ABSTRACT
Fixed-effects linear models are general linear models where the response is modeled as a linear function of categorical predictors (factors) with fixed levels. Inference for such models usually follows inference for the less-than-full rank GLM. In Chapter 4, we introduced the least squares approach for balanced fixed-effects models, and discussed the ANOVA decomposition and F-test under normality in Chapter 7. In section 10.1, we describe parametric inference for unbalanced models, including one-way models and higher-order cross-classified models and nested models. This is followed by a description of two nonparametric procedures in section 10.2. Section 10.3 describes analysis of covariance procedures. In Section 10.4, we discuss concepts and procedures of multiple testing.
