Techniques for adjusting for the effect of running multiple tests on data. If you do ten tests you are ten times as likely to see an apparently positive result by pure chance; if you do 1000 tests then it starts to become highly likely that something will look promising even if there is no real effect. Some form of multiplicity adjustment or multiplicity control is therefore critical to prevent false positive results.
Defined on pages 80, 80
Used on page 80
Also known as multiplicity adjustments