Exploring the impact of truck traffic on road segment-based severe crash proportion using extensive weigh-in-motion data

•Used extensive Weigh-in-Motion (WIM) data to measure truck traffic characteristics, focusing particularly on vehicle weight distribution and truck traffic volume.•Constructed models of severe crash proportions based on road segments, utilizing the concept of homogeneous sections.•Established a link between vehicle weight, as determined by WIM data, and the severity of crashes on particular road segments.•Employed Fractional Regression Models (FRMs) to manage ratio response variables effectively.•Implemented a systematic approach to choose the optimal link function and establish the structure of the Fractional Regression Models. Fixed proportions by severity assumption in Highway Safety Manual could be violated since the proportions of severe crashes are likely to be affected by truck traffic characteristics. Previous studies often used truck proportion as the key indicator of truck traffic. However, it considered different trucks the same regardless of their actual weight. Therefore, this paper aimed to explore the impact of truck traffic characteristics, especially actual weight, on the proportions of severe crashes on road segments while controlling for other contributing factors. Extensive Weigh-in-Motion (WIM) data from five-year (2011–2015) 88 WIM stations in New Jersey were utilized to capture detailed vehicle weight information and other truck traffic-related characteristics. Road features, traffic volume, and crash data were also collected and aggregated for road segments. To account for the bounded nature of Fatality and Injury Proportion (FIP), one-part and two-part Fractional Regression Models (FRMs) were developed, and the link functions were appropriately selected based on corresponding statistical tests. The results show that the mean of vehicle weight was significant and positively related to the FIP of nonzero-FIP road segments while controlling for other contributing factors. For the road segment with a nonzero FIP, if the mean of vehicle weight increased by 1 kip, the total crash FIP, single-vehicle crash FIP, and multiple-vehicle crash FIP for the road segment with nonzero FIP increased by 3.3%, 3.4%, 2.2% respectively. This study contributes to the literature by building a link between actual vehicle weight measured in the traffic flow and road segment crash severity.