Analysis on Influencing Factors of Rear-end Crash Severity Based on Ordered Probit Model

Peng WANG,Xiao-zhao LU,Zhang-cun YAN,Min-jie ZHANG
DOI: https://doi.org/10.3969/j.issn.1002-0268.2018.04.013
2018-01-01
Journal of Highway and Transportation Research and Development
Abstract:To reduce traveler casualties and property damages result from road crashes, and provide administrative department with theoretical support of decisions,it is critical to analyze the characteristic rule of rear-end crash and explore the causational relationships between environment factor, roadway factor, vehcle factor,driver factor and rear-end crash severity. Crash records from North Carolina in 2010-2014 are screened and analysed, and finally 1 315 complete rear-end crash records are remained and recoded. It is believed that the causation of crash severity is inherently ordinal, so an Ordered Probit (ORP) model is adopted for regression analysis of various factors affecting rear-end crash severities. A backward variable elimination method is adopted to find out the statistically significant factors, and a model containing all the significant factors is obtained. Finally, the linear hypotheses test and likelihood ratio test are conducted for the model to examine its statistic rationality, and a series of indices are selected to evaluate the fitness of model data. The result shows that (1) the using of restraint, lighting condition, road alignment, traffic control condition and annual average daily traffic (AADT) is significantly related to rear-end crash severity. The model marginal effect value indicates that strict restraint usage, good light condition, rational road alignment,proper traffic control can significantly reduce the severity of rear-end crash; (2) because of the continuous collisions,heavy traffic can increase the crash severity. The significant test proves that ORP model is obviously instrumental for analyzing the rear-end crash severity and the influencing factors.
What problem does this paper attempt to address?