The role of reduced visibility in nighttime road fatalities

Abstract

Two quasi-experiments investigated the contribution of reduced visibility to fatal accidents recorded by the U.S. Fatal Accident Reporting Systems from 1980 through 1990. Quasi-Experiment 1 evaluated 104,235 accidents that occurred during morning and evening time periods, called Twilight Zones, during which natural illumination varied systematically in conjunction with the annual solar cycle. Fatal accidents were found to be overrepresented during darker portions of the Twilight Zones. This finding was not related to time of day, day of week, or drivers’ consumption of alcohol. The contribution of reduced visibility was also indicated by higher overrepresentation of fatal accidents in low illumination under adverse atmospheric conditions and with pedestrians and pedalcyclists as opposed to all other accidents. Reduced visibility was more important than drivers’ drinking as a contributor to fatal pedestrian and pedalcycle accidents, while the reverse pattern was found for all other fatal traffic accidents. Quasi-Experiment 2 assessed the role of seasonal variables other than natural illumination by comparing the monthly distributions of 337,726 accidents recorded during three time periods: the Twilight Zones plus equal-duration control periods of Daylight and Darkness. The distribution of fatal nonpedestrian accidents exhibited no substantial variation across months in any of the test periods. The incidence of fatal pedestrian/pedalcycle accidents covaried with natural illumination during the Twilight Zones, while showing no variation during Daylight and Nighttime Control periods, confirming that visibility is a key factor in pedestrian and pedal cycle accidents. The present findings provide new evidence for both (1) the importance of visibility as a major contributing factor in fatal pedestrian and pedal cycle accidents, and (2) the relative success of previous efforts to deal with the difficulties of nighttime driving in other classes of fatal accidents.