Statistical power

Power is the probability of rejecting the null hypothesis when it is false (1 - beta). Conventionally fixed at 80% for confirmatory studies. Depends on expected effect size, variance, and sample size.