Other's On-Line Resources (Chapter 18)


Wilcoxon's Matched-Pairs Signed-Ranks Test
  • Description:
    In using this interactive on-line resource, you'll be able to enter hypothetical sets of matched data and then see the p-level that's computed for your data. By varying the nature of your data sets, you'll come to understand why the Wilcoxon test sometimes leads to a statistically significant result.

  • What to Do:
    1. Click on the colored title of this on-line resource: "Wilcoxon's Matched-Pairs Signed-Ranks Test."
    2. On the screen that pops up, Not that there's a white box with two columns of scores.
    3. In the white box, imagine that each row of scores corresponds to a different high school students who wants to go to college. Further imagine that the specific numbers are SAT scores and that 2 numbers appear on each row because each of our 9 hypothetical students took the SAT twice.
    4. Change the last student's second score from 582 to 537.
    5. Click on the "Submit" button, and then examine the p-level that's presented above the box containing the paired data. Originally the p-level was 0.0396
    6. Originally the p-level was 0.03906; now, it's equal to .09766. Think about the one change in the data you made, consider all 9 pairs of scores, and ask yourself whether it makes sense that the p-level got larger because of the change you made.
    7. Make other changes in the data, each time checking to see what happens to the resulting p-level. (To make the original 9 pairs of scores to appear in the white data box, click on "Return to Statistics" in the upper left-hand corner of the screen, and then--on the new screen that pops up--click on "Wilcoxon Matched-Pairs Signed-Ranks Test" under the heading "One-Sample and Matched-Pairs Tests.")

  • Sky Huck's Puzzle Question:
    If you start with the original 9 pairs of scores (by clicking on the "Reset" button, you'll note that you can increase the first student's 2nd score (of 594) by 1 point, by 10 points, or by 100 points, and yet these changes do not alter the p-level. Why does the p-level remain constant even though this score is changed? (If you answer this question correctly, you'll probably be able to sense how extensively you could LOWER 594 without affecting the p-level.)


Copyright © 2012

Schuyler W. Huck
All rights reserved.

| Book Info | Author Info |

Site URL: www.readingstats.com

Top | Site Map
Site Design: John W. Taylor V