Posts

2021

  • Simulation-based power curves

    Introduction In a previous post I covered how to perform simulation-based power analyses. A limitation in that post is that usually the question is not what our power is, but rather which sample size gives us the desired power. With the code from my previous post you can get to the right sample size by changing the sample size parameters and then checking whether this gives you enough power.

  • Simulation-based power analyses

    Introduction Setup One Sample t-test Welch’s Two Sample t-test Two Sample t-test Paired t-test Correlation 2 t-tests Regression (2 x 2 interaction) Conclusion Introduction Doing power analyses is hard. I know this from experience, both as a researcher and as a reviewer. As a researcher, I have found power analyses to be difficult because performing a good power analysis requires a thorough understanding of the (hypothesized) data.