ABSTRACT
This chapter gives a brief review of statistical reasoning, including frequentist and Bayesian viewpoints, mostly in the context of binary outcomes. We discuss the p-value controversy and the need for interval estimation to complement results of testing. We illustrate the ideas with exact inference to introduce Clopper-Pearson confidence intervals. We also briefly summarize two major results that facilitate asymptotic inference, the central limit theorem and the delta method. We emphasize the importance of robust statistical analyses in clinical trials, including nonparametric methods.
