Exploration of the Heterogeneity among Elderly Drivers by Analyzing Traffic Crash Data: A Case Study in Pennsylvania, USA

Abstract

With population aging and life expectancy increasing, elderly drivers have been increasing quickly in the United States and the heterogeneity among them with age is also increasingly non-ignorable. Based on traffic crash data of Pennsylvania from 2011 to 2019, this study was designed to identify this heterogeneity by quantifying the relationship between age and crash characteristics using linear regression. It is found that for elderly driver-involved crashes, the proportion leading to casualties significantly increases with age. Meanwhile, the proportions at night, on rainy days, on snowy days, and involving driving under the influence (DUI) decrease linearly with age, implying that elderly drivers tend to avoid traveling in risky scenarios. Regarding collision types, elderly driver-involved crashes are mainly composed of angle, rear-end, and hit-fixed-object collisions, proportions of which increase linearly, decrease linearly, and keep consistent with age, respectively. The increase in angle collisions is primarily attributed to more crashes at stop-controlled intersections. The findings suggest that it may be inappropriate to take elderly drivers as homogeneous or simply categorize them into several age groups. Instead, regarding elderly drivers, age should be taken as continuous in future studies to display their linearly changing trends. This is one of the pioneering studies exploring the heterogeneity across elderly drivers with age with solid data analysis. The findings are expected to provide new insights for agencies to develop customized countermeasures regarding elderly traffic safety in the aging society.