs when we observe the outcome of interest.
2. All repair jobs are considered to be equivalent in the work required.
3. The outcome of a repair rework or no rework is independent of whether the preceding
repair required rework or not.
4. The probability of rework being needed p 0.10 remains constant from car to car.
5. The probability of a car not needing rework q 1  p 0.90 remains constant from car
to car.
To determine whether the Lake City mechanics are continuing to function at the stand
ard level of performance the shop foreman randomly selects four cars from the week’s list
of repaired vehicles and tracks those to see whether they need rework or not note the need
for rework is known within a few hours of the work being completed. Because the sample
size is small 14300 0.0133 or 1.332 relative to the size of the population 300 units per
week the conditions of independence and constant probability will be approximately satis
fied because the sample is less than 5 of the population.
We let the number of reworked cars be the random variable of interest. The number of
reworked units is limited to discrete values x 0 1 2 3 or 4. We can determine the prob
ability that the random variable will have any of the discrete values. One way is to list the
sample space as shown in Table 2. We can find the probability of zero cars needing rework
for instance by employing the Multiplication Rule for Independent Events.
P1x 0 reworks2 P1G and G and G and G2
where:
G Car does not require rework
Here
P1G2 0.90
and we have assumed the repair jobs are independent. Using the Multiplication Rule for
Independent Events
P1G and G and G and G2 P1G2P1G2P1G2P1G2 10.90210.90210.90210.902
0.90
4
0.6561
Dmitry V ereshchagin/Fotolia
www.downloadslide.comslide 213:
Discrete Probability Distributions
We can also find the probability of exactly one reworked car in a sample of four. This is
accomplished using both the Multiplication Rule for Independent Events and the Addition
Rule for Mutually Exclusive Events:
D car needs rework
P1x 1 rework2 P1G and G and G and D2 + P1G and G and D and G2
+ P1G and D and G and G2 + P1D and G and G and G2
where:
P1G and G and G and D2 P1G2P1G2P1G2P1D2 10.90210.90210.90210.102
10.90
3
210.102
Likewise:
P1G and G and D and G2 10.90
3
210.102
P1G and D and G and G2 10.90
3
210.102
P1D and G and G and G2 10.90
3
210.102
Then:
P11 rework2 10.90
3
210.102 + 10.90
3
210.102 + 10.90
3
210.102 + 10.90
3
210.102
14210.90
3
210.102
0.2916
Note that each of the four possible ways of finding one rework car has the same probabil
ity 10.90
3
210.102. We determine the probability of one of the ways to obtain one rework
car and multiply this value by the number of ways four of obtaining one reworked car. This
produces the overall probability of one reworked car.
Combinations In this relatively simple application we can fairly easily list the sample
space and from that count the number of ways that each possible number of reworked cars
can occur. However for examples with larger sample sizes this approach is inefficient. A
more effective method exists for counting the number of ways binomial events can occur. This
method is called the counting rule for combinations. This rule is used to find the number of
outcomes from an experiment in which x objects are to be selected from a group of n objects.
Equation 3 is used to find the combinations.
TABLE 2  Sample Space for Lake City Automotive
Results No. of Reworked Cars No. of Ways
GGGG 01
GGGD 1
GGDG 14
GDGG 1
DGGG 1
GGDD 2
GDGD 2
DGGD 26
GDDG 2
DGDG 2
DDGG 2
DDDG 3
DDGD 34
DGDD 3
GDDD 3
DDDD 41
www.downloadslide.comslide 214:
Discrete Probability Distributions
Using Equation 3 we find the number of ways that x 2 reworked cars can occur in a sam
ple of n 4 as
C
n
xn x
x
n
55 5
−−
4
24 2
43 2 1
21 21
24
4
6 55 ways
Refer to Table 2 to see that this is the same value for two reworked cars in a sample of four
that was obtained by listing the sample space.
Now we can find the probabilities of two reworked cars.
P12 reworks2 16210.90
2
210.10
2
2
0.0486
Use this method to verify the following:
P13 reworks2 14210.90210.10
3
2
0.0036
P14 reworks2 11210.10
4
2
0.0001
The key to developing the probability distribution for a binomial process is first to
determine the probability of any one way the event of interest can occur and then to multi
ply this probability by the number of ways that event can occur. Table 3 shows the binomial
probability distribution for the number of reworked cars in a sample size of four when
the probability of any individual car requiring rework is 0.10. The probability distribution
is graphed in Figure 2. Most samples would contain zero or one reworked car when the
mechanics at Lake City are performing the work to standard.
Counting Rule for Combinations
C
n
xn x
x
n
3
where:
Cx
x
n
Number of combinations of objects sel lected from objects n
nnn n ... 12 2
1
0 1 by defnition
TABLE 3  Binomial
Distribution for Lake City
Automotive: n 4 p 0.10
x of Reworks P1x2
0 0.6561
1 0.2916
2 0.0486
3 0.0036
4 0.0001
g 1.0000
Probability
04 3 2 1
x 5 Number of Reworks
Px
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
FIGURE 2 
Binomial Distribution for Lake
City Automotive
www.downloadslide.comslide 215:
Discrete Probability Distributions
Binomial Formula The steps that we have taken to develop this binomial probability
distribution can be summarized through a formula called the binomial formula shown
as Equation 4. Note this formula is composed of two parts: the combinations of x items
selected from n items and the probability that x items can occur.
Binomial Formula
Px
n
xn x
xn x
pq
4
where:
n Random sample size
x Number of successes 1when a success is defined as what we are looking for2
n  x Number of failures
p Probability of a success
q 1  p Probability of a failure
n n1n  121n  221n  32 . . . 122112
0 1 by definition
Applying Equation 4 to the Lake City Automotive example for n 4 p 0.10 and x 2
reworked cars we get
Px
n
xn x
pq
xn x
5
−
−
P . . . . 2 0 10 0 90 6 0 10 0 90 0
22 22
55 5
4
22
. .0486
This is the same value we calculated earlier when we listed out the sample space above.
EXAMPLE 2 USING THE BINOMIAL FORMULA
Creative Style and Cut Creative Style and Cut an upscale beauty
salon in San Francisco offers a full refund to anyone who is not satis
fied with the way his or her hair looks after it has been cut and styled.
The owners believe the hairstyle satisfaction from customer to customer
is independent and that the probability a customer will ask for a refund
is 0.20. Suppose a random sample of six customers is observed. In four
instances the customer has asked for a refund. The owners might be
interested in the probability of four refund requests from six customers.
If the binomial distribution applies the probability can be found using
the following steps:
Step 1 Define the characteristics of the binomial distribution.
In this case the characteristics are
n 6 p 0.20 q 1  p 0.80
Step 2 Determine the probability of x successes in n trials using the binomial
formula Equation 4.
In this case n 6 p 0.20 q 0.80 and we are interested in the
probability of x 4 successes.
Px
n
xn x
pq
P
xn x
. 4
6
46 4
020
4
.
. .
.
080
4 15 0 20 0 80
4 0 0154
64
42
P
P
SzaszFabian Ilka Erika/Shutterstock
www.downloadslide.comslide 216:
Discrete Probability Distributions
There is only a 0.0154 chance that exactly four customers will want a refund in a
sample of six if the chance that any one of the customers will want a refund is 0.20.
END EXAMPLE
TRY PROBLEM 24
Using the Binomial Distribution Table Using Equation 4 to develop the binomial dis
tribution is not difficult but it can be time consuming. To make binomial probabilities easier
to find you can use the binomial table in Cumulative Binomial Distribution Table. This table
is constructed to give cumulative probabilities for different sample sizes and probabilities of
success. Each column is headed by a probability p which is the probability associated with
a success. The column headings correspond to probabilities of success ranging from 0.01 to
1.00. Down the left side of the table are integer values that correspond to the number of suc
cesses x for the specified sample size n. The values in the body of the table are the cumula
tive probabilities of x or fewer successes in a random sample of size n.
BUSINESS APPLICATION BINOMIAL DISTRIBUTION TABLE
BIG O TIRE COMPANY Big O Tire Company operates tire and repair stores throughout
the western United States. Upper management is considering offering a moneyback warranty
if one of their tires blows out in the first 20000 miles of use. The managers are willing to
make this warranty if 2 or fewer of the tires they sell blow out within the 20000mile limit.
The company plans to test 10 randomly selected tires over 20000 miles of use in test
conditions. The number of possible blowouts will be x 0 1 2 c10. We can use the bino
mial table in Cumulative Binomial Distribution Table to develop the probability distribution.
This table is called a cumulative probability table. Go to the table for n 10 and the column
for p 0.02. The values of x are listed down the left side of the table. For example the cumu
lative probability of x … 2 occurrences is 0.9991. This means that it is extremely likely that 2
or fewer tires in a sample of 10 would blow out in the first 20000 miles of use if the overall
fraction of tires that will blow out is 0.02. The probability of 3 or more blowouts in the sample
of n 10 is
P1x Ú 32 1  P1x … 22 1  0.9991 0.0009
There are about 9 chances in 10000 that we would fnd 3 or more tires in a sample of 10 that
will blow out if the probability of it happening for any one tire is p 0.02. If the test did
show that 3 tires blew out the true rate of tire blowouts likely exceeds 2 and the company
should have serious doubts about making the warranty. Note the decision about this will
depend on which sample outcome occurs.
EXAMPLE 3 USING THE BINOMIAL TABLE
Nielsen Television Ratings The Nielsen Media Group is
the bestknown television ratings company. On Monday Feb
ruary 6 2012 the day after the 2012 Super Bowl between the
Giants and the Patriots Nielsen reported that the game was the
mostwatched program in history with just over 40 of all
televisions in the United States tuned to the game on Sunday.
Assuming that the 40 rating is correct what is the probabil
ity that in a random sample of 20 television sets 2 or fewer would have been tuned to the
Super Bowl This question can be answered assuming that the binomial distribution applies
using the following steps:
Step 1 Define the characteristics of the binomial distribution.
In this case the characteristics are
n 20 p 0.40 q 1  p 0.60
Step 2 Define the event of interest.
We are interested in knowing
P1x … 22 P102 + P112 + P122
Ronen/Shutterstock
www.downloadslide.comslide 217:
Discrete Probability Distributions
Step 3 Go to the binomial table in Cumulative Binomial Distribution Table to
find the desired probability.
In this case we locate the section of the table corresponding to sample size
equal to n 20 and go to the column headed p 0.40 and the row labeled
x 2. The cumulative P1x … 22 listed in the table is 0.0036.
Thus there is only a 0.0036 chance that 2 or fewer sets in a random
sample of 20 were tuned to the Super Bowl. Thus it is unlikely that 2 or
fewer sets in a sample of 20 TV sets would have been tuned to the Super
Bowl in 2012.
END EXAMPLE
TRY PROBLEM 28
EXAMPLE 4 USING THE BINOMIAL DISTRIBUTION
Clearwater Research Clearwater Research is a fullservice
marketing research consulting firm. Recently it was retained to
do a project for a major U.S. airline. The airline was considering
changing from an assignedseating reservation system to one in
which fliers would be able to take any seat they wished on a
firstcome firstserved basis. The airline believes that 80 of
its fliers would like this change if it was accompanied with a
reduction in ticket prices. Clearwater Research will survey a large number of customers
on this issue but prior to conducting the full research it has selected a random sample of
20 customers and determined that 12 like the proposed change. What is the probability of
finding 12 or fewer who like the change if the probability is 0.80 that a customer will like
the change
If we assume the binomial distribution applies we can use the following steps to answer
this question:
Step 1 Define the characteristics of the binomial distribution.
In this case the characteristics are
n 20 p 0.80 q 1  p 0.20
Step 2 Define the event of interest.
We are interested in knowing
P1x … 122
Step 3 Go to the binomial table in Cumulative Binomial Distribution Table to
find the desired probability.
Locate the table for the sample size n. Locate the column for p 0.80. Go to
the row corresponding to x 12 and the column for p 0.80 in the section of
the table for n 20 to get
P1x … 122 0.0321
Thus it is quite unlikely that if 80 of customers like the new seating plan 12 or fewer
in a sample of 20 would like it. The airline may want to rethink its plan.
END EXAMPLE
TRY PROBLEM 29
Mean and Standard Deviation of the Binomial Distribution In Section 1 we stated
the mean of a discrete probability distribution is also referred to as the expected value. The
expected value of a discrete random variable x is found using Equation 1.
m
x
E1x2 ΣxP1x2
Fernando Blanco Calzada/
Shutterstock
www.downloadslide.comslide 218:
Discrete Probability Distributions
MEAN OF A BINOMIAL DISTRIBUTION This equation for the expected value can be
used with any discrete probability distribution including the binomial. However if we are
working with a binomial distribution the expected value can be found more easily by using
Equation 5.
Expected Value of a Binomial Distribution
m
x
E1x2 np 5
where:
n Sample size
p Probability of a success
BUSINESS APPLICATION BINOMIAL DISTRIBUTION
CATALOG SALES Catalog sales have been a part of the U.S. economy for many years and
companies such as Lands’ End L.L. Bean and Eddie Bauer have enjoyed increased business.
One feature that has made mailorder buying so popular is the ease with which customers can
return merchandise. Nevertheless one mailorder catalog has the goal of no more than 11
of all purchased items returned.
The binomial distribution can describe the number of items returned. For instance in a
given hour the company shipped 300 items. If the probability of an item being returned is
p 0.11 the expected number of items mean to be returned is
m
x
E1x2 np
m
x
E1x2 1300210.112 33
Thus the average number of returned items for each 300 items shipped is 33.
Suppose the company sales manager wants to know if the return rate is stable at 11. To
test this she monitors a random sample of 300 items and finds that 44 have been returned.
This return rate exceeds the mean of 33 units which concerns her. However before reaching
a conclusion she will be interested in the probability of observing 44 or more returns in a
sample of 300.
P1x Ú 442 1  P1x … 432
The Cumulative Binomial Distribution Table does not contain sample sizes as large as
300. Instead we can use Excel’s BINOM.DIST function to find the probability. The Excel
output in Figure 3 shows the cumulative probability of 43 or fewer is equal to
P1x … 432 0.97
Then the probability of 44 or more returns is
P1x Ú 442 1  0.97 0.03
There is only a 3 chance of 44 or more items being returned if the 11 return rate
is still in effect. This low probability suggests that the return rate may have increased
above 11 because we would not expect to see 44 returned items. The probability is
very small.
Excel Tutorial
Excel
tutorials
www.downloadslide.comslide 219:
Discrete Probability Distributions
Excel 2010 Instructions:
1. Open blank worksheet.
2. Click on f
x
function
wizard.
3. Select Statistical category.
4. Select the BINOM.DIST
function.
5. Fill in the requested
information in the template.
6. True indicates cumulative
probabilities.
Minitab Instructions for similar results:
1. Choose Calc Probability
Distribution Binomial.
2. Choose Cumulative probability.
3. In Number of trials enter sample size.
4. In Probability of success enter p.
5. In Input constant enter
the number of successes: x.
6. Click OK.
Cumulative Probability
FIGURE 3 
Excel 2010 Binomial
Distribution Output for
Catalog Sales
EXAMPLE 5 FINDING THE MEAN OF THE BINOMIAL DISTRIBUTION
Clearwater Research In Example 4 Clearwater Research had
been hired to do a study for a major airline that is planning to change
from a designatedseat assignment plan to an open seating system.
The company believes that 80 of its customers approve of the idea.
Clearwater Research interviewed a sample of n 20 and found 12
who like the proposed change. If the airline is correct in its assess
ment of the probability what is the expected number of people in a
sample of n 20 who will like the change We can find this using the following steps:
Step 1 Define the characteristics of the binomial distribution.
In this case the characteristics are
n 20 p 0.80 q 1  p 0.20
Step 2 Use Equation 5 to find the expected value.
m
x
E1x2 np
E1x2 2010.802 16
The average number who would say they like the proposed change is 16 in a
sample of 20.
END EXAMPLE
TRY PROBLEM 33a
STANDARD DEVIATION OF A BINOMIAL DISTRIBUTION The standard deviation for
any discrete probability distribution can be calculated using Equation 2. We show this again as
x
xEx Px ∑
2
Fernando Blanco Calzada/
Shutterstock
www.downloadslide.comslide 220:
Discrete Probability Distributions
If a discrete probability distribution meets the binomial distribution conditions the stand
ard deviation is more easily computed by Equation 6.
Standard Deviation of the Binomial Distribution
npq
6
where:
n Sample size
p Probability of a success
q 1  p Probability of a failure
Probability Px
14 18 22 26 30 34 38
Number of Successes x c
0.12
0.10
0.08
0.06
0.04
0.02
0
Probability Px
010 8 6 429 7 5 3 1
Number of Successes x b
0.30
0.25
0.20
0.15
0.10
0.05
0
Probability Px
05 4 3 2 1
Number of Successes x a
n 5 p 0.50 n 10 p 0.50 n 50 p 0.50
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0
FIGURE 4 
The Binomial Distribution
with Varying Sample Sizes
1p 0.502
EXAMPLE 6 FINDING THE STANDARD DEVIATION OF A BINOMIAL
DISTRIBUTION
Clearwater Research Refer to Examples 4 and 5 in which
Clearwater Research surveyed a sample of n 20 airline custom
ers about changing the way seats are assigned on flights. The airline
believes that 80 of its customers approve of the proposed change.
Example 5 showed that if the airline is correct in its assessment the
expected number in a sample of 20 who would like the change is
16. However there are other possible outcomes if 20 customers are
surveyed. What is the standard deviation of the random variable x in this case We can find the
standard deviation for the binomial distribution using the following steps:
Step 1 Define the characteristics of the binomial distribution.
In this case the characteristics are
n 20 p 0.80 q 1  p 0.20
Step 2 Use Equation 6 to calculate the standard deviation.
npq 20 0 80 0 20 1 7889 . . .
END EXAMPLE
TRY PROBLEM 33b
Additional Information about the Binomial Distribution At this point several com
ments about the binomial distribution are worth making. If p the probability of a success
is 0.50 the binomial distribution is symmetrical and bellshaped regardless of the sam
ple size. This is illustrated in Figure 4 which shows frequency histograms for samples of
n 5 n 10 and n 50. Notice that all three distributions are centered at the expected
value E1x2 np.
When the value of p differs from 0.50 in either direction the binomial distribution is
skewed. The skewness will be most pronounced when n is small and p approaches 0 or 1.
However the binomial distribution becomes more bell shaped as n increases. The frequency
histograms shown in Figure 5 bear this out.
Fernando Blanco Calzada/
Shutterstock
www.downloadslide.comslide 221:
Discrete Probability Distributions
Probability Px
011 8910 6 4 27 5 3 1
Number of Successes x c
0.30
0.25
0.20
0.15
0.10
0.05
0
05 4 3 2 1
Number of Successes x a
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
Probability Px
Probability Px
08 6 4 27 5 3 1
Number of Successes x b
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0
n 10 p 0.05 n 20 p 0.05 n 50 p 0.05 FIGURE 5 
The Binomial Distribution
with Varying Sample Sizes
1p 0.052
Skill Development
522. The manager for State Bank and Trust has recently
examined the credit card account balances for the
customers of her bank and found that 20 have an
outstanding balance at the credit card limit. Suppose
the manager randomly selects 15 customers and finds
4 that have balances at the limit. Assume that the
properties of the binomial distribution apply.
a. What is the probability of finding 4 customers in
a sample of 15 who have “maxed out” their credit
cards
b. What is the probability that 4 or fewer customers
in the sample will have balances at the limit of the
credit card
523. For a binomial distribution with a sample size equal
to 10 and a probability of a success equal to 0.30
what is the probability that the sample will contain
exactly three successes Use the binomial formula to
determine the probability.
524. Use the binomial formula to calculate the following
probabilities for an experiment in which n 5 and
p 0.4:
a. the probability that x is at most 1
b. the probability that x is at least 4
c. the probability that x is less than 1
525. If a binomial distribution applies with a sample size of
n 20 find
a. the probability of 5 successes if the probability of a
success is 0.40
b. the probability of at least 7 successes if the
probability of a success is 0.25
c. the expected value n 20 p 0.20
d. the standard deviation n 20 p 0.20
526. A report issued by the American Association of
Building Contractors indicates that 40 of all home
buyers will do some remodeling to their home within
the first five years of home ownership. Assuming this
is true use the binomial distribution to determine the
probability that in a random sample of 20 homeowners
2 or fewer will remodel their homes. Use the binomial
table.
527. Find the probability of exactly 5 successes in a sample
of n 10 when the probability of a success is 0.70.
528. Assuming the binomial distribution applies with a
sample size of n 15 find
a. the probability of 5 or more successes if the
probability of a success is 0.30
b. the probability of fewer than 4 successes if the
probability of a success is 0.75
c. the expected value of the random variable if the
probability of success is 0.40
d. the standard deviation of the random variable if the
probability of success is 0.40
529. A random variable follows a binomial distribution with
a probability of success equal to 0.65. For a sample
size of n 7 find
a. the probability of exactly 3 successes
b. the probability of 4 or more successes
c. the probability of exactly 7 successes
d. the expected value of the random variable
530. A random variable follows a binomial distribution with
a probability of success equal to 0.45. For n 11 find
a. the probability of exactly 1 success
b. the probability of 4 or fewer successes
c. the probability of at least 8 successes
531. Use the binomial distribution table to determine the
following probabilities:
a. n 6 p 0.08 find P1x 22
b. n 9 p 0.80 determine P1x 6 42
c. n 11 p 0.65 calculate P12 6 x … 52
d. n 14 p 0.95 find P1x Ú 132
e. n 20 p 0.50 compute P1x 7 32
532. Use the binomial distribution in which n 6 and
p 0.3 to calculate the following probabilities:
a. x is at most 1.
b. x is at least 2.
c. x is more than 5.
d. x is less than 6.
5 Exercises
MyStatLab
www.downloadslide.comslide 222:
Discrete Probability Distributions
b. If Intel and Dell agree that Intel will not provide
more than 5 defective chips calculate the
probability that the entire shipment will be returned
even though only 5 are defective.
c. Calculate the probability that the entire shipment
will be kept by Dell even though the shipment has
10 defective microprocessors.
537. In his article titled “Acceptance Sampling Solves
Drilling Issues: A Case Study” published in
Woodworking Magazine author Ken Wong discusses
a problem faced by furniture manufacturing
companies dealing with the quality of the drilling
of dowel holes. Wong states “Incorrect sizing and
distances with respect to dowel holes can cause
many problems for the rest of the process especially
when drilling is conducted early in the production
process.”
Consider the case of Dragon Wood Furniture
in Bismarck North Dakota which believes
that when the drilling process is operating
at an acceptable rate the upper limit on the
percentage of incorrectly drilled dowel holes is
4. To monitor its drilling process Dragon Wood
Furniture randomly samples 20 products each hour
and determines if the dowel hole in each product
is correctly drilled or not. If in the sample of
20 holes 1 or more incorrectly drilled holes is
discovered the production process is stopped and
the drilling process is recalibrated.
a. If the process is really operating correctly
1p 0.042 what is the probability that the
sampling effort will produce x 0 defective holes
and thus the process will properly be left to continue
running
b. Suppose the true defect rate has risen to 0.10. What
is the probability the sample will produce results
that properly tell the managers to halt production to
recalibrate the drilling machine
c. Prepare a short letter to the manufacturing
manager at Dragon Wood Furniture discussing
the effectiveness of the sampling process that her
company is using. Base your response on the results
to parts a and b.
538. Mooney Hileman Jones a marketing agency
located in Cleveland has created an advertising
campaign for a major retail chain which the
agency’s executives believe is a winner. For an ad
campaign to be successful at least 80 of those
seeing a television commercial must be able to recall
the name of the company featured in the commercial
one hour after viewing the commercial. Before
distributing the ad campaign nationally the company
plans to show the commercial to a random sample
of 20 people. It will also show the same people two
additional commercials for different products or
businesses.
a. Assuming that the advertisement will be successful
80 will be able to recall the name of the company
533. Given a binomial distribution with n 8 and
p 0.40 obtain the following:
a. the mean
b. the standard deviation
c. the probability that the number of successes is
larger than the mean
d. the probability that the number of successes is
within 2 standard deviations of the mean
Business Applications
534. Magic Valley Memorial Hospital administrators
have recently received an internal audit report that
indicates that 15 of all patient bills contain an error
of one form or another. After spending considerable
effort to improve the hospital’s billing process
the administrators are convinced that things have
improved. They believe that the new error rate is
somewhere closer to 0.05.
a. Suppose that recently the hospital randomly
sampled 10 patient bills and conducted a thorough
study to determine whether an error exists. It found
3 bills with errors. Assuming that managers are
correct that they have improved the error rate to
0.05 what is the probability that they would find 3
or more errors in a sample of 10 bills
b. Referring to part a what conclusion would you
reach based on the probability of finding 3 or more
errors in the sample of 10 bills
535. The Center for the Study of the American Electorate
indicated that 64 of the votingage voters cast ballots
in the 2008 presidential election. It also indicated in the
west 34.6 of votingage voters voted for a Democrat
as a Representative an increase from 30.8 in 2004.
A startup company in San Jose California has 10
employees.
a. How many of the employees would you expect to
have voted for a Democrat as a Representative
b. All of the employees indicated that they voted
in the 2008 presidential election. Determine the
probability of this assuming they followed the
national trend.
c. Eight of the employees voted for a Democratic
Representative. Determine the probability that
at least 8 of the employees would vote for the
Democrat if they followed the national trend.
d. Based on your calculations in parts b and c do the
employees reflect the national trend Support your
answer with statistical calculations and reasoning.
536. Dell Computers receives large shipments of micro
processors from Intel Corp. It must try to ensure the
proportion of microprocessors that are defective is
small. Suppose Dell decides to test five microprocessors
out of a shipment of thousands of these microprocessors.
Suppose that if at least one of the microprocessors is
defective the shipment is returned.
a. If Intel Corp.’s shipment contains 10 defective
microprocessors calculate the probability the entire
shipment will be returned.
www.downloadslide.comslide 223:
Discrete Probability Distributions
a. Suppose that the producers of NCIS commissioned
a study that called for the consultants to randomly
call 25 people immediately after the NCIS time
slot and interview those who said that they had just
watched NCIS. Suppose the consultant submits a
report saying that it found no one in the sample
of 25 homes who claimed to have watched the
program and therefore did not do any surveys. What
is the probability of this happening assuming that
the Nielsen ratings for the show are accurate
b. Assume the producers for The Big Bang Theory
planned to survey 1000 people on the day following
the broadcast of the program. The purpose of the
survey was to determine what the reaction would
be if one of the leading characters was retired from
the show. Based on the Nielsen ratings what would
be the expected number of people who would end
up being included in the analysis assuming that all
1000 people could be reached
541. A small hotel in a popular resort area has 20 rooms.
The hotel manager estimates that 15 of all confirmed
reservations are “noshows.” Consequently the hotel
accepts confirmed reservations for as many as 25
rooms. If more confirmed reservations arrive than there
are rooms the overbooked guests are sent to another
hotel and given a complementary dinner. If the hotel
currently has 25 confirmed reservations find
a. the probability that no customers will be sent to
another hotel
b. the probability that exactly 2 guests will be sent to
another hotel
c. the probability that 3 or more guests will be sent to
another hotel
542. A manufacturing firm produces a product that has
a ceramic coating. The coating is baked on to the
product and the baking process is known to produce
15 defective items for example cracked or chipped
finishes. Every hour 20 products from the thousands
that are baked hourly are sampled from the ceramic
coating process and inspected.
a. What is the probability that 5 defective items will be
found in the next sample of 20
b. On average how many defective items would be
expected to occur in each sample of 20
c. How likely is it that 15 or more nondefective
good items would occur in a sample due to
chance alone
543. The Employee Benefit Research Institute reports that
69 of workers reported that they and/or their spouse
had saved some money for retirement.
a. If a random sample of 30 workers is taken what
is the probability that fewer than 17 workers and/
or their spouses have saved some money for
retirement
b. If a random sample of 50 workers is taken what
is the probability that more than 40 workers and/
or their spouses have saved some money for
retirement
in the ad what is the expected number of people in
the sample who will recall the company featured in
the Mooney Hileman Jones commercial one hour
after viewing the three commercials
b. Suppose that in the sample of 20 people 11
were able to recall the name of the company in
the Mooney Hileman Jones commercial one
hour after viewing. Based on the premise that the
advertising campaign will be successful what is
the probability of 11 or fewer people being able to
recall the company name
c. Based on your responses to parts a and b what
conclusion might Mooney Hileman Jones
executives make about this particular advertising
campaign
539. A survey by KRC Research for U.S. News reported that
37 of people plan to spend more on eating out after
they retire. If eight people are randomly selected then
determine the
a. expected number of people who plan to spend more
on eating out after they retire
b. standard deviation of the individuals who plan to
spend more on eating out after they retire
c. probability that two or fewer in the sample indicate
that they actually plan to spend more on eating out
after retirement
540. Nielsen is the major media measurement company
and conducts surveys to determine household
viewing choices. The following table shows the
top 10 broadcast television programs for the week
of January 23 2012.
Rank Program Network HH Rating Viewers
1 American Idol—
W ednesday
FOX 11.1 19.671
2 American Idol—
Thursday
FOX 10.0 17.141
3 Big Bang
Theory
CBS 9.7 16.130
4 CSI CBS 9.0 14.257
5 Criminal Minds CBS 8.7 13.815
6 NCIS CBS 8.1 12.548
7 Undercover
Boss
CBS 7.9 13.151
8 AFCNFC
ProBowl
NBC 7.3 12.498
9 The Good Wife CBS 7.2 11.083
10 60 Minutes CBS 7.1 11.188
Rank is based on U.S. Household Rating from Nielsen’s National People Meter
Sample.
A household rating is the estimate of the size of a television audience relative to
the total universe expressed as a percentage. As of August 27 2012 there are an
estimated 114200000 television households in the United States. A single national
household ratings point represents 1 or 1142000 households.
Measured in millions includes all persons over the age of two.
Source: www.nielsenmedia.com
www.downloadslide.comslide 224:
Discrete Probability Distributions
the percent reported by Ms. Borrus is correct a sample
of 20 shareholders was asked if they held their stock in
“street names.” Seventeen responded that they did.
a. Supposing the true proportion of shareholders that
hold stock under street names is 0.80 calculate
the probability that 17 or more of the sampled
individuals hold their stock under street names.
b. Repeat the calculation in part a using proportions of
0.70 and 0.90.
c. Based on your calculations in parts a and b which
proportion do you think is most likely true Support
your answer.
Computer Database Exercises
547. USA Today has reported on the gender gap that exists
between married spouses. One of the measures is
the number of women who outearn their husbands.
According to a study conducted by the Bureau of
Labor Statistics 32.5 of female spouses outearn their
male counterparts. The file titled Gendergap contains
the incomes of 150 married couples in Utah.
a. Determine the number of families in which the
female outearns her husband.
b. Calculate the expected number of female spouses
who outearn their male counterparts in the sample
of 150 married couples based on the Bureau of
Labor Statistics study.
c. If the percentage of married women in Utah
who outearn their male spouses is the same as
that indicated by the Bureau of Labor Statistics
determine the probability that at least the number
found in part a would occur.
d. Based on your calculation in part c does the Bureau
of Labor Statistics’ percentage seem plausible if
Utah is not different from the rest of the United
States
548. Tony Hsieh is CEO of etailer Zappos.com. His
company sells shoes online. It differentiates itself
by its selection of shoes and a devotion to customer
service. It offers free shipping and free return shipping.
