Modelling Suburban Travel Demand: Application of Mixed Logit with Models Incorporating RP/SP Data

This paper will analyze the choice of mode in suburban corridors using nested and mixed logit specifications and incorporating revealed and stated preference data. The latter were obtained from a choice experiment between car and bus, which allowed for interactions among the main policy variables: travel cost, travel time and frequency. The experiment also included parking cost and comfort attributes. The attribute levels in the experiment were adapted to travelers’ experience using information from the revealed preferences. Different model specifications are presented accounting for the presence of income effect, analyzing random and systematic taste variations and incorporating the effect of latent variables. We derived different willingness-to pay measures, such as the subjective value of time, that vary among individuals. Finally, the paper concluded by looking at the sensitivity of individual behavior to various policy scenarios. In general, demand was shown to be more sensitive to policies that penalize the private car than those improving public transport.