Wilcoxon Rank-Sum Tests

Mann-Whitney U or Wilcoxon Rank-Sum Tests

Test Assumptions When running a Mann-Whitney-Wilcoxon test, the basic assumptions are random sampling from each of the two populations and that the measurement scale is at least ordinal. These assumptions are sufficient for testing whether the two populations are different. If we can additionally assume that the two populations are identical except possible for a difference in location, then this test can be used as a test of equal means or medians.