An area in which costs could be cut back is the
shipping charges for return shipping specifically those
that result from the wrong size of shoes being sent.
Zappos may try to keep the percentage of returns due
to incorrect size to no more than 5. The file titled
Shoesize contains a sample of 125 shoe sizes that
were sent to customers and the sizes that were actually
ordered.
a. Determine the number of pairs of wrongsize shoes
that were delivered to customers.
b. Calculate the probability of obtaining at least that
many pairs of wrongsized shoes delivered to
customers if the proportion of incorrect sizes is
actually 0.05.
c. On the basis of your calculation determine whether
Zappos has kept the percentage of returns due to
incorrect size to no more than 5. Support your
answer with statistical reasoning.
544. Radio frequency identification RFID is an electronic
scanning technology that can be used to identify items
in a number of ways. One advantage of RFID is that
it can eliminate the need to manually count inventory
which can help improve inventory management. The
technology is not infallible however and sometimes
errors occur when items are scanned. If the probability
that a scanning error occurs is 0.0065 use either Excel
or Minitab to find
a. the probability that exactly 20 items will be scanned
incorrectly from the next 5000 items scanned
b. the probability that more than 20 items will be
scanned incorrectly from the next 5000 items
scanned
c. the probability that the number of items scanned
incorrectly is between 10 and 25 from the next
5000 items scanned
d. the expected number of items scanned incorrectly
from the next 5000 items scanned
545. Peter S. Kastner director of the consulting firm
Vericours Inc. reported that 40 of all rebates are
not redeemed because consumers either fail to apply
for them or their applications are rejected. TCA
Fulfillment Services published its redemption rates:
50 for a 30 rebate on a 100 product 10 for a 10
rebate on a 100 product and 35 for a 50 rebate on
a 200 product.
a. Calculate the weighted average proportion of
redemption rates for TCA Fulfillment using the size
of the rebate to establish the weights. Does it appear
that TCA Fulfillment has a lower rebate rate than
that indicated by Vericours Explain.
b. To more accurately answer the question posed in
part a a random sample of 20 individuals who
purchased an item accompanied by a rebate could
be asked if they submitted their rebate. Suppose
four of the questioned individuals said they did
redeem their rebate. If Vericours’ estimate of the
redemption rate is correct determine the expected
number of rebates that would be redeemed. Does it
appear that Vericours’ estimate may be too high
c. Determine the likelihood that such an extreme
sample result as indicated in part b or something
more extreme would occur if the weighted average
proportion provides the actual rebate rate.
d. Repeat the calculations of part c assuming that
Vericours’ estimate of the redemption rate is
correct.
e. Are you convinced that the redemption rate is
smaller than that indicated by Vericours Explain.
546. Business Week reported that business executives
want to break down the obstacles that keep them
from communicating directly with stock owners. Ms.
Borrus reports that 80 of shareholders hold stock
in “street names” which are registered with their
bank or brokerage. If the brokerage doesn’t furnish
these names to the corporations executives cannot
communicate with their shareholders. To determine if
www.downloadslide.comslide 225:
Discrete Probability Distributions
d. If Zappos sells 5 million pairs of shoes in one
year and it costs an average of 4.75 a pair to
return them calculate the expected cost associated
with wrongsized shoes being returned using the
probability calculated from the sample data.
549. International Data Corp. IDC has shown that the
average return on business analytics projects was
almost fourandahalf times the initial investment.
Analytics consists of tools and applications that present
better metrics to the user and to the probable future
outcome of an event. IDC looked at how long it takes a
typical company to recoup its investment in analytics.
It determined that 29 of the U.S. corporations that
adopted analytics took six months or less to recoup
their investment. The file titled Analytics contains a
sample of the time it might have taken 35 corporations
to recoup their investment in analytics.
a. Determine the number of corporations that
recovered their investment in analytics in six
months or less.
b. Calculate the probability of obtaining at most the
number of corporations that you determined in part
a if the percent of those recovering their investment
is as indicated by IDC.
c. Determine the 70th percentile of the number of the
35 corporations that recovered their investment in
analytics in six months or less. Hint: Recall and use
the definition of percentiles from Section 1.
END EXERCISES 52
Other Discrete Probability
Distributions
The binomial distribution is very useful in many business situations as indicated by the
examples and applications presented in the previous section. However as we pointed out
there are several requirements that must hold before we can use the binomial distribution to
determine probabilities. If those conditions are not satisfied there may be other theoretical
probability distributions that could be employed. In this section we introduce two other very
useful discrete probability distributions: the Poisson distribution and the hypergeometric
distribution.
The Poisson Distribution
To use the binomial distribution we must be able to count the number of successes and the
number of failures. Although in many situations you may be able to count the number of
successes you often cannot count the number of failures. For example suppose a company
builds freeways in Vermont. The company could count the number of potholes that develop
per mile here a pothole is referred to as a success because it is what we are looking for but
how could it count the number of nonpotholes Or what about a hospital supplying emer
gency medical services in Los Angeles It could easily count the number of emergencies its
units respond to in one hour but how could it determine how many calls it did not receive
Obviously in these cases the number of possible outcomes of 1successes + failures2 is
difficult if not impossible to determine. If the total number of possible outcomes cannot be
determined the binomial distribution cannot be applied. In these cases you may be able to use
the Poisson distribution.
Characteristics of the Poisson Distribution The Poisson distribution
1
describes a
process that extends over time space or any welldefined unit of inspection. The outcomes of
interest such as emergency calls or potholes occur at random and we count the number of
outcomes that occur in a given segment of time or space. We might count the number of emer
gency calls in a onehour period or the number of potholes in a twomile stretch of freeway.
As we did with the binomial distribution we will call these outcomes successes even though
like potholes they might be undesirable.
The possible counts are the integers 0 1 2 c and we would like to know the probabil
ity of each of these values. For example what is the chance of getting exactly four emergency
calls in a particular hour What is the chance that a chosen twomile stretch of freeway will
contain zero potholes
Poisson Distribution
The Poisson distribution describes a process
that extends over space time or any well
defined interval or unit of inspection in which
the outcomes of interest occur at random and
the number of outcomes that occur in any given
interval are counted. The Poisson distribution
rather than the binomial distribution is used
when the total number of possible outcomes
cannot be determined.
1
The Poisson distribution can be derived as the limiting distribution of the binomial distribution as the number of
trials n tends to infinity and the probability of success decreases to zero. It serves as a good approximation to the
binomial when n is large.
Chapter Outcome 3.
www.downloadslide.comslide 226:
Discrete Probability Distributions
We can use the Poisson probability distribution to answer these questions if we make the
following assumptions:
1. We know l the average number of successes in one segment. For example we know
that there is an average of 8 emergency calls per hour 1l 82 or an average of 15 pot
holes per mile of freeway 1l 152.
2. The probability of x successes in a segment is the same for all segments of the same
size. For example the probability distribution of emergency calls is the same for any
onehour period of time at the hospital.
3. What happens in one segment has no influence on any nonoverlapping segment. For
example the number of calls arriving between 9:30 p.m. and 10:30 p.m. has no influence
on the number of calls between 11:00 p.m. and 12:00 midnight.
4. We imagine dividing time or space into tiny subsegments. Then the chance of more than
one success in a subsegment is negligible and the chance of exactly one success in a tiny
subsegment of length t is lt. For example the chance of two emergency calls in the same
second is essentially 0 and if l 8 calls per hour the chance of a call in any given second
is 18211/36002 ≈ 0.0022.
Once l has been determined we can calculate the average occurrence rate for any number
of segments t. This is lt. Note that l and t must be in compatible units. If we have l 20
arrivals per hour the segments must be in hours or fractional parts of an hour. That is if we have
l 20 per hour and we wish to work with halfhour time periods the segment would be
t 5
1
2
hour
not t 30 minutes.
Although the Poisson distribution is often used to describe situations such as the number
of customers who arrive at a hospital emergency room per hour or the number of calls the
HewlettPackard LaserJet printer service center receives in a 30minute period the segments
need not be time intervals. Poisson distributions are also used to describe such random vari
ables as the number of knots in a sheet of plywood or the number of contaminants in a gallon
of lake water. The segments would be the sheet of plywood and the gallon of water.
Another important point is that lt the average number in t segments is not necessarily
the number we will see if we observe the process for t segments. We might expect an aver
age of 20 people to arrive at a checkout stand in any given hour but we do not expect to find
exactly that number arriving every hour. The actual arrivals will form a distribution with an
expected value or mean equal to lt. So for the Poisson distribution
E3x4 m
x
lt
Once l and t have been specified the probability for any discrete value in the Poisson
distribution can be found using Equation 7.
Poisson Probability Distribution
Px
te
x
xt
−
7
where:
t Number of segments of interest
x Number of successes in t segments
l Expected number of successes in one segment
e Base of the natural logarithm system 12.71828 c2
BUSINESS APPLICATION POISSON DISTRIBUTION
WHOLE FOODS GROCERY A study conducted at Whole Foods Grocery shows that the
average number of arrivals to the checkout section of the store per hour is 16. Further the
distribution for the number of arrivals is considered to be Poisson distributed. Figure 6 shows
www.downloadslide.comslide 227:
Discrete Probability Distributions
the shape of the Poisson distribution for l 16. The probability of each possible number of
customers arriving can be computed using Equation 7. For example we can find the probability
of x 12 customers in one hour 1t 12 as follows:
Px
te
x
e
xt
. 12
16
12
0 0661
12 16
Poisson Probability Distribution Table As was the case with the binomial distribu
tion a table of probabilities exists for the Poisson distribution. The Poisson table appears in
Cumulative Poisson Probability Distribution Table. The Poisson table shows the cumulative
probabilities for x or fewer occurrences for different lt values. We can use the following busi
ness application to illustrate how to use the Poisson table.
BUSINESS APPLICATION USING THE POISSON DISTRIBUTION TABLE
WHOLE FOODS GROCERY CONTINUED At Whole Foods Grocery customers are
thought to arrive at the checkout section according to a Poisson distribution with l 16
customers per hour. See Figure 6. Based on previous studies the store manager believes that
the service time for each customer is quite constant at six minutes. Suppose during each six
minute time period the store has three checkers available. This means that three customers
can be served during each sixminute segment. The manager is interested in the probability
that one or more customers will have to wait for service during a sixminute period.
To determine this probability you will need to convert the mean arrivals from l 16
customers per hour to a new average for a sixminute segment. Six minutes corresponds to
0.10 hours so you will change the segment size t 0.10. Then the mean number of arrivals
in six minutes is lt 1610.102 1.6 customers.
Now because there are three checkers any time four or more customers arrive in a six
minute period at least one customer will have to wait for service. Thus
P11 or more customers wait2 P142 + P152 + P162 + c
or you can use the Complement Rule as follows:
P11 or more customers wait2 1  P1x … 32
The Poisson table in Cumulative Poisson Probability Distribution Table can be used to
find the necessary probabilities. To use the table first go across the top of the table until you
find the desired value of lt. In this case look for lt 1.6. Next go down the lefthand side
to find the value of x corresponding to the number of occurrences of interest. For example
consider x 3 customer arrivals. Because Cumulative Poisson Probability Distribution Table
Poisson Probability
0
Number of Customers x
Px Mean 16
0.1200
0.1000
0.0800
0.0600
0.0400
0.0200
0.0000
123 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
FIGURE 6 
Poisson Distribution for Whole
Foods Checkout Arrivals with
l 16
www.downloadslide.comslide 228:
Discrete Probability Distributions
is a cumulative Poisson table the probabilities are for x or fewer occurrences. This the prob
ability of x … 3 is given in the table as 0.9212. Thus
P1x … 32 0.9212
Then the probability of four or more customers arriving is
P14 or more customers2 1  P1x … 32
P14 or more customers2 1  0.9212 0.0788
Given the store’s capacity to serve three customers in a sixminute period the probability of
one or more customers having to wait is 0.0778.
Suppose that when the store manager sees this probability she is somewhat concerned.
She states that she wants enough checkout stands open so that the chance of a customer wait
ing does not exceed 0.05. To determine the appropriate number of checkers you can use the
Poisson table to find the following:
P14 or more customers2 1  P1x … 2 … 0.05
In other words a customer will have to wait if more customers arrive than there are checkers.
As long as the number of arrivals x is less than or equal to the number of checkers no one
will wait. Then what value of x will provide the following
1  P1x … 2 … 0.05
Therefore you want
P1x … 2 Ú 0.95
You can go to the table for lt 1.6 and scan down the column starting with
P1x 02 0.2019 until the cumulative probability listed is 0.95 or higher. When you reach
x 4 the cumulative probability P1x … 42 Ú 0.9763. Then
P14 or more customers2 1  P1x … 42 1  0.9763 0.0237
Because 0.0237 is less than or equal to the 0.05 limit imposed by the manager she would
have to schedule four checkers.
EXAMPLE 7 USING THE POISSON DISTRIBUTION
Fashion Leather Products Fashion Leather Products headquartered in Argentina makes
leather clothing for export to many other countries around the world. Before shipping quality
managers perform tests on the leather products. The industry standards call for the average
number of defects per square meter of leather to not exceed five. During a recent test the
inspector selected 3 square meters finding 18 defects. To determine the probability of this
event occurring if the leather meets the industry standards assuming that the Poisson distribu
tion applies the company can perform the following steps:
Step 1 Define the segment unit.
Because the mean was stated as five defects per square meter the segment unit
in this case is one meter.
Step 2 Determine the mean of the random variable.
In this case if the company meets the industry standards the mean will be
l 5
Step 3 Determine the segment size t.
The company quality inspectors analyzed 3 square meters which is equal to
3 units. So t 3.0 Then
lt 15213.02 15.0
When looking at 3 square meters the company would expect to find at most
15.0 defects if the industry standards are being met.
Step 4 Define the event of interest and use the Poisson formula or the Poisson
tables to find the probability.
Example 7
Using the Poisson Distribution
The following steps are used to find
probabilities using the Poisson
distribution:
Define the segment units.
The segment units are usually
blocks of time areas of space or
volume.
Determine the mean of the
random variable.
The mean is the parameter that
defines the Poisson distribution
and is referred to as l. It is the
average number of successes in
a segment of unit size.
Determine t the number of the
segments to be considered and
then calculate lt.
Define the event of interest
and use the Poisson formula
or the Poisson table to find the
probability.
www.downloadslide.comslide 229:
Discrete Probability Distributions
In this case 18 defects were observed. Because 18 exceeds the expected
number 1lt 15.02 the company would want to find
P1x Ú 182 P1x 182 + P1x 192 + c
The Poisson table in Cumulative Poisson Probability Distribution Table is
used to determine these probabilities. Locate the desired probability under
the column headed lt 15.0 Then find the values of x down the lefthand
column.
P1x Ú 182 1  P1x … 172
1  0.7489
0.2511
There is about a 0.25 chance of finding 18 or more defects in 3 square meters of leather prod
ucts made by Fashion Leather if they are meeting the quality standard.
END EXAMPLE
TRY PROBLEM 50
The Mean and Standard Deviation of the Poisson Distribution The mean of
the Poisson distribution is lt. This is the value we use to specify which Poisson distribu
tion we are using. We must know the mean before we can find probabilities for a Poisson
distribution.
Figure 6 illustrated that the outcome of a Poisson distributed variable is subject to vari
ation. Like any other discrete probability distribution the standard deviation for the Poisson
can be computed using Equation 2:
x
xEx Px ∑−
2
However for a Poisson distribution the standard deviation also can be found using
Equation 8.
Standard Deviation of the Poisson Distribution
t
8
The standard deviation of the Poisson distribution is simply the square root of the mean.
Therefore if you are working with a Poisson process reducing the mean will reduce the vari
ability also.
BUSINESS APPLICATION THE POISSON PROBABILITY DISTRIBUTION
HERITAGE TILE To illustrate the importance of the relationship between the mean and
standard deviation of the Poisson distribution consider Heritage Tile in New York City. The
company makes ceramic tile for kitchens and bathrooms. The quality standards call for the
number of imperfections in a tile to average 3 or fewer. The distribution of imperfections
is thought to be Poisson. Many software packages including Excel will generate Poisson
probabilities in much the same way as for the binomial distribution which was discussed in
Section 2. If we assume that the company is meeting the standard Figure 7 shows the Poisson
probability distribution generated using Excel when lt 3.0. Even though the average
number of defects is 3 the manager is concerned about the high probabilities associated with
the number of imperfections equal to 4 5 6 or more on a tile. The variability is too great.
Using Equation 5 the standard deviation for this distribution is
30 1732 ..
This large standard deviation means that although some tiles will have few if any imperfec
tions others will have several causing problems for installers and unhappy customers.
Excel Tutorial
Excel
tutorials
www.downloadslide.comslide 230:
Discrete Probability Distributions
Minitab Instructions for similar results:
1. Create column Number with integers
from 0 to 10.
2. Choose Calc Probability
Distributions Poisson.
3. Select Probability.
4. In Mean enter 3.
5. Select Input column.
6. In Input column enter the column of
integers.
7. In Optional storage enter column
Probability.
8. Click OK.
9. Choose Graph Bar Chart.
10. In Bars represent select Values from
a Table select Simple.
11. Click OK.
12. In Graph variables Insert Probability.
13. In Categorical variable insert Number.
14. Click OK.
Excel 2010 Instructions:
1. Enter values for x ranging
from 0 to 10.
2. Place the cursor in the
frst blank cell in the next
column.
3. Click on f
x
function
wizard and then select
the Statistical category.
4. Select the POISSON.
DIST function.
5. Reference the cell with
the desired x value and
enter the mean. Enter
False to choose
noncumulative
probabilities.
6. Copy function down for
all values of x.
7. Graph using Insert
Column. Remove gaps
and add lines to the bars.
Label axes and add titles.
FIGURE 7 
Excel 2010 Output for
Heritage Tile Example
A qualityimprovement effort directed at reducing the average number of imperfections
to 2.0 would also reduce the standard deviation to
20 1414 ..
Further reductions in the average would also reduce variation in the number of imperfec
tions between tiles. This would mean more consistency for installers and higher customer
satisfaction.
The Hypergeometric Distribution
Although the binomial and Poisson distributions are very useful in many business
decisionmaking situations they both require that the trials be independent. For instance
in binomial applications the probability of a success in one trial must be the same as the
probability of a success in any other trial. Although there are certainly times when this
assumption can be satisfied or at least approximated in instances in which the population
is fairly small and we are sampling without replacement the condition of independence
will not hold. In these cases a discrete probability distribution referred to as the hypergeo
metric distribution can be useful.
Chapter Outcome 3.
www.downloadslide.comslide 231:
Discrete Probability Distributions
BUSINESS APPLICATION THE HYPERGEOMETRIC PROBABILITY DISTRIBUTION
DOLBY INDUSTRIES Dolby Industries contracts with a
Chinese manufacturer to make women’s handbags. Because
of the intense competition in the marketplace for handbags
Dolby has made every attempt to provide highquality products.
However a recent production run of 20 handbags of a particular
model contained 2 units that tested out as defective. The problem
was traced to a shipment of defective latches that Dolby’s
Chinese partner received shortly before the production run started.
The production manager ordered that the entire batch of 20 handbags be isolated from
other production output until further testing could be completed. Unfortunately a new ship
ping clerk packaged 10 of these isolated handbags and shipped them to a California retailer to
fill an order that was already overdue. By the time the production manager noticed what had
happened the handbags were already in transit.
The immediate concern was whether one or more of the defectives had been included
in the shipment. The new shipping clerk thought there was a good chance that no defectives
were included. Short of reinspecting the remaining handbags how might Dolby Industries
determine the probability that no defectives were actually shipped
At first glance it might seem that the question could be answered by employing the bino
mial distribution with n 10 p 220 0.10 and x 0. Using the binomial distribution
table in Cumulative Binomial Distribution Table we get
P1x 02 0.3487
There is a 0.3487 chance that no defectives were shipped assuming the selection process satisfied
the requirements of a binomial distribution. However for the binomial distribution to be applica
ble the trials must be independent and the probability of a success p must remain constant from
trial to trial. In order for this to occur when the sampling is from a “small” finite population as is
the case here the sampling must be performed with replacement. This means that after each item is
selected it is returned to the population and therefore may be selected again later in the sampling.
In the Dolby example the sampling was performed without replacement because each
handbag could only be shipped one time. Also the population of handbags is finite with size
N 20 which is a “small” population. Thus p the probability of a defective handbag does
not remain equal to 0.10 on each trial. The value of p on any particular trial depends on what
has already been selected on previous trials.
The event of interest is
G G G G G G G G G G
The probability that the first item selected for shipment would be good would be 18/20
because there were 18 good handbags in the batch of 20. Now assuming the first unit selected
was good the probability the second unit was good is 17/19 because we then had only 19
handbags to select from and 17 of those would be good. The probability that all 10 items
selected were good is
18
20
17
19
16
18
15
17
14
16
13
15
12
14
11
13
10
1
×××× ×× × ×
2 2
9
11
0 2368 × 5 .
This value is not the same as the 0.3847 probability we got when the binomial distri
bution was used. This demonstrates that when sampling is performed without replacement
from finite populations the binomial distribution produces inaccurate probabilities. To pro
tect against large inaccuracies the binomial distribution should only be used when the sample
is small relative to the size of the population. Under that circumstance the value of p will not
change very much as the sample is selected and the binomial distribution will be a reasonable
approximation to the actual probability distribution.
In cases in which the sample is large relative to the size of the population a discrete prob
ability distribution called the hypergeometric distribution is the correct distribution for
computing probabilities for the random variable of interest.
We use combinations see Section 2 to form the equation for computing probabilities
for the hypergeometric distribution. When each trial has two possible outcomes success and
failure hypergeometric probabilities are computed using Equation 9.
Hypergeometric Distribution
The hypergeometric distribution is formed by the
ratio of the number of ways an event of interest
can occur over the total number of ways any
event can occur.
Discovod/Fotolia
www.downloadslide.comslide 232:
Discrete Probability Distributions
Notice that the numerator of Equation 9 is the product of the number of ways you can select
x successes in a random sample out of the X successes in the population and the number of
ways you can select n  x failures in a sample from the N  X failures in the population.
The denominator in the equation is the number of ways the sample can be selected from the
population.
In the earlier Dolby Industries example the probability of zero defectives being shipped
1x 02 is
Px
CC
C
Px
CC
C
55
55
0
0
10 0
20 2
0
2
10
20
10
18
0
2
1
−
−
.
.
0 0
20
Carrying out the arithmetic we get
Px
. 0
43 758 1
184 756
0 2368 55
As we found before the probability that zero defectives were included in the shipment is
0.2368 or approximately 24.
The probabilities of x 1 and x 2 defectives can also be found by using Equation 9
as follows:
Px
CC
C
.
−
−
1 0 5264
10 1
20 2
1
2
10
20
55
.
and
Px
CC
C
.
−
−
2 0 2368
10 2
20 2
2
2
10
20
55
.
Thus the hypergeometric probability distribution for the number of defective handbags in a
random selection of 10 is
xPx
0 0.2368
1 0.5264
2 0.2368
ΣP1x2 1.0000
Recall that when we introduced the hypergeometric distribution we said that it is used
in situations when we are sampling without replacement from a finite population. However
when the population size is large relative to the sample size decision makers typically use the
binomial distribution as an approximation of the hypergeometric. This eases the computational
Hypergeometric Distribution Two Possible Outcomes per Trial
Px
CC
C
nx
NX
x
X
n
N
.
9
where:
N Population size
X Number of successes in the population
n Sample size
x Number of successes in the sample
n  x Number of failures in the sample
www.downloadslide.comslide 233:
Discrete Probability Distributions
burden and provides useful approximations in those cases. Although there is no exact rule for
when the binomial approximation can be used we suggest that the sample should be less than
5 of the population size. Otherwise use the hypergeometric distribution when sampling is
done without replacement from the finite population.
EXAMPLE 8 THE HYPERGEOMETRIC DISTRIBUTION ONE OF TWO
POSSIBLE OUTCOMES PER TRIAL
Gender Equity One of the biggest changes in U.S. business practice in the past few decades
has been the inclusion of women in the management ranks of companies. Tom Peters manage
ment consultant and author of such books as In Search of Excellence has stated that one of
the reasons the Middle Eastern countries have suffered economically compared with countries
such as the United States is that they have not included women in their economic system. How
ever there are still issues in U.S. business. Consider a situation in which a Maryland company
needed to downsize one department having 30 people—12 women and 18 men. Ten people
were laid off and upper management said the layoffs were done randomly. By chance alone
40 12/30 of the layoffs would be women. However of the 10 laid off 8 were women. This
is 80 not the 40 due to chance. A labor attorney is interested in the probability of eight or
more women being laid off by chance alone. This can be determined using the following steps:
Step 1 Determine the population size and the combined sample size.
The population size and sample size are
N 30 and n 10
Step 2 Define the event of interest.
The attorney is interested in the event:
P1x Ú 82
What are the chances that eight or more women would be selected
Step 3 Determine the number of successes in the population and the number of
successes in the sample.
In this situation a success is the event that a woman is selected. There are
X 12 women in the population and x Ú 8 in the sample. We will break this
down as x 8 x 9 x 10.
Step 4 Compute the desired probabilities using Equation 9.
Px
CC
C
nx
NX
x
X
n
N
5
−
−
.
We want:
2
Px Px Px Px
Px
C
88 9 10
8
10 8
30
−
−1 12
8
12
10
30
2
18
8
12
10
30
0 0025
9
.. C
C
CC
C
Px
C
.
1 1
18
9
12
10
30
0
18
10
12
10
3
0 0001
10
.
.
C
C
Px
CC
C
.
0 0
0 0000 .
Therefore P1x Ú 82 0.0025 + 0.0001 + 0.0000 0.0026
The chances that 8 or more women would have been selected among the 10
people chosen for layoff strictly due to chance is 0.0026. The attorney will
likely wish to challenge the layoffs based on this extremely low probability.
END EXAMPLE
TRY PROBLEM 53
2
Note you can use Excel’s HYPGEOM.DIST function to compute these probabilities.
www.downloadslide.comslide 234:
Discrete Probability Distributions
The Hypergeometric Distribution with More Than Two Possible Outcomes
per Trial Equation 9 assumes that on any given sample selection or trial only one of two pos
sible outcomes will occur. However the hypergeometric distribution can easily be extended
to consider any number of possible categories of outcomes on a given trial by employing
Equation 10.
Hypergeometric Distribution k Possible Outcomes per Trial
Px x x
CC C C
k
x
X
x
X
x
X
x
X
k
K
...
...
12
1
1
2
2
3
3
5
.. . .
C C
n
N
10
where:
XN
xn
i
i
k
i
i
k
5
5
5
5
1
1
∑
∑
N Population size
n Total sample size
X
i
Number of items in the population with outcome i
x
i
Number of items in the sample with outcome i
EXAMPLE 9 THE HYPERGEOMETRIC DISTRIBUTION FOR MULTIPLE
OUTCOMES
Breakfast Cereal Preferences Consider a marketing study that involves placing break
fast cereal made by four different companies in a basket at the exit to a food store. A sign on
the basket invites customers to take one box of cereal free of charge. At the beginning of the
study the basket contains the following:
5 brand A
4 brand B
6 brand C
4 brand D
The researchers were interested in the brand selection patterns for customers who could
select without regard to price. Suppose six customers were observed and three selected brand
B two selected brand D and one selected brand C. No one selected brand A. The probabil
ity of this selection mix assuming the customers were selecting entirely at random without
replacement from a finite population can be found using the following steps:
Step 1 Determine the population size and the combined sample size.
The population size and sample size are
N 19 and n 6
Step 2 Define the event of interest.
The event of interest is
P1x
1
0 x
2
3 x
3
1 x
4
22
Step 3 Determine the number in each category in the population and the number
in each category in the sample.
X
1
5 brand A x
1
0
X
2
4 brand B x
2
3
X
3
6 brand C x
3
1
X
4
4 brand D x
4
2
N 19 n 6
www.downloadslide.comslide 235:
Discrete Probability Distributions
Skill Development
550. The mean number of errors per page made by a
member of the wordprocessing pool for a large
company is thought to be 1.5 with the number of
errors distributed according to a Poisson distribution. If
three pages are examined what is the probability that
more than three errors will be observed
551. Arrivals to a bank automated teller machine ATM are
distributed according to a Poisson distribution with a
mean equal to three per 15 minutes.
a. Determine the probability that in a given
15minute segment no customers will arrive at the
ATM.
b. What is the probability that fewer than four
customers will arrive in a 30minute segment
552. Consider a situation in which a usedcar lot contains
five Fords four General Motors GM cars and five
Toyotas. If five cars are selected at random to be placed
on a special sale what is the probability that three are
Fords and two are GMs
553. A population of 10 items contains 3 that are red and 7
that are green. What is the probability that in a random
sample of 3 items selected without replacement 2 red
and 1 green items are selected
554. If a random variable follows a Poisson distribution
with l 20 and t
1
2
find the
a. expected value variance and standard deviation of
this Poisson distribution
b. probability of exactly 8 successes
555. A corporation has 11 manufacturing plants. Of these
seven are domestic and four are located outside the
United States. Each year a performance evaluation is
conducted for 4 randomly selected plants.
a. What is the probability that a performance
evaluation will include exactly 1 plant outside the
United States
b. What is the probability that a performance
evaluation will contain 3 plants from the United
States
c. What is the probability that a performance
evaluation will include 2 or more plants from
outside the United States
556. Determine the following values associated with a
Poisson distribution with l t equal to 3:
a. P1x … 32
b. P1x 7 32
c. P12 6 x … 52
d. Find the smallest x′ so that P1x … x′2 7 0.50.
557. A random variable x has a hypergeometric distribution
with N 10 X 7 and n 4. Calculate the
following quantities:
a. P1x 32
b. P1x 52
c. P1x Ú 42
d. Find the largest x′ so that P1x 7 x′2 7 0.25.
Business Applications
558. A new phoneanswering system installed by the Ohio
Power Company is capable of handling five calls
every 10 minutes. Prior to installing the new system
company analysts determined that the incoming calls
to the system are Poisson distributed with a mean
5 Exercises
MyStatLab
Step 4 Compute the desired probability using Equation 10
Px x x x
CC C C
k
x
X
x
X
x
X
x
...
...
12 3
1
1
2
2
3
3
5
.. . .
k k
k
X
n
N
C
P
CCC C
C
0312
14
0
5
3
4
1
6
2
4
6
19
5
5
.. .
.
66
27 132
144
27 132
0 0053
5
5
There are slightly more than 5 chances in 1000 of this exact selection
occurring by random chance.
END EXAMPLE
TRY PROBLEM 52
www.downloadslide.comslide 236:
Discrete Probability Distributions
hardware software or computer peripheral crises.
