Understanding the drowsy driving crash patterns from correspondence regression analysis

Abstract

Abstract: Drowsy driving-related crashes have been a key concern in transportation safety. In Louisiana, 14% (1,758 out of 12,512) of police-reported drowsy driving-related crashes during 2015–2019 resulted in injury (fatal, severe, or moderate). Amid the calls for action against drowsy driving by national agencies, it is of paramount importance to explore the key reportable attributes of drowsy driving behaviors and their potential association with crash severity. Method: This study used 5-years (2015–2019) of crash data and utilized the correspondence regression analysis method to identify the key collective associations of attributes in drowsy driving-related crashes and interpretable patterns based on injury levels. Results: Several drowsy driving-related crash patterns were identified through crash clusters – afternoon fatigue crashes by middle-aged female drivers on urban multilane curves, crossover crashes by young drivers on low-speed roadways, crashes by male drivers during dark rainy conditions, pickup truck crashes in manufacturing/industrial areas, late-night crashes in business and residential districts, and heavy truck crashes on elevated curves. Several attributes – scattered residential areas indicating rural areas, multiple passengers, and older drivers (aged more than 65 years) – showed a strong association with fatal and severe injury crashes. Practical Applications: The findings of this study are expected to help researchers, planners, and policymakers in understanding and developing strategic mitigation measures to prevent drowsy driving.