Modeling E-Bike Crash Severity by Accounting for Unobserved Heterogeneity in China

This study investigates factors that significantly contribute to the severity of electric bike crashes. Two months e-bike crash data were collected in city of Ningbo, China. A random parameters multinomial logit model is developed to account for the potential unobserved heterogeneous effects. The Markov chain Monte Carlo simulation-based full Bayesian approach is employed to estimate the model parameters. Both parameter estimates and odds ratio are developed and used to interpret the model. The estimation results show that the impacts of contributing factors which significantly affect the severity of the e-bike crash differ across severity categories. Modeling results show that age, gender, e-bike behavior, license plate use, bicycle type, location, and speed limit are statistically significant contribute to the severity of electric bike crashes. The variables of gender, e-bike behavior, bicycle type, and speed limit are found to have heterogeneous effects, appearing in the form of random parameters in the statistical model.