ABSTRACT
This chapter reviews the statistical concepts underlying sample size calculation. In most settings, the null hypothesis is that there is “no effect”, such as no difference between the groups, or no association between the variables. In clinical trials, this is the case for a superiority study, where the objective is to show that one treatment is better than another, and the null hypothesis is that there is no difference. As an alternative to sample size calculations based on statistical power, one may calculate sample sizes based on confidence interval estimation. Hypothesis testing and confidence interval estimation are related; a confidence interval can be derived by inversion of the hypothesis test. Power analysis or sample size estimation is relevant at the planning stage. Post hoc power calculations are futile, although it has been recommended by some journals.
