Analyzing the severity of motorcycle crashes in North Carolina using highway safety information systems data

A useful insight into single-vehicle, single-rider motorcycle crash injury severity is discussed in an effort to develop effective strategies and countermeasures. The dependent variable, i.e., the severity sustained by the rider given a motorcycle crash has happened, is modeled by considering the ordered nature of crash severity and using the partial proportional odds (PPO) model. According to this assumption, the effect of the explanatory variables entered into the model is assumed to be constant across each ordinal category, and the only difference between the regression lines is the cut-off point for the threshold. In terms of the temporal characteristics of crashes, the analysis of the results showed that having a crash during the summer season is associated with higher likelihood of severe injuries and, conversely, in winter is associated with lower likelihood of severe injuries. This variation in the severity of crashes can be attributed to the motorcycle traffic volume fluctuation during various seasons. As for the type of setting, the analysis results indicated that the crashes occurred in rural areas, compared to those that happened in urban areas, and are more severe, with an increase in the likelihood of severe crashes by 47.5 percent. Driving under adverse weather conditions reduces the likelihood of severe crashes by 27.4 percent and increases the possibility of no/possible injuries crashes by 20.3 percent. For wet roadway surface conditions, these numbers dropped to 25.5 percent and 18.8 percent, respectively.