DISCRETE CHOICE MODELLING FOR TRAVEL MODE AND RESIDENTIAL LOCATION CHOICES IN AN ADELAIDE RAIL CORRIDOR

This paper uses discrete choice models to study how travel behaviour affects people's choices of travel mode and residential location, using a rail corridor case study. The study conducted a series of site observations to define the attributes and attribute's levels to be used in the models. The levels were then assessed to formulate Bayesian prior parameters for allocation by Bayesian efficient design in a stated preference experiment. Two sets of hypothetic choice scenarios were produced and included in a household questionnaire focusing on station access mode choice and residential location choice for data collection. A Random Parameter model, an Error Component model and a Latent Class model were applied to test the heterogeneity preference around the mean and standard deviation of random parameters and to specify important attributes for 'latent' heterogeneity of respondents. Modelling results revealed that the travel distance from home to the railway station as well as house type are important factors for local travel choices. The factors of daily activity, age, house affordability, time of day and safety are strongly influenced by heterogeneity preferences. The findings of this study should help to provide valuable information for local transport and land use planning.