ABSTRACT
This chapter provides some practical issues regarding sample size calculation for comparing proportions in clinical research. It focuses on comparing proportions between treatment groups with binary responses. The chapter discusses the general procedure of power analysis for sample size calculation for testing one-sample problem. It summarizes statistical procedures for sample size calculation for a two-sample problem under a parallel-group design and a crossover design. The chapter also discusses statistical procedures for testing a multiple-sample problem under a parallel design and a crossover design (Williams design). It also provides formulas for sample size calculation for comparing relative risks between treatment groups under a parallel design and a crossover design. The formulas for sample size calculation for testing equality, noninferiority/superiority, and equivalence are derived. The formulas provide required sample sizes for achieving a desired power under the alternative hypothesis. Once the sample size is determined, it can be justified under appropriate statistical assumption for other comparisons of interest.
