Justifying toll payment with biased travel time estimates: Behavioral findings and route choice modeling

Travel time is considered a main factor in travel demand modeling, taken almost exclusively in its objective form. Given that people base their decisions on perceptions rather than on objective attributes, this study aims to examine whether forecasting could benefit from the inclusion of subjective time. A second objective is to test and disentangle drivers' tendency to estimate toll trips as shorter than non-toll trips. In a field experiment, shoppers departing a mall described their intended route and an alternative route, one of which via a nearby toll road. Participants provided time estimates for the two routes. Objective times were collected via smartphone tracking apps and Waze. All 386 participants were paid 10 NIS. To test the effects of toll self-payment, some participants were told that this sum was to cover the toll payment, and others that it was a participation fee. Consequently, some participants who had not intended to drive via the toll road were paid to do so. Results showed that drivers who intended to drive via the toll road exaggerated their time savings compared to drivers who did not intend to drive it but eventually did, suggesting drivers' time estimates reflected an attempt to justify their route choice. Self-payment decreased estimated toll time savings. Drivers' toll-route choice was estimated using binomial logit models, revealing better fit for models based on estimated, rather than objective, time. We concluded that estimated time data entails unique valuable information regarding drivers' preferences, rendering its integration in toll-route modeling constructive and beneficial.