1 / 17

NON-PARAMETRIC SUBSTITUTES FOR t-TESTS

NON-PARAMETRIC SUBSTITUTES FOR t-TESTS. Parametric Non-Parametric Independent t Wilcoxon Rank-Sum or Mann-Whitney Dependent t Wilcoxon T. WILCOXON RANK-SUM TEST. Purpose: Compare the medians of two groups. Design: between subjects Assumptions: independent observations

clover
Télécharger la présentation

NON-PARAMETRIC SUBSTITUTES FOR t-TESTS

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. NON-PARAMETRIC SUBSTITUTES FOR t-TESTS Parametric Non-Parametric Independent tWilcoxon Rank-Sum or Mann-Whitney Dependent t Wilcoxon T

  2. WILCOXON RANK-SUM TEST • Purpose: Compare the medians of two groups. • Design: between subjects • Assumptions: • independent observations • at least ordinal level data

  3. Example Four students with ADD were assigned to receive counseling, and four others with ADD were assigned to receive Ritalin. The number of minutes out of their seats per day was recorded.

  4. COUNSELING RITALIN 31 41 60 25 72 31 19 54

  5. STEP 1: Rank all scores from lowest to highest. COUNSELING RITALIN 31 Rank =3.5 41 Rank = 5 60 Rank = 7 25 Rank = 2 72 Rank = 8 31 Rank = 3.5 19 Rank = 1 54 Rank = 6

  6. COUNSELING RITALIN 31 Rank = 3.5 41 Rank = 5 60 Rank = 7 25 Rank = 2 72 Rank = 8 31 Rank = 3.5 19 Rank = 1 54 Rank = 6 SR = 19.5 SR = 16.5 STEP 2: Sum the ranks for each group.

  7. STEP 3: W is the smaller SR. W = 16.5

  8. STEP 4: Compare to critical value of W. Observed W must be EQUAL OR LESS THAN W-critical to be significant. For N1 = 4, N2 = 4, a = .05 two-tailed W-crit = 10 W = 16.5, not significant

  9. APA Format Sentence A Wilcoxon Rank-Sum test showed that the Ritalin and Counseling groups were not significantly different, W (n1 = 4, n2 = 4) = 16.50, p > .05.

  10. WILCOXON T • Purpose: Test whether two distributions are different • Design: within subjects or matched • Assumptions: • at least ordinal level data • populations are identical except for means • minimum N of 6

  11. Wilcoxon T Calculation Example: Eight patients are exposed to a placebo and an experimental treatment (at different times). They are measured on severity of symptoms. Was there a significant difference between conditions? (scores on next page)

  12. Patient P E 1 10 8 2 14 10 3 12 13 4 15 15 5 6 4 6 8 11 7 9 6 8 14 12

  13. Patient P E Diff 1 10 8 +2 2 14 10 +4 3 12 13 -1 4 15 15 0 5 6 4 +2 6 8 11 -3 7 9 6 +3 8 14 12 +2 STEP 1: Determine difference scores

  14. Patient P E Diff Rank 1 10 8 +2 3 2 14 10 +4 7 3 12 13 -1 1 4 15 15 0 5 6 4 +2 3 6 8 11 -3 5.5 7 9 6 +3 5.5 8 14 12 +2 3 STEP 2: Rank the difference scores from smallest to largest, ignoring differences of zero.

  15. STEP 3: Compute the sum of the positive ranks and the sum of the negative ranks. SR positive = 3 + 7 + 3+ 5.5 + 3 = 21.5 SR negative = 1 + 5.5 = 6.5 STEP 4: The Wilcoxon T is the smaller SR. T = 6.5

  16. STEP 5: Compare to critical value. N = number of NONZERO differences N = 7 Tcrit = 2 Your T must be LESS THAN or equal to Tcrit to be significant!

  17. APA Format Sentence A Wilcoxon T showed no significant difference between the placebo and experimental treatments, T (N = 7) = 6.5, p > .05 .

More Related