The discrete choice analysis of residential location choice forms an important part of land use-transport modelling systems but gives rise to a number of significant modelling challenges, one of which is the choice set formation problem. Choice set formation is an important aspect of modelling any kind of spatial choice (e.g., residential location, destination, route) both from the cognitive/behavioural and the computational efficiency perspectives. Most of the previous studies in residential location choice modelling use one of several simple approaches to address the choice set formation problem, such as random sampling. To the best of our knowledge, there have been no efforts to empirically test or compare different choice set formation techniques within the context of modelling residential location.

In this paper, we propose to address this gap by examining the performance of several choice set formation methods within an empirical context of modelling residential location choice. Further, we propose to implement a recently developed hazard-based model of housing search choice set formation (Rashidi and Mohammadian, 2010). The hazard-based screening model will be implemented in two different ways: (a) using deterministic thresholds on different attributes, and (b) using a probabilistic, importance sampling approach. Moreover, the structure of the hazard-based screening model enables us to model systematic and random inter-individual heterogeneity in choice set formation.

This paper first presents a systematic comparison, using real rather than simulated data, between the various choice set formation approaches and their impacts on the residential location choice model. The paper presents a brief review of the relevant theoretical and empirical literature on the treatment of the choice set formation problem in the context of spatial choice and sets out the characteristics of the specific modelling approaches implemented in the empirical research. The empirical research uses a specially constructed dataset of residential location choice, developed by the authors for the Greater London area, which draws on information from a number of separate data sources (including revealed preference data from the London Area Travel Survey). The alternative choice set formation approaches are assessed in terms of the estimated model parameters and implied model elasticities. Model predictions are also compared using a validation sample. Further, in order to undertake a more rigorous comparison of the choice set formation models and their effectiveness in modelling residential location choice, we propose to generate a synthetic dataset based on the deterministic threshold choice set formation model. The different approaches will then be compared within the context of the synthetic data in order to quantify the effects of using a mis-specified model.

Rashidi, T., and A. Mohammadian, “Behavioral Housing Search Choice Set Formation: Hazard-Based Screening Model of Property Value and Work Distance”, Proceedings of the 89th Annual Meeting of the Transportation Research Board (DVD), Washington, D.C., January 11-15, 2010.