Objective: Crash injury results from complex interaction among factors related to at-fault driver's behavior, vehicle characteristics, and road conditions. Identifying the significance of these factors which affect crash injury severity is critical for improving traffic safety. A method was developed to explore the relationship based on crash data collected on rural two-lane highways in China.
Methods: There were 673 crash records collected on rural two-lane highways in China. A partial proportional odds model was developed to examine factors influencing crash injury severity owing to its high ability to accommodate the ordered response nature of injury severity. An elasticity analysis was conducted to quantify the marginal effects of each contributing factor.
Results: The results show that nine explanatory variables, including at-fault driver's age, at-fault driver having a license or not, alcohol usage, speeding, pedestrian involved, type of area, weather condition, pavement type, and collision type, significantly affect injury severity. In addition to alcohol usage and pedestrian involved, others violate the proportional odds assumption. At-fault driver's age of 25-39years, alcohol usage, speeding, pedestrian involved, pavement type of asphalt, and collision type of angle are found to be increased crash injury severity.
Practical applications: The developed logit model has demonstrated itself efficient in identifying the effect of contributing factors on the crash injury severity.
Keywords: Crash injury severity; Elasticity analysis; Partial proportional odds model; Rural two-lane highway; Traffic safety.
Copyright © 2015 Elsevier Ltd and National Safety Council. All rights reserved.