Ali Tavakoli Kashani - Assistant Professor, School of Civil Engineering, Iran University of Science and Technology, Tehran, Iran
Mohammad Mehdi Besharati - Ph.D. Candidate, School of Civil Engineering, Iran University of Science and Technology, Tehran, Iran
Ahmad Mohamadian - M.Sc. in Transportation Engineering, School of Civil Engineering, Iran University of Science and Technology, Tehran, Iran,

Motorcycle crashes constitute a significant proportion of traffic accidents all over the world. The aim of this paper was to examine the motorcycle crash patterns and rider fault status across the provinces of Iran. For this purpose, 6638 motorcycle crashes occurred in Iran through 2009-2012 were used as the analysis data and a two-step clustering approach was adopted as the analysis framework. Firstly, hierarchical clustering (HC) was applied to group the provinces into homogenous clusters, based on the distribution of crash characteristics in each province. In the second step, the latent class clustering (LCC) was employed to investigate the crash patterns and rider fault status among the provinces. The provincial groupings were found to be an influential factor in the final crash clusters implying the effectiveness of the proposed framework. Results of LCC also indicated that Cluster 8 with the highest percentages of not wearing helmet, unlicensed and under 21 years old riders, had the highest percentage of fatal crashes. In addition, the motorcyclists seemed to be less responsible in the pedestrian-motorcycle crashes. Accordingly, training programs for the riders in the license issuance process about the risk of pedestrian-motorcycle crashes could help mitigate this type of crashes. Generally, analyzing the culpability in pedestrian-motorcycle crashes might be a good topic for future research. Further discussions on the crash patterns are provided. Finally, the combined use of HC and LCC should not be regarded as an alternative to the other more qualitative predictive methods, but as a preliminary analysis tool to provide insights over the road safety condition at the national level.

Hierarchical clustering, Latent class clustering, Motorcycle crashes, Motorcyclists’ fault status