Data-Driven Decision Making: The Case of Ridesharing With Implications for Engineering Managers

Abstract

As data-driven decision making becomes prevalent, research needs to provide more evidence to direct user decision making, particularly concerning transportation systems. In concurrence, ridesharing has been touted to reduce driving-while-intoxicated fatalities, albeit prior studies have provided inconsistent findings. A limitation of prior research on this topic is lacking adequate experimental controls while addressing the impact of potential confounds. This issue may affect potential assumptions and conclusions on whether the deployment of ridesharing services has led to a considerable reduction in driving-while-intoxicated fatalities. The present article leverages statistical modeling to control age, education, vehicle miles traveled, and metropolitan size. It reveals that ridesharing represented a 13.8% decline in driving-while-intoxicated fatalities among youths’ ages 17–34, but without significantly affecting drivers’ ages 35–65. Also, the results suggest that city population, vehicle miles traveled, and educational attainment can affect younger adults, whereas the same features were not significant for older adults. Furthermore, the article suggests that the initiation of UberX can serve as a ride-planning option to reduce driving-while-intoxicated fatalities among younger rather than older drivers. Based on the analysis results, multiple implications for transportation platform and software engineering managers are provided, especially in the areas of dispatch algorithms, requirement analysis, and ridesharing security and safety.