MastersatWork became highly successful with
branches throughout the South and was purchased by
Best Buy but continued to operate under the Masters
atWork name. A shipment of 20 Intel
®
Pentium
®
4 processors was sent to MastersatWork. Four of
them were defective. One of the MastersatWork
technicians selected 5 of the processors to put in his
parts inventory and went on three service calls.
a. Determine the probability that only 1 of the 5
processors is defective.
b. Determine the probability that 3 of the 5 processors
are not defective.
c. Determine the probability that the technician will
have enough processors to replace 3 defective
processors at the repair sites.
564. John Thurgood founded a company that translates
Chinese books into English. His company is currently
testing a computerbased translation service. Since
Chinese symbols are difficult to translate John
assumes the computer program will make some errors
but then so do human translators. The computer error
rate is supposed to be an average of 3 per 400 words
of translation. Suppose John randomly selects a 1200
word passage. Assuming that the Poisson distribution
applies if the computer error rate is actually 3 errors
per 400 words
a. determine the probability that no errors will be
found.
b. calculate the probability that more than 14 errors
will be found.
c. find the probability that fewer than 9 errors will be
found.
d. If 15 errors are found in the 1200word passage
what would you conclude about the computer
company’s claim Why
565. Beacon Hill Trees Shrubs currently has an inventory
of 10 fruit trees 8 pine trees and 14 maple trees. It
plans to give 4 trees away at next Saturday’s lawn
and garden show in the city park. The 4 winners can
select which type of tree they want. Assume they select
randomly.
a. What is the probability that all 4 winners will select
the same type of tree
b. What is the probability that 3 winners will select
pine trees and the other tree will be a maple
c. What is the probability that no fruit trees and 2 of
each of the others will be selected
566. Fasteners used in a manufacturing process are
shipped by the supplier to the manufacturer in boxes
that contain 20 fasteners. Because the fasteners are
critical to the production process their failure will
cause the product to fail. The manufacturing firm
and the supplier have agreed that a random sample
of 4 fasteners will be selected from every box and
tested to see if the fasteners meet the manufacturer’s
specifications. The nature of the testing process is
equal to two every 10 minutes. If this incoming call
distribution is what the analysts think it is what is
the probability that in a 10minute period more calls
will arrive than the system can handle Based on this
probability comment on the adequacy of the new
answering system.
559. The Weyerhauser Lumber Company headquartered in
Tacoma Washington is one of the largest timber and
woodproduct companies in the world. Weyerhauser
manufactures plywood at one of its Oregon plants.
Plywood contains minor imperfections that can be
repaired with small “plugs.” One customer will accept
plywood with a maximum of 3.5 plugs per sheet on
average. Suppose a shipment was sent to this customer
and when the customer inspected two sheets at random
10 plugged defects were counted. What is the probability
of observing 10 or more plugged defects if in fact the
3.5 average per sheet is being satisfied Comment on
what this probability implies about whether you think the
company is meeting the 3.5 per sheet defect rate.
560. When things are operating properly EBank United
an Internet bank can process a maximum of 25
electronic transfers every minute during the busiest
periods of the day. If it receives more transfer requests
than this then the bank’s computer system will
become so overburdened that it will slow to the point
that no electronic transfers can be handled. If during
the busiest periods of the day requests for electronic
transfers arrive at the rate of 170 per 10minute period
on average what is the probability that the system will
be overwhelmed by requests Assume that the process
can be described using a Poisson distribution.
561. A stock portfolio contains 20 stocks. Of these stocks
10 are considered “largecap” stocks 5 are “mid cap”
and 5 are “small cap.” The portfolio manager has been
asked by his client to develop a report that highlights 7
randomly selected stocks. When she presents her report
to the client all 7 of the stocks are largecap stocks.
The client is very suspicious that the manager has not
randomly selected the stocks. She believes that the
chances of all 7 of the stocks being large cap must be
very low. Compute the probability of all 7 being large
cap and comment on the concerns of the client.
562. CollegePro Painting does home interior and exterior
painting. The company uses inexperienced painters
that do not always do a highquality job. It believes
that its painting process can be described by a Poisson
distribution with an average of 4.8 defects per 400
square feet of painting.
a. What is the probability that a 400squarefoot
painted section will have fewer than 6 blemishes
b. What is the probability that six randomly sampled
sections of size 400 square feet will each have 7 or
fewer blemishes
563. MastersatWork was founded by two brothers in
Atlanta to provide inhome computer and electronic
installation services as well as tech support to solve
www.downloadslide.comslide 237:
Discrete Probability Distributions
such that tested fasteners become unusable and must
be discarded. The supplier and the manufacturer have
agreed that if 2 or more fasteners fail the test the entire
box will be rejected as being defective. Assume that a
new box has just been received for inspection. If the
box has 5 defective fasteners what is the probability
that a random sample of 4 will have 2 or more
defective fasteners What is the probability the box
will be accepted
567. Lucky Dogs sells spicy hot dogs from a pushcart.
The owner of Lucky Dogs is open every day between
11:00 a.m. and 1:00 p.m. Assume the demand for spicy
hot dogs follows a Poisson distribution with a mean of
50 per hour.
a. What is the probability the owner will run out of
spicy dogs over the twohour period if he stocks his
cart with 115 spicy dogs every day
b. How many spicy hot dogs should the owner stock
if he wants to limit the probability of being out of
stock to less than 2.5
Hint: to solve this problem use Excel’s Formulas
Statistics 7 POISSON.DIST function or Minitab’s
Calc 7 Probability Distributions 7 Poisson
option.
568. USA Today recently reported that about onethird of
eligible workers haven’t enrolled in their employers’
401k plans. Costco has been contemplating new
incentives to encourage more participation from its
employees. Of the 12 employees in one of Costco’s
automotive departments 5 have enrolled in Costco’s
401k plan. The store manager has randomly selected
7 of the automotive department employees to receive
investment training.
a. Calculate the probability that all of the employees
currently enrolled in the 401k program are
selected for the investment training.
b. Calculate the probability that none of the employees
currently enrolled in the 401k program is selected
for the investment training.
c. Compute the probability that more than half of the
employees currently enrolled in the 401k program
are selected for the investment training.
569. The Small Business Administration’s Center for
Women’s Business Research indicated 30 of private
firms had female owners 52 had male owners
and 18 had male and female coowners. In one
community there are 50 privately owned firms. Ten
privately owned firms are selected to receive assistance
in marketing their products. Assume the percentages
indicated by the Small Business Administration apply
to this community.
a. Calculate the probability that onehalf of the firms
selected will be solely owned by a woman 3 owned
by men and the rest coowned by women and men.
b. Calculate the probability that all of the firms
selected will be solely owned by women.
c. Calculate the probability that 6 will be owned by a
woman and the rest coowned.
Computer Database Exercises
570. The National Federation of Independent Business
NFIB survey contacted 130 small firms. One of
the many inquiries was to determine the number of
employees the firms had. The file titled Employees
contains the responses by the firms. The number of
employees was grouped into the following categories:
1 fewer than 20 2 20–99 3 100–499 and
4 500 or more.
a. Determine the number of firms in each of these
categories.
b. If the NFIB contacts 25 of these firms to gather
more information determine the probability that it
will choose the following number of firms in each
category: 1 22 2 2 3 1 and 4 0.
c. Calculate the probability that it will choose all of
the firms from those businesses with fewer than
20 workers.
571. Cliff Summey is the qualityassurance engineer for
Sticks and Stones Billiard Supply a manufacturer
of billiard supplies. One of the items that Sticks and
Stones produces is sets of pocket billiard balls. Cliff
has been monitoring the finish of the pocket billiard
balls. He is concerned that sets of billiard balls have
been shipped with an increasing number of scratches.
The company’s goal is to have no more than an
average of one scratch per set of pocket billiard balls.
A set contains 16 balls. Over the last week Cliff
selected a sample of 48 billiard balls and inspected
them to determine the number of scratches. The data
collected by Cliff are displayed in the file called
Poolball.
a. Determine the number of scratches in the sample.
b. Calculate the average number of scratches for 48
pocket billiard balls if Sticks and Stones has met its
goal.
c. Determine the probability that there would be
at least as many scratches observed per set of
pocket billiard balls if Sticks and Stones has met
its goal.
d. Based on the sample evidence does it appear
that Sticks and Stones has met its goal Provide
statistical reasons for your conclusion.
END EXERCISES 53
www.downloadslide.comslide 238:
Discrete Probability Distributions
1 Introduction to Discrete Probability Distributions
2 The Binomial Probability Distribution
Outcome 2. Be able to apply the binomial distribution to business decisionmaking situations
3 Other Discrete Probability Distributions
Outcome 3. Be able to compute probabilities for the Poisson and hypergeometric
distributions and apply these distributions to decisionmaking situations
Summary
Although the binomial distribution may be the most often applied discrete distribution
for business decision makers the Poisson distribution and the hypergeometric
distribution are also frequently employed. The Poisson distribution is used in situations
where the value of the random variable is found by counting the number of occurrences
within a defned segment of time or space. If you know the mean number of occurrences
per segment you can use the Poisson formula the Poisson tables in software such as
Excel or Minitab to fnd the probability of any specifc number of occurrences within the
segment. The Poisson distribution is often used to describe the number of customers who
arrive at a service facility in a specifc amount of time. The hypergeometric distribution is
used in situations where the sample size is large relative to the size of the population and
the sampling is done without replacement.
Conclusion
Business applications involving discrete random variables are
very common in business situations. The probabilities for each
possible outcome of the discrete random variable form the
discrete probability distribution. The expected value of a discrete
probability distribution is the mean and represents the longrun
average value of the random variable. This chapter has
introduced three specifc discrete random variables that are
frequently used in business situations: binomial distribution
Poisson distribution and the hypergeometric distribution.
A random variable can take on values that are either discrete or continuous. This chapter
has focused on discrete random variables where the potential values are usually integer
values. Examples of discrete random variables include the number of defects in a sample of
twenty parts the number of customers who purchase CocaCola rather than Pepsi when 100
customers are observed the number of days late a shipment will be when the product is
shipped from India to the United States or the number of female managers who are
promoted from a pool of 30 females and 60 males at a Fortune 500 company. The
probabilities associated with the individual values of a random variable form the probability
distribution. The most frequently used discrete probability distributions are the binomial
distribution and the Poisson distribution.
Summary
A discrete random variable can assume only a fnite number of values or an infnite sequence of values
such as 0 1 2…. The mean of a discrete random variable is called the expected value and
represents the longrun average value for the random variable. The graph of a discrete random variable
looks like a histogram with the values of the random variable presented on the horizontal axis and the bars
above the values having heights corresponding to the probability of the outcome occurring. The sum of the
individual probabilities sum to one.
Outcome 1. Be able to calculate and interpret the expected value of a discrete random variable
Summary
The binomial distribution applies when an experimental trial has only two possible outcomes called success and
failure the probability of success remains constant from trial to trial the trials are independent and there are a
fxed number of identical trials being considered. The probabilities for a binomial distribution can be calculated using
Equation 4 or found using Excel or Minitab. The expected value of the binomial distribution is found by multiplying
n the number of trials by p the probability of a success on any one trial. The shape of a binomial distribution
depends on the sample size number of trials and p the probability of a success. When p is close to .50 the
binomial distribution will be fairly symmetric and bell shaped. Even when p is near 0 or 1 if n the sample size is
large the binomial distribution will still be fairly symmetric and bell shaped.
www.downloadslide.comslide 239:
Discrete Probability Distributions
7 Poisson Probability Distribution
Px
te
x
xt
−
8 Standard Deviation of the Poisson Distribution
t
9 Hypergeometric Distribution Two Possible Outcomes per
Trial
Px
CC
C
nx
NX
x
X
n
N
.
10 Hypergeometric Distribution k Possible Outcomes per Trial
Px x x
CC C C
k
x
X
x
X
x
X
x
X
k
K
...
...
12
1
1
2
2
3
3
5
.. . .
C C
n
N
1 Expected Value of a Discrete Probability Distribution
E1x2 ΣxP1x2
2 Standard Deviation of a Discrete Probability
Distribution
x
xEx Px ∑−
2
3 Counting Rule for Combinations
C
n
xn x
x
n
4 Binomial Formula
Px
n
xn x
xn x
pq
5 Expected Value of a Binomial Distribution
m
x
E1x2 np
6 Standard Deviation of the Binomial Distribution
npq
Equations
Key Terms
Binomial Probability Distribution
Characteristics
Continuous random variable
Counting rule for combinations
Discrete random variable
Expected value
Hypergeometric distribution
Random variable
Poisson distribution
Chapter Exercises
MyStatLab
Conceptual Questions
572. Three discrete distributions were discussed in this
chapter. Each was defined by a random variable that
measured the number of successes. To apply these
distributions you must know which one to use.
Describe the distinguishing characteristics for each
distribution.
573. How is the shape of the binomial distribution changed
for a given value of p as the sample size is increased
Discuss.
574. Discuss the basic differences and similarities between
the binomial distribution and the Poisson distribution.
575. Beginning statistics students are often puzzled by two
characteristics of distributions in this chapter: 1 The
trials are independent and 2 the probability of a
success remains constant from trial to trial. Students
often think these two characteristics are the same.
The questions in this exercise point out the difference.
Consider a hypergeometric distribution where
N 3 X 2 and n 2.
a. Mathematically demonstrate that the trials for
this experiment are dependent by calculating the
probability of obtaining a success on the second
trial if the first trial resulted in a success. Repeat this
calculation if the first trial was a failure. Use these two
probabilities to prove that the trials are dependent.
b. Now calculate the probability that a success is
obtained on each of the three respective trials and
therefore demonstrate that the trials are dependent
but that the probability of a success is constant from
trial to trial.
576. Consider an experiment in which a sample of size n 5
is taken from a binomial distribution.
a. Calculate the probability of each value of the random
variable for the probability of a success equal to 1
0.1 2 0.25 3 0.50 4 0.75 and 5 0.9.
b. Which probabilities produced a rightskewed
distribution Why
c. Which probability of a success yielded a symmetric
distribution Why
d. Which probabilities produced a leftskewed
distribution Discuss why.
Business Applications
577. The McMillan Newspaper Company sometimes makes
printing errors in its advertising and is forced to provide
www.downloadslide.comslide 240:
Discrete Probability Distributions
company concerning the return rate Make sure you
support your statement with something other than
opinion.
581. The Defense Department has recently advertised for
bids for producing a new nightvision binocular. Vista
Optical has decided to submit a bid for the contract.
The first step was to supply a sample of binoculars for
the army to test at its Kentucky development grounds.
Vista makes a superior nightvision binocular. However
the 4 sent to the army for testing were taken from a
developmentlab project of 20 units that contained
4 defectives. The army has indicated it will reject
any manufacturer that submits 1 or more defective
binoculars. What is the probability that this mistake has
cost Vista any chance for the contract
582. VERCOR provides merger and acquisition consultants
to assist corporations when owners decide to offer their
business for sale. One of its news releases “Tax Audit
Frequency Is Rising” written by David L. Perkins Jr.
a VERCOR partner and which originally appeared in
The Business Owner indicated that the proportion of the
largest businesses those corporations with assets of 10
million and over that were audited was 0.17.
a. One member of VERCOR’s board of directors
is on the board of directors of four other large
corporations. Calculate the expected number of
these five corporations that should get audited
assuming selection is random.
b. Three of the five corporations were actually audited.
Determine the probability that at least three of the
five corporations would be audited if 17 of large
corporations are audited. Assume random selection.
c. The board member is concerned that the
corporations have been singled out to be audited
by the Internal Revenue Service IRS. Respond to
these thoughts using probability and statistical logic.
583. Stafford Production Inc. is concerned with the quality
of the parts it purchases that will be used in the end
items it assembles. Part number 3478D is used in the
company’s new laser printer. The parts are sensitive
to dust and can easily be damaged in shipment even if
they are acceptable when they leave the vendor’s plant.
In a shipment of four parts the purchasing agent has
assessed the following probability distribution for the
number of defective products:
xP x
0 0.20
1 0.20
2 0.20
3 0.20
4 0.20
a. What is the expected number of defectives in a
shipment of four parts Discuss what this value
really means to Stafford Production Inc.
corrected advertising in the next issue of the paper. The
managing editor has done a study of this problem and
found the following data:
No. of Errors x Relative Frequency
0 0.56
1 0.21
2 0.13
3 0.07
4 0.03
a. Using the relative frequencies as probabilities what
is the expected number of errors Interpret what this
value means to the managing editor.
b. Compute the variance and standard deviation for
the number of errors and explain what these values
measure.
578. The Ziteck Corporation buys parts from international
suppliers. One part is currently being purchased from a
Malaysian supplier under a contract that calls for at most
5 of the 10000 parts to be defective. When a shipment
arrives Ziteck randomly samples 10 parts. If it finds 2
or fewer defectives in the sample it keeps the shipment
otherwise it returns the entire shipment to the supplier.
a. Assuming that the conditions for the binomial
distribution are satisfied what is the probability that
the sample will lead Ziteck to keep the shipment if
the defect rate is actually 0.05
b. Suppose the supplier is actually sending Ziteck 10
defects. What is the probability that the sample will
lead Ziteck to accept the shipment anyway
c. Comment on this sampling plan sample size and
accept/reject point. Do you think it favors either
Ziteck or the supplier Discuss.
579. Californiabased Wagner Foods Inc. has a process that
inserts fruit juice into 24ounce containers. When the
process is in control half the cans actually contain more
than 24 ounces and half contain less. Suppose a quality
inspector has just randomly sampled nine cans and
found that all nine had more than 24 ounces. Calculate
the probability that this result would occur if the filling
process was actually still in control. Based on this
probability what conclusion might be reached Discuss.
580. Your company president has told you that the company
experiences product returns at the rate of two per month
with the number of product returns distributed as a
Poisson random variable. Determine the probability that
next month there will be
a. no returns
b. one return
c. two returns
d. more than two returns
e. In the last three months your company has had only
one month in which the number of returns was at
most two. Calculate the probability of this event
occurring. What will you tell the president of your
www.downloadslide.comslide 241:
Discrete Probability Distributions
c. Recalling that the coefficient of variation is
determined by the ratio of the standard deviation
over the mean compute the coefficient of variation
for each random variable.
d. Referring to part c suppose the seminar director
said that the first stock was riskier since its standard
deviation was greater than the standard deviation of
the second stock. How would you respond Hint:
What do the coefficients of variation imply
586. Simmons Market Research conducted a national
consumer study of 13787 respondents. The respondents
were asked to indicate the primary source of the
vitamins or mineral supplements they consume. Thirty
five percent indicated a multiple formula was their
choice. A subset of 20 respondents who used multiple
vitamins was selected for further questioning. Half of
them used a One A Day vitamin the rest used generic
brands. Of this subset 4 were asked to fill out a more
complete health survey.
a. Calculate the probability that the final selection
of 4 subset members were all One A Day multiple
vitamin users.
b. Compute the number of One A Day users expected
to be selected.
c. Calculate the probability that fewer than half of the
final selection were One A Day users.
587. The 700room Westin Charlotte offers a premiere
uptown location in the heart of the city’s financial
district. On a busy weekend the hotel has 20 rooms that
are not occupied. Suppose that smoking is allowed in 8
of the rooms. A small tour group arrives which has four
smokers and six nonsmokers. The desk clerk randomly
selects 10 rooms and gives the keys to the tour guide to
distribute to the travelers.
a. Compute the probability that the tour guide
will have the correct mix of rooms so that all
members of the tour group will receive a room that
accommodates their smoking preferences.
b. Determine the probability that the tour guide will have
to assign at least one nonsmoker to a smoking room.
c. Determine the probability that the tour guide will
have to assign at least one smoker to a nonsmoking
room.
Computer Database Exercises
588. A 23mile stretch of a twolane highway east of Paso
Robles California was once considered a “death trap”
by residents of San Luis Obispo County. Formerly
known as “Blood Alley” Highway 46 gained notoriety
for the number of fatalities 29 and crashes over
a 240 week period. More than twothirds involved
headon collisions. The file titled Crashes contains
the simulated number of fatal crashes during this time
period.
a. Determine the average number of crashes in the 240
weeks.
b. Compute and interpret the standard deviation of the
number of defective parts in a shipment of four.
c. Examine the probabilities as assessed and indicate
why this probability distribution might be called a
uniform distribution. Provide some reasons why the
probabilities might all be equal as they are in this
case.
584. Bach Photographs takes school pictures and charges
only 0.99 for a sitting which consists of six poses.
The company then makes up three packages that are
offered to the parents who have a choice of buying 0
1 2 or all 3 of the packages. Based on his experience
in the business Bill Bach has assessed the following
probabilities of the number of packages that might be
purchased by a parent:
No. of Packages xP x
0 0.30
1 0.40
2 0.20
3 0.10
a. What is the expected number of packages to be
purchased by each parent
b. What is the standard deviation for the random
variable x
c. Suppose all of the picture packages are to be priced
at the same level. How much should they be priced
if Bach Photographs wants to break even Assume
that the production costs are 3.00 per package.
Remember that the sitting charge is 0.99.
585. The managing partner for Westwood One Investment
Managers Inc. gave a public seminar in which she
discussed a number of issues including investment risk
analysis. In that seminar she reminded people that the
coefficient of variation often can be used as a measure
of risk of an investment. To demonstrate her point she
used two hypothetical stocks as examples. She let x
equal the change in assets for a 1000.00 investment in
stock 1 and y reflect the change in assets for a 1000.00
investment in stock 2. She showed the seminar
participants the following probability distributions:
x P x YP y
+1000.00 0.10 +1000.00 0.20
0.00 0.10 0.00 0.40
500.00 0.30 500.00 0.30
1000.00 0.30 1000.00 0.05
2000.00 0.20 2000.00 0.05
a. Compute the expected values for random variables x
and y.
b. Compute the standard deviations for random
variables x and y.
www.downloadslide.comslide 242:
Discrete Probability Distributions
c. To make sure that the proportion of defectives
does not change the quality control manager
wants to establish control limits that are 3 standard
deviations above the mean and 3 standard deviations
below the mean. Calculate these limits.
d. Determine the probability that a randomly chosen
set of 20 smoke alarms would have a number
of defectives that was beyond the control limits
established in part c.
590. Covercraft manufactures covers to protect automobile
interiors and finishes. Its BlockIt 200 Series fabric has
a limited twoyear warranty. Periodic testing is done
to determine if the warranty policy should be changed.
One such study examined those covers that became
unserviceable while still under warranty. Data that
could be produced by such a study are contained in the
file titled Covers. The data represent the number of
months a cover was used until it became unserviceable.
Covercraft might want to examine more carefully the
covers that became unserviceable while still under
warranty. Specifically it wants to examine those that
became unserviceable before they had been in use one
year.
a. Determine the number of covers that became
unserviceable before they had been in use less than
a year and a half.
b. If Covercraft qualitycontrol staff selects 20 of the
covers at random determine the probability that
none of them will have failed before they had been
in service a year and a half.
c. If Covercraft quality control staff needs to examine
at least 5 of the failed covers determine the
probability that they will obtain this many.
b. Calculate the probability that at least 19 crashes
would occur over the 240week period if the
average number of crashes per week was as
calculated in part a.
c. Calculate the probability that at least 19 crashes would
occur over a fiveyear period if the average number of
crashes per week was as calculated in part a.
d. A coalition of state local and private organizations
devised a coordinated and innovative approach to
dramatically reduce deaths and injuries on this road.
During the 16 months before and after completion
of the project fatal crashes were reduced to zero.
Calculate the probability that there would be no fatal
crashes if the mean number of fatal crashes was not
changed by the coalition. Does it appear that the aver
age number of fatal accidents has indeed decreased
589. American Household SM Inc. produces a wide array of
home safety and security products. One of its products
is the First Alert SA302 Dual Sensor Remote Control
Smoke Alarm. As part of its quality control program it
constantly tests to assure that the alarms work. A change
in the manufacturing process requires the company to
determine the proportion of alarms that fail the quality
control tests. Each day 20 smoke alarms are taken
from the production line and tested and the number of
defectives is recorded. A file titled Smokeless contains
the possible results from the last 90 days of testing.
a. Compute the proportion of defective smoke alarms.
b. Calculate the expected number and the standard
deviation of defectives for each day’s testing. Assume
the proportion of defectives is what was computed in
part a. Hint: Recall the formulas for the mean and
the standard deviation for a binomial distribution.
Case 1
SaveMor Pharmacies
A common practice now is for large retail pharmacies to buy the
customer base from smaller independent pharmacies. The way
this works is that the buyer requests to see the customer list along
with the buying history. The buyer then makes an offer based on its
projection of how many of the seller’s customers will move their
business to the buyer’s pharmacy and on how many dollars of new
business will come to the buyer as a result of the purchase. Once
the deal is made the buyer and seller usually send out a joint letter
to the seller’s customers explaining the transaction and informing
them that their prescription files have been transferred to the pur
chasing company.
The problem is that there is no guarantee regarding what
proportion of the existing customers will make the switch to the
buying company. That is the issue facing Heidi Fendenand acqui
sitions manager for SaveMor Pharmacies. SaveMor has the oppor
tunity to purchase the 6780person customer base from Hubbard
Pharmacy in San Jose California. Based on previous acquisitions
Heidi believes that if 70 or more of the customers will make the
switch then the deal is favorable to SaveMor. However if 60 or
less make the move to SaveMor then the deal will be a bad one
and she would recommend against it.
Quincy Kregthorpe a research analyst who works for Heidi
has suggested that SaveMor take a new approach to this acquisi
tion decision. He has suggested that SaveMor contact a random
sample of 20 Hubbard customers telling them of the proposed sale
and asking them if they will be willing to switch their business to
SaveMor. Quincy has suggested that if 15 or more of the 20 cus
tomers indicate that they would make the switch then SaveMor
should go ahead with the purchase. Otherwise it should decline
the deal or negotiate a lower purchase price.
Heidi liked this idea and contacted Cal Hubbard Hubbard’s
owner to discuss the idea of surveying 20 randomly selected cus
tomers. Cal was agreeable as long as only these 20 customers
would be told about the potential sale.
Before taking the next step Heidi met with Quincy to
discuss the plan one more time. She was concerned that the
www.downloadslide.comslide 243:
Discrete Probability Distributions
even if the true proportion of all customers who would
switch is actually 0.70.
2. Compute the probability that the sampling plan will provide
a result that suggests that SaveMor should accept the deal
even if the true proportion of all customers who would
switch is actually only 0.60.
3. Write a short report to Heidi outlining the sampling plan the
assumptions on which the evaluation of the sampling plan
has been based and the conclusions regarding the potential
effectiveness of the sampling plan. The report should make
a recommendation about whether Heidi should go through
with the idea of using the sampling plan.
proposed sampling plan might have too high a probability of
rejecting the purchase deal even if it was a positive one from
SaveMor’s viewpoint. On the other hand she was concerned
that the plan might also have a high probability of accept
ing the purchase deal when in fact it would be unfavorable to
SaveMor. After discussing these concerns for over an hour
Quincy finally offered to perform an evaluation of the sam
pling plan.
Required Tasks:
1. Compute the probability that the sampling plan will provide
a result that suggests that SaveMor should reject the deal
Arrowmark Vending
Arrowmark Vending has the contract to supply pizza at all home
football games for a university in the Big 12 athletic conference. It
is a constant challenge at each game to determine how many pizzas
to have available at the games. Tom Kealey operations manager for
Arrowmark has determined that his fixed cost of providing piz
zas whether he sells 1 pizza or 4000 pizzas is 1000. This cost
includes hiring employees to work at the concession booths hiring
extra employees to cook the pizzas the day of the game delivering
them to the game and advertising during the game. He believes that
this cost should be equally allocated between two types of pizzas.
Tom has determined that he will supply only two types of piz
zas: plain cheese and pepperoniandcheese combo. His cost to
make a plain cheese pizza is 4.50 each and his cost to make pep
peroniandcheese combo is 5.00 each. Both pizzas will sell for
9.00 at the game. Unsold pizzas have no value and are donated to
a local shelter for the homeless.
Experience has shown the following demand distributions for
the two types of pizza at home games:
Plain Cheese Demand Probability PepperoniandCheese Demand Probability
200 0.10 300 0.10
300 0.15 400 0.20
400 0.15 500 0.25
500 0.20 600 0.25
600 0.20 700 0.15
700 0.10 800 0.05
800 0.05
900 0.05
Required Tasks:
1. For each type of pizza determine the profit or loss
associated with producing at each possible demand level. For
instance determine the profit if 200 plain cheese pizzas are
produced and 200 are demanded. What is the profit if 200
plain cheese pizzas are produced but 300 were demanded
and so on
2. Compute the expected profit associated with each possible
production level assuming Tom will only produce at one of
the possible demand levels for each type of pizza.
3. Prepare a short report that provides Tom with the information
regarding how many of each type of pizza he should produce
if he wants to achieve the highest expected profit from pizza
sales at the game.
scale stations to have open during various times of the day. If he
has too many stations open the scalers will have excessive idle
time and the cost of scaling will be unnecessarily high. On the
other hand if too few scale stations are open some log trucks will
have to wait.
Boise Cascade Corporation
At the Boise Cascade Corporation lumber mill logs arrive by
truck and are scaled measured to determine the number of board
feet before they are dumped into a log pond. Figure C3 illus
trates the basic flow. The mill manager must determine how many
www.downloadslide.comslide 244:
Discrete Probability Distributions
0 to 6 trucks: open 1 scale station
7 to 12 trucks: open 2 scale stations etc.
However the number of trucks is a random variable and is
uncertain. Your task is to provide guidance for the decision.
The manager has studied the truck arrival patterns and has
determined that during the first open hour 7:00 a.m.–8:00 a.m.
the trucks randomly arrive at 12 per hour on average. Each scale
station can scale 6 trucks per hour 10 minutes each. If the man
ager knew how many trucks would arrive during the hour he
would know how many scale stations to have open.
FIGURE C3 
Truck Flow for Boise Cascade
Mill Example
Pond
Scale Stations
Trucks
Enter
Trucks Exit
www.downloadslide.comslide 245:
Discrete Probability Distributions
Answers to Selected OddNumbered Problems
This section contains summary answers to most of the oddnumbered problems in the text. The Student Solutions Manual contains fully developed
solutions to all oddnumbered problems and shows clearly how each answer is determined.
31. a. 0.0688
b. 0.0031
c. 0.1467
d. 0.8470
e. 0.9987
33. a. 3.2
b. 1.386
c. 0.4060
d. 0.9334
35. a. 3.46
b. 0.012
c. 0.004
d. It is quite unlikely.
37. a. 0.5580
b. 0.8784
c. An increase in sample size would be required.
39. a. 2.96
b. Variance 1.8648 Standard deviation 1.3656
c. 0.3811
41. a. 0.3179
b. 0.2174
c. 0.25374
43. a. 0.051987
b. 0.028989
45. a. 0.372
b. 12 estimate may be too high.
c. 0.0832
d. 0.0003
e. Redemption rate is lower than either Vericours or TCA
Fulfillment estimate.
49. a. 9 corporations
b. 0.414324
c. 70th percentile is 12.
51. a. 0.0498
b. 0.1512
53. 0.175
55. a. 0.4242
b. 0.4242
c. 0.4696
57. a. Px 3 0.5
b. Px 5 0
c. 0.6667
d. Since 0.6667 7 0.25 then x 2.
