Power (statistical power)

Power (1 − β) is the probability of detecting a truly existing effect of a given size as significant. Targets of .80 to .95 are usual; the larger the sample and the effect, the higher the power.

Example: with two groups of 64 people each, a t test finds a true effect of d = 0.50 with a probability of .80 (α = .05, two-tailed).

Caution: power belongs before the study. An "observed power" computed from your own results is only a transformation of the p value.