This is a project designed to develop a new class of loglinear and logit models for examining structural effects in social science research. Work on the development of these models will include efforts to examine the robustness of the models and their ability to deal with a range of empirical phenomena. The project will culminate in the the development of user-friendly computer software to make the methods developed available to other researchers. Loglinear and logistic models have been widely used in recent years by sociologists and others to study phenomena which cannot be represented accurately by continuous variables. Existing versions of these models suffer from a number of assumptions and restrictions that limit their utility in social science research. By developing a new and broader range of such models and preparing computer software to make the newly developed models available to other researchers, this project will make an important contribution to sociological research on a wide range of topics.
