Substitution or Complementarity? A Latent-Class Cluster Analysis of Ridehailing Impacts on the Use of Travel Modes in Southern U.S. Cities

Abstract

In this study, the authors explore the heterogeneous impacts of ridehailing in a sample (N=1,439) of users from three regions in southern U.S. states: Phoenix, Arizona; Atlanta, Georgia; and Austin, Texas. In so doing, they apply a latent-class cluster analysis on indicators of changes in the use of various travel modes as a result of ridehailing adoption, with covariates of socioeconomics, demographics, a land-use attribute, and individual attitudes. The authors find four distinctive classes of behavioral changes in response to ridehailing. About half of the sample (50.3%) belong to middle-class suburban families, whose members use ridehailing rarely, and do not make much change to their travel routines as a result of its adoption. The second largest class is moderate-income urbanites (24.6% of the sample), whose members undergo many changes in their mode use, but for other reasons. Young affluent urbanites (14.3%) frequently hail a ride, and as a result, reduce their use of private vehicles while making more trips by public transit and active modes. Interestingly, car-deficient suburbanites (10.8%) often use ridehailing, mainly for trips that would otherwise have been made by public transit or active modes, and consequently reduce their use of these less-polluting modes while enjoying enhanced mobility. This study reveals substantial heterogeneity in ridehailing impacts, which were masked in previous studies that focused on average impacts, and it suggests that policy responses should be customized by users’ socioeconomics and residential neighborhoods where their trips happen.