Investigating factors affecting injury severity in bicycle–vehicle crashes: a day-of-week analysis with partial proportional odds logit models

Abstract

Cyclists are vulnerable road users and prone to experience severe injury when accidents occur. Both driving and riding behaviors of drivers and cyclists could vary in different days of week, which further influences the injury severity of crashes involving cyclists. This study aims to investigate the factors that affect injury severity in crashes involving cyclists on weekdays and weekends separately using police-reported data ranging from 2007 to 2018 in North Carolina. The impact of cyclist, driver, vehicle, road, environment, and crash characteristics on injury severities are explored. Ordered logit model and partial proportional odds logit models are developed respectively for the injury severities of crashes on weekdays and weekends. Different sets of significant factors are identified for weekdays and weekends. For the common factors identified for both weekdays and weekends, the influencing extent of significant variables varies significantly between weekdays and weekends. Older-aged cyclists, riding direction, pickup, older-aged drivers, male drivers and periods of 0–5:59 and 10–14:59 are only found significant on weekdays while speed limits of 45–55 mph, piedmont areas, commercial development, head-on, and non-roadway locations are only found significant for injury severities of crashes on weekends. Speed limits, time of day, alcohol usage are found to have different or even opposite impacts on the injury severities of crashes on weekdays and weekends.