Wilcoxon rank sum test

[Visiting]

If sample size is not equal, is it possible to conduct comparison by wilcoxon rank sum test? Thank you!

[art297 · Accepted solution]

Yes. You can find a discussion (with references) at: https://stats.stackexchange.com/questions/40342/mann-whitney-u-test-with-unequal-sample-sizes

Art, CEO, AnalystFinder.com

[Visiting]

Thank you!

[mkeintz]

Looking at the wikipedia entry for Wilcoxon Signed Rank Test, it states that one of the assumption is "Data is paired and comes from the same population", i.e. the two "samples" must be of the same size.  But it also includes a reference to the related Mann-Whitney U Test (also call the Mann-Whitney-Wilcoxon Rank-sum test).  which does NOT require identical sample size.

Take a look at the NPAR1WAY procedure, example 83.3 "Exact Wilcoxon Two-Sample Test".

Note: This is not based on my expertise - other than using wikipedia, and searching for 'mann-whitney' in sas documentation.

[Visiting]

Thank you for the response!