Transportation Research: An exploration of pedestrian fatalities by race in the United States

Traffic Safety Pulse News


(Rebecca L.Sanders and Robert J.Schneider) Black and Native American pedestrians are disproportionately killed in the US, yet relatively little is known about how fatal crash patterns differ between races. Our multinomial logit analysis of six years of US pedestrian fatality data (2012–2017) and built environment and census data reveals notable differences between races compared to the baseline of White pedestrians, including that Black and Native American pedestrians were significantly more likely to have been killed in darkness, Black and Hispanic pedestrians under age 16 were significantly more likely to have been killed, and Asian pedestrians age 65 or older were significantly more likely to have been killed. Importantly, models with crash, built environment, and population data suggest critical connections between roadway design and population patterns that are risk factors for all pedestrians, but disproportionately affect certain races. Our findings highlight important risk factors for pedestrian safety and provide several areas for future research.

Section snippets

Introduction and background Pedestrian fatalities increased at an alarming rate over the last decade, culminating in a nearly thirty-year high in 2018 (Retting, 2020) and remaining high in 2019 and 2020 (National Highway Traffic Safety Administration (NHTSA), 2021). In the same timeframe (2009–2018), the proportion of fatalities comprised of pedestrians increased from 12% to 17%, another disturbing trend. Moreover, the population burden of pedestrian fatalities is not equally distributed: Black Americans and Native

Data sources

This analysis combined data on pedestrian fatalities with data on sociodemographic characteristics, commuting behavior, and built environment characteristics from the surrounding neighborhoods at the census tract level.

National overview of pedestrian fatalities by race

Pedestrian fatalities increased substantially starting in 2015, as Fig. 1 shows. Within that time period, White pedestrians comprised approximately 50% of those killed, although that percentage dipped slightly toward the end of the period. Percentages rose slightly for Black pedestrians and fluctuated for pedestrians classified as “other” for the purposes of this study, but otherwise held fairly steady.

Discussion of either raw numbers or percentages obscures the dramatic differences by race

Key findings

Our estimates of pedestrian fatality risk corroborate research showing that Black and Native Americans have higher rates of pedestrian fatalities on a per-capita and per-trip basis (Bellis et al., 2021). Our bivariate and multivariate analyses underscore the need to investigate pedestrian fatalities across all race categories, particularly given how results change when census variables are added to crash data. The following section discusses key findings and what they suggest about potential


This paper used multinomial logit modeling to compare key correlates of pedestrian fatalities and race. We hope our findings establish a baseline for future research and help policymakers and practitioners respond more proactively to dynamics that are harmful in general and seem to be particularly harmful for pedestrians of certain races. While aggregate statistics are helpful for raising alarm bells, disaggregation can be critical for informing next steps – particularly if, as this analysis


This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Credit authorship contribution statement

Rebecca L. Sanders: Conceptualization, Methodology, Data curation, Formal analysis, Investigation, Writing – original draft. Robert J. Schneider: Conceptualization, Methodology, Data curation, Writing – review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. To read the whole report, click here.