• Applies a random parameters logit model with heterogeneous means and variances. • Examines effects of rider, vehicle, environment, and crash factors on severity. • Reveals how interaction effects amplify crash severity under multiple risk factors. • Evaluates policy interventions with counterfactual to reduce crash severity. • Provides empirical evidence to guide targeted electric motorcycle safety policies. Electric motorcycles are highly vulnerable road users in complex roadside environments, yet the mechanisms governing their crash severity remain insufficiently understood. Using crash data from the UK STATS19 database (2020–2024), this study applies a random-parameters binary logit model with heterogeneity in means and variances (RPBL-HMV) to analyze the determinants of electric motorcycle crash severity, classified as serious injury or fatality versus slight injury. The results reveal substantial unobserved heterogeneity in environmental effects, with dry road surface identified as a key random parameter whose influence varies across crashes. Male riders, young riders, and high-class roads significantly increase the likelihood of severe injuries, whereas male casualties and junction locations are more strongly associated with slight injuries. Multi-vehicle crashes pose a higher severity risk than single-vehicle crashes due to more complex crash interactions. Electric motorcycle crashes also exhibit a distinct low-energy injury pattern, in which skidding, overturning, and crashes with roadway objects are frequent but generally associated with slight injury. Interaction effects further indicate elevated severity risks for crashes involving young male riders and male riders on high-class roads. Counterfactual simulations based on the variation amplitude of crash severity probability (VACSP) demonstrate that temporal interventions yield the greatest benefits, with morning-period risk management reducing severe injury probabilities, while police presence provides a stable mitigating effect. By accounting for unobserved heterogeneity and interaction effects, this study improves understanding of electric motorcycle crash severity and provides evidence to guide targeted traffic safety management and policy.
Wang et al. (Thu,) studied this question.