Investigating the heterogeneity of bus users' preferences through discrete choice modelling

In this paper we investigate differences in bus passengers' perceptions about transit service quality through the calibration of different discrete choice logit models (multinomial, mixed and latent class models) in which users' heterogeneity is introduced. The application of these different approaches to the same experimental context is proposed, highlighting the findings emerging from the analysis. The importance of investigating passengers' perceptions can help transit operators and transportation planners prepare better investment plans; therefore, to adopt tools able to take into account the heterogeneity among users is very important for obtaining as reliable as possible service quality measures. We find that there are observed and unobserved groups of users who perceive service quality differently, and that there is heterogeneity among users in perceiving certain bus service characteristics.