Latent class analysis of the effects of age, gender, and alcohol consumption on driver-injury severities

• We examined the differences in driver-injury severity between drivers impaired and not-alcohol-impaired, while taking into consideration the role of age and gender.
• A latent class multinomial logit modeling approach was used to capture unobserved heterogeneity.
• Considerable heterogeneity was found across the sub-populations.
• The estimated parameters differed in sign and magnitude across classes.

This study explores the differences in driver-injury severity between drivers impaired and not-alcohol-impaired, while taking into consideration the role of age and gender. Using data from single-vehicle crashes from Illinois' Cook County over an eight-year period from January 1, 2004 to December 31, 2011, separate alcohol-impaired and not-alcohol-impaired models of driver-injury severity (with possible outcomes of no injury, minor injury, and severe injury) were estimated for younger male, older male, younger female, and older female drivers (those younger than 31 years old were considered younger drivers, and those 31 years old and older were considered older drivers). In addition to considering driver age, alcohol condition, and gender, a wide range of variables potentially affecting crash severity was considered, including a number of variables relating to highway attributes, vehicle characteristics, and environmental conditions. Using a latent class multinomial logit modeling approach to capture unobserved heterogeneity, estimation results show that there were substantial differences across age/gender groups in the absence/presence of alcohol. In addition, among others, particularly complex relationships were uncovered with regard to the impact of alcohol consumption, safety-belt effectiveness, roadway type, distracted driving, vehicle occupancy, and the effects of airbag deployment on injury severity.