Heterogeneity in travel episode satisfaction of public transport trips: Results of a latent class regression model

Rasouli (2013) formulated and estimated models that predict judgments (satisfaction) of public transport travel episodes as a function of trip characteristics, activity envelopes, travel party and the nature and extent of multitasking. In this sequel, the topic of heterogeneity in travel episode satisfaction of public transport trips will be addressed. A sample of about 1000 respondents carried a GPS device up to a maximum of three months. A Bayesian Belief Network was used to classify epochs into travel and activity episodes, and to impute transport model and activity type. A Web-based prompted recall instrument was then used to invite respondents to verify and if needed correct the resulting imputed activity-travel diaries. At the same time, using the same instrument, they were asked to rate their satisfaction with each travel episode on a 10-point scale, and provide additional information about travel party, and the nature and extent of multitasking. Satisfaction ratings related to public transport were selected for further analysis. Heterogeneity was examined by estimating a latent class model. Results suggest that a substantial amount of heterogeneity exists in the data. Distinct classes can be identified