A Temporal Analysis of Driver-Injury Severities in Crashes Involving Aggressive and Non-aggressive Driving

Document Type


Publication Date



Aggressive driving, Temporal stability, Injury severity, Mixed logit model with heterogeneity in means and variances

Digital Object Identifier (DOI)



Aggressive driving has become a national traffic-safety concern, with increasing congestion and other stress-inducing factors making it more likely drivers take out their frustrations by driving aggressively. Looking at single-vehicle crashes, this study investigates differences between resulting crash-injury severities when aggressive and non-aggressive driving behavior is observed, and how these differences changed over time. Using three years of crash data from 2015 to 2017 (inclusive), random parameters multinomial logit models with unobserved heterogeneity in means and variances were estimated. The available crash data included a wide variety of factors known to influence driver-injury severity including data related to the crash, vehicle, driver, spatial and temporal characteristics, roadway attributes, and traffic volume. Model estimates show that there were significant differences in driver-injury severities resulting from aggressive and non-aggressive driving, and that the effect of factors that determine injury severities changed significantly over time (statistically significant temporal instability). However, it is noteworthy that crashes involving non-aggressive drivers had many explanatory variables that produced temporally stable marginal effects, whereas crashes involving aggressive drivers had only one such variable (restraint belt usage). This suggests the possibility that temporal instability found in many recent safety studies may be driven by a subset of the crash population, and that there may be temporal stability in many crashes. Exploring this possibility is a promising direction for future empirical investigation.

Was this content written or created while at USF?


Citation / Publisher Attribution

Analytic Methods in Accident Research, v. 27, art. 100128