Investigation of preferences for ride-sourcing taxis: Compromise effect and regret-based latent class modeling

This study explores preferences for ride-sourcing taxis, which represent a compromise option between traditional taxis and ride-sourcing services. To capture the additional market share for the ride-sourcing taxi in line with the theory of compromise effects, random regret minimization (RRM) models are applied. We considered both the classical RRM (C-RRM) model and the generalized RRM (G-RRM) model. We also proposed a new perspective on the derivation of the G-RRM model, in that this model is derived from the parallel translation of the C-RRM model. The data used in this study is based on a stated preference survey that measured not only walking time, fare, and cancellation fees but also uncertain waiting time regarding the ride-sourcing taxi. Also, the effects of trip contexts and latent attitudes toward emerging transportation technology and innovative service were explored. Our results, based on a latent class G-RRM model, reveal that most respondents are either high compromisers or low compromisers in the choice context. The high compromisers and low compromisers showed regret-aversion behavior to process the information of the explanatory variables in the survey. The low compromisers showed the greatest degree of regret-aversion behavior for uncertain waiting time.