sampling

Views:
 
Category: Education
     
 

Presentation Description

No description available.

Comments

By: aaartii (23 month(s) ago)

This is really nice one

By: mansi123m (39 month(s) ago)

gud wrk

By: arunbhuyan (41 month(s) ago)

good work..

By: joyfulljeni (41 month(s) ago)

hi can you please do me a favor. . . pls allow us to download this ppt.....................

By: ravikabduwal (42 month(s) ago)

Good work..like like

See all

Presentation Transcript

Sampling: 

Sampling

Steps in the Sampling Process: 

Steps in the Sampling Process Define the population Frame the population Choose a sample design Draw the sample Execute the research

Designing the Sampling Procedure: 

Designing the Sampling Procedure Probability (random) samples simple random sampling; systematic sampling stratified sampling cluster sampling Nonprobability samples judgment sampling convenience sampling volunteer sampling quota sampling

Probability Samples: 

Probability Samples When to use stratified sampling over SRS: Primary research objective is to compare groups There are separate confidence interval objectives by strata Variances differ by strata Costs differ by strata Prior information differs by strata

Probability Samples: 

Probability Samples Examples of cluster sampling: Blocks (for consumers); dormitories or classes(for college students) When to use cluster sampling over SRS: When travel costs can be reduced as a result When there are substantial fixed costs associated with each data collection location When there is a list of clusters but not of individual population members

Probability vs. Nonprobability Samples: 

Probability vs. Nonprobability Samples Probability samples allow hypotheses tests through inferential statistics Biases of nonprobability samples: Subjects not typical of population Easily manipulated Biased toward well-known population members Biased against unusual population members

Probability vs. Nonprobability Samples: 

Probability vs. Nonprobability Samples Nonprobability samples are adequate when: the sample is very small the population is homogenous on the variable being studied informal, exploratory research is being done it is desired to screen out “loser” products

Discussion: 

Discussion Provide the marketing researcher’s definitions for each of the following populations: Columbia House, a mail order house specializing in tapes and CDs, wants to determine interest in a 10-for-1 offer on hard rock CDs. The manager of your student union is interested in determining if students desire a “universal checking account ID card” that will be accepted anywhere on campus and in many stores off campus. Joy Mfg. Co. decides to conduct a survey to determine the sales potential of a new type of air compressor used by construction companies.

Discussion: 

Discussion Taco Bell approaches an official at your university and proposes to locate one of its restaurants on campus. Because it would be the first commercial interest of this sort on this campus, the administration requires Taco Bell to conduct a survey of students to assess the desirability of this operation. Analyze the practical difficulties with doing a census in this situation, and provide specific examples of each one. For example, how long might such a census take? How much might it cost? What types of students are more accessible than others?...

Sampling and Nonsampling Errors: 

Sampling and Nonsampling Errors Nonsampling error: Unrelated to the sampling of respondents Sampling error Samples do not always reflect a populations’ true characteristics. Controlled by sample size. Sample bias Members of sample differ from larger population in some systematic fashion. Controlled by defining population and selective representative sample.

Sampling Error: 

Sampling Error Sampling error is the standard deviation of the distribution of sample means. The larger the sample, the smaller the sampling error

The Confidence Interval Approach: 

The Confidence Interval Approach Confidence intervals Example: 95% of values in normal distribution fall within ± 1.96 standard deviations from the mean of the distribution. Therefore, for the population mean, 95% of values in normal distribution will fall within the interval

The Confidence Interval Approach: 

The Confidence Interval Approach To determine sample size using a mean: Use the formula For 95% CI, z = 1.96 Must estimate s Must assign a value to e

The Confidence Interval Approach: 

The Confidence Interval Approach To determine sample size using a percentage: Use the formula For 95% CI, z = 1.96 Must estimate p. Then, q = 1 - p Must assign a value to e

The Confidence Interval Approach: 

How to estimate variability in a population: Prior research (secondary sources,…) Experience Intuition Range divided by 6 (for means) Worst possible scenario = .5 (for percentages) How to determine amount of precision desired: Cost and feasibility of having a large sample Client’s tolerance for uncertainty The Confidence Interval Approach

Discussion: 

Discussion Last year, Lipton Tea Company conducted a mall-intercept study at six regional malls in the U.S. and found that 20% of the public preferred tea over coffee as a mid-afternoon hot drink. This year, Lipton wants to have a nationwide telephone survey performed with random digit dialing. What sample size should be used in this year’s study in order to achieve an accuracy level of ± 2.5% at the 99% level of confidence? What about at the 95% level of confidence?

Value of Information Approach: 

Value of Information Approach CI method ignores: Relative importance of project New versus old products Cost per observation Value of information is more decision-oriented: best for “problem solving” research

Value of Information Approach: 

Value of Information Approach Factors relating to the value of information: Uncertainty about the proper course of action Gains or losses available from the decision Nearness to breakeven Calculating sample size. Two main considerations: marginal value of additional units decreases marginal cost of additional units increases

Nonstatistical Approaches: 

Nonstatistical Approaches Using previous sample sizes Using typical sample sizes Using a “magic number” Anticipation of subgroup analyses Use of resource limitations Use of expert guidance

Discussion: 

Discussion What sample size would you use? Why? Design for a new microwave oven Major new product introduction for company Research objective: what would the market share for this new product be? Previous introductions: Similarly-designed conventional oven: 36% of those questioned prefer new design over any other Similarly-designed deep fryer: 17% of those questioned prefer new design over any other 1% market share = $5,000,000

authorStream Live Help