Aggregated North American safety performance functions for signalized and stop-controlled intersections

Abstract

The statistical analysis of intersection collisions has allowed practitioners to develop reliable local models for collision prediction. While many North American jurisdictions have developed such models, a gap remains for the development of safety performance functions that represent the average North American intersection collision expectation. Such models could be used in the development of national guidelines, benchmarking local models and hotspots, and by jurisdictions lacking the capacity to develop their own models. This research bridged that gap by developing aggregate models of collision expectations at stop-controlled and signalized intersections in North America. In analyzing the results, it was found that the Highway Safety Manual predictive equations are not a good representation of the average intersection collision expectation. Further, it was found that the aggregate models are particularly useful to practitioners looking to estimate the change in collisions resulting from signalization given the partial cancelling out of jurisdiction-level effects.