Skip to content

What n do I need for this experiment?

Work out how many samples you need

Power analysis, done before the experiment rather than after.

Power analysis connects four quantities: the effect size you care about, your significance threshold, the power you want (conventionally 80%), and n. Fix any three and the fourth follows. The hard part is the effect size, which has to come from pilot data, a published study, or an honest decision about the smallest difference that would matter biologically.

That last option is underrated. "A 20% change is the smallest I would act on" is a defensible basis for a sample size calculation, and it makes the experiment answer a question you actually have.

Post-hoc power — computing power from the effect you observed, after a non-significant result — is circular and uninformative. It is a recognised statistical error, not a rescue for an underpowered study.

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 →