Novel Methodology for Safety and Economic Evaluation of Red-light Cameras at Signalized Intersections

Red-light running is a phenomenon that has led to frequent collisions causing fatalities, injuries, and property damage. Red-light cameras (RLCs) have been used in an attempt to reduce the frequency of these collisions. The main objectives achieved in this thesis are: (1) The site-selection bias of traditional methodologies were corrected by using a novel methodology (propensity score matching); (2) The fixed and random-effect panel regressions were utilized to account for the spatial and temporal correlations in the data; (3) The identification of how RLC effectiveness varies by site characteristics through the use of interaction terms; (4) The determination of spillover distances and times resulting from the presence of RLCs; (5) The benefit-cost methodology supported by economic analysis and sensitivity analysis using the novel methodology; and (6) The establishment of guidelines for effective implementation of the RLC treatment.The research used field data from the City of Ottawa (Canada) involving 34 RLC intersections and 14 control intersections observed for the period 1999-2012. The results from this thesis indicate a consistent significant reduction in angle and turning movement collisions of 19% and 21%, respectively, and an increase in rear-end and sideswipe collisions of 12% and 6.7%, respectively. The magnitude and the direction of these effects are comparable to results of previous studies in the literature. The benefit-cost analysis, based on social costs of collisions avoided, the spillover effects and fine revenue, yielded an overall annual cost savings to the community of over $4.4 million CAD across the 34 RLC intersections; with a benefit-cost ratio of 4.50. An extended cost sensitivity analysis was incorporated to quantify the robustness of the base-case conclusions.RLC implementation decisions represent a major policy action, with serious repercussions for public safety. This thesis enhances such decisions by establishing effective implementation guidelines using rigorous, novel statistical analysis methods aiming to assist in this key policy effort. Furthermore, this thesis contributes to quantifying the RLC spillover effects as well as identifying critical road factors that the RLC treatment is likely to affect and benefit from. These contributions are all critical components of the traffic safety management process.