CHAPTER 7 One-Way and Higher Designs for Independent Groups This chapters describes techniques for testing hypotheses in one-way and higher designs involving independent groups. Included are random effects models plus methods for performing multiple comparisons. This chapter makes no attempt at covering all the designs that are encountered in practice, but it does cover many of the more common designs that are used. In this chapter, only heteroscedastic methods are considered. It might be hoped that as the number of groups increases, problems associated with homoscedastic methods, described in Chapter 5, might be reduced. In one-way designs, the exact opposite seems to be true. For example, even under normality with equal sample sizes but unequal variances, problems controlling the probability of a type I error can arise. With four independent groups, each