LAUSR

Abstract Transit use in many major US cities has recently been declining, often entailing shifts from transit to less sustainable modes of travel, such as private automobiles or ride-hailing. To address this challenge, it is critical to reduce the attrition of existing riders, which requires understanding the determinants of transit user satisfaction and the link between dissatisfaction and attrition. To date, few studies have examined the effect of satisfaction on observed transit use over time. We use a unique panel data set where satisfaction with various factors and changes in transit use were measured one year apart, allowing us to quantify revealed behavior changes and determine which aspects of satisfaction were most predictive. We present two integrated choice and latent variable models (a panel model and a predictive model) to describe the relationship between satisfaction and transit user loyalty, measured as transit use frequency and retention rate over a year. We found that satisfaction with operations significantly affect the level of transit use, but satisfaction with the travel environment and life event do not have a significant impact. Users’ self-reported reasons for attrition corroborates the above findings and offers additional insights on observed mode shifts, such as the effect of competing ride-hailing services and bicycling on transit use. Our predictive model, together with an accompanying sensitivity analysis, can be used to forecast attrition as a function of satisfaction. We conclude by recommending strategies to increase user retention and reduce shifts to less sustainable modes.