ABSTRACT
This chapter covers the more basic methods for means that are typically used in applied research, and then describes methods that address the practical problems associated with these techniques. It presents a one-way repeated measures design, also called a within-subjects design, where the goal is to compare more than two dependent groups. The percentile bootstrap methods do an excellent job of avoiding Type I errors greater than the nominal level, provided that the estimator being used has a reasonably high breakdown point. The two best methods for controlling Type I error probabilities and simultaneously providing reasonably high power are the bootstrap-t method based on 20" trimmed means and the percentile bootstrap method used in conjunction with M-estimator. An important feature of the rank-based method for analyzing a between-by-within design is that it is designed to perform well, in terms of controlling the probability of a Type I error, when tied values occur.
