Effect of Quantitative and Qualitative Factors on Elder Driver Intersection Crashes

Past research has shown that intersections are particularly difficult for older drivers to negotiate. To investigate potential contributing factors, standard database elements were paired with visual inspections of video log images to evaluate factors such as visual clutter, approach sight distances, and the condition and visibility of pavement markings. For comparative purposes, the paper also examines the effect of the same factors on crashes in which non-elderly drivers were found to be at fault. The study set includes all intersections on state roads in Florida that had undergone at least one fatal traffic crash in the year 2000, primarily because of information available from a companion study. Thirty-two independent variables were selected to describe the roadway geometrics, traffic characteristics, pavement, signage, and control devices. A variety of regression models were investigated, and best-fit measures, including Akaike and Bayesian Information Criteria, favored negative binomial models reduced by forward stepwise selection for both elder and non-elder crashes. Fourteen variables were found to have significant effect on crash likelihood. Major operation and design aspects, such as high traffic counts and absence of signalization and medians, appear to be much more significant measures of crash likelihood for both older and younger drivers than other features or enhancements. Older drivers had more crashes when pavements were not marked with advanced lane assignment marking and when skid resistance was lower. However, younger drivers were more prone to traffic crashes at intersections with lower sign reflectivity and hanging signals, factors which did not impact older drivers. Further, several qualitative safety features, including advanced warning signs recommended by the Florida Elder Road User Program, had greater impacts on non-elderly drivers. Potentially counter-intuitive results, such as decreased crash likelihoods with higher speed limits, decreased turning sight distances, and presence of raised pavement markings, might be caused by selflimiting behavior among drivers when presented with safety concerns; however, additional research is recommended.