59. Px Ú 10 1  0.8305 0.1695
61. 0.0015
63. a. Px 4 0.4696
b. Px 3 0.2167
c. 0.9680
65. a. 0.0355
b. 0.0218
c. 0.0709
67. a. 0.0632
b. 120 Spicy Dogs
69. a. 0.0274
b. 0.0000
c. 0.0001
1. a. discrete random variable
b. The possible values for x are x 0 1 2 3 4 5 6.
3. a. number of children under 22 living in a household
b. discrete
5. 3.7 days
7. a. 130
b. 412.50
c. 1412.50 20.31
9. a. 15.75
b. 20.75
c. 78.75
d. increases the expected value by an amount equal to the
constant added
e. the expected value being multiplied by that same
constant
11. a. 3.51
b. s
2
1.6499 s 1.2845
13. a. 2.87 days
b. s 11.4931 1.22
c. 1.65 days to 4.09 days
15. a. 58300
b. 57480
17. a. Small firm profits +135000
Midsized profits +155000
Large firm profits +160000
b. Small firm: s +30000
Midsized firm: s +90000
Large firm: s +156604.60
c. The large firm has the largest expected profit.
21. a.
b. 19.168 s 2s
2
23.1634 1.7787
c. Median is 19 and the quality control department is correct.
23. 0.2668
25. a. Px 5 0.0746
b. Px Ú 7 0.2143
c. 4
d. s 1npq 120.20.80 1.7889
27. 0.1029
29. a. 0.1442
b. 0.8002
c. 4.55
x Px
14 0.008
15 0.024
16 0.064
17 0.048
18 0.184
19 0.216
20 0.240
21 0.128
22 0.072
23 0.016
www.downloadslide.comslide 246:
Discrete Probability Distributions
79. 0.0020
81. 0.6244
83. a. 2.0
b. 1.4142
c. because outcomes equally likely
85. a. Ex 750 Ey 100
b. StDevx 844.0972 StDevy 717.635
c. CVx 844.0972750 1.1255
CVy 717.635100 7.1764
87. a. 0.3501
b. 0.3250
89. a. 0.02
b. EX 0.6261
c. 0 2.2783
d. 1  0.9929 0.0071
71. a. 8
b. lt 13 3
c. 0.0119
d. It is very unlikely. Therefore we believe that the goal has
not been met.
75. a. This means the trials are dependent.
b. does not imply that the trials are independent
77. a.
b. Standard deviation 1.0954 Variance 1.20
X Px xPx
0 0.56 0.00
1 0.21 0.21
2 0.13 0.26
3 0.07 0.21
4 0.03 0.12
0.80
Continuous Random Variables Random variables that can
assume an uncountably infinite number of values.
Discrete Random Variable A random variable that can only
assume a finite number of values or an infinite sequence of
values such as 0 1 2c.
Expected Value The mean of a probability distribution. The
average value when the experiment that generates values for
the random variable is repeated over the long run.
Hypergeometric Distribution The hypergeometric distribu
tion is formed by the ratio of the number of ways an event of
interest can occur over the total number of ways any event
can occur.
Random Variable A variable that takes on different numerical
values based on chance.
Binomial Probability Distribution Characteristics A distri
bution that gives the probability of x successes in n trials in a
process that meets the following conditions:
1. A trial has only two possible outcomes: a success or a
f ailure.
2. There is a fixed number n of identical trials.
3. The trials of the experiment are independent of each
other. This means that if one outcome is a success this
does not influence the chance of another outcome being a
success.
4. The process must be consistent in generating successes
and failures. That is the probability p associated with a
success remains constant from trial to trial.
5. If p represents the probability of a success then 1 p q
is the probability of a failure.
Glossary
DeVeaux Richard D. Paul F. Velleman and David E. Bock
Stats Data and Models 3rd ed. New York: AddisonWesley
2012.
Hogg R. V . and Elliot A. Tanis Probability and Statistical Infer
ence 8th ed. Upper Saddle River NJ: Prentice Hall 2010.
Larsen Richard J. and Morris L. Marx An Introduction to
Mathematical Statistics and Its Applications 5th ed. Upper
Saddle River NJ: Prentice Hall 2012.
Microsoft Excel 2010 Redmond WA: Microsoft Corp. 2010.
Siegel Andrew F. Practical Business Statistics 5th ed. Burr
Ridge IL: Irwin 2002.
References
www.downloadslide.comslide 247:
www.downloadslide.comslide 248:
Introduction to Continuous
Probability Distributions
From Chapter 6 of Business Statistics A DecisionMaking Approach Ninth Edition. David F. Groebner
Patrick W. Shannon and Phillip C. Fry. Copyright © 2014 by Pearson Education Inc. All rights reserved.
www.downloadslide.comslide 249:
Why you need to know
You will encounter many business situations in which the random variable of interest is discrete and in which prob
ability distributions such as the binomial the Poisson or the hypergeometric will be useful for analyzing decision
situations. But there will also be many situations in which the random variable of interest is continuous rather than
discrete. For instance the managers at Harley Davidson might be interested in a measure called throughput time
which is the time it takes from when a motorcycle is started on the manufacturing line until it is completed. Lots
of factors can affect the throughput time including breakdowns need for rework the type of accessories added
to the motorcycle and worker productivity. The managers might be interested in determining the probability that
the throughput time will be between 3.5 and 5.0 hours. In this case time is the random variable of interest and
is continuous.
A pharmaceutical company may be interested in the probability that a new drug will reduce blood pressure by
more than 20 points for patients. Blood pressure is the continuous random variable of interest. The Kellogg’s Cereal
company could be interested in the probability that cereal boxes labeled as containing 16 ounces will actually con
tain at least that much cereal. Here the variable of interest is weight which can be measured on a continuous scale.
In each of these examples the value of the variable of interest is determined by measuring measuring the
time required to make a motorcycle measuring the blood pressure reading measuring the weight of cereal in a
box. In every instance the number of possible values for the variable is limited only by the capacity of the meas
uring device. The constraints imposed by the measuring devices produce a finite number of outcomes. In these
and similar situations a continuous probability distribution can be used to approximate the distribution of possible
outcomes for the random variables. The approximation is appropriate when the number of possible outcomes is
large. This text introduces three specific continuous probability distributions of particular importance for decision
making and the study of business statistics. The first of these the normal distribution is by far the most important
because a great many applications involve random variables that possess the characteristics of the normal distri
bution. In addition many topics dealing with statistical estimation and hypothesis testing are based on the normal
distribution.
Review the concept of zscores. Review the discussion of the mean and
standard deviation.
Quick Prep Links
Review the methods for determining the
probability for a discrete random variable.
Outcome 4. Calculate probabilities associated with a
uniformly distributed random variable.
Outcome 5. Determine probabilities using an exponential
probability distribution.
Outcome 1. Convert a normal distribution to a standard
normal distribution.
Outcome 2. Determine probabilities using the standard
normal distribution.
Outcome 3. Calculate values of the random variable
associated with specified probabilities from a normal
distribution.
1 The Normal Probability
Distribution
Introduction to Continuous
Probability Distributions
2 Other Continuous Probability
Distributions
Rafa Irusta/Shutterstock
www.downloadslide.comslide 250:
Introduction to Continuous Probability Distributions
In addition to the normal distribution you will be introduced to the uniform distribution and the exponential
distribution. Both are important continuous probability distributions and have many applications in business decision
making. You need to have a firm understanding and working knowledge of all three continuous probability distributions
introduced in this chapter.
1 The Normal Probability Distribution
A PepsiCola can is supposed to contain 12 ounces but it might actually contain any
amount between 11.90 and 12.10 ounces such as 11.9853 ounces. When the variable
of interest such as the volume of soda in a can is approximately continuous the prob
ability distribution associated with the random variable is called a continuous probability
distribution.
One important difference between discrete and continuous probability distributions
involves the calculation of probabilities associated with specific values of the random varia
ble. For instance in a market research example in which 100 people are surveyed and asked
whether they have a positive view of a product we could use the binomial distribution to find
the probability of any specific number of positive reviews such as P1x 752 or P1x 762.
Although these individual probabilities may be small values they can be computed because
the random variable is discrete. However if the random variable is continuous as in the
PepsiCola example there is an uncountable infinite number of possible outcomes for the
random variable. Theoretically the probability of any one of these individual outcomes is
zero. That is P1x 11.922 0 or P1x 12.052 0. Thus when you are working with
continuous distributions you will need to find the probability for a range of possible values
such as P1x … 11.922 or P1 11.92 … x … 12.02. You can also conclude that
P1x … 11.922 P1x 6 11.922
because we assume that P1x 11.922 0.
There are many different continuous probability distributions but the most important of
these is the normal distribution.
The Normal Distribution
1
Figure 1 illustrates a typical normal distribution and highlights the normal distribution’s
characteristics. All normal distributions have the same general shape as the one shown
in Figure 1. However they can differ in their mean value and their variation depending
on the situation being considered. The process being represented determines the scale
of the horizontal axis. It may be pounds inches dollars or any other attribute with a
Normal Distribution
The normal distribution is a bellshaped
distribution with the following properties:
1. It is unimodal that is the normal distribu
tion peaks at a single value.
2. It is symmetrical this means that the two
areas under the curve between the mean
and any two points equidistant on either
side of the mean are identical. One side of
the distribution is the mirror image of the
other side.
3. The mean median and mode are equal.
4. The normal approaches the horizontal axis
on either side of the mean toward plus
and minus infinity ∞. In more formal
terms the normal distribution is asymp
totic to the x axis.
5. The amount of variation in the random vari
able determines the height and spread of the
normal distribution.
1
It is common to refer to the very large family of normal distributions as “the normal distribution.”
x
Mean
Median
Mode
Probability 0.50 Probability 0.50
FIGURE 1 
Characteristics of the Normal
Distribution
www.downloadslide.comslide 251:
Introduction to Continuous Probability Distributions
FIGURE 2 
Difference between Normal
Distributions
x
x
x
Area 0.50 Area 0.50
Area 0.50 Area 0.50
Area 0.50 Area 0.50
a
b
c
Normal Probability Density Function
fx e
x
1 2
2
2
2
1
where:
x Any value of the continuous random variable
s Population standard deviation
p 3.14159
e Base of the natural log 2.71828
m Population mean
To graph the normal distribution we need to know the mean m and the standard devia
tion s. Placing m s and a value of the variable x into the probability density function we
can calculate a height fx of the density function. If we could try enough x values we could
construct curves like those shown in Figures 1 and 2.
The area under the normal curve corresponds to probability. Because x is a continuous
random variable the probability Px is equal to 0 for any particular x. However we can find
the probability for a range of values between x
1
and x
2
by finding the area under the curve
between these two values. A special normal distribution called the standard normal distribu
tion is used to find areas probabilities for all normal distributions.
The Standard Normal Distribution
The trick to finding probabilities for a normal distribution is to convert the normal distribution
to a standard normal distribution.
To convert a normal distribution to a standard normal distribution the values x of the
random variable are standardized. The conversion formula is shown as Equation 2.
Standard Normal Distribution
A normal distribution that has a mean 0.0
and a standard deviation 1.0 The horizontal
axis is scaled in zvalues that measure the
number of standard deviations a point is from
the mean. Values above the mean have positive
zvalues. Values below the mean have negative
zvalues.
Chapter Outcome 1.
continuous measurement. Figure 2 shows several normal distributions with different
centers and different spreads. Note that the total area probability under each normal
curve equals 1.
The normal distribution is described by the rather complicatedlooking probability den
sity function shown in Equation 1.
www.downloadslide.comslide 252:
Introduction to Continuous Probability Distributions
Standardized Normal zValue
z
x
2
where:
z Scaled value 1 the number of standard deviations a point x is from the mean2
x Any point on the horizontal axis
m Mean of the specific normal distribution
s Standard deviation of the specific normal distribution
x Ounces
16
4
fx
FIGURE 3 
Distribution of Grapefruit
Weights
Equation 2 scales any normal distribution axis from its true units time weight dollars
volume and so forth to the standard measure referred to as a zvalue. Thus any value of
the normally distributed continuous random variable can be represented by a unique zvalue.
Positive zvalues represent corresponding values of the random variable x that are higher
than the population mean. Values of x that are less than the population mean will have cor
responding zvalues that are negative.
BUSINESS APPLICATION STANDARD NORMAL DISTRIBUTION
FRUIT PRODUCTION Fruit growers in California and Florida
strive for consistency in their products in terms of quality
and size. For example a grower in Florida that specializes in
grapefruit has determined that in one orchard the mean weight of
his “King” brand grapefruit is 16 ounces. Suppose after careful
analysis he has determined the grapefruit weight distribution is
approximated by a normal distribution with a standard deviation
of 4 ounces. Figure 3 shows this normal distribution with m 16 and s 4.
Three grapefruit were selected from a case in the grower’s cold storage. The weights of
these three grapefruit were
Grapefruit 1: x 16 ounces
Grapefruit 2: x 18.5 ounces
Grapefruit 3: x 9 ounces
Equation 2 is used to convert these values from a normally distributed population
with m 16 and s 4 to corresponding zvalues in a standard normal distribution. For
Grapefruit 1 we get
z
x 16 16
4
0
Note Grapefruit 1 weighed 16 ounces which happens to be equal to the population mean for
all grapefruit. The standardized zvalue corresponding to the population mean is zero. This
indicates that the population mean is 0 standard deviations from itself.
atoss/Fotolia
www.downloadslide.comslide 253:
Introduction to Continuous Probability Distributions
FIGURE 4 
Standard Normal Distribution
z
–3.0 –2.5 –2.0 –1.5 –1.0 –0.5 1.0 0.5 1.5 2.0 2.5 3.0 0
fz
Chapter Outcome 2.
For Grapefruit 2 we get
z
x−− 18 5 16
4
063
.
.
Thus a grapefruit that weighs 18 ounces is 0.63 standard deviations heavier than the mean for
all grapefruit. The standardized zvalue for Grapefruit 3 is
z
x−−
−
916
4
175 .
This means a grapefruit that weighs only 9 ounces is 1.75 standard deviations below the popu
lation mean. Note a negative zvalue always indicates the xvalue is less than the mean m.
The zvalue represents the number of standard deviations a point is above or below
the population mean. Equation 2 can be used to convert any specified value x from the
population distribution to a corresponding zvalue. If the population distribution is normally
distributed as shown in Figure 3 then the distribution of zvalues will also be normally
distributed and is called the standard normal distribution. Figure 4 shows a standard normal
distribution where the horizontal axis represents zvalues.
Y ou can convert the normal distribution to a standard normal distribution and use the standard
normal table to find the desired probability. Example 1 shows the steps required to do this.
Using the Standard Normal Table The standard normal table provides probabilities or
areas under the normal curve associated with many different zvalues. The standard normal
table is constructed so that the probabilities provided represent the chance of a value being
between a positive zvalue and its population mean 0.
The standard normal table is also reproduced in Table 1. This table provides probabilities
for zvalues between z 0.00 and z 3.09. Note because the normal distribution is sym
metric the probability of a value being between a positive zvalue and its population mean 0
is the same as that of a value being between a negative zvalue and its population mean 0. So
we can use one standard normal table for both positive and negative zvalues.
EXAMPLE 1 USING THE STANDARD NORMAL TABLE
Airline Passenger Loading Times After completing a
study the Chicago O’Hare Airport managers have concluded
that the time needed to get passengers loaded onto an airplane
is normally distributed with a mean equal to 15 minutes and a
standard deviation equal to 3.5 minutes. Recently one airplane
required 22 minutes to get passengers on board and ready for
takeoff. To find the probability that a flight will take 22 or more
minutes to get passengers loaded you can use the following steps:
Step 1 Determine the mean and standard deviation for the random variable.
The parameters of the probability distribution are
m 15 and s 3.5
How to do it Example 1
Using the Normal Distribution
If a continuous random variable is
distributed as a normal distribution
the distribution is symmetrically
distributed around the mean and is
described by the mean and standard
deviation. T o find probabilities associ
ated with a normally distributed ran
dom variable use the following steps:
1. Determine the mean m and the
standard deviation s.
2. Define the event of interest
such as P1x Ú x
1
2.
3. Convert the normal distribution
to the standard normal distribu
tion using Equation 2:
4. Use the standard normal
distribution table to find the
probability associated with the
calculated zvalue. The table
gives the probability between
the zvalue and the mean.
5. Determine the desired probabil
ity using the knowledge that the
probability of a value being on
either side of the mean is 0.50
and the total probability under
the normal distribution is 1.0.
claudiozacc / Fotolia
www.downloadslide.comslide 254:
Introduction to Continuous Probability Distributions
Step 2 Define the event of interest.
The flight load time is 22 minutes. We wish to find
P1x Ú 222
Step 3 Convert the random variable to a standardized value using Equation 2.
z
x 22 15
35
200
.
.
Step 4 Find the probability associated with the zvalue in the standard normal
distribution table.
To find the probability associated with z 2.00 3i.e. P10 … z … 2.0024 do
the following:
1. Go down the lefthand column of the table to z 2.0.
2. Go across the top row of the table to the column 0.00 for the second decimal
place in z 2.00.
3. Find the value where the row and column intersect.
The value 0.4772 is the probability that a value in a normal distribution will
lie between the mean and 2.00 standard deviations above the mean.
Step 5 Determine the probability for the event of interest.
P1x Ú 222
We know that the area on each side of the mean under the normal distribution
is equal to 0.50. In Step 4 we computed the probability associated with
z 2.00 to be 0.4772 which is the probability of a value falling between the
mean and 2.00 standard deviations above the mean. Then the probability we
are looking for is
P1x Ú 222 P1z Ú 2.002 0.5000  0.4772 0.0228
END EXAMPLE
TRY PROBLEM 2
BUSINESS APPLICATION THE NORMAL DISTRIBUTION
FRUIT PRODUCTION CONTINUED Earlier we dis
cussed the situation involving the fruit grower in Florida. The
grapefruit for this grower’s orchard have a mean weight of 16
ounces. We assumed the distribution for grapefruit weight was
normally distributed with m 16 and s 4. A local television
station that runs a consumer advocacy program reported that a
grapefruit from this grower was selected and weighed only 14
ounces. The reporter said she thought it should have been heavier if the mean weight is
supposed to be 16 ounces. The grower when interviewed said that he thought the probability
was quite high that a grapefruit would weigh 14 or more ounces. To check his statement out
we want to find
P1x Ú 142
This probability corresponds to the area under a normal distribution to the right of x 14
ounces. This will be the sum of the area between x 14 and m 16 plus the area to the
right of m 16. Refer to Figure 5.
atoss/Fotolia
www.downloadslide.comslide 255:
Introduction to Continuous Probability Distributions
TABLE 1  Standard Normal Distribution Table
z
0 0.52
0.1985
Example:
z 0.52 or – 0.52
P0 z 0.52 0.1985 or 19.85
z .00 .01 .02 .03 .04 .05 .06 .07 .08 .09
0.0 .0000 .0040 .0080 .0120 .0160 .0199 .0239 .0279 .0319 .0359
0.1 .0398 .0438 .0478 .0517 .0557 .0596 .0636 .0675 .0714 .0753
0.2 .0793 .0832 .0871 .0910 .0948 .0987 .1026 .1064 .1103 .1141
0.3 .1179 .1217 .1255 .1293 .1331 .1368 .1406 .1443 .1480 .1517
0.4 .1554 .1591 .1628 .1664 .1700 .1736 .1772 .1808 .1844 .1879
0.5 .1915 .1950 .1985 .2019 .2054 .2088 .2123 .2157 .2190 .2224
0.6 .2257 .2291 .2324 .2357 .2389 .2422 .2454 .2486 .2517 .2549
0.7 .2580 .2611 .2642 .2673 .2704 .2734 .2764 .2794 .2823 .2852
0.8 .2881 .2910 .2939 .2967 .2995 .3023 .3051 .3078 .3106 .3133
0.9 .3159 .3186 .3212 .3238 .3264 .3289 .3315 .3340 .3365 .3389
1.0 .3413 .3438 .3461 .3485 .3508 .3531 .3554 .3577 .3599 .3621
1.1 .3643 .3665 .3686 .3708 .3729 .3749 .3770 .3790 .3810 .3830
1.2 .3849 .3869 .3888 .3907 .3925 .3944 .3962 .3980 .3997 .4015
1.3 .4032 .4049 .4066 .4082 .4099 .4115 .4131 .4147 .4162 .4177
1.4 .4192 .4207 .4222 .4236 .4251 .4265 .4279 .4292 .4306 .4319
1.5 .4332 .4345 .4357 .4370 .4382 .4394 .4406 .4418 .4429 .4441
1.6 .4452 .4463 .4474 .4484 .4495 .4505 .4515 .4525 .4535 .4545
1.7 .4554 .4564 .4573 .4582 .4591 .4599 .4608 .4616 .4625 .4633
1.8 .4641 .4649 .4656 .4664 .4671 .4678 .4686 .4693 .4699 .4706
1.9 .4713 .4719 .4726 .4732 .4738 .4744 .4750 .4756 .4761 .4767
2.0 .4772 .4778 .4783 .4788 .4793 .4798 .4803 .4808 .4812 .4817
2.1 .4821 .4826 .4830 .4834 .4838 .4842 .4846 .4850 .4854 .4857
2.2 .4861 .4864 .4868 .4871 .4875 .4878 .4881 .4884 .4887 .4890
2.3 .4893 .4896 .4898 .4901 .4904 .4906 .4909 .4911 .4913 .4916
2.4 .4918 .4920 .4922 .4925 .4927 .4929 .4931 .4932 .4934 .4936
2.5 .4938 .4940 .4941 .4943 .4945 .4946 .4948 .4949 .4951 .4952
2.6 .4953 .4955 .4956 .4957 .4959 .4960 .4961 .4962 .4963 .4964
2.7 .4965 .4966 .4967 .4968 .4969 .4970 .4971 .4972 .4973 .4974
2.8 .4974 .4975 .4976 .4977 .4977 .4978 .4979 .4979 .4980 .4981
2.9 .4981 .4982 .4982 .4983 .4984 .4984 .4985 .4985 .4986 .4986
3.0 .4987 .4987 .4987 .4988 .4988 .4989 .4989 .4989 .4990 .4990
To illustrate: 19.85 of the area under a normal curve lies between the mean m and a point 0.52 standard deviation
units away.
www.downloadslide.comslide 256:
Introduction to Continuous Probability Distributions
To find this probability you first convert x 14 ounces to its corresponding zvalue.
This is equivalent to determining the number of standard deviations x 14 is from the popu
lation mean of m 16. Equation 2 is used to do this as follows:
z
x 14 16
4
050 .
Because the normal distribution is symmetrical even though the zvalue is –0.50 we find
the desired probability by going to the standard normal distribution table for a positive
z 0.50. The probability in the table for z 0.50 corresponds to the probability of a zvalue
occurring between z 0.50 and z 0.00. This is the same as the probability of a z v alue
falling between z 0.50 and z 0.00. Thus from the standard normal table Table 1
we get
P1 0.50 … z … 0.002 0.1915
This is the area between x 14 and m 16 in Figure 5. We now add 0.1915 to 0.5000
3P1x 7 16 0.500024. Therefore the probability that a grapefruit will weigh 14 or more
ounces is
P1x Ú 142 0.1915 + 0.5000 0.6915
This is illustrated in Figure 5. Thus there is nearly a 70 chance that a grapefruit will weigh
at least 14 ounces.
BUSINESS APPLICATION USING THE NORMAL DISTRIBUTION
GENERAL ELECTRIC COMPANY Several states includ ing
California have passed legislation requiring automakers to sell
a certain percentage of zeroemissions cars within their borders.
One current alternative is batterypowered cars. The major
problem with batteryoperated cars is the limited time they can
be driven before the batteries must be recharged. Suppose that
General Electric GE has developed a Longlife battery pack it
claims will power a car at a sustained speed of 45 miles per hour for an average of 8 hours.
But of course there will be variations: Some battery packs will last longer and some less than
8 hours. Current data indicate that the standard deviation of battery operation time before a
charge is needed is 0.4 hours. Data show a normal distribution of uptime on these battery
FIGURE 5 
Probabilities from the Normal
Curve for Fruit Production
x
x 14
z
z –.50
x 14
0.50 0.1915
Pixel Embargo / Fotolia
www.downloadslide.comslide 257:
Introduction to Continuous Probability Distributions
packs. Automakers are concerned that batteries may run short. For example drivers might
find an “8hour” battery that lasts 7.5 hours or less unacceptable. What are the chances of this
happening with the Longlife battery pack
To calculate the probability the batteries will last 7.5 hours or less find the appropriate
area under the normal curve shown in Figure 6. There is approximately 1 chance in 10 that a
battery will last 7.5 hours or less when the vehicle is driven at 45 miles per hour.
Suppose this level of reliability is unacceptable to the automakers. Instead of a 10
chance of an “8hour” battery lasting 7.5 hours or less the automakers will accept no more
than a 2 chance. GE managers ask what the mean uptime would have to be to meet the 2
requirement.
Assuming that uptime is normally distributed we can answer this question by using the
standard normal distribution. However instead of using the standard normal table to find a
probability we use it in reverse to find the zvalue that corresponds to a known probability.
Figure 7 shows the uptime distribution for the battery packs. Note the 2 probability is shown
in the left tail of the distribution. This is the allowable chance of a battery lasting 7.5 hours or
less. We must solve for m the mean uptime that will meet this requirement.
1. Go to the body of the standard normal table where the probabilities are located and
find the probability as close to 0.48 as possible. This is 0.4798.
2. Determine the zvalue associated with 0.4798. This is z 2.05. Because we are below
the mean the z is negative. Thus z 2.05.
FIGURE 6 
Longlife Battery
z
z –1.25
x 7.5
0.3944
0.1056
z
From the normal table P–1.25 z 0 0.3944
Ten we fnd Px 7.5 hours 0.5000 – 0.3944 0.1056
–1.25
7.5 – 8
0.4
FIGURE 7 
Longlife Battery Solving for
the Mean
x Battery
uptime
hours
7.5
z –2.05
0.48
0.02
fx
z
–2.05
www.downloadslide.comslide 258:
Introduction to Continuous Probability Distributions
3. The formula for z is
z
x
4. Substituting the known values we get
205
75
.
.
0.4
5. Solve for m:
m 7.5 12.05210.42 8.32 hours
General Electric will need to increase the mean life of the battery pack to 8.32 hours to meet
the automakers’ requirement that no more than 2 of the batteries fail in 7.5 hours or less.
BUSINESS APPLICATION USING THE NORMAL DISTRIBUTION
STATE BANK AND TRUST The director of operations for the State Bank and Trust
recently performed a study of the time bank customers spent from when they walk into the
bank until they complete their banking. The data file State Bank contains the data for a
sample of 1045 customers randomly observed over a fourweek period. The customers in the
survey were limited to those who were there for basic bank business such as making a deposit
or a withdrawal or cashing a check. The histogram in Figure 8 shows that the banking times
are distributed as an approximate normal distribution.
2
Minitab Instructions for similar results:
1. Open fle: State Bank. MTW.
2. Choose Graph Histogram.
3. Click Simple.
4. Click OK.
5. In Graph Variables enter
data column Service Time.
6. Click OK.
Excel 2010 Instructions:
1. Open fle:
State Bank.xlsx.
2. Create bins upper limit
of each class.
3. Select Data Data
Analysis.
4. Select Histogram.
5. Defne data and bin
ranges.
6. Check Chart Output.
7. Defne Output Location
and click OK.
8. Select the chart and
right click.
9. Click on Format Data
Series and set gap width
to zero. Add lines to the
bars and label axes and
title appropriately.
FIGURE 8 
Excel 2010 Output for State
Bank and Trust Service Times
Chapter Outcome 3.
Excel
tutorials
Excel Tutorial
2
A statistical technique known as the chisquare goodnessoffit test can be used to determine statistically whether
the data follow a normal distribution.
www.downloadslide.comslide 259:
Introduction to Continuous Probability Distributions
The mean service time for the 1045 customers was 22.14 minutes with a standard
deviation equal to 6.09 minutes. On the basis of these data the manager assumes that
the service times are normally distributed with m 22.14 and s 6.09. Given these
assumptions the manager is considering providing a gift certificate to a local restaurant to
any customer who is required to spend more than 30 minutes to complete basic bank business.
Before doing this she is interested in the probability of having to pay off on this offer.
Figure 9 shows the theoretical distribution with the area of interest identified. The
manager is interested in finding
P1x 7 30 minutes2
This can be done manually or with Excel. Figure 10 shows the computer output for Excel. The
cumulative probability is
P1x … 302 0.9016
Then to find the probability of interest we subtract this value from 1.0 giving
P1x … 30 minutes2 1.0  0.9016 0.0984
FIGURE 9 
Normal Distribution for the
State Bank and Trust Example
x Time
x 30
Area of interest 0.0984
FIGURE 10 
Excel 2010 Output for State
Bank and Trust
Minitab Instructions for similar results:
1. Choose Calc Probability
Distribution Normal.
2. Choose Cumulative probability.
3. In Mean enter m.
4. In Standard deviation enter s.
5. In Input constant enter x.
6. Click OK.
Excel 2010 Instructions:
1. Open a blank worksheet.
2. Select Formulas.
3. Click on f
x
function
wizard.
4. Select the Statistical
category.
5. Select the NORM.DIST
function.
6. Fill in the requested
information in the
template.
7. True indicates
cumulative probabilities.
8. Click OK.
www.downloadslide.comslide 260:
Introduction to Continuous Probability Distributions
Thus there are just under 10 chances in 100 that the bank would have to give out a gift
certificate. Suppose the manager believes this policy is too liberal. She wants to set the time
limit so that the chance of giving out the gift is at most only 5. You can use the standard
normal table or the NORM.INV function in Excel to find the new limit.
3
To use the table we
first consider that the manager wants a 5 area in the upper tail of the normal distribution.
This will leave
0.50  0.05 0.45
between the new time limit and the mean. Now go to the body of the standard normal table
where the probabilities are and locate the value as close to 0.45 as possible 0.4495 or 0.4505.
Next determine the zvalue that corresponds to this probability. Because 0.45 lies midway
between 0.4495 and 0.4505 we interpolate halfway between z 1.64 and z 1.65 to get
z 1.645
Now we know
z
x
We then substitute the known values and solve for x:
1 645
22 14
22 14 1 645 6 09
32 15
.
.
.. .
.
x
x
x
6.09
8 8 minutes
Therefore any customer required to spend more than 32.158 minutes will receive the gift.
This should result in no more than 5 of the customers getting the restaurant certificate.
Obviously the bank will work to reduce the average service time or standard deviation so
even fewer customers will have to be in the bank for more than 32 minutes.
EXAMPLE 2 USING THE NORMAL DISTRIBUTION
Lockheed Martin Lockheed Martin the defense contractor has a project underway involv
ing the design and construction of communication satellite systems to be used by the U.S.
military. Because of the very high cost more than 1 billion each the company performs
numerous tests on every component. These tests tend to extend the component assembly time.
