ABSTRACT
This chapter discusses some practical issues regarding sample size calculation for comparing means in clinical research, including sample size reduction when switching from a two-sided test to a one-sided test or from a parallel design to a crossover design, sensitivity analysis with respect to change in variability, and a brief discussion regarding Bayesian approach. It considers testing in one-sample problems and summarizes procedures for sample size calculation in two-sample problems under a parallel design and a crossover design. The chapter presents procedures in multiple-sample problems under a parallel design (one-way analysis of variance [ANOVA]) and a crossover design (Williams design), respectively. Under a valid design, methods for sample size calculation are derived for achieving a desired power of statistical tests for appropriate hypotheses. A typical approach for sample size calculation is to reject the null hypothesis of no treatment difference and conclude that there is a significant difference between treatment groups.
