How to: Kruskal-Wallis test

The Kurskal-Wallis test is a nonparametric test that compares three or more unmatched groups.

1. Enter data

From the Welcome (or New Table and graph) dialog, choose any graph from the One-way tab. We suggest plotting a scatter graph showing every point, with a line at the median.

If you aren't ready to enter your own data, choose to use sample data, and choose: One-way ANOVA - Ordinary.

Enter the data for each group into a separate column. The two groups do not have be the same size (it's OK to leave some cells empty). Since the data are unmatched, it makes no sense to enter any row titles.

2. Choose the Kruskal-Wallis test

1.	From the data table, click on the toolbar.

2.	Choose one-way ANOVA from the list of column analyses.

3.	On the ANOVA dialog, choose the Kruskal-Wallis test.

3. Choose a post test

Prism only offers Dunn's post test, either to compare all pairs of groups, or just selected pairs.

4. Review the results and inspect the graph

ANOVA investigates the likelihood that the difference among the groups could have been caused by chance. The most important results are the P value and the results of post tests.

Before accepting the results, review the analysis checklist.