The time required to construct and test called build time a particular component part is
thought to be normally distributed with a mean equal to 30 hours and a standard deviation
equal to 4.7 hours. To keep the assembly flow moving on schedule this component needs to
have a build time of between 26 to 35 hours. To determine the probability of this happening
use the following steps:
Step 1 Determine the mean m and the standard deviation s.
The mean build time for this step in the process is thought to be 30 hours and
the standard deviation is thought to be 4.7 hours.
Step 2 Define the event of interest.
We are interested in determining the following:
P126 … x … 352
Step 3 Convert values of the specified normal distribution to corresponding
values of the standard normal distribution using Equation 2:
z
x
3
The function is NORM.INV1probability mean standard deviation2. For this example NORM.INV
.9522.146.09 32.157.
www.downloadslide.comslide 261:
Introduction to Continuous Probability Distributions
We need to find the zvalue corresponding to x 26 and to x 35.
z
x
z
26 30
47
085
35 30
47 .
.
.
and106 .
Step 4 Use the standard normal table to find the probabilities associated with
each z value.
For z 0.85 the probability is 0.3023.
For z 1.06 the probability is 0.3554.
Step 5 Determine the desired probability for the event of interest.
P126 … x … 352 0.3023 + 0.3554 0.6577
Thus there is a 0.6577 chance that the build time will be such that assembly
will stay on schedule. Using Excel’s NORM.DIST function we find NORM.
DIST35304.7true  NORM.DIST26304.7true 0.6589. The
difference in the two probabilities is due to rounding of the calculated zvalues
when done using the table.
END EXAMPLE
TRY PROBLEM 13
Approximate Areas under the Normal Curve For the normal distribution we
can make this rule more precise. Knowing the area under the normal curve between
1s 2s and 3s provides a useful benchmark for estimating probabilities and checking
reasonableness of results. Figure 11 shows these benchmark areas for any normal distribution.
FIGURE 11 
Approximate Areas under the
Normal Curve
68.26
95.44
99.74
MyStatLab
61: Exercises
Skill Development
61. For a normally distributed population with m 200
and s 20 determine the standardized zvalue for
each of the following:
a. x 225
b. x 190
c. x 240
62. For a standardized normal distribution calculate the
following probabilities:
a. P1z 6 1.52
b. P1z Ú 0.852
c. P11.28 6 z 6 1.752
63. For a standardized normal distribution calculate the
following probabilities:
a. P10.00 6 z … 2.332
b. P11.00 6 z … 1.002
c. P11.78 6 z 6 2.342
www.downloadslide.comslide 262:
Introduction to Continuous Probability Distributions
a. What is the probability that a randomly selected
value will be greater than 1550
b. What is the probability that a randomly selected
value will be less than 1485
c. What is the probability that a randomly selected
value will be either less than 1475 or greater than
1535
612. A random variable is normally distributed with a mean
of 25 and a standard deviation of 5. If an observation is
randomly selected from the distribution
a. What value will be exceeded 10 of the time
b. What value will be exceeded 85 of the time
c. Determine two values of which the smallest has
25 of the values below it and the largest has 25
of the values above it.
d. What value will 15 of the observations be below
613. A random variable is normally distributed with a mean
of 60 and a standard deviation of 9.
a. What is the probability that a randomly selected
value from the distribution will be less than 46.5
b. What is the probability that a randomly selected
value from the distribution will be greater than 78
c. What is the probability that a randomly selected
value will be between 51 and 73.5
Business Applications
614. A global financial institution transfers a large data
file every evening from offices around the world to its
London headquarters. Once the file is received it must
be cleaned and partitioned before being stored in the
company’s data warehouse. Each file is the same size
and the time required to transfer clean and partition a
file is normally distributed with a mean of 1.5 hours
and a standard deviation of 15 minutes.
a. If one file is selected at random what is the
probability that it will take longer than 1 hour and
55 minutes to transfer clean and partition the file
b. If a manager must be present until 85 of the files
are transferred cleaned and partitioned how long
will the manager need to be there
c. What percentage of the data files will take between
63 minutes and 110 minutes to be transferred
cleaned and partitioned
615. Doggie Nuggets Inc. DNI sells large bags of dog
food to warehouse clubs. DNI uses an automatic filling
process to fill the bags. Weights of the filled bags are
approximately normally distributed with a mean of 50
kilograms and a standard deviation of 1.25 kilograms.
a. What is the probability that a filled bag will weigh
less than 49.5 kilograms
b. What is the probability that a randomly sampled
filled bag will weigh between 48.5 and 51
kilograms
c. What is the minimum weight a bag of dog food
could be and remain in the top 15 of all bags
filled
d. DNI is unable to adjust the mean of the filling
process. However it is able to adjust the standard
64. For a standardized normal distribution determine a
value say z
0
so that
a. P10 6 z 6 z
0
2 0.4772
b. P1 z
0
… z 6 02 0.45
c. P1z
0
… z … z
0
2 0.95
d. P1z 7 z
0
2 0.025
e. P1z … z
0
2 0.01
65. Consider a random variable z that has a standardized
normal distribution. Determine the following
probabilities:
a. P10 6 z 6 1.962
b. P1z 7 1.6452
c. P11.28 6 z … 2.332
d. P12 … z … 32
e. P1z712
66. A random variable x has a normal distribution with
m 13.6 and s 2.90. Determine a value x
0
so that
a. P1x 7 x
0
2 0.05.
b. P1x … x
0
2 0.975.
c. P1m  x
0
… x …m + x
0
2 0.95.
67. For the following normal distributions with parameters
as specified calculate the required probabilities:
a. m 5 s 2 calculate P10 6 x 6 82.
b. m 5 s 4 calculate P10 6 x 6 82.
c. m 3 s 2 calculate P10 6 x 6 82.
d. m 4 s 3 calculate P1x 7 12.
e. m 0 s 3 calculate P1x 7 12.
68. A population is normally distributed with m 100 and
s 20.
a. Find the probability that a value randomly selected
from this population will have a value greater than
130.
b. Find the probability that a value randomly selected
from this population will have a value less than 90.
c. Find the probability that a value randomly selected
from this population will have a value between 90
and 130.
69. A random variable is known to be normally distributed
with the following parameters:
m 5.5 and s 0.50
a. Determine the value of x such that the probability of
a value from this distribution exceeding x is at most
0.10.
b. Referring to your answer in part a what must the
population mean be changed to if the probability of
exceeding the value of x found in part a is reduced
from 0.10 to 0.05
610. A randomly selected value from a normal distribution
is found to be 2.1 standard deviations above its mean.
a. What is the probability that a randomly selected
value from the distribution will be greater than 2.1
standard deviations above the mean
b. What is the probability that a randomly selected
value from the distribution will be less than 2.1
standard deviations from the mean
611. Assume that a random variable is normally distributed
with a mean of 1500 and a variance of 324.
www.downloadslide.comslide 263:
Introduction to Continuous Probability Distributions
of operating hours before failure is approximately
normally distributed.
a. What is the probability that a washdown motor will
have to be replaced free of charge
b. What percentage of FG washdown motors can be
expected to operate for more than 17500 hours
c. If FG wants to design a washdown motor so that
no more than 1 are replaced free of charge what
would the average hours of operation before failure
have to be if the standard deviation remains at 1250
hours
620. A private equity firm is evaluating two alternative
investments. Although the returns are random each
investment’s return can be described using a normal
distribution. The first investment has a mean return of
2000000 with a standard deviation of 125000. The
second investment has a mean return of 2275000
with a standard deviation of 500000.
a. How likely is it that the first investment will return
1900000 or less
b. How likely is it that the second investment will
return 1900000 or less
c. If the firm would like to limit the probability of a
return being less than 1750000 which investment
should it make
621. L.J. Raney Associates is a financial planning group
in Kansas City Missouri. The company specializes in
financial planning for schoolteachers in the Kansas
City area. As such it administers a 403b tax shelter
annuity program in which public schoolteachers can
participate. The teachers can contribute up to 20000
per year on a pretax basis to the 403b account. Very
few teachers have incomes sufficient to allow them
to make the maximum contribution. The lead analyst
at L.J. Raney Associates has recently analyzed
the company’s 403b clients and determined that
the annual contribution is approximately normally
distributed with a mean equal to 6400. Further he
has determined that the probability a customer will
contribute more than 13000 is 0.025. Based on
this information what is the standard deviation of
contributions to the 403b program
622. No Leak Plumbing and Repair provides customers with
firm quotes for a plumbing repair job before actually
starting the job. To be able to do this No Leak has been
very careful to maintain time records over the years.
For example it has determined that the time it takes to
remove a broken sink disposal and to install a new unit
is normally distributed with a mean equal to 47 minutes
and a standard deviation equal to 12 minutes. The
company bills at 75.00 for the first 30 minutes and
2.00 per minute for anything beyond 30 minutes.
Suppose the going rate for this procedure by other
plumbing shops in the area is 85.00 not including
the cost of the new equipment. If No Leak bids the
disposal job at 85 on what percentage of such jobs
will the actual time required exceed the time for which
it will be getting paid
deviation of the filling process. What would the
standard deviation need to be so that no more than
2 of all filled bags weigh more than 52 kilograms
616. LaCrosse Technology is one of many manufacturers
of atomic clocks. It makes an atomic digital watch
that is radio controlled and that maintains its accuracy
by reading a radio signal from a WWVB radio signal
from Colorado. It neither loses nor gains a second
in 20 million years. It is powered by a 3volt lithium
battery expected to last three years. Suppose the life of
the battery has a standard deviation of 0.3 years and is
normally distributed.
a. Determine the probability that the watch’s battery
will last longer than 3.5 years.
b. Calculate the probability that the watch’s battery
will last more than 2.75 years.
c. Compute the lengthoflife value for which 10 of
the watch’s batteries last longer.
617. The average number of acres burned by forest and
range fires in a large Wyoming county is 4300 acres
per year with a standard deviation of 750 acres. The
distribution of the number of acres burned is normal.
a. Compute the probability that more than 5000 acres
will be burned in any year.
b. Determine the probability that fewer then 4000
acres will be burned in any year.
c. What is the probability that between 2500 and
4200 acres will be burned
d. In those years when more than 5500 acres are
burned help is needed from easternregion fire
teams. Determine the probability help will be
needed in any year.
618. An Internet retailer stocks a popular electronic toy at
a central warehouse that supplies the eastern United
States. Every week the retailer makes a decision
about how many units of the toy to stock. Suppose that
weekly demand for the toy is approximately normally
distributed with a mean of 2500 units and a standard
deviation of 300 units.
a. If the retailer wants to limit the probability of being
out of stock of the electronic toy to no more than
2.5 in a week how many units should the central
warehouse stock
b. If the retailer has 2750 units on hand at the start
of the week what is the probability that weekly
demand will be greater than inventory
c. If the standard deviation of weekly demand for
the toy increases from 300 units to 500 units how
many more toys would have to be stocked to ensure
that the probability of weekly demand exceeding
inventory is no more than 2.5
619. FG Industries manufactures a washdown motor
that is used in the food processing industry. The motor
is marketed with a warranty that guarantees it will
be replaced free of charge if it fails within the first
13000 hours of operation. On average FG wash
down motors operate for 15000 hours with a standard
deviation of 1250 hours before failing. The number
www.downloadslide.comslide 264:
Introduction to Continuous Probability Distributions
usage would have enough capacity with the new 1GB
player
626. According to the Federal Reserve Board the average
credit card debt per U.S. household was 8565 in
2008. Assume that the distribution of credit card debt
per household has a normal distribution with a standard
deviation of 3000.
a. Determine the percentage of households that have a
credit card debt of more than 13000.
b. One household has a credit card debt that is at the
95th percentile. Determine its credit card debt.
c. If four households were selected at random
determine the probability that at least half of them
would have credit card debt of more than 13000.
627. GeorgiaPacific is a major forest products company
in the United States. In addition to timberlands the
company owns and operates numerous manufacturing
plants that make lumber and paper products. At one
of their plywood plants the operations manager
has been struggling to make sure that the plywood
thickness meets quality standards. Specifically all
sheets of their 3/4inch plywood must fall within the
range 0.747 to 0.753 inches in thickness. Studies have
shown that the current process produces plywood that
has thicknesses that are normally distributed with a
mean of 0.751 inches and a standard deviation equal
to 0.004 inches.
a. Use Excel to determine the proportion of plywood
sheets that will meet quality specifications 0.747 to
0.753 given how the current process is performing.
b. Referring to part a suppose the manager is
unhappy with the proportion of product meeting
specifications. Assuming that he can get the mean
adjusted to 0.75 inches what must the standard
deviation be if he is going to have 98 of his
product meet specifications
628. A senior loan officer for Whitney National Bank has
recently studied the bank’s real estate loan portfolio
and found that the distribution of loan balances is
approximately normally distributed with a mean of
155600 and a standard deviation equal to 33050.
As part of an internal audit bank auditors recently
randomly selected 100 real estate loans from the
portfolio of all loans and found that 80 of these
loans had balances below 170000. The senior loan
officer is concerned that the sample selected by the
auditors is not representative of the overall portfolio.
In particular he is interested in knowing the expected
proportion of loans in the portfolio that would have
balances below 170000. You are asked to conduct
an appropriate analysis and write a short report to the
senior loan officers with your conclusion about the
sample.
Computer Database Exercises
629. The PricewaterhouseCoopers Human Capital Index
Report indicated that the average cost for an American
company to fill a job vacancy during the study period
623. According to Business Week Maternity Chic a
purveyor of designer maternity wear sells dresses and
pants priced around 150 each for an average total sale
of 1200. The total sale has a normal distribution with
a standard deviation of 350.
a. Calculate the probability that a randomly selected
customer will have a total sale of more than 1500.
b. Compute the probability that the total sale will be
within 2 standard deviations of the mean total sales.
c. Determine the median total sale.
624. The Aberdeen CocaCola Bottling plant located in
Aberdeen North Carolina is the bottler and distributor
for CocaCola products in the Aberdeen area. The
company’s product line includes 12ounce cans of
Coke products. The cans are filled by an automated
filling process that can be adjusted to any mean fill
volume and that will fill cans according to a normal
distribution. However not all cans will contain the
same volume due to variation in the filling process.
Historical records show that regardless of what the
mean is set at the standard deviation in fill will be
0.035 ounces. Operations managers at the plant know
that if they put too much Coke in a can the company
loses money. If too little is put in the can customers
are short changed and the North Carolina Department
of Weights and Measures may fine the company.
a. Suppose the industry standards for fill volume call
for each 12ounce can to contain between 11.98
and 12.02 ounces. Assuming that the Aberdeen
manager sets the mean fill at 12 ounces what is the
probability that a can will contain a volume of Coke
product that falls in the desired range
b. Assume that the Aberdeen manager is focused on an
upcoming audit by the North Carolina Department
of Weights and Measures. She knows the process
is to select one Coke can at random and that if it
contains less than 11.97 ounces the company will
be reprimanded and potentially fined. Assuming
that the manager wants at most a 5 chance of this
happening at what level should she set the mean fill
level Comment on the ramifications of this step
assuming that the company fills tens of thousands of
cans each week.
625. MP3 players and most notably the Apple iPod have
become an industry standard for people who want to
have access to their favorite music and videos in a
portable format. The iPod can store massive numbers
of songs and videos with its 120GB hard drive.
Although owners of the iPod have the potential to store
lots of data a recent study showed that the actual disk
storage being used is normally distributed with a mean
equal to 1.95 GB and a standard deviation of 0.48 GB.
Suppose a competitor to Apple is thinking of entering
the market with a lowcost iPod clone that has only 1.0
GB of storage. The marketing slogan will be “Why Pay
for Storage Capacity that You Don’t Need”
Based on the data from the study of iPod owners
what percentage of owners based on their current
www.downloadslide.comslide 265:
Introduction to Continuous Probability Distributions
randomly chosen transaction would yield a price of
2.12 or smaller even though the population mean
was 2.51.
631. USA Today’s annual survey of public flagship
universities Arienne Thompson and Breanne
Gilpatrick “DoubleDigit Hikes Are Down” October
5 2005 indicates that the median increase in instate
tuition was 7 for the 2005–2006 academic year. A
file titled Tuition contains the percentage change for
the 67 flagship universities.
a. Produce a relative frequency histogram for these
data. Does it seem plausible that the data are from a
population that has a normal distribution
b. Suppose the decimal point of the three largest
numbers had inadvertently been moved one
place to the right in the data. Move the decimal
point one place to the left and reconstruct the
relative frequency histogram. Now does it seem
plausible that the data have an approximate normal
distribution
c. Use the normal distribution of part b to approximate
the proportion of universities that raised their
instate tuition more than 10. Use the appropriate
parameters obtained from this population.
d. Use the normal distribution of part b to approximate
the fifth percentile for the percent of tuition
increase.
was 3270. Sample data similar to those used in the
study are in a file titled Hired.
a. Produce a relative frequency histogram for these
data. Does it seem plausible the data were sampled
from a normally distributed population
b. Calculate the mean and standard deviation of the
cost of filling a job vacancy.
c. Determine the probability that the cost of filling a
job vacancy would be between 2000 and 3000.
d. Given that the cost of filling a job vacancy was
between 2000 and 3000 determine the
probability that the cost would be more than 2500.
630. A recent article in USA Today discussed prices for
the 200 brandname drugs most commonly used by
Americans over age 50. Atrovent a treatment for lung
conditions such as emphysema was one of the drugs.
The file titled Drug contains daily cost data similar to
those obtained in the research.
a. Produce a relative frequency histogram for these
data. Does it seem plausible the data were sampled
from a population that was normally distributed
b. Compute the mean and standard deviation for the
sample data in the file Drug.
c. Assuming the sample came from a normally
distributed population and the sample standard
deviation is a good approximation for the population
standard deviation determine the probability that a
END EXERCISES 61
2 Other Continuous Probability
Distributions
This section introduces two additional continuous probability distributions that are used in
business decision making: the uniform distribution and the exponential distribution.
Uniform Probability Distribution
The uniform distribution is sometimes referred to as the distribution of little information
because the probability over any interval of the continuous random variable is the same as for
any other interval of the same width.
Equation 3 defines the continuous uniform density function.
Chapter Outcome 4.
Continuous Uniform Density Function
fx ba
axb
1
0
if
otherwise
3
where:
f1x2 Value of the density function at any xvalue
a The smallest value assumed by the uniform random variable of interest
b The largest value assumed by the uniform random variable of interest
www.downloadslide.comslide 266:
Introduction to Continuous Probability Distributions
Figure 12 illustrates two examples of uniform probability distributions with different a to
b intervals. Note the height of the probability density function is the same for all values of
x between a and b for a given distribution. The graph of the uniform distribution is a rec
tangle.
EXAMPLE 3 USING THE UNIFORM DISTRIBUTION
Georgia Pacific The Georgia Pacific Company owns and operates several tree farms in dif
ferent parts of the United States and South America. The lead botanist for the company has
stated that pine trees on one parcel of land will increase in diameter between one and four
inches per year according to a uniform distribution. Suppose the company is interested in the
probability that a given tree will have an increased diameter of more than 2 inches. The prob
ability can be determined using the following steps:
Step 1 Define the density function.
The height of the probability rectangle fx for the tree growth interval of one
to four inches is determined using Equation 3 as follows:
fx
ba
fx
.
1
1
41
1
3
033
Step 2 Define the event of interest.
The botanist is specifically interested in a tree that has increased by more than
two inches in diameter. This event of interest is x 7 2.0.
Step 3 Calculate the required probability.
We determine the probability as follows:
P1x 7 2.02 1  P1x … 2.02
1  f1x212.0  1.02
1  0.331 1.02
1  0.33
0.67
Thus there is a 0.67 probability that a tree will increase by more than two inches in diameter.
END EXAMPLE
TRY PROBLEM 32
Like the normal distribution the uniform distribution can be further described by
specifying the mean and the standard deviation. These values are computed using
Equations 4 and 5.
FIGURE 12 
Uniform Distributions
fx
0.50
0.25
2
a
5
b
fx
0.50
0.25
3
a
8
b
fx
xx
1
5 – 2
0.33 for 2 x 5
1
3
fx
1
8 – 3
0.2 for 3 x 8
1
5
a b
www.downloadslide.comslide 267:
Introduction to Continuous Probability Distributions
EXAMPLE 4 THE MEAN AND STANDARD DEVIATION
OF A UNIFORM DISTRIBUTION
Surgery Recovery The chief administrator of a San
Francisco–area surgical center has analyzed data from a large
number of shoulder surgeries conducted by her center and others
in a medical association in California. The analysis shows that
the recovery time for shoulder surgery ranges between 15 and 45
weeks. Without any further information the administrator will
apply a uniform distribution to surgery times. Based on this she
can determine the mean and standard deviation for the recovery duration using the following
steps:
Step 1 Define the density function.
Equation 3 can be used to define the distribution:
fx
ba
.
11
45 15
1
30
0 0333
Step 2 Compute the mean of the probability distribution using Equation 4.
ab
2
15 45
2
30
Thus the mean recovery time is 30 weeks.
Step 3 Compute the standard deviation using Equation 5.
.
ba
22
12
45 15
12
75 8 66
The standard deviation is 8.66 weeks.
END EXAMPLE
TRY PROBLEM 34
The Exponential Probability Distribution
Another continuous probability distribution frequently used in business situations is the
exponential distribution. The exponential distribution is used to measure the time that
elapses between two occurrences of an event such as the time between “hits” on an Internet
home page. The exponential distribution might also be used to describe the time between
arrivals of customers at a bank drivein teller window or the time between failures of an
electronic component. Equation 6 shows the probability density function for the exponen
tial distribution.
Mean and Standard Deviation of a Uniform Distribution
Mean Expected Value:
Ex
ab
2
4
Standard Deviation:
ba
2
12
5
where:
a The smallest value assumed by the uniform random variable of interest
b The largest value assumed by the uniform random variable of interest
Chapter Outcome 5.
araraadt/Fotolia
www.downloadslide.comslide 268:
Introduction to Continuous Probability Distributions
Note the parameter that defines the exponential distribution is l lambda. You should
recall that l is the mean value for the Poisson distribution. If the number of occurrences per
time period is known to be Poisson distributed with a mean of l then the time between occur
rences will be exponentially distributed with a mean time of 1l.
If we select a value for l we can graph the exponential distribution by substituting l
and different values for x into Equation 6. For instance Figure 13 shows exponential density
functions for l 0.5 l 1.0 l 2.0 and l 3.0. Note in Figure 13 that for any expo
nential density function f1x2 f102 l as x increases fx approaches zero. It can also be
shown that the standard deviation of any exponential distribution is equal to the mean 1l.
As with any continuous probability distribution the probability that a value will fall
within an interval is the area under the graph between the two points defining the interval.
Equation 7 is used to find the probability that a value will be equal to or less than a particular
value for an exponential distribution.
Exponential Density Function
A continuous random variable that is exponentially distributed has the probability
density function given by
fx le
lx
x Ú 0 6
where:
e 2.71828c
1l The mean time between events1l 7 02
fx Probability Density Function
3.0
2.5
2.0
1.5
1.0
0.5
0
0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5
Values of x
l 3.0 Mean 0.3333
l 2.0 Mean 0.50
l 1.0 Mean 1.0
l 0.50 Mean 2.0
FIGURE 13 
Exponential Distributions
Exponential Probability
P10 … x … a2 1  e
la
7
where:
a the value of interest
1l Mean
e Base of natural log 2.71828
www.downloadslide.comslide 269:
Introduction to Continuous Probability Distributions
Exponential Distribution Table contains a table of e
la
values for different values of l a.
You can use this table and Equation 7 to find the probabilities when the l a of interest is con
tained in the table. You can also use Excel to find exponential probabilities as the following
application illustrates.
BUSINESS APPLICATION USING EXCEL TO CALCULATE EXPONENTIAL
PROBABILITIES
HAINES INTERNET SERVICES The Haines Internet Services Company has determined
that the number of customers who attempt to connect to the Internet per hour is Poisson
distributed with l 30 per hour. The time between connect requests is exponentially
distributed with a mean time between requests of 2.0 minutes computed as follows:
l 30 attempts per 60 minutes 0.50 attempts per minute
The mean time between attempted connects then is
1
1
050
20 /
.
. minutes
Because of the system that Haines uses if customer requests are too close together—45
seconds 0.75 minutes or less—the connection will fail. The managers at Haines are analyzing
whether they should purchase new equipment that will eliminate this problem. They need to
know the probability that a customer will fail to connect. Thus they want
P1x … 0.75 minutes2
Excel
tutorials
Excel Tutorial
FIGURE 14 
Excel 2010 Exponential
Probability Output for Haines
Internet Services
Minitab Instructions for similar results:
1. Choose Calc Probability Distributions
Exponential.
2. Choose Cumulative probability.
3. In Scale enter m .
4. In Input constant enter value for x.
5. Click OK.
Excel 2010 Instructions:
1. On the Formula tab click
on f
x
function wizard.
2. Select Statistical category.
3. Select EXPON.DIST
function.
4. Supply x and l.
5. Set Cumulative TRUE
for cumulative probability.
Inputs:
x 0.75 minutes
45 seconds
l 0.50 per minute
True output is the
cumulative
probability
www.downloadslide.comslide 270:
Introduction to Continuous Probability Distributions
To find this probability using a calculator we need to first determine la. In this example
l 0.50 and a 0.75. Then
l a 10.50210.752 0.3750
We find that the desired probability is
1  e
la
1  e
 0.3750
0.3127
The managers can also use the EXPON.DIST function in Excel to compute the precise
value for the desired probability.
4
Using Excel Figure 14 shows that the chance of failing
to connect is 0.3127. This means that nearly one third of the customers will experience a
problem with the current system.
4
The Excel EXPON.DIST function requires that l be inputted rather than 1l.
Skill Development
632. A continuous random variable is uniformly distributed
between 100 and 150.
a. What is the probability a randomly selected value
will be greater than 135
b. What is the probability a randomly selected value
will be less than 115
c. What is the probability a randomly selected value
will be between 115 and 135
633. Determine the following:
a. the probability that a uniform random variable
whose range is between 10 and 30 assumes a value
in the interval 10 to 20 or 15 to 25
b. the quartiles for a uniform random variable whose
range is from 4 to 20
c. the mean time between events for an exponential
random variable that has a median equal to 10
d. the 90th percentile for an exponential random
variable that has the mean time between events
equal to 0.4.
634. Suppose a random variable x has a uniform
distribution with a 5 and b 9.
a. Calculate P15.5 … x … 82.
b. Determine P1x 7 72.
c. Compute the mean m and standard deviation s of
this random variable.
d. Determine the probability that x is in the interval
1m 2s2.
635. Let x be an exponential random variable with l 0.5.
Calculate the following probabilities:
a. P1x 6 52
b. P1x 7 62
c. P15 … x … 62
d. P1x Ú 22
e. the probability that x is at most 6
636. The useful life of an electrical component is
exponentially distributed with a mean of 2500
hours.
MyStatLab
62: Exercises
a. What is the probability the circuit will last more
than 3000 hours
b. What is the probability the circuit will last between
2500 and 2750 hours
c. What is the probability the circuit will fail within
the first 2000 hours
637. The time between telephone calls to a cable television
payment processing center follows an exponential
distribution with a mean of 1.5 minutes.
a. What is the probability that the time between the
next two calls will be 45 seconds or less
b. What is the probability that the time between the
next two calls will be greater than 112.5 seconds
Business Applications
638. Suppose you are traveling on business to a foreign
country for the first time. You do not have a bus
schedule or a watch with you. However you have been
told that buses stop in front of your hotel every 20
minutes throughout the day. If you show up at the bus
stop at a random moment during the day determine the
probability that
a. you will have to wait for more than 10 minutes
b. you will only have to wait for 6 minutes or less
c. you will have to wait between 8 and 15 minutes
639. When only the valueadded time is considered the
time it takes to build a laser printer is thought to be
uniformly distributed between 8 and 15 hours.
a. What are the chances that it will take more than 10
valueadded hours to build a printer
b. How likely is it that a printer will require less than 9
valueadded hours
c. Suppose a single customer orders two printers.
Determine the probability that the first and second
printer each will require less than 9 valueadded
hours to complete.
640. The time required to prepare a dry cappuccino using
whole milk at the Daily Grind Coffee House is
www.downloadslide.comslide 271:
Introduction to Continuous Probability Distributions
a. What is the probability that a customer will arrive
within the next 3 minutes
b. What is the probability that the time between the
arrivals of customers is 12 minutes or more
c. What is the probability that the next customer will
arrive within 4 and 6 minutes
645. The average amount spent on electronics each year in
U.S. households is 1250 according to an article in
USA Today Michelle Kessler “Gadget Makers Make
Mad Dash to Market” January 4 2006. Assume
that the amount spent on electronics each year has an
exponential distribution.
a. Calculate the probability that a randomly chosen
U.S. household would spend more than 5000 on
electronics.
b. Compute the probability that a randomly chosen
U.S. household would spend more than the average
amount spent by U.S. households.
c. Determine the probability that a randomly chosen
U.S. household would spend more than 1 standard
deviation below the average amount spent by U.S.
households.
646. Charter Southeast Airlines states that the flight
between Fort Lauderdale Florida and Los Angeles
takes 5 hours and 37 minutes. Assume that the actual
flight times are uniformly distributed between 5 hours
and 20 minutes and 5 hours and 50 minutes.
a. Determine the probability that the flight will be
more than 10 minutes late.
b. Calculate the probability that the flight will be more
than 5 minutes early.
c. Compute the average flight time between these two
cities.
d. Determine the variance in the flight times between
these two cities.
647. A corrugated container company is testing whether
a computer decision model will improve the uptime
of its box production line. Currently knives used in
the production process are checked manually and
replaced when the operator believes the knives are dull.
Knives are expensive so operators are encouraged not
to change the knives early. Unfortunately if knives
are left running for too long the cuts are not made
properly which can jam the machines and require
that the entire process be shut down for unscheduled
maintenance. Shutting down the entire line is costly
in terms of lost production and repair work so
the company would like to reduce the number of
shutdowns that occur daily. Currently the company
experiences an average of 0.75 kniferelated shutdowns
per shift exponentially distributed. In testing the
computer decision model reduced the frequency of
kniferelated shutdowns to an average of 0.20 per
shift exponentially distributed. The decision model is
expensive but the company will install it if it can help
achieve the target of four consecutive shifts without a
kniferelated shutdown.
uniformly distributed between 25 and 35 seconds.
Assuming a customer has just ordered a wholemilk
dry cappuccino
a. What is the probability that the preparation time
will be more than 29 seconds
b. What is the probability that the preparation time
will be between 28 and 33 seconds
c. What percentage of wholemilk dry cappuccinos
will be prepared within 31 seconds
d. What is the standard deviation of preparation times
for a dry cappuccino using whole milk at the Daily
Grind Coffee House
641. The time to failure for a power supply unit used in a
particular brand of personal computer PC is thought
to be exponentially distributed with a mean of 4000
hours as per the contract between the vendor and
the PC maker. The PC manufacturer has just had a
warranty return from a customer who had the power
supply fail after 2100 hours of use.
a. What is the probability that the power supply would
fail at 2100 hours or less Based on this probability
do you feel the PC maker has a right to require that
the power supply maker refund the money on this
unit
b. Assuming that the PC maker has sold 100000
computers with this power supply approximately
how many should be returned due to failure at 2100
hours or less
642. A delicatessen located in the heart of the business
district of a large city serves a variety of customers.
