ABSTRACT
Procedures for sample size calculation for testing hypotheses of interest are obtained under appropriate nonparametric statistical methods. This chapter discusses nonparametric methods for testing differences in location for one-sample and two-sample problems. It examines the loss in power due to the violation of the normality assumption. The chapter provides nonparametric tests for independence and the corresponding procedure for sample size calculation and presents some practical issues. Nonparametric methods for comparing variabilities between treatment groups, however, are much more complicated and require further research. When a nonparametric method is used, a formula for sample size calculation may not be available or may not exist as a closed form, especially when the study design/objective is rather complicated. In this case, the technique of bootstrapping may be applied. Unlike the parametric approach, formulas or procedures for sample size calculation for testing difference in multiple-sample locations using nonparametric methods are much more complicated.
