Formula Power = Phi( |delta| / (SD sqrt(1/n1 + 1/n2)) - Z(a/2) ).
Assumptions Normally distributed outcome, equal SDs, two-sided test. Limitations Power calculated after a study from the observed effect adds no information beyond the p-value. Use it for planning, not to explain a null result. References Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale: Lawrence Erlbaum; 1988. Lwanga SK, Lemeshow S. Sample Size Determination in Health Studies: A Practical Manual. Geneva: World Health Organization; 1991. Runs in your browser with fixed formulas. No AI, no AI credits, and your inputs are not sent anywhere. This is a planning aid and not a substitute for review by a qualified statistician.