Trends in highway safety: effects of an aging population on accident propensity

Aging of the United States population has a potentially adverse effect on highway safety. A key question is whether the known deterioration of driving skills with aging will be compensated by other factors, especially improved learning and attitudinal experiences of more recent cohorts. We investigate effects of driver age, cohort, and gender on accident propensity and evaluate confounding effects of year, location, and lighting. The accident propensity of different groups of drivers is measured, using a database of two-vehicle accidents, by the ratio of the number of at-fault drivers of a specific group to the corresponding number of not-at-fault drivers. Logistic regression modelling determines the statistical significance of the findings. The analysis reveals the following statistically significant effects: (a) middle-aged drivers are safer than younger drivers who, in turn, are safer than older drivers; (b) female drivers are safer on average than male drivers; (c) younger female drivers are safer than younger male drivers; (d) older male drivers are safer than older female drivers; (e) more recent cohorts of older drivers are safer than more distant cohorts; and (f) more distant cohorts of younger drivers are safer than more recent cohorts. We conclude that driver cohort provides a plausible explanation for many of the measurable, time-related accident trends that are observed. Accordingly, older drivers will be safer in the future because they will have acquired basic driving skills and attitudes in increasingly more automobile-dominated times. At the same time, the cohort effect appears small relative to other time related effects, notably aging, and older drivers will continue to be a high-risk component of the driving population and to require special consideration in driver education and licensing and in highway design and operations.