Assessing the potential for a new suburban rail: Do user characteristics matter?

Abstract

This article examines the role user-specific characteristics can play on assessing a new public transit service. The case under consideration is the new Western Suburban Rail service in Thessaloniki, Greece, which will cover a suburban area. Binary logit models were developed in order to investigate the potential demand for this project. A revealed and stated preference questionnaire survey was conducted. The sample of the survey was stratified according to the three current travel options, namely car, regular-fare bus and low-fare bus. The revealed preference survey included questions about trip and socio-economic characteristics of each user. The stated preference part examined the potential for the new mode in comparison to the current travel modes through changes in travel cost, travel time and number of transfers. Six binary logit choice models were developed in total, two for each user group. For each group, a simplified model was initially calibrated only with the mode-specific variables; at a next step user-specific variables were added and a second extended model was created. The results show significant relationship of mode- and user-specific variables to the final choice for all the three user groups. Likelihood ratio tests were used to examine whether user-specific attributes improve the overall fit of the models or not, compared to traditional choice models of travel time, travel cost and number of transfers. The outcomes show that the inclusion of user-specific variables improved the overall fit and the explanatory power of the models.