A generalized ordered probit model for analyzing driver injury severity of head-on crashes on two-lane rural highways in Malaysia

Abstract

The number of fatal crashes in Malaysia has steadily been increasing, exceeding 7,000 deaths every year since 2016. Compared to other collision types, head-on crashes account for about 12% of total crashes in Malaysia yet are responsible for 17% of total fatal crashes. This indicates that head-on crashes are among the most fatal collision types. However, very few studies have been focused on the injury severity of head-on crashes. Therefore, this study aims to investigate the factors that are hypothesized to be associated with driver injury severity in head-on crashes. To do this, a generalized ordered probit (GOP) model was applied using 4-year (2009–2012) crash data collected on two-lane rural highways in Malaysia. The results show that the factors of dark not-lighted condition, horizontal curvature, heavy vehicle percentage, light vehicle traffic, posted speed limit, undulating terrain, improper overtaking maneuvers, not wearing seatbelt, and vehicle age contributed to more severe injuries, whereas rainy weather condition was found to be the only factor that was associated with less severe outcome. The findings of this study can help safety authorities understand the relationship between various risk factors and head-on crash severity.