Chapter 13 discusses the use of weighted least squares in modeling categorical data. The first step of this methodology is to arrange the data from the experiment or study as an s × r contingency table. In this general WLS framework, there are s groups defined by the cross-classification of the factors of interest (explanatory variables) and r response profiles. Chapter 13 considers situations in which there is a single outcome (dependent) variable, so that the response profiles are defined by the r possible levels of the dependent variable. In each group, at most r – 1 linearly independent response functions can be analyzed. Thus, in the applications in which the response variable is dichotomous, ...

Start Free Trial

No credit card required