I was telling a colleague that the best and most popular nonparametric alternative to the unpaired student t test is the Mann Whitney test. Ifigured that is probably what they use in clinical trials and other social science environment. But, he suggested an alternative the two-sample Kolmogorov-Smirnov (KS) test. I never heard of people using the KS test for this purpose. But, I wondered why not.
Can you help me figure out the relative strength and weaknesses of each test. Thanks in advance for any info.
Gaetan Lion