This paper presents a quasi-induced exposure analysis of 4 years of crashes involving older drivers in Kentucky. Single- and multivehicle crashes were disaggregated according to the number of passengers: 1) no passenger, 2) 1 passenger, and 3) 2 or more passengers. Results show the presence of 2 or more passengers was found to negatively impact the probability that drivers 75 years of age or older were at fault in crashes. Several potential factors were studied for interactive effects with passengers: vehicle occupant gender mix, time of the day, road curvature, grade, and number of lanes. The negative impact of passengers increased for some geometric road conditions. However, older drivers were found to be safer at night when carrying 2 or more passengers. The presence or absence of passengers was not found to affect the 65- to 74-year-old driver group. Groups of male vehicle occupants with a 75+ male driver were found to have high single-vehicle crash rates. These results are among the first to directly consider the effect of passengers on the crash-causing propensity of older drivers and the findings suggest more work is warranted to consider causes for the crash rate differences.
Abstract