Verlag:
Lund University, School of Economics and Management, Department of Economics, Lund
Zusammenfassung:
You want to run an experiment where you anticipate finding a treatment effect. How large should your sample size be to have a reasonable chance of detecting significant results? In this manuscript, I present and explain the Stata code I use to address this question. The code uses simulations to conduct power analyses, offering a flexible alternative to the commonly used analytical tools. Unlike traditional methods, this approach can accommodate any experimental design and statistical test that Stata supports. The code is straightforward, user-friendly, and can be used effectively with minimal coding experience.