OPEN A NEW WINDOW IN RATIONAL RESEARCH PLANNING: ADJUST ALPHA TO MAXIMIZE STATISTICAL POWER

Alternative strategies for optimizing statistical power in applied psychological research are considered. Increasing sample size and combining predictors in order to yield a useful effect size are well-known tactics for increasing power. A third approach, increasing alpha, is rarely used because of zealous adherence to convention. There are two related aspects in setting the alpha level. First, the relative seriousness of Type I and Type II errors must be considered. This assessment must then be qualified and redetermined after taking into account the prior probability that an effect exists. Procedures that make these processes objective are demonstrated. When sample size and effect size are both fixed, increasing alpha may be the only feasible strategy for maximizing power. It is concluded that a priori power analysis should be a major consideration in any test of an hypothesis, and that alpha level adjustment should be viewed as a useful strategy for increasing power.