The delicatessen is open 24 hours a day every day of
the week. In an effort to speed up takeout orders the
deli accepts orders by fax. If on the average 20 orders
are received by fax every two hours throughout the day
find the
a. probability that a faxed order will arrive within the
next 9 minutes
b. probability that the time between two faxed orders
will be between 3 and 6 minutes
c. probability that 12 or more minutes will elapse
between faxed orders
643. Dennis Cauchon and Julie Appleby reported in
USA Today that the average patient cost per stay in
American hospitals was 8166. Assume that this cost
is exponentially distributed.
a. Determine the probability that a randomly selected
patient’s stay in an American hospital will cost more
than 10000.
b. Calculate the probability that a randomly selected
patient’s stay in an American hospital will cost less
than 5000.
c. Compute the probability that a randomly selected
patient’s stay in an American hospital will cost
between 8000 and 12000.
644. During the busiest time of the day customers arrive
at the Daily Grind Coffee House at an average of
15 customers per 20minute period.
www.downloadslide.comslide 272:
Introduction to Continuous Probability Distributions
e. Using only the information obtained in parts c and
d describe the shape of this distribution. Does this
agree with the findings in part a
649. Although some financial institutions do not charge
fees for using ATMs many do. A recent study found
the average fee charged by banks to process an ATM
transaction was 2.91. The file titled ATM Fees
contains a list of ATM fees that might be required by
banks.
a. Produce a relative frequency histogram for these
data. Does it seem plausible the data came from a
population that has an exponential distribution
b. Calculate the mean and standard deviation of the
ATM fees.
c. Assume that the distribution of ATM fees is
exponentially distributed with the same mean as
that of the sample. Determine the probability that a
randomly chosen bank’s ATM fee would be greater
than 3.00.
650. The San Luis Obispo California Transit Program
provides daily fixedroute transit service to the
general public within the city limits and to Cal
Poly State University’s staff and students. The
most heavily traveled route schedules a city bus to
arrive at Cal Poly at 8:54 a.m. The file titled Late
lists plausible differences between the actual and
scheduled time of arrival rounded to the nearest
minute for this route.
a. Produce a relative frequency histogram for these
data. Does it seem plausible the data came from a
population that has a uniform distribution
b. Provide the density for this uniform distribution.
c. Classes start 10 minutes after the hour and classes
are a 5minute walk from the dropoff point.
Determine the probability that a randomly chosen
bus on this route would cause the students on board
to be late for class. Assume the differences form a
continuous uniform distribution with a range the
same as the sample.
d. Determine the median difference between the actual
and scheduled arrival times.
a. Under the current system what is the probability
that the plant would run four or more consecutive
shifts without a kniferelated shutdown
b. Using the computer decision model what is the
probability that the plant could run four or more
consecutive shifts without a kniferelated shutdown
Has the decision model helped the company achieve
its goal
c. What would be the maximum average number of
shutdowns allowed per day such that the probability
of experiencing four or more consecutive shifts
without a kniferelated shutdown is greater than or
equal to 0.70
Computer Database Exercises
648. RollsRoyce PLC provides forecasts for the business
jet market and covers the regional and major aircraft
markets. In a recent release Rolls Royce indicated
that in both North America and Europe the number
of delayed departures has declined since a peak in
1999/2000. This is partly due to a reduction in the
number of flights at major airports and the younger
aircraft fleets but it also results from improvements in
air traffic management capacity especially in Europe.
Comparing January–April 2003 with the same period
in 2001 for similar traffic levels the average en route
delay per flight was reduced by 65 from 2.2 minutes
to 0.7 minutes. The file titled Delays contains a
possible sample of the en route delay times in minutes
for 200 flights.
a. Produce a relative frequency histogram for this
data. Does it seem plausible the data come from a
population that has an exponential distribution
b. Calculate the mean and standard deviation of the en
route delays.
c. Determine the probability that this exponential
random variable will be smaller than its mean.
d. Determine the median time in minutes for the en
route delays assuming they have an exponential
distribution with a mean equal to that obtained in
part b.
END EXERCISES 62
www.downloadslide.comslide 273:
Introduction to Continuous Probability Distributions
Visual Summary
1 The Normal Probability Distribution
2 Other Continuous Probability Distributions
Outcome 4. Calculate probabilities associated with a uniformly distributed random variable.
Outcome 5. Determine probabilities for an exponential probability distribution.
Conclusion
The normal distribution has wide application throughout the study of business statistics. You will
be making use of the normal distribution. The normal distribution has very special properties. It
is a symmetric bellshaped distribution. To fnd probabilities for a normal distribution you will
frst standardize the distribution by converting values of the random variable to standardized
zvalues. Other continuous distributions introduced in this chapter are the exponential
distribution and the uniform distribution. Figure 15 summarizes the discrete probability
distributions and the continuous probability distributions introduced in this chapter.
A random variable can take on values that are either discrete or continuous. This
chapter has focused on continuous random variables where the potential values of the
variable can be any value on a continuum. Examples of continuous random variables
include the time it takes a worker to assemble a part the weight of a potato the
distance it takes to stop a car once the brakes have been applied and the volume of
waste water emitted from a food processing facility. Values of a continuous random
variable are generally determined by measuring. One of the most frequently used
continuous probability distributions is called the normal distribution.
Summary
The normal distribution is a symmetric bellshaped probability distribution. Half the probability lies to
the right and half lies to the left of the mean. To fnd probabilities associated with a normal distribution you
will want to convert to a standard normal distribution by frst converting values of the random
variables to standardizedzvalues. The probabilities associated with a range of values for the random
variable are found using Excel.
Outcome 1. Convert a normal distribution to a standard normal distribution.
Outcome 2. Determine probabilities using the standard normal distribution.
Outcome 3. Calculate values of the random variable associated with specifed probabilities from
a normal distribution.
Summary
Although the normal distribution is by far the most frequently used continuous probability distribution two other
continuous distributions are introduced in this section. These are the uniform distribution and the exponential
distribution. With the uniform distribution the probability over any interval is the same as any other interval of the
same width. The probabilities for the uniform distribution are computed using Equation 3. The exponential
distribution is based on a single parameter lambda and is often used to describe random service times or the time
between customer arrivals in waiting line applications. The probability over a range of values for an exponential
distribution can be computed using Equation 7. Also Excel and Minitab have functions for calculating the exponential
probabilities.
www.downloadslide.comslide 274:
Introduction to Continuous Probability Distributions
4 Mean of the Uniform Distribution
Ex
ab
2
5 Standard Deviation of the Uniform Distribution
ba
2
12
6 Exponential Density Function
f1x2 l e
lx
x Ú 0
7 Exponential Probability
P10 … x … a2 1  e
la
1 Normal Probability Density Function
fx e
x
1 2
2
2
2
2 Standardized NormalzValue
z
x
3 Continuous Uniform Density Function
fx ba
axb
1
0
if
otherwise
FIGURE 15 
Probability Distribution
Summary
Exponential
Distribution
Hypergeometric
Distribution
Random Variable
Values Are
Determined by
Counting
Random Variable
Values Are
Determined by
Measuring
Random
Variable
Binomial
Distribution
Discrete Continuous
Normal
Distribution
Uniform
Distribution
Poisson
Distribution
Equations
Key Terms
Normal distribution Standard normal distribution
Chapter Exercises
MyStatLab
Conceptual Questions
651. Discuss the difference between discrete and continuous
probability distributions. Discuss two situations in
which a variable of interest may be considered either
continuous or discrete.
652. Recall the Empirical Rule. It states that if the data
distribution is bell shaped then the interval m s
contains approximately 68 of the values m 2s
contains approximately 95 and m 3s contains
virtually all of the data values. The bellshaped
distribution referenced is the normal distribution.
a. Verify that a standard normal distribution contains
approximately 68 of the values in the interval
m s.
b. Verify that a standard normal distribution contains
approximately 95 of the values in the interval
m 2s.
c. Verify that a standard normal distribution contains
virtually all of the data in the interval m 3s.
www.downloadslide.comslide 275:
Introduction to Continuous Probability Distributions
Business Applications
659. The manager for SelectaSeat a company that sells
tickets to athletic games concerts and other events
has determined that the number of people arriving
at the Broadway location on a typical day is Poisson
distributed with a mean of 12 per hour. It takes
approximately four minutes to process a ticket request.
Thus if customers arrive in intervals that are less than
four minutes they will have to wait. Assuming that a
customer has just arrived and the ticket agent is starting
to serve that customer what is the probability that the
next customer who arrives will have to wait in line
660. The Four Brothers Lumber Company is considering
buying a machine that planes lumber to the correct
thickness. The machine is advertised to produce “6inch
lumber” having a thickness that is normally distributed
with a mean of 6 inches and a standard deviation of
0.1 inch.
a. If building standards in the industry require a
99 chance of a board being between 5.85 and
6.15 inches should Four Brothers purchase this
machine Why or why not
b. To what level would the company that
manufactures the machine have to reduce the
standard deviation for the machine to conform to
industry standards
661. Two automatic dispensing machines are being
considered for use in a fastfood chain. The first
dispenses an amount of liquid that has a normal
distribution with a mean of 11.9 ounces and a standard
deviation of 0.07 ounces. The second dispenses an
amount of liquid that has a normal distribution with a
mean of 12.0 ounces and a standard deviation of 0.05
ounces. Acceptable amounts of dispensed liquid are
between 11.9 and 12.0 ounces. Calculate the relevant
probabilities and determine which machine should be
selected.
662. A small private ambulance service in Kentucky has
determined that the time between emergency calls is
exponentially distributed with a mean of 41 minutes.
When a unit goes on call it is out of service for 60
minutes. If a unit is busy when an emergency call is
received the call is immediately routed to another
service. The company is considering buying a second
ambulance. However before doing so the owners are
interested in determining the probability that a call
will come in before the ambulance is back in service.
Without knowing the costs involved in this situation
does this probability tend to support the need for a
second ambulance Discuss.
663. Assume that after the first 12 hours the average
remaining useful life of a particular battery before
recharging is required is 9 hours and that the remaining
time is exponentially distributed. What is the probability
that a randomly sampled battery of this type will last
between 15 and 17 hours
653. The probability that a value from a normally distributed
random variable will exceed the mean is 0.50. The same
is true for the uniform distribution. Why is this not
necessarily true for the exponential distribution Discuss
and show examples to illustrate your point.
654. Suppose you tell one of your fellow students that when
working with a continuous distribution it does not make
sense to try to compute the probability of any specific
value since it will be zero. She says that when the
experiment is performed some value must occur the
probability can’t be zero. Y our task is to respond to her
statement and in doing so explain why it is appropriate
to find the probability for specific ranges of values for a
continuous distribution.
655. The exponential distribution has a characteristic that
is called the “memoryless” property. This means
P1X 7 x2 P1X 7 x + x
0
X 7 x
0
2. To illustrate
this consider the calls coming into 911. Suppose that
the distribution of the time between occurrences has an
exponential distribution with a mean of one half hour
1 0.52.
a. Calculate the probability that no calls come in
during the first hour.
b. Now suppose that you are monitoring the call
frequency and you note that a call does not come in
during the first two hours. Determine the probability
that no calls will come in during the next hour.
656. Revisit Problem 55 but examine whether it would
matter when you started monitoring the 911 calls if the
time between occurrences had a uniform distribution
with a mean of 2 and a range of 4.
a. Calculate the probability that no call comes in
during the first hour.
b. Now suppose that you are monitoring the call
frequency and you note that no call comes in during
the first two hours. Determine the probability that
no calls will arrive during the next hour.
657. Suppose that on average 20 customers arrive every
hour at a twentyfourhour coffee shop. Assume that
the time between customer arrivals is exponentially
distributed. Determine
a. The probability that a customer arrives within the
next 2 minutes.
b. The probability that the time between two arriving
customers will be between 1 and 4 minutes.
c. The probability that 5 or more minutes will pass
between customer arrivals.
658. Assume that the time required to receive a confirmation
that an electronic transfer has occurred is uniformly
distributed between 30 and 90 seconds.
a. What is the probability that a randomly selected
transfer will take between 30 and 45 seconds
b. What is the probability that a randomly selected
transfer will take between 50 and 90 seconds
c. What proportion of transfers will take between 40
and 75 seconds
www.downloadslide.comslide 276:
Introduction to Continuous Probability Distributions
b. How likely is it that no traffic violations will be
recorded within the next 7 minutes
668. The St. Maries plywood plant is part of the Potlatch
Corporation’s Northwest Division. The plywood
superintendent organized a study of the tree diameters
that are being shipped to the mill. After collecting a
large amount of data on diameters he concluded that
the distribution is approximately normally distributed
with a mean of 14.25 inches and a standard deviation
of 2.92 inches. Because of the way plywood is made
there is a certain amount of waste on each log because
the peeling process leaves a core that is approximately
3 inches thick. For this reason he feels that any log less
than 10 inches in diameter is not profitable for making
plywood.
a. Based on the data the superintendent has
collected what is the probability that a log will be
unprofitable
b. An alternative is to peel the log and then sell the
core as “peeler logs.” These peeler logs are sold
as fence posts and for various landscape projects.
There is not as much profit in these peeler logs
however. The superintendent has determined that
he can make a profit if the peeler log’s diameter is
not more than 32 of the diameter of the log. Using
this additional information calculate the proportion
of logs that will be unprofitable.
669. The personnel manager for a large company is interested
in the distribution of sickleave hours for employees of
her company. A recent study revealed the distribution to
be approximately normal with a mean of 58 hours per
year and a standard deviation of 14 hours.
An office manager in one division has reason to
believe that during the past year two of his employees
have taken excessive sick leave relative to everyone else.
The first employee used 74 hours of sick leave and the
second used 90 hours. What would you conclude about
the office manager’s claim and why
670. If the number of hours between servicing required
for a particular snowmobile engine is exponentially
distributed with an average of 118 hours determine the
probability that a randomly selected engine
a. Will run at least 145 hours before servicing is
needed.
b. Will run at most 161 hours before servicing is
needed.
671. Assume that the amount of time in minutes for eighth
graders to compete an assessment examination is 78
minutes with a standard deviation of 12 minutes.
a. What proportion of eighth graders completes the
assessment examination in 72 minutes or less
b. What proportion of eighth graders completes the
assessment examination in 82 minutes or more
c. For what number of minutes would 90 of
all eighth graders complete the assessment
examination
664. An online article http://beauty.about.com by Julyne
Derrick “Shelf Lives: How Long Can Y ou Keep
Makeup” suggests that eye shadow and eyeliner each
have a shelf life of up to three years. Suppose the shelf
lives of these two products are exponentially distributed
with an average shelf life of one year.
a. Calculate the probability that the shelf life of eye
shadow will be longer than three years.
b. Determine the probability that at least one of these
products will have a shelf life of more than three
years.
c. Determine the probability that a purchased eyeliner
that is useful after one year will be useful after three
years.
665. The Shadow Mountain Golf Course is preparing for a
major LPGA golf tournament. Since parking near the
course is extremely limited room for only 500 cars
the course officials have contracted with the local
community to provide parking and a bus shuttle service.
Sunday the final day of the tournament will have the
largest crowd and the officials estimate there will be
between 8000 and 12000 cars needing parking spaces
but think no value is more likely than another. The
tournament committee is discussing how many parking
spots to contract from the city. If they want to limit the
chance of not having enough provided parking to 10
how many spaces do they need from the city on Sunday
666. One of the products of Pittsburg Plate Glass Industries
PPG is laminated safety glass. It is made up of two
pieces of glass 0.125 inch thick with a thin layer of
vinyl sandwiched between them. The average thickness
of the laminated safety glass is 0.25 inch. The thickness
of the glass does not vary from the mean by more than
0.10 inch. Assume the thickness of the glass has a
uniform distribution.
a. Provide the density for this uniform distribution.
b. If the glass has a thickness that is more than 0.05
inch below the mean it must be discarded for safety
considerations. Determine the probability that a
randomly selected automobile glass is discarded due
to safety considerations.
c. If the glass is more than 0.075 above the mean
it will create installation problems and must be
discarded. Calculate the probability that a randomly
selected automobile glass will be rejected due to
installation concerns.
d. Given that a randomly selected automobile glass is
not rejected for safety considerations determine the
probability that it will be rejected for installation
concerns.
667. A traffic control camera at a busy intersection records
on average 5 traffic violations per hour. Assume that the
random variable number of recorded traffic violations
follow a Poisson distribution.
a. What is the probability that the next recorded
violation will occur within 5 minutes
www.downloadslide.comslide 277:
Introduction to Continuous Probability Distributions
1 users can expect to lose as much as 12 pounds in
four weeks Discuss.
674. Midwest Fan Manufacturing Inc. was established
in 1986 as a manufacturer and distributor of quality
ventilation equipment. Midwest Fan’s products include
the AXC range hood exhaust fans. The file titled Fan
Life contains the length of life of 125 randomly chosen
AXC fans that were used in an accelerated lifetesting
experiment.
a. Produce a relative frequency histogram for the
data. Does it seem plausible the data came from a
population that has an exponential distribution
b. Calculate the mean and standard deviation of the
fans’ length of life.
c. Calculate the median length of life of the fans.
d. Determine the probability that a randomly chosen
fan will have a life of more than 25000 hours.
675. Team Marketing Report TMR produces the Fan Cost
Index™ FCI survey now in its 16th year which tracks
the cost of attendance for a family of four at National
Football League NFL games. The FCI includes four
averageprice tickets four small soft drinks two small
beers four hot dogs two game programs parking and
two adultsize caps. The league’s average FCI in 2008
was 396.36. The file titled NFL Price is a sample of
175 randomly chosen fans’ FCIs.
a. Produce a relative frequency histogram for these
data. Does it seem plausible the data were sampled
from a population that was normally distributed
b. Calculate the mean and standard deviation of the
league’s FCI.
c. Calculate the 90th percentile of the league’s fans’
FCI.
d. The San Francisco 49ers had an FCI of 376.71.
Determine the percentile of the FCI of a randomly
chosen family whose FCI is the same as that of the
49ers’ average FCI.
676. The FutureVision Cable TV Company recently
surveyed its customers. A total of 548 responses were
received. Among other things the respondents were
asked to indicate their household income. The data from
the survey are found in a file named FutureVision.
a. Develop a frequency histogram for the income
variable. Does it appear from the graph that income
is approximately normally distributed Discuss.
b. Compute the mean and standard deviation for the
income variable.
c. Referring to parts a and b and assuming that
income is normally distributed and the sample
mean and standard deviation are good substitutes
for the population values what is the probability
that a FutureVision customer will have an income
exceeding 40000
d. Suppose that FutureVision managers are thinking
about offering a monthly discount to customers
who have a household income below a certain level.
Computer Database Exercises
672. The Cozine Corporation runs the landfill operation
outside Little Rock Arkansas. Each day each of the
company’s trucks makes several trips from the city to
the landfill. On each entry the truck is weighed. The
data file Cozine contains a sample of 200 truck weights.
Determine the mean and standard deviation for the
garbage truck weights. Assuming that these sample
values are representative of the population of all Cozine
garbage trucks and assuming that the distribution is
normally distributed
a. Determine the probability that a truck will arrive at
the landfill weighing in excess of 46000 pounds.
b. Compare the probability in part a to the proportion
of trucks in the sample that weighed more than
46000 pounds. What does this imply to you
c. Suppose the managers are concerned that trucks are
returning to the landfill before they are fully loaded.
If they have set a minimum weight of 38000
pounds before the truck returns to the landfill what
is the probability that a truck will fail to meet the
minimum standard
673. The Hydronics Company is in the business of
developing health supplements. Recently the
company’s research and development department
came up with two weightloss products that included
products produced by Hydronics. To determine
whether these products are effective the company
has conducted a test. A total of 300 people who were
30 pounds or more overweight were recruited to
participate in the study. Of these 100 people were
given a placebo supplement 100 people were given
product 1 and 100 people were given product 2.
As might be expected some people dropped out
of the study before the fourweek study period
was completed. The weight loss or gain for each
individual is listed in the data file called Hydronics.
Note positive values indicate that the individual
actually gained weight during the study period.
a. Develop a frequency histogram for the weight
loss or gain for those people on product 1.
Does it appear from this graph that weight loss is
approximately normally distributed
b. Referring to part a assuming that a normal
distribution does apply compute the mean and
standard deviation weight loss for the product
1 subjects.
c. Referring to parts a and b assume that the weight
change distribution for product 1 users is normally
distributed and that the sample mean and standard
deviation are used to directly represent the
population mean and standard deviation. What is
the probability that a plan 1 user will lose over 12
pounds in a fourweek period
d. Referring to your answer in part c would it be
appropriate for the company to claim that plan
www.downloadslide.comslide 278:
Introduction to Continuous Probability Distributions
a. Produce a relative frequency histogram for these
data. Does it seem plausible the data came from a
population that has a uniform distribution
b. Provide the density f x for this uniform distribution.
c. A billiard retailer Sticks Stones Billiard Supply
is going to recover the pool tables in the local
college pool hall which has eight tables. It takes
approximately 3.8 yards per table. If Championship
ships a randomly chosen halfbolt determine the
probability that it will contain enough cloth to
recover the eight tables.
If the management wants to grant discounts to no
more than 7 of the customers what income level
should be used for the cutoff
677. Championship Billiards owned by D R Industries
in Lincolnwood Illinois provides some of the finest
billiard fabrics cushion rubber and component parts in
the industry. It sells billiard cloth in bolts and halfbolts.
A halfbolt of billiard cloth has an average length of
35 yards with widths of either 62 or 66 inches. The file
titled Half Bolts contains the lengths of 120 randomly
selected halfbolts.
Case 1
State Entitlement Programs
Franklin Joiner director of health education and welfare had just
left a meeting with the state’s newly elected governor and several
of the other recently appointed department heads. One of the gov
ernor’s campaign promises was to try to halt the rising cost of a
certain state entitlement program. In several speeches the gover
nor indicated the state of Idaho should allocate funds only to those
individuals ranked in the bottom 10 of the state’s income dis
tribution. Now the governor wants to know how much one could
earn before being disqualified from the program and he also wants
to know the range of incomes for the middle 95 of the state’s
income distribution.
Frank had mentioned in the meeting that he thought incomes
in the state could be approximated by a normal distribution and
that mean per capita income was about 33000 with a standard
deviation of nearly 9000. The governor was expecting a memo in
his office by 3:00 p.m. that afternoon with answers to his questions.
Required Tasks:
1. Assuming that incomes can be approximated using a normal
distribution with the specified mean and standard deviation
calculate the income that cut off the bottom 10 of incomes.
2. Assuming that incomes can be approximated using a normal
distribution with the specified mean and standard deviation
calculate the middle 95 of incomes. Hint: This requires
calculating two values.
3. Write a short memo describing your results and how they
were obtained. Your memo should clearly state the income
that would disqualify people from the program as well as
the range of incomes in the middle 95 of the state’s income
distribution.
Case 2
Credit Data Inc.
Credit Data Inc. has been monitoring the amount of time its bill
collectors spend on calls that produce contacts with consumers.
Management is interested in the distribution of time a collector
spends on each call in which he or she initiates contact informs
a consumer about an outstanding debt discusses a payment plan
and receives payments by phone. Credit Data is mostly interested
in how quickly a collector can initiate and end a conversation to
move on to the next call. For employees of Credit Data time is
money in the sense that one account may require one call and 2
minutes to collect whereas another account may take five calls
and 10 minutes per call to collect. The company has discovered
that the time collectors spend talking to consumers about accounts
is approximated by a normal distribution with a mean of 8 min
utes and a standard deviation of 2.5 minutes. The managers believe
that the mean is too high and should be reduced by more efficient
phone call methods. Specifically they wish to have no more than
10 of all calls require more than 10.5 minutes.
Required Tasks:
1. Assuming that training can affect the average time but not the
standard deviation the managers are interested in knowing to
what level the mean call time needs to be reduced in order to
meet the 10 requirement.
2. Assuming that the standard deviation can be affected by
training but the mean time will remain at 8 minutes to what
level must the standard deviation be reduced in order to meet
the 10 requirement
3. If nothing is done what percent of all calls can be expected
to require more than 10.5 minutes
www.downloadslide.comslide 279:
Introduction to Continuous Probability Distributions
the equipment’s ability to detect oil. The enhancement requires
800 capacitors which must operate within ±0.50 microns from the
specified standard of 12 microns.
The problem is that the supplier can provide capacitors that
operate according to a normal distribution with a mean of 12
microns and a standard deviation of 1 micron. Thus Chad knows
that not all capacitors will meet the specifications required by the
new piece of exploration equipment. This will mean that to have at
least 800 usable capacitors American Oil will have to order more
than 800 from the supplier. However these items are very expen
sive so he wants to order as few as possible to meet their needs.
At the meeting the group agreed that they wanted a 98 chance
that any order of capacitors would contain the sufficient number
of usable items. If the project is to remain on schedule Chad must
place the order by tomorrow. He wants the new equipment ready to
go by the time he leaves for an exploration trip in Australia. As he
reclined in his seat sipping a cool lemonade he wondered whether
a basic statistical technique could be used to help determine how
many capacitors to order.
American Oil Company
Chad Williams field geologist for the American Oil Company
settled into his firstclass seat on the SunAir flight between Los
Angeles and Oakland California. Earlier that afternoon he had
attended a meeting with the design engineering group at the Los
Angeles New Product Division. He was now on his way to the
home office in Oakland. He was looking forward to the onehour
flight because it would give him a chance to reflect on a prob
lem that surfaced during the meeting. It would also give him a
chance to think about the exciting opportunities that lay ahead
in Australia.
Chad works with a small group of highly trained people at
American Oil who literally walk the earth looking for new sources
of oil. They make use of the latest in electronic equipment to take
a wide range of measurements from many thousands of feet below
the earth’s surface. It is one of these electronic machines that is the
source of Chad’s current problem. Engineers in Los Angeles have
designed a sophisticated enhancement that will greatly improve
Case 3
www.downloadslide.comslide 280:
Introduction to Continuous Probability Distributions
33. a. 0.75
b. Q
1
4.250.0625 8 Q
2
4.500.0625 12
Q
3
4.750.0625 16
c. 14.43
d. 0.92
35. a. 0.9179
b. 0.0498
c. 0.0323
d. 0.9502
37. a. 0.3935
b. 0.2865
39. a. 0.7143
b. 0.1429
c. 0.0204
41. a. 0.4084 yes
b. 40840
43. a. 0.2939
b. 0.4579
c. 0.1455
45. a. 0.0183
b. 0.3679
47. a. 0.0498
b. 0.4493
c. approximately l 0.08917
49. a. positively skewed
b. Descriptive Statistics: ATM FEES
Variable Mean StDev
ATM FEES 2.907 2.555
c. 1  0.6433 0.3567
55. a. 0.1353
b. 0.1353
57. a. 0.486583
b. 0.452934
c. 0.18888
59. 0.5507
61. Machine 1: 0.4236
Machine 2: 0.4772
63. 0.142778
65. Need an additional 11500  400 11100 parking spaces.
Px 6 60 0.7686 shou
67. a. 0.3406
b. 0.5580
71. a. 0.3085
b. 0.3707
c. 93.36 minutes.
73. a. approximately normally distributed
b. Mean 2.453 Standard deviation 4.778
c. 0.0012
d. No
1. a.
225  200
20
25
20
1.25
b.
190  200
20
10
20
0.50
c.
240  200
20
40
20
2.00
3. a. 0.4901
b. 0.6826
c. 0.0279
5. a. 0.4750
b. 0.05
c. 0.0904
d. 0.97585
e. 0.8513
7. a. 0.9270
b. 0.6678
c. 0.9260
d. 0.8413
e. 0.3707
9. a. x 1.290.50 + 5.5 6.145
b. m 6.145  1.650.50 5.32
11. a. 0.0027
b. 0.2033
c. 0.1085
13. a. 0.0668
b. 0.228
c. 0.7745
15. a. 0.3446
b. 0.673
c. 51.30
d. 0.9732
17. a. 0.1762
b. 0.3446
c. 0.4401
d. 0.0548
19. The mean and standard deviation of the random variable are
15000 and 1250 respectively.
a. 0.0548
b. 0.0228
c. m 15912 approximately
21. about 3367.35
23. a. 0.1949
b. 0.9544
c. Mean Median symmetric distribution
25. Px 6 1.0 0.5000  0.4761 0.0239
27. a. P0.747 … x … 0.753 0.6915  0.1587 0.5328
b. s
0.753  0.75
2.33
0.001
31. a. skewed right
b. approximate normal distribution
c. 0.1230
d. 2.034
Answers to Selected OddNumbered Problems
This section contains summary answers to most of the oddnumbered problems in the text. The Student Solutions Manual contains fully developed
solutions to all oddnumbered problems and shows clearly how each answer is determined.
www.downloadslide.comslide 281:
Introduction to Continuous Probability Distributions
77. a. Uniform distribution. Sampling error could account for
differences in this sample.
b. fx
1
b  a
1
35  24.8
0.098
c. 0.451
75. a. The histogram seems to be “bell shaped.”
b. Mean 396.36 Standard Deviation 112.41
c. The 90th percentile is 540.419.
d. 376.71 is the 43rd percentile.
Albright Christian S. Wayne L. Winston and Christopher
Zappe Data Analysis for Managers with Microsoft Excel
Pacific Grove CA: Duxbury 2003.
DeVeaux Richard D. Paul F. Velleman and David E. Bock
Stats Data and Models 3rd ed. New York: AddisonWesley
2012.
Hogg R. V. and Elliot A. Tanis Probability and Statistical
Inference 8th ed. Upper Saddle River NJ: Prentice Hall
2010.
Larsen Richard J. and Morris L. Marx An Introduction to
Mathematical Statistics and Its Applications 5th ed. Upper
Saddle River NJ: Prentice Hall 2012.
Microsoft Excel 2010 Redmond WA: Microsoft Corp. 2010.
Siegel Andrew F. Practical Business Statistics 5th ed. Burr
Ridge IL: Irwin 2002.
References
Normal Distribution The normal distribution is a bellshaped
distribution with the following properties:
1. It is unimodal that is the normal distribution peaks at a
single value.
2. It is symmetrical this means that the two areas under the
curve between the mean and any two points equidistant
on either side of the mean are identical. One side of the
distribution is the mirror image of the other side.
3. The mean median and mode are equal.
4. The normal approaches the horizontal axis on either side
of the mean toward plus and minus infinity 1∞2. In more
formal terms the normal distribution is asymptotic to the
x axis.
5. The amount of variation in the random variable deter
mines the height and spread of the normal distribution.
Standard Normal Distribution A normal distribution that has
a mean 0.0 and a standard deviation 1.0. The horizontal
axis is scaled in zvalues that measure the number of stand
ard deviations a point is from the mean. Values above the
mean have positive zvalues. Values below the mean have
negative zvalues.
