Exploring Factors Influencing Injury Severity of Bicyclists in Crashes with Motor Vehicles – Combined Latent Class Clustering and MNL Model

With the increase in travel costs brought about by motor vehicles and the enhancement of environmental awareness among travellers, the rise in bicycle usage has grown globally. However, cyclists are more vulnerable in the process of a crash with motor vehicles, and tend to have more serious injuries compared to motorists after a crash. Therefore, cycling safety needs to be taken seriously, and it is crucial to analyse the factors that influence the severity of injuries sustained by cyclists in collisions. In this paper, we investigated the bicycle-motor vehicle (BMV) crash statistics in North Carolina, U.S.A., conducted preliminary screening of the variables covered in the data, classified the data into clusters by applying the latent cluster analysis, and used the multinomial logit (MNL) model based on the latent category to explore the main influences on the severity of rider injuries in BMV crashes in each cluster and the mechanism of the factors’ effects. The results of this paper explain the contribution of the influencing factors to the severity of cyclists’ injuries under each cluster characteristic. The results can be used for the special management and optimisation of the causative factors affecting the severity of cyclists’ injuries under different cluster categories.