Individual Level Models for Mobility Tables and other Cross-Classifications

Abstract

In recent years a number of methods have been developed for introducing continuous covariates into mobility analyses and into the study of cross-classified data more generally. Drawing on the work of Logan, this article shows how McFadden's Conditional Logit Model can be used to fit existing log-linear models and can incorporate into them the effects of continuous covariates. The method is illustrated using a simple example.