Glossary
www.downloadslide.comslide 282:
1 Sampling Error: What It Is and
Why It Happens
2 Sampling Distribution of the
Mean
3 Sampling Distribution of a
Proportion
Outcome 2. Determine the mean and standard deviation for
the sampling distribution of the sample mean x.
Outcome 3. Understand the importance of the Central Limit
Theorem.
Why you need to know
The Jamaica Director of Tourism has recently conducted a study that shows that the mean daily expenditure for
adult visitors to the country is 318.69. The mean value is based on a statistical sample of 780 adult visitors to
Jamaica. The 318.69 is a statistic not a parameter because it is based on a sample rather than an entire popu
lation. If you were this official you might have several questions:
Is the actual population mean equal to 318.69
If the population mean is not 318.69 how close is 318.69 to the true population mean
Is a sample of 780 taken from a population of almost 2 million annual visitors to the country sufficient to
provide a “good” estimate of the population mean
A furniture manufacturer that makes madetoassemble furniture kits
selects a random sample of kits boxed and ready for shipment to custom
ers. These kits are unboxed and inspected to see whether what is in the box
matches exactly what is supposed to be in the box. This past week 150 kits
were sampled and 15 had one or more discrepancies. This is a 10 defect
rate. Should the quality engineer conclude that exactly 10 of the 6900 fur
niture kits made since the first of the year reached the customer with one or
more order discrepancies Is the actual percentage higher or lower than 10
and if so by how much Should the quality engineer request that more furni
ture kits be sampled
The questions facing the tourism director and the furniture quality engi
neer are common to those faced by people in business everywhere. You will
almost assuredly find yourself in a similar situation many times in the future. To
help answer these questions you need to have an understanding of sampling
distributions. Whenever decisions are based on samples rather than an entire
Outcome 4. Determine the mean and standard deviation for
the sampling distribution of the sample proportion p.
Quick Prep Links
Review the discussion of random sampling.
Review the steps for computing means and
standard deviations.
Make sure you are familiar with the normal
distribution and how to compute standardized
zvalues.
Review the concepts associated with
finding probabilities with a standard normal
distribution.
Introduction to Sampling
Distributions
Outcome 1. Understand the concept of sampling error.
gdvcom/Shutterstock
From Chapter 7 of Business Statistics A DecisionMaking Approach Ninth Edition. David F. Groebner
Patrick W. Shannon and Phillip C. Fry. Copyright © 2014 by Pearson Education Inc. All rights reserved.
www.downloadslide.comslide 283:
Introduction to Sampling Distributions
population questions about the sample results exist. Anytime we sample from a population there are many many
possible samples that could have been selected. Each sample will contain different items. Because of this the sample
means for each possible sample can be different or the sample percentages can be different. The sampling distribu
tion describes the distribution of possible sample outcomes. Knowing what this distribution looks like will help you
understand the specific result you obtained from the one sample you selected.
This chapter introduces you to the important concepts of sampling error and sampling distributions and dis
cusses how you can use this knowledge to help answer the questions facing the tourism director and the quality
engineer. The information presented here provides an essential building block to understanding statistical estimation
and hypothesis testing.
1 Sampling Error: What It Is
and Why It Happens
You will encounter many situations in business in which a sample will be taken from a popu
lation and you will be required to analyze the sample data. You have learned about several
different statistical sampling techniques including random sampling. The objective of ran
dom sampling is to gather data that accurately represent a population. Then when analysis
is performed on the sample data the results will be as though we had worked with all the
population data.
Calculating Sampling Error
Regardless of how careful we are in using random sampling methods the sample may not be
a perfect representation of the population. For example a statistic such as x might be com
puted for sample data. Unless the sample is a perfect replication of the population the statistic
will likely not equal the parameter m. In this case the difference between the sample mean
and the population mean is called sampling error. In the case in which we are interested in
the mean value the sampling error is computed using Equation 1.
Sampling Error
The difference between a measure
computed from a sample a statistic and the
corresponding measure computed from the
population a parameter.
Sampling Error of the Sample Mean
Sampling error x
1
where:
x 5
5
Sample mean
Population mean
BUSINESS APPLICATION SAMPLING ERROR
HUMMEL DEVELOPMENT CORPORATION The Hummel Development Corporation
has built 12 office complexes. Table 1 shows a list of the 12 projects and the total square
footage of each project.
Because these 12 projects are all the office complexes the company has worked on the
squarefeet area for all 12 projects shown in Table 1 is a population. Equation 2 is used to
compute the mean square feet in the population of projects.
Chapter Outcome 1.
Population Mean
5
x
N
2
where:
5
5
Population mean
Values in the population x
N N 5 Population size
TABLE 1  Square Feet for
Office Complex Projects
Complex Square Feet
1 114560
2 202300
3 78600
4 156700
5 134600
6 88200
7 177300
8 155300
9 214200
10 303800
11 125200
12 156900
www.downloadslide.comslide 284:
Introduction to Sampling Distributions
The mean square feet for the 12 office complexes is
114 560 125 200 156 900
12
15
202300
...
8 8 972 square feet
The average square footage of the offices built by the firm is 158972 square feet. This value
is a parameter. No matter how many times we compute the value assuming no arithmetic
mistakes we will get the same value for the population mean.
Hummel is a finalist to be the developer of a new office building in Madison Wisconsin.
The client who will hire the firm plans to select a simple random sample of n 5 projects
from those the finalists have completed. The client plans to travel to these office buildings to
see the quality of the construction and to interview owners and occupants. You may want to
review the material on simple random samples.
Referring to the office complex data in Table 1 suppose the client randomly selects the
following five Hummel projects from the population:
Complex Square Feet
5 134600
4 156700
1 114560
8 155300
9 214200
Key in the selection process is the finalists’ past performance on large projects so the client
might be interested in the mean size of the office buildings that the Hummel Development
Company has built. Equation 3 is used to compute the sample mean.
Parameter
A measure computed from the entire population.
As long as the population does not change the
value of the parameter will not change.
Simple Random Sample
A sample selected in such a manner that each
possible sample of a given size has an equal
chance of being selected.
Sample Mean
x
x
n
5
3
where:
x
x
5
5
Sample mean
Sample values selected from t the population
Sample size n 5
The sample mean is
x
134 600 156 700 114 560 155 300 214 200
5
7 775 360
5
155 072
The average number of square feet in the random sample of five Hummel office buildings
selected by the client is 155072. This value is a statistic based on the sample.
Recall the mean for the population:
m 158972 square feet
The sample mean is
x 5155 072 square feet
As you can see the sample mean does not equal the population mean. This difference is called
the sampling error. Using Equation 1 we compute the sampling error as follows.
Sampling error
s
x
155 072 158 972 3 900 q quare feet
The sample mean for the random sample of n 5 office buildings is 3900 square feet less
than the population mean. Regardless of how carefully you construct your sampling plan you
www.downloadslide.comslide 285:
Introduction to Sampling Distributions
can expect to see sampling error. A random sample will almost never be a perfect representa
tion of its population. The sample value and the population value will most likely be different.
Suppose the client who selected the random sample throws these five projects back into
the stack and selects a second random sample of five as follows:
Complex Square Feet
9 214200
6 88200
5 134600
12 156900
10 303800
The mean for this sample is
x
214 200 88 200 134 600 156 900 303 800
5
8 9 97 700
5
179 540
square feet
This time the sample mean is larger than the population mean. This time the sampling error is
x 179 540 158 972
20 568
square feet
This illustrates some useful fundamental concepts:
● The size of the sampling error depends on which sample is selected.
● The sampling error may be positive or negative.
● There is potentially a different x for each possible sample.
If the client wanted to use these sample means to estimate the population mean in one case they
would be 3900 square feet too small and in the other they would be 20568 square feet too large.
EXAMPLE 1 COMPUTING THE SAMPLING ERROR
HighDefinition Televisions The website for a major seller of
consumer electronics has 10 different brands of HD televisions avail
able. The stated prices for the 46inch size for the 10 brands are listed
as follows:
479 569 599 649 649 699 699 749 799 799
Suppose a competitor who is monitoring this company has randomly sampled n 4 HD
brands and recorded the prices from the population of N 10. The selected HD prices
were
569 649 799 799
The sampling error can be computed using the following steps:
Step 1 Determine the population mean using Equation 2.
x
N
479 569 599 799 799
10
6 690
10
66
...
9 9
Step 2 Compute the sample mean using Equation 3.
x
x
n
569 649 799 799
4
2 816
4
704
Step 3 Compute the sampling error using Equation 1.
x 704 669 35
nexusseven/Fotolia
www.downloadslide.comslide 286:
Introduction to Sampling Distributions
This sample of four has a sampling error of 35. The sample of TV prices has a
slightly larger mean price than the mean for the population.
END EXAMPLE
TRY PROBLEM 1
The Role of Sample Size in Sampling Error
BUSINESS APPLICATION SAMPLING ERROR
HUMMEL DEVELOPMENT CORPORATION CONTINUED We selected random
samples of 5 office complexes from the 12 projects Hummel Development Corporation has
built. We then computed the resulting sampling error. There are actually 792 possible samples
of size 5 taken from 12 projects. This value is found using the counting rule for combinations.
1
In actual situations only one sample is selected and the decision maker uses the sample
measure to estimate the population measure. A “small” sampling error may be acceptable. How
ever if the sampling error is too “large” conclusions about the population could be misleading.
We can look at the extremes on either end to evaluate the potential for extreme sampling
error. The population of square feet for the 12 projects is
Complex Square Feet Complex Square Feet
1 114560 7 177300
2 202300 8 155300
3 78600 9 214200
4 156700 10 303800
5 134600 11 125200
6 88200 12 156900
Suppose by chance the developers ended up with the fve smallest offce complexes in their
sample. These would be
Complex Square Feet
3 78600
6 88200
1 114560
11 125200
5 134600
The mean of this sample is
x 5108 232 square feet
Of all the possible random samples of fve this one provides the smallest sample mean. The
sampling error is
x 108 232 158 972 50 740 squarefeet
On the other extreme suppose the sample contained the five largest office complexes as
follows:
Complex Square Feet
10 303800
9 214200
2 202300
7 177300
12 156900
1
The number of combinations of items from a sample of is
n
xn x
2
www.downloadslide.comslide 287:
Introduction to Sampling Distributions
The mean for this sample is x 210900. This is the largest possible sample mean from all
the possible samples of fve complexes. The sampling error in this case would be
x 210 900 158 972 51 928 squarefeet
The potential for extreme sampling error ranges from
50740 to +51928 square feet
The remaining possible random samples of fve will provide sampling errors between these
limits.
What happens if the size of the sample selected is larger or smaller Suppose the client
scales back his sample size to n 3 office complexes. Table 2 shows the extremes.
By reducing the sample size from five to three the range of potential sampling error has
increased from
150740 to +51928 square feet2
to
165185.33 to +81128 square feet2
This illustrates that the potential for extreme sampling error is greater when smallersized
samples are used. Likewise larger sample sizes will reduce the range of potential sampling
error.
Although larger sample sizes reduce the potential for extreme sampling error there is no
guarantee that a larger sample size will always give a smaller sampling error. For example
Table 3 shows two further applications of the office complex data. As illustrated this random
sample of three has a sampling error of  2672 square feet whereas the larger random sample
of size five has a sampling error of 16540 square feet. In this case the smaller sample was
TABLE 2  Hummel Office Building Example for n 5 3 Extreme Samples
Smallest Office Buildings Largest Office Buildings
Complex Square Feet Complex Square Feet
3 78600 10 303800
6 88200 9 214200
1 114560 2 202300
x 93786.67 sq. feet x 240100 sq. feet
Sampling Error: Sampling Error:
93786.67  15897265185.33 square feet 240100  158972 81128 square feet
TABLE 3  Hummel Office Building Example with Different Sample Sizes
n 5 n 3
Complex Square Feet Complex Square Feet
4 156700 12 156900
1 114560 8 155300
7 177300 4 156700
11 125200
10 303800
x 175512 sq. feet x 156300 sq. feet
Sampling Error: Sampling Error:
175512  158972 16540 square feet 156300  158972 2672 square feet
www.downloadslide.comslide 288:
Introduction to Sampling Distributions
“better” than the larger sample. However in Section 2 you will learn that on average the sam
pling error produced by large samples will be less than the sampling error from small samples.
Skill Development
71. A population has a mean of 125. If a random sample
of 8 items from the population results in the following
sampled values what is the sampling error for the
sample
103 123 99 107 121 100 100 99
72. The following data are the 16 values in a population:
10 5192010 8 10 2
14 18 7 814 2 310
a. Compute the population mean.
b. Suppose a simple random sample of 5 values
from the population is selected with the following
results:
5 10202 3
Compute the mean of this sample.
c. Based on the results for parts a and b compute the
sampling error for the sample mean.
73. The following population is provided:
17 15 812 9 7 911
12 14 16 9 5101413
12 12 11 914 81412
Further a simple random sample from this population
gives the following values:
12 9 5101411
Compute the sampling error for the sample mean in
this situation.
74. Consider the following population:
18 26 32 17 34 17 17 22
29 24 24 35 13 29 38
The following sample was drawn from this population:
35 18 24 17 24 32 17 29
a. Determine the sampling error for the sample
mean.
b. Determine the largest possible sampling error for
this sample of n 8.
75. Assume that the following represent a population of
N 24 values:
10 14 32 9 34 19 31 24
33 11 14 30 6 27 33 32
28 30 10 31 19 13 6 35
a. If a random sample of n 10 items includes the
following values compute the sampling error for
the sample mean:
32 19 6 11 10
19 28 9 13 33
b. For a sample of size n 6 compute the range
for the possible sampling error. Hint: Find the
sampling error for the 6 smallest sample values and
the 6 largest sample values.
c. For a sample of size n 12 compute the range
for the possible sampling error. How does sample
size affect the potential for extreme sampling
error
76. Assume that the following represent a population of
N 16 values.
25 12 21 13 19 17 15 18
23 16 18 15 22 14 23 17
a. Compute the population mean.
b. If a random sample of n 9 includes the following
values
12 18 13 17 23 14 16 25 15
compute the sample mean and calculate the
sampling error for this sample.
c. Determine the range of extreme sampling error for
a sample of size n 4. Hint: Calculate the lowest
possible sample mean and highest possible sample
mean.
77. Consider the following population:
369
a. Calculate the population mean.
b. Select with replacement and list each possible
sample of size 2. Also calculate the sample mean
for each sample.
71: Exercises
MyStatLab
www.downloadslide.comslide 289:
Introduction to Sampling Distributions
c. Calculate the sampling error associated with each
sample mean.
d. Assuming that each sample is equally likely
produce the distribution of the sampling errors.
Business Applications
78. Hillman Management Services manages apartment
complexes in Tulsa Oklahoma. They currently have
30 units available for rent. The monthly rental prices
in dollars for this population of 30 units are
455 690 450 495 550 780 800 395 500 405
675 550 490 495 700 995 650 550 400 750
600 780 650 905 415 600 600 780 575 750
a. What is the range of possible sampling error if a
random sample of size n 6 is selected from the
population
b. What is the range of possible sampling error if a
random sample of size n 10 is selected Compare
your answers to parts a and b and explain why the
difference exists.
79. A previous report from the Centers for Disease Control
and Prevention CDC indicates that smokers on
average miss 6.16 days of work per year due to sickness
including smokingrelated acute and chronic conditions.
Nonsmokers miss an average of 3.86 days of work per
year. If two years later the CDC believes that the average
days of work missed by smokers has not changed it could
confirm this by sampling. Consider the following sample:
13445 1289 11 1 5
69 14635 10 7 0 14
6 150253 10 8 6 7
00 15 146221 4 15
10 12 3 0 14 10 0 1 9 14
Determine the sampling error of this sample assuming
that the CDC supposition is correct.
710. An Internet service provider states that the average number
of hours its customers are online each day is 3.75. Suppose
a random sample of 14 of the company’s customers is
selected and the average number of hours that they are
online each day is measured. The sample results are
3.11 1.97 3.52 4.56 7.19 3.89 7.71
2.12 4.68 6.78 5.02 4.28 3.23 1.29
Based on the sample of 14 customers how much
sampling error exists Would you expect the sampling
error to increase or decrease if the sample size was
increased to 40
711. The Anasazi Real Estate Company has 20 listings
for homes in Santa Fe New Mexico. The number of
days each house has been on the market without selling
is as follows:
26 45 16 77 33 50 19 23 55 107
88 15 7193060806631 17
a. Considering these 20 values to be the population of
interest what is the mean of the population
b. The company is making a sales brochure and wishes
to feature 5 homes selected at random from the list.
The number of days the 5 sampled homes have been
on the market is
77 60 15 31 23
If these 5 houses were used to estimate the mean for
all 20 what would the sampling error be
c. What is the range of possible sampling error
if 5 homes are selected at random from the
population
712. The administrator at Saint Frances Hospital is
concerned about the amount of overtime the nursing
staff is incurring and wonders whether so much
overtime is really necessary. The hospital employs 60
nurses. Following is the number of hours of overtime
reported by each nurse last week. These data are the
population of interest.
Nurse Overtime Nurse Overtime Nurse Overtime
1 2 214413
2 1 222423
3 7 233432
4 0 245441
5 4 255453
6 2 266463
76 27 2 47 3
8 4 282483
9 2 297494
10 5304506
11 5314510
12 4323523
13 5333534
14 0344546
15 6355550
16 0365563
17 2370573
18 4380587
19 2394595
20 5403607
Using the Random Numbers Table with a starting
point in column digit 14 and row 10 select a
random sample of 6 nurses. Go down the table
from the starting point. Determine the mean hours
of overtime for these 6 nurses and calculate the
sampling error associated with this particular sample
mean.
713. Princess Cruises recently offered a 16day voyage
from Beijing to Bangkok during the time period from
May to August. The announced price excluding
airfare for a room with an ocean view or a balcony
was listed as 3475. Cruise fares usually are quite
variable due to discounting by the cruise line and
travel agents. A sample of 20 passengers who
www.downloadslide.comslide 290:
Introduction to Sampling Distributions
c. What is the range of possible sampling error if
a random sample size of 7 computers is taken to
estimate the mean scan time for all 25 machines
Computer Database Exercises
716. USA Today reports salaries for National Football
League NFL teams. The file Jaguars contains the
salaries for the 2011 Jacksonville Jaguars.
a. Calculate the average total salary for the
Jacksonville Jaguars for 2011.
b. Calculate the smallest sample mean for total salary
and the largest sample mean for total salary using a
sample size of 10. Calculate the sampling error for
each sample mean.
c. Repeat the calculations in part b for samples of size
5 and 2.
d. What effect does a change in the sample size appear
to have on the dispersion of the sampling errors
717. The file titled Clothing contains the monthly retail
sales millions of U.S. women’s clothing stores for
70 months. A sample taken from this population to
estimate the average sales in this time period follows:
2942 2574 2760 2939 2642 2905 2568
2677 2572 3119 2697 2884 2632 2742
2671 2884 2946 2825 2987 2729 2676
2846 3112 2924 2676
a. Calculate the population mean.
b. Calculate the sample mean.
c. How much sampling error is present in this
sample
d. Determine the range of possible sampling error if
25 sales figures are sampled at random from this
population.
718. The DowJones Industrial Average DJIA Index is
a wellknown stock index. The index was originally
developed in 1884 and has been in place ever since as a
gauge of how the U.S. stock market is performing. The
file Dow Jones contains date open high low close
and volume for the DJIA for almost eight years of the
trading days.
a. Assuming that the data in the file Dow Jones
constitute the population of interest what is the
population mean closing value for the DJIA
b. Using Excel select a random sample of 10 days’
closing values make certain not to include
duplicate days and calculate the sample mean and
the sampling error for the sample.
c. Repeat part b with a sample size of 50 days’ closing
values.
d. Repeat part b with a sample size of 100 days’
closing values.
e. Write a short statement describing your results.
Were they as expected Explain.
719. Welco Lumber Company is based in Shelton
Washington and is a privately held company
that makes cedar siding cedar lumber and
purchased this cruise paid the following amounts in
dollars:
3559 3005 3389 3505 3605 3545 3529 3709 3229 3419
3439 3375 3349 3559 3419 3569 3559 3575 3449 3119
a. Calculate the sample mean cruise fare.
b. Determine the sampling error for this sample.
c. Would the results obtained in part b indicate that the
average cruise fare during this period for this cruise
is different from the listed price Explain your
answer from a statistical point of view.
714. An investment advisor has worked with 24 clients
for the past five years. Following are the percentage
rates of average fiveyear returns that these 24 clients
experienced over this time frame on their investments:
11.2 11.2 15.9 2.7 4.6 7.6 15.6 1.3 3.3 4.8 12.8 14.9
10.1 10.9 4.9 2.1 12.5 3.7 7.6 4.9 10.2 0.4 9.6 0.5
This investment advisor plans to introduce a new
investment program to a sample of his customers
this year. Because this is experimental he plans to
randomly select 5 of the customers to be part of the
program. However he would like those selected to
have a mean return rate close to the population mean
for the 24 clients. Suppose the following 5 values
represent the average fiveyear annual return for the
clients that were selected in the random sample:
11.2 2.1 12.5 1.3 3.3
Calculate the sampling error associated with the mean
of this random sample. What would you tell this
advisor regarding the sample he has selected
715. A computer lab at a small college has 25 computers.
Twice during the day a full scan for viruses is
performed on each computer. Because of differences in
the configuration of the computers the times required
to complete the scan are different for each machine.
Records for the scans are kept and indicate that the
time in seconds required to perform the scan for each
machine is as shown here.
Time in Seconds to Complete Scan
1500 1347 1552 1453 1371
1362 1447 1362 1216 1378
1647 1093 1350 1834 1480
1522 1410 1446 1291 1601
1365 1575 1134 1532 1534
a. What is the mean time required to scan all
25 computers
b. Suppose a random sample of 5 computers is taken
and the scan times for each are as follows: 1534
1447 1371 1410 and 1834. If these 5 randomly
sampled computers are used to estimate the mean
scan time for all 25 computers what would the
sampling error be
www.downloadslide.comslide 291:
Introduction to Sampling Distributions
cedar fencing products for sale and distribution
throughout North America. The major cost of
production is the cedar logs that are the raw
material necessary to make the finished cedar
products. Thus it is very important to the company
to get the maximum yield from each log. Of course
the dollar value to be achieved from a log depends
initially on the diameter of the log. Each log is 8
feet long when it reaches the mill. The file called
Welco contains a random sample of logs of various
diameters and the potential value of the finished
products that could be developed from the log if it
is made into fence boards.
a. Calculate the sample mean potential value for each
diameter of logs in the sample.
b. Discuss whether there is a way to determine how
much sampling error exists for a given diameter log
based on the sample. Can you determine whether
the sampling error will be positive or negative
Discuss.
720. Maher Barney and White LLC is a legal firm with
40 employees. All of the firm’s employees are
eligible to participate in the company’s 401k plan
and the firm is proud of its 100 participation rate.
The file MBW 401 contains the most recent year
end 401k account balance for each of the firm’s
40 employees.
a. Compute the population mean and population
standard deviation for the most recent yearend
401k account balances at Maher Barney and
White.
b. Suppose that an audit of the firm’s 401k plan
is being conducted and 12 randomly selected
employee account balances are to be examined. If
the following employees indicated by employee
number are randomly selected to be included in
the study what is the estimate for the most recent
yearend mean 401k account balance How much
sampling error is present in this estimate
Employee
26 831 3383017 921391811
c. Calculate the range of possible sampling error if a
random sample of 15 employees is used to estimate the
most recent yearend mean 401k account balance.
721. The Badke Foundation was set up by the Fred Badke
family following his death in 2001. Fred had been a
very successful heart surgeon and real estate investor
in San Diego and the family wanted to set up an
organization that could be used to help less fortunate
people. However one of the concepts behind the
Badke Foundation is to use the Badke money as seed
money for gathering contributions from middleclass
families. To help in the solicitation of contributions
the foundation was considering the idea of hiring a
consulting company that specialized in this activity.
Leaders of the consulting company maintained in their
presentation that the mean contribution from families
who actually contribute after receiving a specially
prepared letter would be 20.00.
Before actually hiring the company the Badke
Foundation sent out the letter and request materials
to many people in the San Diego area. They received
contributions from 166 families. The contribution
amounts are in the data file called Badke.
a. Assuming that these data reflect a random sample
of the population of contributions that would be
received compute the sampling error based on the
claim made by the consulting firm.
b. Comment on any issues you have with the
assumption that the data represent a random sample.
Does the calculation of the sampling error matter if
the sample is not a random sample Discuss.
END EXERCISES 71
2 Sampling Distribution of the Mean
Section 1 introduced the concept of sampling error. A random sample selected from a popula
tion will not perfectly match the population. Thus the sample statistic likely will not equal
the population parameter. If this difference arises because the random sample is not a perfect
representation of the population it is called sampling error.
In business applications decision makers select a single random sample from a popula
tion. They compute a sample measure and use it to make decisions about the entire population.
For example Nielsen Media Research takes a single random sample of television viewers to
determine the percentage of the population who are watching a particular program during a
particular week. Of course the sample selected is only one of many possible samples that
could have been selected from the same population. The sampling error will differ depending
on which sample is selected. If in theory you were to select all possible random samples of
a given size and compute the sample means for each one these means would vary above and
below the true population mean. If we graphed these values as a histogram the graph would
be the sampling distribution.
Sampling Distribution
The distribution of all possible values of a
statistic for a given sample size that has been
randomly selected from a population.
www.downloadslide.comslide 292:
Introduction to Sampling Distributions
In this section we introduce the basic concepts of sampling distributions. We will use an
Excel tool to select repeated samples from the same population for demonstration purposes
only.
Simulating the Sampling Distribution for x
BUSINESS APPLICATION SAMPLING DISTRIBUTIONS
AIMS INVESTMENT COMPANY Aims Investment
Company handles employee retirement funds primarily for
small companies. The file named AIMS contains data on the
number of mutual funds in each client’s portfolio. The file
contains data for all 200 Aims customers so it is considered
a population. Figure 1 shows a histogram for the population.
The mean number of mutual funds in a portfolio is 2.505
funds. The standard deviation is 1.507 funds. The graph in Figure 1 indicates that the popula
tion is spread between zero and six funds with more customers owning two funds than any
other number.
Suppose the controller at Aims plans to select a random sample of 10 accounts. In Excel
we can use the Sampling tool to generate the random sample.
2
Figure 2 shows the num
ber of mutual funds owned for a random sample of 10 clients. The sample mean of 2.1 is
also shown. To illustrate the concept of a sampling distribution we repeat this process 500
times generating 500 different random samples of 10. For each sample we compute the
sample mean. Figure 3 shows the frequency histogram for these sample means. Note that
the horizontal axis represents the xvalues. The graph in Figure 3 is not a complete sampling
distribution because it is based on only 500 samples out of the many possible samples that
could be selected. However this simulation gives us an idea of what the sampling distribu
tion looks like.
Look again at the population distribution in Figure 1 and compare it with the shape
of the frequency histogram in Figure 3. Although the population distribution is some
what skewed the distribution of sample means is taking the shape of a normal distribu
tion.
Note also the population mean number of mutual funds owned by the 200 Aims Invest
ment customers in the population is 2.505. If we average the 500 sample means in Figure 3
we get 2.41. This value is the mean of the 500 sample means. It is reasonably close to the
population mean.
2
The same thing can be achieved in Minitab by using the Sample from Columns option under the Calc 7
Probability Data command.
Number of Customers
06 5 4 3 2 1
Number of Mutual Funds
60
50
40
30
20
10
0
POPULATION OF FUNDS OWNED
FIGURE 1 
Distribution of Mutual Funds
for the Aims Investment
Company
Excel Tutorial
Excel
tutorials
Chapter Outcome 2.
Gunnar Pippel/Shutterstock
www.downloadslide.comslide 293:
Introduction to Sampling Distributions
Frequency
1.4
x
Sample Mean
100
80
60
40
20
0
1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6
DISTRIBUTION OF SAMPLE MEANS
n 10
FIGURE 3 
Aims Investment Company
Histogram of 500 Sample
Means from Sample Size
n 10
Excel 2010 Instructions:
1. Open fle: Aims.xlsx.
2. Select Data Data
Analysis.
3. Select Sampling.
4. Defne the population data
range B2:B201.
5. Select Random Number
of Samples: 10.
6. Select Output Range: D2.
7. Compute sample mean
using Excel Average
function using the range
D2:D11.
Minitab Instructions for similar results:
1. Open fle: AIMS.MTW.
2. Choose Calc Random Data Sample
From Columns.
3. In Number of rows to Sample enter the
sample size.
4. In box following From columns
enter data column: Number of Mutual
Fund Accounts.
5. In Store Samples in enter
sample’s storage column.
6. Click OK.
7. Choose Calc Calculator.
8. In Store Result in Variable
enter column to store mean.
9. Choose Mean from
Functions. Expression:
Mean Sample Column.
Repeat steps 3–10 to
form sample.
Click OK.
10.
11.
FIGURE 2 
Excel 2010 Output for the
Aims Investment Company
First Sample Size n 10
www.downloadslide.comslide 294:
Introduction to Sampling Distributions
Had we selected all possible random samples of size 10 from the population and com
puted all possible sample means the average of all the possible sample means would be equal
to the population mean. This concept is expressed as Theorem 1.
Theorem 1
For any population the average value of all possible sample means computed from all
possible random samples of a given size from the population will equal the population
mean. This is expressed as
x
5
When the average of all possible values of the sample statistic equals the corresponding parameter
no matter the value of the parameter we call that statistic an unbiased estimator of the parameter.
Also the population standard deviation is 1.507 mutual funds. This measures the variation
in the number of mutual funds between individual customers. When we compute the standard
deviation of the 500 sample means we get 0.421 which is considerably smaller than the popu
lation standard deviation. If all possible random samples of size n are selected from the popula
tion the distribution of possible sample means will have a standard deviation that is equal to the
population standard deviation divided by the square root of the sample size as Theorem 2 states.
Unbiased Estimator
A characteristic of certain statistics in which
the average of all possible values of the sample
statistic equals a parameter no matter the value
of the parameter.
Theorem 2
For any population the standard deviation of the possible sample means computed from
all possible random samples of size n is equal to the population standard deviation
divided by the square root of the sample size. This is shown as
x
n
5
Recall the population standard deviation is s 1.507. Then based on Theorem 2 had we
selected all possible random samples of size n 10 rather than only 500 samples the stand
ard deviation for the possible sample means would be
x
n
55 5
1 507
10
0 477
.
.
Our simulated value of 0.421 is fairly close to 0.477.
The standard deviation of the sampling distribution will be less than the popula
tion standard deviation. To further illustrate suppose we increased the sample size from
n 10 to n 20 and selected 500 new samples of size 20. Figure 4 shows the distribution
of the 500 different sample means.
The distribution in Figure 4 is even closer to a normal distribution than what we observed
in Figure 3. As sample size increases the distribution of sample means will become shaped
more like a normal distribution. The average sample mean for these 500 samples is 2.53 and
the standard deviation of the different sample means is 0.376. Based on Theorems 1 and 2 for
a sample size of 20 we would expect the following:
x
x
n
55 5 5 5 2 505
1 507
20
0 337 .
.
