LAUSR

ABSTRACT This study aimed to explore the combination of factors and circumstances that could increase the likelihood of a pedestrian crash to be fatal. Six pedestrian crash clusters identified in our previous study on 34,178 crashes occurred in Iran, were employed in the current work and the association rules technique was used as the analysis tool. Based on the results, jaywalking or waiting beside the road on poorly illuminated locations of those segments of rural roads that pass through residential/agricultural districts substantially increased the fatality risk of pedestrian (Lift = 3.4). Therefore, appropriately illuminated locations need to be designed on these segments of rural roads for pedestrians to wait beside or cross the road safely. In case of motorcycle–pedestrian crashes, inattention to yield sign, especially in residential areas, might significantly increase the fatality risk of pedestrian (by a factor of 8.1). Also, in those crashes where the motorcyclist is under 19 years old (i.e. unlicensed) and the pedestrian is over 60, the crash could be 5.3 times more likely to be fatal. More strict law enforcement over unlicensed motorcyclists is required as they were found to involve in more severe pedestrian crashes. The combined use of clustering analysis and association rules discovery should not be considered as an alternative to other modelling techniques, but as a complementary descriptive method to be used in road safety research.