Formula n = ((Z(a/2) + Z(b)) x SD_diff / delta)^2 pairs.
Assumptions Differences are approximately normal. Pairs are independent of each other. Limitations If you only know the SD of each measurement, you also need their correlation to get SD_diff. 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.