. and
Thus our simulated values are quite close to the theoretical values we would expect had we
selected all possible random samples of size 20.
Sampling from Normal Populations The previous discussion began with the population
of mutual funds shown in Figure 1. The population was not normally distributed but as we
increased the sample size the sampling distribution of possible sample means began to approach
a normal distribution. We will return to this situation shortly but what happens if the population
itself is normally distributed To help answer this question Theorem 3 can be applied.
www.downloadslide.comslide 295:
Introduction to Sampling Distributions
Frequency
1.4
Sample Means
140
120
100
80
60
40
20
0
1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6
DISTRIBUTION OF SAMPLE MEANS
n 20
x
FIGURE 4 
Aims Investment Company
Histogram of Sample Means
from Sample Size n5 20
Theorem 3
If a population is normally distributed with mean m and a standard deviation s the
sampling distribution of the sample mean x is also normally distributed with a mean
equal to the population mean 1m
x
m2 and a standard deviation equal to the popula
tion standard deviation divided by the square root of the sample size 1s
x
s1n2.
In Theorem 3 the quantity 1s
x
s1n2 is the standard deviation of the sampling dis
tribution. Another term that is given to this is the standard error of x because it is the measure
of the standard deviation of the potential sampling error.
We can again use simulation to demonstrate Theorem 3. We begin by using Excel to
generate a normally distributed population.
3
Figure 5 shows a simulated population that is
approximately normally distributed with a mean equal to 1000 and a standard deviation equal
to 200. The data range is from 250 to 1800.
Next we simulate the selection of 2000 random samples of size n 10 from the
normally distributed population and compute the sample mean for each sample. These sample
means can then be graphed as a frequency histogram as shown in Figure 6. This histogram
represents the sampling distribution. Note that it too is approximately normally distributed.
3
The same task can be performed in Minitab using the Calc. 7 Random Data command. However you will have to
generate each sample individually which will take time.
x
1600
1400
1200
1000
800
600
400
200
250 1800
0
Frequency
1000
200
FIGURE 5 
Simulated Normal Population
Distribution
www.downloadslide.comslide 296:
Introduction to Sampling Distributions
We next compute the average of the 2000 sample means and use it to approximate m
x
as follows:
x
x
2 000
2 000 178
2 000
1000
5
The mean of these sample means is approximately 1000. This is the same value as the popu
lation mean.
We also approximate the standard deviation of the sample means as follows:
x
x
x
.
2
2 000
62 10
We see the standard deviation of the sample means is 62.10. This is much smaller than the
population standard deviation which is 200. The largest sample mean was just more than
1212 and the smallest sample mean was just less than 775. Recall however that the popula
tion ranged from 250 to 1800. The variation in the sample means always will be less than the
variation for the population as a whole. Using Theorem 3 we would expect the sample means
to have a standard deviation equal to
x
n
55 5
200
10
63 25 .
Our simulated standard deviation of 62.10 is fairly close to the theoretical value of 63.25.
Suppose we again use the simulated population shown in Figure 5 with m 1 000
and s 200. We are interested in seeing what the sampling distribution will look like for
different size samples. For a sample size of 5 Theorem 3 indicates that the sampling distri
bution will be normally distributed and have a mean equal to 1000 and a standard deviation
equal to
x
55
200
5
89 44 .
If we were to take a random sample of 10 as simulated earlier Theorem 3 indicates the
sampling distribution would be normal with a mean equal to 1000 and a standard deviation
equal to
x
55
200
10
63 25 .
500 750 1250 1500
x
62.10
140
120
100
80
60
40
20
0
Frequency
x
1000
x
Sample Means
DISTRIBUTION OF SAMPLE MEANS
FIGURE 6 
Approximated Sampling
Distribution n5 10
www.downloadslide.comslide 297:
Introduction to Sampling Distributions
For a sample size of 20 the sampling distribution will be centered at m
x
1000 with a
standard deviation equal to
x
55
200
20
44 72 .
Notice as the sample size is increased the standard deviation of the sampling distribution is
reduced. This means the potential for extreme sampling error is reduced when larger sam
ple sizes are used. Figure 7 shows sampling distributions for sample sizes of 5 10 and 20.
However when the population is normally distributed the sampling distribution of x will
always be normal and centered at the population mean. Only the spread in the distribution will
change as the sample size changes.
This illustrates a very important statistical concept referred to as consistency. Earlier
we defined a statistic as unbiased if the average value of the statistic equals the parameter
to be estimated. Theorem 1 asserted that the sample mean is an unbiased estimator of the
population mean no matter the value of the parameter. However just because a statistic
is unbiased does not tell us whether the statistic will be close in value to the parameter.
But if as the sample size is increased we can expect the value of the statistic to become
closer to the parameter then we say that the statistic is a consistent estimator of the
parameter. Figure 7 illustrates that the sample mean is a consistent estimator of the popu
lation mean.
The sampling distribution is composed of all possible sample means of the same
size. Half the sample means will lie above the center of the sampling distribution and
half will lie below. The relative distance that a given sample mean is from the center can
be determined by standardizing the sampling distribution. A standardized value is deter
mined by converting the value from its original units into a zvalue. A zvalue measures
the number of standard deviations a value is from the mean. This same concept can be
used when working with a sampling distribution. Equation 4 shows how the zvalues are
computed.
Consistent Estimator
An unbiased estimator is said to be a consistent
estimator if the difference between the estimator
and the parameter tends to become smaller as
the sample size becomes larger.
FIGURE 7 
Theorem 3 Examples
Population
s 200
Sample Size
n 5
200
5
89.44 s
x
s
x
Sample Size
n 10
200
10
63.25
s
x
Sample Size
n 20
200
20
44.72
x
x
x
x
www.downloadslide.comslide 298:
Introduction to Sampling Distributions
Note if the sample being selected is large relative to the size of the population greater
than 5 of the population size of if the sampling is being done without replacement we need
to modify how we compute the standard deviation of the sampling distribution and zvalue
using what is known as the finite population correction factor as shown in Equation 5.
zValue for Sampling Distribution of x
z
x
n
4
where:
x 5
5
5
Sample mean
Population mean
Population s standard deviation
Sample size n 5
zValue Adjusted for the Finite Population Correction Factor
z
x
n
Nn
N 1
5
where:
N
n
Nn
N
Population size
Sample size
Finite
1
population correction factor
The finite population correction factor is used to calculate the standard deviation of the sam
pling distribution when the sampling is performed without replacement or when the sample
size is greater than 5 of the population size.
EXAMPLE 2 FINDING THE PROBABILITY THAT X IS IN A GIVEN RANGE
Scribner Products Scribner Products manufactures floor
ing materials for the residential and commercial construction
industries. One item they make is a mosaic tile for bathrooms
and showers. When the production process is operating accord
ing to specifications the diagonal dimension of a tile used for
decorative purposes is normally distributed with a mean equal
to 1.5 inches and a standard deviation of 0.05 inches. Before
shipping a large batch of these tiles Scribner quality analysts have selected a random sample
of eight tiles with the following diameters:
1.57 1.59 1.48 1.60 1.59 1.62 1.55 1.52
The analysts want to use these measurements to determine if the process is no longer operat
ing within the specifications. The following steps can be used:
Step 1 Determine the mean for this sample.
x
n
55 5
12 52
8
1 565
.
. inches
x
Wollwerth Imagery/Fotolia
www.downloadslide.comslide 299:
Introduction to Sampling Distributions
Step 2 Define the sampling distribution for x using Theorem 3.
Theorem 3 indicates that if the population is normally distributed the sampling
distribution for x will also be normally distributed with
xx
n
55 and
Thus in this case the mean of the sampling distribution should be 1.50 inches
and the standard deviation should be 0.0518 0.0177 inches.
Step 3 Define the probability statement of interest.
Because the sample mean is x 1.565 which is greater than the mean of the
sampling distribution we want to find
Px. 1 565 inches
Step 4 Convert the sample mean to a standardized zvalue using Equation 4.
z
x
n
1 565 1 50
005
8
0 065
0 0177
367
..
.
.
.
.
Step 5 Use the standard normal distribution table to determine the desired
probability.
P1z Ú 3.672
The Standard Normal Distribution Table does not show zvalues as high as
3.67. This implies that P1z Ú 3.672 ≈ 0.00. So if the production process is
working properly there is virtually no chance that a random sample of eight
tiles will have a mean diameter of 1.565 inches or greater. Because the analysts
at Scribner Products did find this sample result there is a very good chance that
something is wrong with the process.
END EXAMPLE
TRY PROBLEM 26
The Central Limit Theorem
Theorem 3 applies when the population is normally distributed. Although there are many situ
ations in business in which this will be the case there are also many situations in which the
population is not normal. For example incomes in a region tend to be right skewed. Some dis
tributions such as people’s weight are bimodal a peak weight group for males and another
peak weight group for females.
What does the sampling distribution of x look like when a population is not normally dis
tributed The answer is . . . it depends. It depends on what the shape of the population is and
what size sample is selected. To illustrate suppose we have a Ushaped population such as
the one in Figure 8 with mean 14.00 and standard deviation 3.00. Now we select 3000
10
xValues
11 12 13 14 15 16 17 18
FIGURE 8 
Simulated Nonnormal
Population
Chapter Outcome 3.
www.downloadslide.comslide 300:
Introduction to Sampling Distributions
simple random samples of size 3 and compute the mean for each sample. These x  v alues are
graphed in the histogram shown in Figure 9.
The average of these 3000 sample means is
x
x
3 000
14 02
.
Notice this value is approximately equal to the population mean of 14.00 as Theorem 1
would suggest.
4
Next we compute the standard deviation as
x
x
x
.
2
3 000
182
The standard deviation of the sampling distribution is less than the standard deviation for the
population which was 3.00. This will always be the case.
The frequency histogram of xvalues for the 3000 samples of 3 looks different from the
population distribution which is Ushaped. Suppose we increase the sample size to 10 and take
3000 samples from the same Ushaped population. The resulting frequency histogram of x
values is shown in Figure 10. Now the frequency distribution looks much like a normal distri
bution. The average of the sample means is still equal to 14.02 which is virtually equal to the
population mean. The standard deviation for this sampling distribution is now reduced to 0.97.
10 11 12 13 14 15 16 17 18
Sample Means
x
FIGURE 9 
Frequency Histogram of
x1n 32
4
Note if we had selected all possible samples of three the average of the samples means would have been equal to
the population mean.
FIGURE 10 
Frequency Histogram of
x1n 102
10 11 12 13 14 15 16 17 18
x
14.02 Average of sample means
0.97 Standard deviation of sample means
www.downloadslide.comslide 301:
Introduction to Sampling Distributions
The Central Limit Theorem is very important because with it we know the shape of the
sampling distribution even though we may not know the shape of the population distribution.
The one catch is that the sample size must be “sufficiently large.” What is a sufficiently large
sample size
The answer depends on the shape of the population. If the population is quite symmetric
then sample sizes as small as 2 or 3 can provide a normally distributed sampling distribution.
If the population is highly skewed or otherwise irregularly shaped the required sample size
will be larger. Recall the example of the Ushaped population. The frequency distribution
obtained from samples of 3 was shaped differently than the population but not like a normal
distribution. However for samples of 10 the frequency distribution was a very close approxi
mation to a normal distribution. Figures 11 12 and 13 show some examples of the Central
Limit Theorem concept. Simulation studies indicate that even for very strangelooking popu
lations samples of 25 to 30 produce sampling distributions that are approximately normal.
Thus a conservative definition of a sufficiently large sample size is n Ú 30. The Central
Limit Theorem is illustrated in the following example.
Theorem 4: The Central Limit Theorem
For simple random samples of n observations taken from a population with mean m and
standard deviation s regardless of the population’s distribution provided the sample
size is sufficiently large the distribution of the sample means x will be approximately
normal with a mean equal to the population mean 1m
x
m2 and a standard deviation
equal to the population standard deviation divided by the square root of the sample size
1s
x
s1n2. The larger the sample size the better the approximation to the normal
distribution.
FIGURE 11 
Central Limit Theorem
With Uniform Population
Distribution
b
Sampling Distribution
n 2
a
x
x
fx
Population
Sampling Distribution
n 5
c
x
fx
fx
This example is not a special case. Instead it illustrates a very important statistical
concept called the Central Limit Theorem.
www.downloadslide.comslide 302:
Introduction to Sampling Distributions
FIGURE 12 
Central Limit Theorem With
Triangular Population
Sampling Distribution
n 5
Sampling Distribution
n 30
a
x
fx
Population
b
x
c
x
fx
fx
FIGURE 13 
Central Limit Theorem With a
Skewed Population
Sampling Distribution
n 4
Sampling Distribution
n 25
a
x
fx
Population
c
x
b
x
fx
fx
www.downloadslide.comslide 303:
Introduction to Sampling Distributions
EXAMPLE 3 FINDING THE PROBABILITY THAT X IS IN A GIVEN RANGE
Westside DriveIn. Past sales records indicate that the dollar value of lunch orders at the
State Street restaurant are right skewed with a population mean of 12.50 per customer and
a standard deviation of 5.50. The Westside DriveIn manager has selected a random sample
of 100 lunch receipts. She is interested in determining the probability that the mean lunch
order for this sample from this population will fall between 12.25 and 13.00. To find this
probability she can use the following steps.
Step 1 Determine the sample mean.
In this case two sample means are being considered:
xx 55 . . 12 25 13 00 and
Step 2 Define the sampling distribution.
The Central Limit Theorem can be used because the sample size is large
enough1n 1002to determine that the sampling distribution will be
approximately normal even though the population is right skewed with
xx
555 .
.
. 12 50
550
100
055 and
Step 3 Define the probability statement of interest.
The manager is interested in
Px . . 12 25 13 00
Step 4 Use the standard normal distribution to find the probability of interest.
Assuming the population of lunch receipts is quite large we use Equation 4 to
convert the sample means to corresponding zvalues.
x
n
x 12 25 12 50
550
100
046
..
.
.and
n
13 00 12 50
550
100
091
..
.
.
From the standard normal table the probability associated with z 0.46 is
0.1772 and the probability for z 0.91 is 0.3186. Therefore
Px P . . . . . 12 25 13 00 0 46 0 91 0 1772 z 0 0 3186 0 4958 ..
There is nearly a 0.50 chance that the sample mean will fall in the range 12.25
to 13.00.
END EXAMPLE
TRY PROBLEM 30
How to do it Example 3
Sampling Distribution of x
To find probabilities associated
with a sampling distribution of x
for samples of size n from a popu
lation with mean m and standard
deviation s use the following steps.
1. Compute the sample mean using
x
x
n
∑
2. Define the sampling distribution.
If the population is normally
distributed the sampling dis
tribution also will be normally
distributed for any size sample.
If the population is not normally
distributed but the sample size is
sufficiently large the sampling
distribution will be approxi
mately normal. In either case the
sampling distribution will have
n
xx
and 55
3. Define the probability statement
of interest.
We are interested in finding the
probability of some range of
sample means such as
Px 25
4. Use the standard normal distri
bution to find the probability of
interest using Equation 4 or 5
to convert the sample mean to a
corresponding zvalue.
z
x
n
z
x
n
Nn
N
m
s
m
s
or
1
Then use the standard normal
table to find the probability asso
ciated with the calculated zvalue.
Skill Development
722. A population with a mean of 1250 and a standard
deviation of 400 is known to be highly skewed to the
right. If a random sample of 64 items is selected from
the population what is the probability that the sample
mean will be less than 1325
723. Suppose that a population is known to be normally
distributed with m 2000 and s 230. If a
random sample of size n 8 is selected calculate the
probability that the sample mean will exceed 2100.
MyStatLab
72: Exercises
724. A normally distributed population has a mean of 500
and a standard deviation of 60.
a. Determine the probability that a random sample of
size 16 selected from this population will have a
sample mean less than 475.
b. Determine the probability that a random sample
of size 25 selected from the population will have a
sample mean greater than or equal to 515.
725. If a population is known to be normally distributed with
m 250 and s 40 what will be the characteristics
www.downloadslide.comslide 304:
Introduction to Sampling Distributions
a. What is the probability of having a sample mean
equal to or smaller than the sample mean for this
sample if the population mean is 12 processed returns
daily with a standard deviation of 3 returns per day
b. What is the probability of having a sample mean
larger than the one obtained from this sample if the
population mean is 12 processed returns daily with
a standard deviation of 3 returns per day
c. Explain how it is possible to answer parts a and b
when the population distribution of daily tax returns
at Many Happy Returns is not known.
732. SeaFair Fashions relies on its sales force of 220 to do
an initial screening of all new fashions. The company
is currently bringing out a new line of swimwear and
has invited 40 salespeople to its Orlando home office.
An issue of constant concern to the SeaFair sales office
is the volume of orders generated by each salesperson.
Last year the overall company average was 417330
with a standard deviation of 45285. Hint: The finite
population correction factor Equation 5 is required.
a. Determine the probability the sample of 40 will
have a sales average less than 400000.
b. What shape do you think the distribution of all
possible sample means of 40 will have Discuss.
c. Determine the value of the standard deviation of
the distribution of the sample mean of all possible
samples of size 40.
d. How would the answers to parts a b and c change
if the home office brought 60 salespeople to
Orlando Provide the respective answers for this
sample size.
e. Each year SeaFair invites the sales personnel
with sales above the 85th percentile to enjoy a
complementary vacation in Hawaii. Determine the
smallest average salary for the sales personnel that
were in Hawaii last year. Assume the distribution
of sales was normally distributed last year.
733. Suppose the life of a particular brand of calculator
battery is approximately normally distributed with a
mean of 75 hours and a standard deviation of 10 hours.
a. What is the probability that a single battery
randomly selected from the population will have a
life between 70 and 80 hours
b. What is the probability that 16 randomly sampled
batteries from the population will have a sample
mean life of between 70 and 80 hours
c. If the manufacturer of the battery is able to reduce
the standard deviation of battery life from 10 to 9
hours what would be the probability that 16 batteries
randomly sampled from the population will have a
sample mean life of between 70 and 80 hours
734. Sands Inc. makes particleboard for the building
industry. Particleboard is built by mixing wood chips
and resins together and pressing the sheets under
extreme heat and pressure to form a 4feet 8feet
sheet that is used as a substitute for plywood. The
strength of the particleboards is tied to the board’s
weight. Boards that are too light are brittle and do not
of the sampling distribution for x based on a random
sample of size 25 selected from the population
726. Suppose nine items are randomly sampled from a
normally distributed population with a mean of 100
and a standard deviation of 20. The nine randomly
sampled values are
125 95 66 116 99
91 102 51 110
Calculate the probability of getting a sample mean that
is smaller than the sample mean for these nine sampled
values.
727. A random sample of 100 items is selected from a
population of size 350. What is the probability that the
sample mean will exceed 200 if the population mean is
195 and the population standard deviation equals 20
Hint: Use the finite population correction factor since
the sample size is more than 5 of the population size.
728. Given a distribution that has a mean of 40 and a
standard deviation of 13 calculate the probability that
a sample of 49 has a sample mean that is
a. greater than 37
b. at most 43
c. between 37 and 43
d. between 43 and 45
e. no more than 35
729. Consider a normal distribution with mean 12 and
standard deviation 90. Calculate P1x 7 362 for
each of the following sample sizes:
a. n 1
b. n 9
c. n 16
d. n 25
Business Applications
730. SeeClear Windows makes windows for use in homes
and commercial buildings. The standards for glass
thickness call for the glass to average 0.375 inches
with a standard deviation equal to 0.050 inches.
Suppose a random sample of n 50 windows yields a
sample mean of 0.392 inches.
a. What is the probability of x Ú 0.392 if the windows
meet the standards
b. Based on your answer to part a what would you
conclude about the population of windows Is it
meeting the standards
731. Many Happy Returns is a tax preparation service with
offices located throughout the western United States.
Suppose the average number of returns processed by
employees of Many Happy Returns during tax season
is 12 per day with a standard deviation of 3 per day. A
random sample of 36 employees taken during tax season
revealed the following number of returns processed daily:
11 17 13 91313131215
15 13 15 10 15 13 13 15 13
9 9 9 15 13 14 9 14 11
11 17 16 9 8 10 15 12 11
www.downloadslide.comslide 305:
Introduction to Sampling Distributions
airline was 34.3 pounds with a standard deviation of 5.7
pounds. Further it stated that the distribution of weights
was approximately normally distributed. This memo was
leaked to a consumers’ group in Atlanta. This group had
selected and weighed a random sample of 14 bags to be
checked on a flight departing from Atlanta. The following
data pounds were recorded:
29 27 40 34 30 30 35
44 33 28 36 33 30 40
What is the probability that a sample mean as small or
smaller than the one for this sample would occur if the
airline’s claims about the population of baggage weight
is accurate Comment on the results.
738. ACNielsen is a New York–based corporation
and a member of the modern marketing research
industry. One of the items that ACNielsen tracks
is the expenditure on overthecounter OTC
cough medicines. ACNielsen recently indicated
that consumers spent 620 million on OTC cough
medicines in the United States. The article also
indicated that nearly 30 million visits for coughs were
made to doctors’ offices in the United States.
a. Determine the average cost of OTC cough
medicines per doctor’s office visit based on 30
million purchases.
b. Assuming that the average cost indicated in part a is the
true average cost of OTC cough medicines per doctor’s
visit and the standard deviation is 10 determine the
probability that the average cost for a random selection
of 30 individuals will result in an average expenditure
of more than 25 in OTC cough medicines.
c. Determine the 90th percentile for the average
cost of OTC cough medicines for a sample of 36
individuals all of whom have visited a doctor’s
office for cough symptoms.
Computer Database Exercises
739. One of the topselling video games continues to be Mad
den NFL 12. While prices vary widely depending on
store or website the suggested retail price for this video
game is 59.95. The file titled Madden contains a ran
dom sample of the retail prices paid for Madden NFL 12.
a. Calculate the sample mean and standard deviation
of retail prices paid for Madden NFL 12.
b. To determine if the average retail price has fallen
assume the population mean is 59.95 calculate the
probability that a sample of size 200 would result in
a sample mean no larger than the one calculated in
part a. Assume that the sample standard deviation is
representative of the population standard deviation.
c. In part b you used 59.95 as the population
mean. Calculate the probability required in part b
assuming that the population mean is 59.50.
d. On the basis of your calculations in parts b and c
does it seem likely that the average retail price for
Madden NFL 12 has decreased Explain.
meet the quality standard for strength. Boards that are
too heavy are strong but are difficult for customers
to handle. The company knows that there will be
variation in the boards’ weight. Product specifications
call for the weight per sheet to average 10 pounds with
a standard deviation of 1.75 pounds. During each shift
Sands employees select and weigh a random sample of
25 boards. The boards are thought to have a normally
distributed weight distribution.
If the average of the sample slips below 9.60
pounds an adjustment is made to the process to add
more moisture and resins to increase the weight and
Sands hopes the strength.
a. Assuming that the process is operating correctly
according to specifications what is the probability that
a sample will indicate that an adjustment is needed
b. Assume the population mean weight per sheet slips
to 9 pounds. Determine the probability that the
sample will indicate an adjustment is not needed.
c. Assuming that 10 pounds is the mean weight what
should the cutoff be if the company wants no more
than a 5 chance that a sample of 25 boards will
have an average weight less than 9.6 lbs
735. The branch manager for United Savings and Loan in
Seaside Virginia has worked with her employees in
an effort to reduce the waiting time for customers at
the bank. Recently she and the team concluded that
average waiting time is now down to 3.5 minutes with
a standard deviation equal to 1.0 minute. However
before making a statement at a managers’ meeting
this branch manager wanted to doublecheck that the
process was working as thought. To make this check
she randomly sampled 25 customers and recorded the
time they had to wait. She discovered that mean wait
time for this sample of customers was 4.2 minutes.
Based on the team’s claims about waiting time what is
the probability that a sample mean for n 25 people
would be as large or larger than 4.2 minutes What
should the manager conclude based on these data
736. Mileage ratings for cars and trucks generally come
with a qualifier stating actual mileage will depend
on driving conditions and habits. Ford is stating the
Ecoboost F150 pickup truck will get 19 miles per
gallon with combined town and country driving.
Assume the mean stated by Ford is the actual average
and the distribution has a standard deviation of 3 mpg.
a. Given the above mean and standard deviation what
is the probability that 100 drivers will get more than
19.2 miles per gallon average
b. Suppose 1000 drivers were randomly selected.
What is the probability the average obtained by
these drivers would exceed 19.2 mpg
737. Airlines have recently toughened their standards for the
weight of checked baggage limiting the weight of a bag
to 50 pounds on domestic U.S. flights. Heavier bags will
be carried but at an additional fee. Suppose that one
major airline has stated in an internal memo to employees
that the mean weight for bags checked last year on the
www.downloadslide.comslide 306:
Introduction to Sampling Distributions
b. Use Excel to select a simple random sample of 30
executive compensation amounts. Compute the
sample mean for this sample. Find the probability
of getting a sample mean as extreme or more
extreme than the one you got. Hint: Use the finite
population correction factor because the sample is
large relative to the size of the population.
743. The data file called CEO Compensation contains data
for the most highly paid CEOs in the nation. Separate
the values into those from 2010 and those from 2011.
Treat the values found for 2010 data as population
values. Assume the value found for the mean of the
2011 data is a sample from the 2010 population. What
is the probability of finding a value this large or larger
744. The file Salaries contains the annual salary for all faculty
at a small state college in the Midwest. Assume that these
faculty salaries represent the population of interest.
a. Compute the population mean and population
standard deviation.
b. Develop a frequency distribution of these data using
10 classes. Do the population data appear to be
normally distributed
c. What is the probability that a random sample of 16
faculty selected from the population would have a
sample mean annual salary greater than or equal to
56650
d. Suppose the following 25 faculty were randomly
sampled from the population and used to estimate
the population mean annual salary:
Faculty ID Number
137 040 054 005 064
134 013 199 168 027
095 065 193 059 192
084 176 029 143 182
009 033 152 068 044
What would the sampling error be
e. Referring to part d what is the probability of
obtaining a sample mean smaller than the one
obtained from this sample
740. Acee Bottling and Distributing bottles and markets
PepsiCola products in southwestern Montana. The
average fill volume for Pepsi cans is supposed to be
12 ounces. The filling machine has a known standard
deviation of 0.05 ounces. Each week the company
selects a simple random sample of 60 cans and carefully
measures the volume in each can. The results of the latest
sample are shown in the file called Acee Bottling. Based
on the data in the sample what would you conclude
about whether the filling process is working as expected
Base your answer on the probability of observing the
sample mean you compute for these sample data.
741. Bruce Leichtman is president of Leichtman Research
Group Inc. LRG which specializes in research and
consulting on broadband media and entertainment
industries. In a recent survey the company determined
the cost of extra highdefinition HD gear needed to
watch television in HD. The costs ranged from 5 a
month for a settop box to 200 for a new satellite. The
file titled HDCosts contains a sample of the cost of the
extras whose purchase was required to watch television
in HD. Assume that the population average cost is
150 and the standard deviation is 50.
a. Create a box and whisker plot and use it and the
sample average to determine if the population from
which this sample was obtained could be normally
distributed.
b. Determine the probability that the mean of a
random sample of size 150 costs for HD extras
would be more than 5 away from the mean of the
sample described above.
c. Given your response to part a do you believe the
results obtained in part b are valid Explain.
742. CEO pay has been a controversial subject for several
years and even came up in the 2012 presidential
election. The file CEO Compensation contains data
on the 100 top paid CEOs in either 2010 or 2011
depending on when the information was made available.
a. Treating the data in the file as the population of
interest compute the population mean and standard
deviation for CEO compensation.
END EXERCISES 72
3 Sampling Distribution of a Proportion
Working with Proportions
In many instances the objective of sampling is to estimate a population proportion. For instance an
accountant may be interested in determining the proportion of accounts payable balances that are
correct. A production supervisor may wish to determine the percentage of product that is defect
free. A marketing research department might want to know the proportion of potential customers
who will purchase a particular product. In all these instances the decision makers could select a
sample compute the sample proportion and make their decision based on the sample results.
Sample proportions are subject to sampling error just as are sample means. The concept
of sampling distributions provides us a way to assess the potential magnitude of the sampling
error for proportions in given situations.
www.downloadslide.comslide 307:
Introduction to Sampling Distributions
BUSINESS APPLICATION SAMPLING DISTRIBUTIONS FOR PROPORTIONS
FLORIDA POWER AND LIGHT Customer service managers at Florida Power and Light
surveyed every customer who had new power service installed in a suburb of Tampa during
the month of September last year. The key question in the survey was “Are you satisfied with
the service received”
The population size was 80 customers. The number of customers who answered “Yes”
to the question was 72. The value of interest in this example is the population proportion.
Equation 6 is used to compute a population proportion.
Population Proportion
The fraction of values in a population that have a
specific attribute.
Population Proportion
X
N
p
6
where:
p Population proportion
X Number of items in the population having the attribute of interest
N Population size
The proportion of customers in the population who are satisfied with the service by Flor
ida Power and Light is
72
80
p
090 .
Therefore 90 of the population responded “Yes” to the survey question. This is the param
eter. It is a measurement taken from the population. It is the “true value.”
Now suppose that Florida Power and Light wishes to do a followup survey for a simple
random sample of n 20 of the same 80 customers. What fraction of this sample will be
people who had previously responded “Yes” to the satisfaction question
The answer depends on which sample is selected. There are many 3.5353 10
18
to
be precise possible random samples of 20 that could be selected from 80 people. How
ever the company will select only one of these possible samples. At one extreme suppose
the 20 people selected for the sample included all 8 who answered “No” to the satisfaction
question and 12 others who answered “Yes.” The sample proportion is computed using
Equation 7.
Sample Proportion
The fraction of items in a sample that have the
attribute of interest.
Sample Proportion
p
x
n
5
7
where:
p Sample proportion
x Number of items in the sample with the attribute of interest
n Sample size
For the Florida Power and Light example the sample proportion of “Yes” responses is
12
20
060 . p55
The sample proportion of “Yes” responses is 0.60 whereas the population proportion is
0.90. The difference between the sample value and the population value is sampling error.
Equation 8 is used to compute the sampling error involving a single proportion.
www.downloadslide.comslide 308:
Introduction to Sampling Distributions
Then for this extreme situation we get
Sampling error 0.60  0.90 0.30
If a sample on the other extreme had been selected and all 20 people came from the original
list of 72 who had responded “Yes” the sample proportion would be
20
20
100 . 55 p
For this sample the sampling error is
Sampling error 1.00  0.90 0.10
Thus the range of sampling error in this example is from  0.30 to 0.10. As with any sam
pling situation you can expect some sampling error. The sample proportion will probably not
equal the population proportion because the sample selected will not be a perfect replica of
the population.
EXAMPLE 4 SAMPLING ERROR FOR A PROPORTION
HewlettPackard–Compaq Merger In 2002 a proxy fight took
place between the management of HewlettPackard HP and Walter
Hewlett the son of one of HP’s founders over whether the merger
between HP and Compaq should be approved. Each outstanding share
of common stock was allocated one vote. After the vote in March 2002
the initial tally showed that the proportion of shares in the approval col
umn was 0.51.