Assessing the effectiveness of built environment-based safety measures in urban and rural areas for reducing the non-motorist crashes

Abstract

Introduction: Built environment (BE) has a well-documented impact on non-motorist crashes. Interestingly, the urban-rural distinction of the impacts received scant attention in the literature. Moreover, the combined effect of these elements are less studied than their standalone effects. Objective: This study explores the combined effectiveness of built environment-based safety measures in urban and rural settings. Data and method: The study uses nine years (2011–2019) of non-motorist (pedestrian and bicyclist) crash data in Florida. It classifies urban and rural areas with the multivariate clustering method and models the crash count with Log-transformed Spatial Error Models. Results: Findings suggest that urban areas, tracts with low median income, a lower percentage of senior citizens, and a higher percentage of black, white, and Hispanic people are significantly associated with a high number of nonmotorist crashes. The percentage of pedestrian and bicyclist commuters is positively associated with pedestrian and bicycle crash count, respectively. Among BE variables, more crashes are observed in tracts with more commercial land use (LU), less recreational LU, higher LU mix, more traffic, signalized intersection, transit stops, and sidewalks. Having more traffic and fewer transit stops pose lesser risk in urban areas than rural areas. The combined effects suggest that increasing commercial LU where LU entropy is high (or vice-versa) will help to reduce nonmotorist crashes. Also, in high entropy areas, increasing rural traffic is riskier whereas increasing urban traffic is safer. Conclusion: This paper documents the need of considering urban-rural differences and interaction effects among BE elements for nonmotorist safety.

Keywords

• Pedestrian crash
• Bicycle crash
• Built environment
• Safety
• Spatial error model