Designing a model of metropolitan migration probabilities

The main purpose of this paper is to design an appropriate logit model for the analysis of interregional migration in Canada at the metropolitan-nonmetropolitan scale of resolution. Our strategy is to construct a nearly self-contained analytical framework in which various types of logit models can be built with relative ease and can be compared to each other. Using this framework, we specify two- and three-level nested logits, and we discuss criteria for deciding which model framework is more suitable in an empirical context. We estimate the migration probabilities for alternative models and, using our criteria, we argue for a three-level logit model as the proper choice in the context of our study. © 1997 OPA (Overseas Publishers Association) Amsterdam B.V. Published in The Netherlands under license by Gordon and Breach Science Publishers.