Identifying distracted driving hotspots using an event-to-crash conversion method: a case study from New Jersey

Abstract

Identifying crash hotspots is a tool for prioritizing safety countermeasures. However, safety–critical human factors like distracted driving or speeding can often happen without leading to a crash. Hence, a matrix is needed that could translate the contributing safety–critical events into crashes. Quantifying distraction events into crashes could bridge this gap, and comprehensive observational data on drivers’ distractions could help make this quantification. This study aims to build a framework for calculating the relationship between crashes and events by collecting distraction events in selected high crash corridors in New Jersey. On-road observational data was collected to capture distracted driving behaviors using a moving test vehicle. Crashes occurring at the study routes during 2015–2019, with ’distraction’ as the only ’driver contributing factor,’ were collected. Afterward, the distraction event-to-crash ratio was calculated for each 5-mile road segment, which helped determine the study routes’ riskiest segments (hotspots). After converting crashes to events, the potential crash cost savings for various levels of reduction in distraction events (5%, 10%, 20%, and 30%) were calculated. The results revealed that US9, US1, US22, and US130 were the riskier corridors, with at least one crash from every 30,000 distraction events. It was also found that roads with signalized intersections, three-lane roads, rural interstates, minor arterials, and positive medians had the highest number of hotspots. Non-parametric tests confirmed that arterial roads and roads with signalized intersections have significantly lower event-to-crash equivalence than the other types of roads. This study will help state and local agencies identify distracted driving hotspots in their jurisdiction and enable them to allocate resources for education, enforcement, or engineering solutions more efficiently.