Discrete choice models as a tool for transit service quality evaluation

Abstract

In this paper discrete choice Logit models for measuring transit service quality are proposed. Multinomial and Mixed Logit models are used as a tool for evaluating the importance of the different transit service aspects on the overall service quality. Particularly, Mixed Logit models are proposed in order to take into account the heterogeneity of perceptions across individuals. The models were calibrated on the basis of Stated Preferences choice experiments, in which decision makers choose among transit services characterized by different quality levels. The research work is supported by a sample survey addressed to the users of an urban bus service in a medium-sized town.