Available online at https://doi.org/10.1016/j.jocm.2018.08.003
Alemi, Farzad, Giovanni Circella, Patricia L. Mokhtarian, Susan L. Handy (2018) Exploring the Latent Constructs Behind the Use of Ridehailing in California. Journal of Choice Modeling 29, 47 - 62
Emerging transportation services are quickly changing the way individuals travel by expanding the set of transportation alternatives available for a trip, allowing for more flexibility in travel schedules and providing access to transportation without incurring the costs of auto ownership. Among the most controversial and rapidly growing shared-mobility services are ridehailing services, such as those offered by Uber and Lyft in the U.S. market. In this paper, we investigate the factors affecting the adoption of ridehailing through the estimation of a latent-class adoption model that captures the heterogeneity in individuals’ tastes and preferences, focusing on members of the millennial generation and the preceding Generation X. We present a 3-class adoption model that provides better goodness of fit and interpretability of the classes compared to other model specifications that were tested. The three distinct classes have expected sample shares of 28%, 34% and 38%, respectively, and can be characterized as follows: (1) a class that is largely composed of more highly-educated, independent (i.e. who have already established their household) millennials, who has the highest adoption rate. Among other factors, the adoption of ridehailing services for the members of this class is influenced by the frequency of long-distance leisure and business-related trips made by non-car modes. The adoption of ridehailing among the members of this group is higher if they live in high-quality transit neighborhoods. (2) The second highest adoption rate is observed among the members of the class that is mainly composed of affluent individuals living with their families who are either dependent millennials or older members of Generation X. The frequency of use of transportation-related smartphone apps and the share of long-distance leisure trips made by plane affect the adoption of ridehailing for the members of this class. The members of this class also tend to adopt ridehailing if they live in neighborhoods with higher land-use mix and if they have used taxi services within the past 12 months. (3) Finally, the lowest adoption rate is observed among the members of the last class, comprising the least affluent individuals with the lowest level of education. The members of this class are more likely to live in rural neighborhoods and they rarely use ridehailing. The adoption of ridehailing among the members of this class is affected by household income, the frequency of long-distance non-car business trips, and transit accessibility as well as the use of taxi and of carsharing.
Keywords: Ridehailing; On-demand ride services; Uber/Lyft; Millennials; Latent-class choice model; Lifestyle