THE ANALYSIS OF A SET OF MULTIDIMENSIONAL CONTINGENCY TABLES USING LOG-LINEAR MODELS, LATENT-CLASS MODELS, AND CORRELATION MODELS: THE SOLOMON DATA REVISITED

Abstract

This chapter reviews the various kinds of models for analyzing the set of data first introduced by Solomon, namely, the analysis of a set of multidimensional contingency tables using log-linear models, latent-class models, and the correlation models. The chapter presents the correlation model described by Lazarsfeld, Bahadur, and Solomon, the logistic models in Bock, the additive-effects model in Martin and Bradley, the log-linear models in Goodman, and the latent-class models in Clogg and Goodman. The Solomon data illustrates the application of a rather wide range of perspectives to a given set of data—the set of multidimensional contingency tables introduced by Solomon.