Train commuters’ scheduling preferences: Evidence from a large-scale peak avoidance experiment

• Published in: Transportation research. Part B: methodological 2016-01, Vol.83, p.314-333
• Main Authors: Peer, Stefanie, Knockaert, Jasper, Verhoef, Erik T.
• Format: Article
• Language: English
• Subjects: Avoidance; Customizing; Delay; Estimates; GPS data; Revealed preference data; Schedules; Scheduling; Scheduling decisions; Smartphone app; Smartphones; Train commuters; Transportation; Valuations of trip attributes
• ISSN: 0191-2615; 1879-2367
• DOI: 10.1016/j.trb.2015.11.017

•Dutch train commuters were offered monetary rewards for traveling off-peak.•Approximately 1000 train commuters participated in the peak avoidance experiment.•GPS data from a customized smartphone app were used to measure travel behavior.•When the rewards were introduced, the relative share of peak trips decreased by 22%.•Plausible estimates for the monetary valuations of the trip attributes are derived. We study the trip scheduling preferences of train commuters in a real-life setting. The underlying data have been collected during large-scale peak avoidance experiment conducted in the Netherlands, in which participants could earn monetary rewards for traveling outside peak hours. The experiment included ca. 1000 participants and lasted for multiple months. Holders of an annual train pass were invited to join the experiment, and a customized smartphone app was used to measure the travel behavior of the participants. We find that compared to the pre-measurement, the relative share of peak trips decreased by 22% during the reward period, and by 10% during the post-measurement. By combining multiple complementary data sources, we are able to specify and estimate (MNL and panel latent class) departure time choice models. These yield plausible estimates for the monetary values that participants attach to reducing travel time, schedule delays, the number of transfers, crowdedness, and unreliability.