Are Drivers with a History of Drunk Driving More Dangerous? Insights From Quasi-Induced Exposure and Random Parameter Models

Alcohol-impaired driving remains a major contributor to traffic fatalities and injuries, yet little is known about whether drivers with a history of driving while intoxicated (DWI) differ from those without such histories in their likelihood of being at fault in crashes. This study examines differences in at-fault outcomes between DWI and non-DWI drivers and identifies behavioral factors influencing at-fault probability. Drivers were classified into DWI and non-DWI cohorts based on prior drink-driving offenses, and a two-stage approach was employed. First, the quasi-induced exposure method compared the influence of risky driving behaviors on crash rates across cohorts. Second, a random parameters logit (RPL) model identified factors affecting the probability of being at fault among drivers with DWI histories. The results indicated that drivers with a history of DWI were more likely to be identified as at-fault drivers in crashes involving risky driving behaviors. In alcohol-related crashes, drivers with DWI histories exhibited a 2.06-fold higher at-fault rate than non-DWI drivers, with alcohol use itself increasing at-fault probability by 5.92%. Similar trends were observed for hit-and-run, speeding, and drug-impaired driving. Furthermore, the RPL analysis revealed that vehicle type (utility vehicles and light trucks), intersections, and left-turn maneuvers significantly contributed to being at fault among drivers with DWI histories. These findings demonstrate that drivers with prior DWI offenses have a higher propensity for risky driving behaviors, elevating their likelihood of being at fault, and provide evidence to support targeted interventions aimed at improving traffic safety among high-risk driver populations.