Multiscale spatial analysis of macro-level determinants of bicycle crash frequencies in the Phoenix metro region

The realization of the many benefits of bicycling will not be achieved in American regions until safer bike infrastructure and bicycling conditions are presented to a more general population. The Phoenix region—one of the nation’s most populous—has sought policies and programs to increase bicycling rates. Yet, the region continues to have a small mode share, underscoring a need to motivate population-level bicycling adoption. This study examines 2015–2019 bicycle-vehicle crash data to identify those macro-level factors associated with bicycle-vehicle crashes and a subset of crashes where a serious injury or fatality occurred. Specifically, the effects of a robust set of socioeconomic and built environment factors, measured at three hexagon spatial extents, in negative binomial and spatial Durbin models were estimated for the two crash outcomes. Results show denser zones with a traditional network design experienced more bicyclist-involved crashes, as did zones with a higher percentage of low-income households and working-age adults. Findings, which also found spatial clustering of total and severe bicyclist-involved crashes, suggest that the targeted provision of safer bike infrastructure and a more complete network in zones exhibiting certain macro-level attributes holds promise in creating bike-friendly conditions that generate more utilitarian and recreational bicycling throughout the region.