Effects of helmet usage on moped riders' injury severity in moped-vehicle crashes: Insights from partially temporal constrained random parameters bivariate probit models

Chenzhu Wang,Mohamed Abdel-Aty,Pengfei Cui,Lei Han
DOI: https://doi.org/10.1016/j.aap.2024.107800
Abstract:Mopeds are small and move unpredictably, making them difficult for other drivers to perceive. This lack of visibility, coupled with the minimal protection that mopeds provide, can lead to serious crashes, particularly when the rider is not wearing a helmet. This paper explores the association between helmet usage and injury severity among moped riders involved in collisions with other vehicles. A series of joint bivariate probit models are employed, with injury severity and helmet usage serving as dependent variables. Data on two-vehicle moped crashes in Florida from 2019 to 2021 are collected and categorized into three periods: before, during, and after the COVID-19 pandemic. Crash involvement ratios are calculated to examine the safety risk elements of moped riders in various categories, while significant temporal shifts are also explored. The correlated joint random parameters bivariate probit models with heterogeneity in means demonstrate their superiority in capturing interactive unobserved heterogeneity, revealing how various variables significantly affect injury outcomes and helmet usage. Temporal instability related to the COVID-19 pandemic is validated through likelihood ratio tests, out-of-sample predictions, and calculations of marginal effects. Additionally, several parameters are noted to remain temporally stable across multiple periods, prompting the development of a partially temporally constrained modeling approach to provide insights from a long-term perspective. Specifically, it is found that male moped riders are less likely to wear helmets and are negatively associated with injury/fatality rates. Moped riders on two-lane roads are also less likely to wear helmets. Furthermore, moped riders face a lower risk of injury or fatality during daylight conditions, while angle crashes consistently lead to a higher risk of injuries and fatalities across the three periods. These findings provide valuable insights into helmet usage and injury severity among moped riders and offer guidance for developing countermeasures to protect them.
What problem does this paper attempt to address?