Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

How do you report Kruskal Wallis One Way Anova and post hoc results according to APA (American Psychological Association) 6th ed. standards? I have tried to find a template for reporting the results for the Kruskal-Wallis one way analysis of variance test and post hoc tests. I have found a number of descriptions but the templates all contained different types of information. Some included the full sample size, some included the sample size for the two groups being compared, some included the k-df, and some included the confidence intervals in text but I could not find anything consistent. The APA 6th edition’s examples are not specific to the Kruskal-Wallis and I don’t want to make assumptions.

Each of the Kruskal Wallis analysis were significant at .000.

I have a total of 392 post hoc comparisons:

8 DAT groups = 28 comparison pairs between these groups.
15 SDRT tests. The SDRT is a Reading comprehension test. The items in this test were divided into 14 sub-tests (plus total score). The sub-tests are not all independent of each other (One item can be present in two or more sub-tests). These represent different text conditions.

Thus 28 group comparisons were run for all 15 SDRTs (total score and 14 sub-tests) which resulted in 392 analyses.

I first reported the trends for the SDRT total score. Then I went through each SDRT sub-test to see if these trends changed when certain SDRT items are included in the assessment but not others. Thus the comparison is between the 8 DAT groups as well as between the 15 SDRTs.

Could you suggest the best way to report this in text or refer me to some examples?

share|improve this question
Could you please try to ask several shorter and more defined questions? – mbq Oct 23 '11 at 21:18
Just curious: How did you protect your post-hoc tests from inflated Type I Error? (especially in light of possible correlations between subtests) – chl Nov 19 '11 at 20:54
I used the Bonferroni adjustment method. The adjusted significance level was equivalent to an unadjusted significance level of .002. I didn't mention that I used the Mann Whitney U for post hoc analysis. (I'm using SPSS 17 so much of this is built in) – Aleks Nov 20 '11 at 14:52
You should not use run of the mill Mann Whitney U for post hoc analysis, but rather Dunn's test (regardless of one's choice of adjustment method for multiple comparisons). See, for example,… – Alexis Apr 26 '14 at 17:19

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.