Marginal Analysis of Longitudinal Data
In the previous chapters, the foundations have been laid for a comprehensive full ML estimation approach for the handling of marginal categorical data. The approach was illustrated by means of research questions that necessitated the use of marginal modeling, but in a relatively simple form. In the next three chapters, it will be shown that marginal modeling can be used to answer important research questions that require more complex analyses and involve longitudinal data (Chapter 4), structural equation modeling and (quasi-)experimental designs (Chapter 5), and latent variable models (Chapter 6). Except for the extension to latent variable models, no really new principles will be involved. However, that does not mean that the application of marginal modeling in the following chapters is always easy and straightforward. The next chapters will try to explain in an accessible way how to translate particular substantive research questions into the sometimes complex language of themarginalmodeling approach. The data and the programs on our website will help the readers to carry out the analyses themselves and in this way learn by doing.
KeywordsMarginal Distribution Loglinear Model Marginal Modeling Parabolic Model Transition Table
Unable to display preview. Download preview PDF.