Analysis of motorcyclist injury severities in motorcyclist violation crash on suburban roads of China: accommodating temporal instability and the unobserved heterogeneity in means and variances.

This study analysed motorcyclist violation (MV) crashes on suburban roads of China to investigate how determinants affect MV crash injury severity and explore the temporal stability of determinants. Crash data from Xi'an, China (2015-2018) were utilized to investigate three MV crash injury categories: no injury, minor injury and severe injury. Motorcyclist-related, crash-related, roadway-related, environment-related and time-related characteristics were analysed utilizing a group of random parameters multinomial logit models with heterogeneity in means and variances. The temporal instability was measured by performing likelihood ratio tests. Marginal effects were calculated to further illustrate the temporal variations of these factors. The study found an overall temporal instability, with some violations like alcohol-impaired riding, speeding, and unlicensed riding having significant effects on MV crash injury severity. Additionally, the study revealed a significant risk compensation mechanism of riders under adverse riding conditions. The findings provided insights and recommendations for suburban motorcycle crash prevention strategies.