Skip to content

I ran thousands of tests — which results are real?

Correct for multiple testing

Bonferroni, Benjamini–Hochberg, and choosing between them.

Testing 20,000 genes at p < 0.05 gives you around 1,000 false positives before any biology happens. Bonferroni controls the chance of even one false positive and is severe; it suits a handful of pre-planned comparisons. Benjamini–Hochberg controls the proportion of false positives among your hits and is the standard for genomics, where a 5% false discovery rate in a hit list is an acceptable trade.

Report adjusted values and say which method produced them. "q < 0.05" and "p < 0.05" are different claims.

What to use

Drafted from the literature and not yet reviewed by a specialist. Treat it as a starting point, and tell us if something here is wrong. Last touched .

All tasks →