If statistical power is .80 (or 80%) and beta is .2 (or 20%), the probability of a type I error is .20 or 20%.
