4.4 Statistical power

As Hayes discussed, power is an important but thorny issue within the frequentist paradigm. Given that we’re not particularly interested in rejecting the point-null hypothesis as Bayesians and that we bring in priors (which we’ve largely avoided explicitly mentioning in his project but have been quietly using all along), the issue is even more difficult for Bayesians. To learn more on the topic, check out Miočević, MacKinnon, and Levy’s paper on power in small-sample Bayesian analyses or Gelman and Carlin’s paper offering an alternative to the power paradigm. You might also look at Matti Vuorre’s Sample size planning with brms project.