Spatial transferability of an ordered response model of trip generation

This paper documents analysis of the spatial transferability of an ordered response model, a type of discrete choice model which maintains the ordinal nature in the dependent variable in situations where there are more than two responses. The analysis focuses on shopping trip generation in Metropolitan Toronto. The paper investigates the performance of a directly transferred ordered response model (without updating the transferred coefficients) and assesses the effectiveness of a technique for revising the constant terms and scalars in the model by using small-sample data from the region to which the model is to be applied. The results of this spatial transferability analysis show that a directly transferred ordered response model performs reasonably well in predicting the aggregate shares in the application (new) context. Revising the constant terms and the scalars in the model substantially improves the predictive ability of the transferred model.