Income-Based Disparities in Fatal Interstate Crashes: An Association Rule Analysis of Contributing Factors from the NHTSA FARS Dataset

Income-based disparities in fatal Interstate crashes across the United States were investigated using data from the NHTSA Fatality Analysis Reporting System (FARS) for 2019–2023. Machine learning and association rule mining techniques were employed to examine how crash characteristics vary across socioeconomic strata. An XGBoost model was used to identify 25 key variables predictive of fatality count, with crash severity indicators, infrastructure characteristics, and emergency response lag times ranking highest in importance. To uncover deeper crash patterns, the Apriori algorithm was applied to crash data stratified by income quintiles, adjusted for regional price parity. Over 15,000 significant rules were generated, with rule complexity increasing in higher-income areas. Rule patterns highlighted income-related differences in crash types, particularly pedestrian-involved crashes, same-direction collisions, and front-to-rear impacts, as well as variations in emergency response times and infrastructure conditions. The findings reveal that fatal crash patterns are not uniform across income groups on Interstate highways. Lower-income areas exhibited concentrated high-risk crash factors, while higher-income regions demonstrated more diverse crash scenarios. Emergency response lag times and infrastructure disparities further indicated systemic inequities that may influence crash outcomes. The value of integrating socioeconomic context into traffic safety research was demonstrated. Association rule mining enabled interpretable, income-specific insights that can inform more equitable policy and resource allocation. Tailored safety interventions based on regional economic characteristics may enhance effectiveness and help address persistent disparities in fatal crash outcomes.