Analysis of Travel Choices and Scenarios for Sharing Rides
Chapter 6. Potential for Future Research
The research presented in this report spans several topics: carpooling apps, ridesharing via transportation network companies (TNC), private vehicle trip pricing, and the impacts of sharing on vehicle miles traveled (VMT). In each of these areas, the report points to several avenues for future research.
Chapter 2 describes this report's methodology for a stated-preference study anchored off real TNC trips with revealed preferences to simulate other potential decisions. Using stated-preference surveys anchored on real trips taken—because they incorporate the genuine context of trips—can likely provide much more accurate results than do stated-preference surveys more generally. While this research takes several steps forward in explaining TNC travel choices, it also points to several avenues for building on and expanding upon its findings.
First, the data presented here is only a snapshot in time of a TNC user base that is growing and changing rapidly. As this user base evolves and TNCs alter their services, this analysis could be updated to reflect the point-in-time reality of travel behavior. Longitudinal/panel research would support an understanding of how sharing behavior changes over time.
Second, this study did not have data to evaluate the effects of price and time differences that exceed those presented in Table 3 (i.e., maximum shared TNC discount of 75 percent). It is not possible, for example, to use data from the TNC survey to analyze the effects of free shared TNC trips on the rate at which people choose to use shared TNCs. Further research could test the impact of truly free shared TNC rides on mode choice.
Third, the TNC survey results do not address interactions across all modes. For that reason, this study can estimate how price and time affect a user's choice between a private and shared TNC ride, but it cannot estimate how price and time affect a user's choice among TNCs, transit, driving, carpooling, walking, bicycling, or any other mode. Similarly, the study provides no data to consider how TNC characteristics affect a user's decision to take a trip in the first place. Additional multimodal discrete choice analysis is necessary to properly nest these decisions within an integrated mode choice model.
Regarding the analysis presented in chapter 3, the exploration of carpooling apps points to the potential benefit of additional research into the impacts of campaigns that promote carpooling, and specifically randomized controlled experiments on the impacts of incentives, and the altering of such incentives, on carpooling rates.
Regarding the analytic model that enables scenario assessments as presented in chapter 4, assessments could be improved and expanded on in several ways. If data and analysis become available, the analytic model could be updated to analyze the effects of more extreme price and travel time differences between private and shared TNC trips (currently, the maximum allowable input values for scenarios 1 and 2 are capped at $2.80/mile and 2.5 minutes/mile, respectively).
Further, if data supporting additional discrete choice analysis were available, the analytic model could be updated to consider interactions across all modes for each scenario. By posting the model on the Intelligent Transportation Systems (ITS) CodeHub,1 FHWA enables collaborative development and improvements to the model. Readers are permitted and encouraged to collaboratively share any updates, modifications, and improvements to the model as new data sources or city-level information changes.
This research report does not explore factors (beyond cost and travel time) that, according to other research, sometimes make people averse to sharing a vehicle with strangers, such as safety, privacy, and convenience. The impact of each of these factors, if better understood, might suggest additional approaches for policymakers to increase the public's willingness to share rides, whether in private vehicles or through TNC services.