This was a demonstration at AEA2018. To speak truth to power, evaluation studies that are adequately statistically powered are needed to detect significant findings. This demonstration, aimed at novices, covers the inputs needed to answer the question, “what sample size do we need?” The presentation will start with the principles behind sample size, including the need for clear hypothesis and primary outcome, and a review of type I and II errors. Next, five main inputs to the sample size calculation will be covered, using examples from published articles. Additionally, for evaluations that occur in health facilities, schools or other groupings, we’ll address additional inputs to sample size calculations for these clustered studies. Example code will be shared in Stata software. Review the slides and feel more confident in sample size!