Investigating temporal stability of risk externalities in traffic collisions

Abstract

In traffic safety, risk externality occurs when the behavior of a certain driver increases the risk of injury to other drivers. The scope of this research is to investigate temporal trends in risk externalities related to traffic collisions by analyzing collision records from North Carolina for all two-vehicle collisions involving light-duty vehicles that occurred between January 1, 2004, and December 31, 2013. Logistic regression was used to evaluate risk externalities by developing two models for each analysis year. One model determined the probability that a driver is at least visibly injured in a collision based on the attributes of the driver, such as age, sex, or whether that driver was impaired or speeding. The other model determined the probability that a driver has caused at least a visible injury to the other driver in a two-vehicle collision based on the same attributes of that driver. The results from different years were compared to identify driving cohorts that pose an increased risk to other drivers. The findings of this study will help decision makers identify driver groups that pose an increased risk to other drivers so that more resources might be allocated to improve awareness programs and licensing procedures dedicated to those groups.