The screening of many endpoints when comparing groups from different strain
s, searching for some statistically significant difference, raises the mult
iple comparisons problem in its most severe form. Using the 0.05 level to d
ecide which of the many endpoints' differences are statistically significan
t, the probability of finding a difference to be significant even though it
is not real increases far beyond 0.05. The traditional approach to this pr
oblem has been to control the probability of making even one such error-the
Bonferroni procedure being the most familiar procedure achieving such cont
rol. However, the incurred loss of power stemming from such control led man
y practitioners to neglect multiplicity control altogether. The False Disco
very Rate (FDR), suggested by Benjamini and Hochberg [J Royal Stat Soc Ser
B 57 (1995) 289], is a new. different, and compromising point of view regar
ding the error in multiple comparisons. The FDR is the expected proportion
of false discoveries among the discoveries, and controlling the FDR goes a
long way towards controlling the increased error from multiplicity while lo
sing less in the ability to discover real differences. In this paper we dem
onstrate the problem in two studies: the study of exploratory behavior [Beh
av Brain Res (2001)], and the study of the interaction of strain difference
s with laboratory environment [Science 284 (1999) 1670]. We explain the FDR
criterion, and present two simple procedures that control the FDR. We demo
nstrate their increased power when used in the above two studies. (C) 2001
Elsevier Science B.V. All rights reserved.