ABSTRACT
In this chapter, we cover some more advanced mathematical topics. We begin with the general techniques and rules for deriving sums of squares, degrees of freedom, and expected mean square terms for a given ANOVA design. We con sider nonorthogonal sums of squares, the situation when predictors are them selves correlated. We examine ANOVA in terms of a general linear model; this is the way computers actually perform ANOVAs these days, and conceptually the most advanced and useful way to think about ANOVA. Then we discuss effect size.
