K Unrelated Samples Tests


The two-sample tests for unpaired samples such as the Wilcoxon rank sum test, or the Mann–Whitney U test for two independent samples can be extended to three or more samples. When there are three or more populations that are not related in any way, the Friedman test is not the appropriate test, because the Friedman test requires matched samples. The appropriate method is to use one of the available nonparametric one-way analysis of variance (ANOVA) procedures. The customary parametric test is based on an F statistic. Contingency tables with three or more variables also belong to this group. Contingency tables are used to test whether multiple samples originated from populations with equal proportions. ...

Get A Primer on Nonparametric Analysis, Volume II now with the O’Reilly learning platform.

O’Reilly members experience live online training, plus books, videos, and digital content from nearly 200 publishers.