0

I am looking at the effect of genotype on a continuous variable. In total 16 comparisons were done involving 9 different SNPs. The number of significant events was 3 out of 16 tests (Mann-Whitney U-test). Is it necessary to correct the p-values to account for multiple comparisons?

chl
  • 50,972
  • 18
  • 205
  • 364
Kams
  • 1
  • 2
    Haven't you find anything relevant by looking at our questions with the [tag:multiple-comparisons] tag? If not, could you clarify where you are having difficulties? – chl Sep 29 '12 at 10:23

1 Answers1

1

Moderator chl is right to say that there are a wealth of discussion of multiple comparisons on this site that should provide a satisfactory answer to your question. But if you are just looking for a simple yes/no answer. The answer is yes. I would simply add that methods for p-value adjustment for the FDR criteria is most commonly used for your problem. See the book by Westfall and Young for resampling approaches to p-value adjustment.

Michael R. Chernick
  • 39,640
  • 28
  • 74
  • 143