Unit_1_lesson_5_redone

Views:
 
Category: Entertainment
     
 

Presentation Description

No description available.

Comments

Presentation Transcript

Slide 1: 

Unit 1: Understanding and Representing DataLesson 5: Score Placement: Percentiles and Percentile Ranks EDER 5210 – Educational Statistics Dr. Robin K. Henson University of North Texas © 2002 University of North Texas Dr. Robin K. Henson © 2002 Next Slide

Distributions characterize entire sets of scores … : 

Distributions characterize entire sets of scores … But what if we want to characterize (talk about) specific scores in a distribution? University of North Texas Dr. Robin K. Henson © 2002 Next Slide

Percentiles : 

Percentiles P50 = 50th percentile = A score value Value on the scale below which a specified % of scores fall. P50 = the median (always) University of North Texas Dr. Robin K. Henson © 2002 Next Slide

Slide 4: 

University of North Texas Dr. Robin K. Henson © 2002 Q: If you know that you scored 75 on a standardized achievement test and that you are at the 50th percentile (P50 = 75), which of the following must be true? You scored better than 50% of the other people taking the test. b. You scored better than 25% of the other people taking the test. c. You scored better than 75% of the other people taking the test. d. There is not enough information to determine the % of people scoring below you. Go to next slide for answer. Click mouse to continue

Answer : 

Answer Yes, because you are at the 50th percentile, then by definition you are at the median with 50% of the score above and below you. The fact that your actual score is 75 is irrelevant. A score at the 50th percentile could be anything depending on the scale used for measurement and how respondents actually perform. b. No, the percentile is not given by the actual score or its difference from some value, such as 100 in this case. c. No, the percentile is not given by the actual score. d. No, this is enough information to determine (a) as correct. Click mouse for next slide.

Percentile Rank : 

Percentile Rank % of scores lower than or equal to a score in question Ex: “If Suzie made on 80 on a test, what % of scores fell at and below an 80? Or, Suzie did as well or better than what % of people?” University of North Texas Dr. Robin K. Henson © 2002 Next Slide

* Differences between raw scores are exaggerated in the middle of the distribution. : 

* Differences between raw scores are exaggerated in the middle of the distribution. Raw Score Difference: 49 – 48 = 1 point PR Difference: About 54% - 50% = 4% University of North Texas Dr. Robin K. Henson © 2002 Next Slide

*Differences between raw scores are underestimated near the tails of the distribution. : 

*Differences between raw scores are underestimated near the tails of the distribution. Raw Score Difference: 71 – 70 = 1 point PR Difference: About 90.5% - 90% = .5% University of North Texas Dr. Robin K. Henson © 2002 Next Slide

What level of scale are PRs? …… ordinal (Inappropriate to do math operations on PRs!) : 

What level of scale are PRs? …… ordinal (Inappropriate to do math operations on PRs!) University of North Texas Dr. Robin K. Henson © 2002 Next Slide

Slide 10: 

Unit 1: Understanding and Representing DataLesson 5: Score Placement: Percentiles and Percentile Ranks EDER 5210 – Educational Statistics Dr. Robin K. Henson University of North Texas © 2002 University of North Texas Dr. Robin K. Henson © 2002 End

authorStream Live Help