Mann-Whitney U test

[ballardw]

---

@docfermi wrote:

Thank you so much!  That's quite helpful.  Besides normality assumption criteria, two sample t-test to look for difference in the means vs. Mann-Whitney U test for difference in the medians?  

---

The Normal approximation and t Approximations are based on the values transformed to ranks. So those tests are of the "mean rank".

The "groups" are ordered by  value of the analysis variable, the ranks are assigned from the combined sample, and the tests are run on the ranks of the values. .

[PGStats]

Use option MEDIAN in proc npar1way to test the difference in medians.

[docfermi]

My mistake, I just realized that sas didn't run exact Wilcoxon due to higher sample frequency than the limit for exact wilcoxon.