查看文章

This study proposes an Empirical Bayesian (EB) before-after analysis method to model the safety effects of Red Light Cameras (RLC) capturing the spillover effect. Zero-Inflated Negative Binomial (ZINB) models are used to estimate Safety Performance Functions for crashes of different types at the intersection, approach level separately for the before and after periods. Two methods are proposed to capture the spillover effect of the behavior at the RLC intersections to the reference intersections. The three models (no spillover, uncontrolled spillover, and controlled spillover) are applied to data from the City of Chicago’s RLC program. The analysis is applied at the intersection approach level, and includes a larger number of intersections than previous studies. The results are largely in agreement with previous RLC studies in the literature.