Injury Severity of Pedestrians at Mid-blocks: A Random Parameter Ordered Probit Approach

Abstract

Improving pedestrian safety is one of the top priorities of different transportation agencies. However, vehicle-pedestrian crashes continue to claim a large toll on fatalities. According to NHSTA, the 5,987 pedestrian fatalities amount to 16 percent of all traffic fatalities in the U.S in 2016. While most of past studies focused on pedestrian crashes at intersections, only 18 percent of the total vehicle-pedestrian crashes happened at the intersections. The vehicle-pedestrian crashes at mid-block crossings were largely ignored. The objective of this study is to examine the contributing factors of pedestrians’ injury severities at mid-blocks and uncover the effects of pedestrians’ pre-crash behavior. To this end, crash records involving pedestrian at the mid-blocks were acquired from the General Estimates System. The pedestrians’ characteristics, their pre-crash behaviors, road geometrics, and crash characteristics were all processed together. A random parameter ordered probit model was developed to account for unobserved heterogeneities in the data. The results show that time of the crashes, crash characteristics, roadway features, and pedestrian characteristic are all significant in determining the injury severity levels of pedestrians. In addition, after accounting for these variables, pedestrians’ pre-crash behavior also plays a significant role in severity outcomes. The findings from this study bears policy potentials to help devise informed injury mitigation strategies at the mid-block crossings.