ABSTRACT

The remaining portion of this book presents statistical models where one or more of the independent variables are metric. This chapter briefly reviews linear regression, highlighting topics that will be important in later chapters (dummy variables and functional form) as well as the assumptions of the general linear model. It then introduces the concept of a generalized linear model (GLM), which forms the basis of logistic, ordinal, and Poisson regression. (As described in Chapters 10–13, these methods are applicable when the dependent variable [DV] is nominal or ordinal or a count.) The final section of this chapter presents GLMs using linear algebra.