This compendium details automated enforcement systems (AES) implemented around the world and characterizes the safety impacts of such deployments, based on available scientific evaluations of the outcome measures. A systematic literature search was conducted. Criteria for selecting key evaluation studies were developed and applied for two AES technologies: speed camera and red light camera (RLC) enforcement systems. For automated speed enforcement, key studies reported significant reductions in estimated crashes following program implementation, but only a few studies were well-controlled. The degree to which reported safety improvements are related to the treatment and desired behavioral change (decrease in speeds) as opposed to regression to the mean (RTM), other confounders, and behavior changes such as choosing alternate routes, cannot be stated with certainty. Most studies attempted to account for unknown and time-related crash trends through the use of comparison groups. Four studies considered RTM, while only one study accounted directly for traffic flow changes and effects on crashes. Speed reductions documented in about half of the studies provided support for a relationship between the treatments and crash reductions. Enforcement intensity of mobile enforcements, site-specific differences (e.g., sample area, roadway characteristics), and statistical anomalies may play a role in AES effects. Recommendations for future studies include the use of empirical Bayes (EB) procedures to control for RTM, careful selection and examination of comparison groups, and consideration of traffic flow effects and possible crash migration due to the treatment, in order to improve the accuracy of treatment safety effect estimates. For red light running enforcement, the key studies support the assertion that RLCs can reduce crash severity at high red light running intersections. Consistent with previous reports, this review attributes a decrease in rightangle crashes with an increase in rear-end crashes to RLC implementations, together with a decrease in red light running violations. Recommendations for future studies of RLC effects include the use of controlled and randomized experiments whenever feasible. Otherwise, adjustments in before- and after- designs with matched control locations (or jurisdictions) to take legal, ethical, and economic barriers to random assignment of RLC installations into account, plus the use of EB procedures to control for RTM and to improve the precision of estimates of treatment safety effects, are strongly recommended. (Author/publisher)
Abstract