Unequal Streets, Unequal Stations: Active Transportation Safety Disparities in the SCAG Region

Abstract

This study uses multivariate regression to determine correlates of pedestrian- and bicycle-involved crashes based on 14 demographic and built environment factors. At the census tract level, higher poverty rate and larger Hispanic/Latino population are statistically significant crash predictors of both pedestrian and bicycle crashes, even when controlling for job density, commute mode, and other variables related to pedestrian and bicycle exposure. Within ¼ mile of Metro stations and within ½ mile of Metrolink stations, the poverty rate is a significant predictor of at least one crash type. Within 1/4 mile of Metro stations, vehicle-miles traveled is a significant crash predictor, although not within ½ mile of Metrolink stations. These findings support increased investment in active transportation, including targeted funds for low-income community-led projects and projects near bus stops.

About the Project

Traffic collisions are just one example of the negative externalities resulting from motorized transportation, along with noise, congestion, localized air pollution, and greenhouse gas emissions. Although some crashes involve only non-motorized modes, most pedestrian and bicycle crashes involve automobiles and other large motorized vehicles. While reducing pedestrian and bicycle collisions should be a priority everywhere, reducing them in Southern California has unique importance. In 2016, California had the 10th most pedestrian fatalities per resident in the United States, and SCAG anticipates an increase in walking and bicycling throughout the region over the next few decades.This study supports the State of California and SCAG’s objectives in decreasing traffic fatalities while increasing active transportation by identifying high-collision areas, ranking which factors predict crashes, and demonstrating that these areas tend to be low-income communities and communities of color. The research team will use linear regression models and geographic information systems software to meet their objectives.