Search

Conventional methods for estimating the effect of an intervention suffer from selection bias, where the units that receive the treatment have different characteristics from those in the control group. This paper proposes a novel method that provides a specific criterion for selecting the control group. The method, called propensity score matching (PSM), was applied to the evaluation of red-light cameras (RLC) and its performance was compared with conventional cross-sectional and empirical Bayes methods. The application was performed using field data from the City of Ottawa involving 30 RLC intersections and 89 non-camera intersections observed for a period of 15 years. All three methods yielded fairly consistent results, indicating an increase in property damage collisions and a decrease in injury and fatal collisions. Given the strong theoretical basis of the PSM method and its ability to produce a more stable and reliable estimator, the method is recommended as a viable alternative to the conventional methods. [ABSTRACT FROM AUTHOR]