I present a new Stata command, simsam, that uses simulation to determine the sample size required to achieve a given statistical power for any hypothesis test under any probability model that can be programmed in Stata. simsam returns the smallest sample size (or smallest multiple of 5, 10, or some other user-specified increment) so that the estimated power exceeds the target. The user controls the precision of the power estimate, and power is reported with a confidence interval. The sample size returned is reliable to the extent that if simsam is repeated, it will, nearly every time, give a sample size no more than one increment away.
CampbellM. J., JuliousS. A., and AltmanD. G.1995. Estimating sample sizes for binary, ordered categorical, and continuous outcomes in two group comparisons. British Medical Journal311: 1145–1148.
4.
DonnerA., BirkettN., and BuckC.1981. Randomization by cluster: Sample size requirements and analysis. American Journal of Epidemiology114: 906–914.
5.
EldridgeS. M., AshbyD., and KerryS.2006. Sample size for cluster randomized trials: Effect of coefficient of variation of cluster size and analysis method. International Journal of Epidemiology35: 1292–1300.
FloreyC. D.1993. Sample size for beginners. British Medical Journal306: 1181–1184.
10.
JungS.-H.2008. Sample size calculation for paired survival data: A simulation method. Journal of Statistical Computation and Simulation78: 85–92.
11.
NewellD. J.1978. Type II errors and ethics (letter). British Medical Journal4: 1789.
12.
WilliamsM. S., EbelE. D., and WagnerB. A.2007. Monte Carlo approaches for determining power and sample size in low-prevalence applications. Preventive Veterinary Medicine82: 151–158.
