Official Title: The Impact of Red Light Cameras (Automated Enforcement) on Safety in Arizona

Authors: Dr. Simon Washington and Mr. Kangwon Shin

This is a 2005 study by the University of Arizona in the cities of Phoenix and Scottsdale, Arizona. This study compares the before and after crash statistics as well as anticipated economic cost differences at 11 RLC intersections in Phoenix and 14 RLC intersections in Scottsdale. The study reports the two cities' data separated without statistically combining them in the final analysis, so this is in effect two separate studies. the Phoenix analysis uses 36 to 39 months of pre-RLC crash data compared to 18 to 21 months of post-RLC data while the Scottsdale analysis uses 83 to 160 months of pre-RLC crash data compared to 79 to 8 months of post-RLC crash data. The statistics used in this analysis are from the cities of Phoenix and Scottsdale which used 100 feet as their Distance-From-Intersection inclusion zone for crashes. There are no significant data or conclusions in this report that specifically suggest a reduction or increase in fatalities due to RLCs. The results show that crashes at RLC-equipped intersections in Phoenix have increased by 5% over the study period while crashes at RLC intersections in Scottsdale decreased by 12%. The study concludes that: "When crash severities and costs are considered and intersections are analyzed..., the benefits of RLCs range small benefits (Phoenix) to relatively large (Scottsdale). For example, the crash costs (frequency and severity considered) of rear end crashes are slightly greater than the reduction in crash costs (benefits) for angle and left-turn crashes. In Scottsdale, in contrast, the cost savings from reducing the severity of angle and left-turn crashes is greater than additional costs of rear-end crashes, and as a result there is an expected cost savings from the RLCs.".

Download this study below (1800k)

This is a detailed 141 page study, the data and statistical analysis appear valid. The selection of 25 RLC intersections is a slightly above-average number of intersections for this kind of study. The use of 100 feet as the Distance-From-Intersection inclusion zone for crashes is close to the average for most states. The before and after RLC date periods are sequential and mostly adequate to provide a valid data set for comparison, although the vast difference in lengths of the Scottsdale data should have been restricted to yield a more uniform set of data to work with. The study mentions the use of non-RLC intersection crash data, but it only mentions a few pieces of this data instead of listing all of the raw statistics on this data, which somewhat degrades the validity of the conclusions. Using the crash-frequency calculations for before and after injury statistics(since there are no raw before/after statistics given for injury crashes) there was a 10% increase in injury crashes in Phoenix and a 21% decrease in injury crashes in Scottsdale during the study period. The final conclusion of the study is that more studies should be conducted with data from more RLC intersections to get a more comprehensive set of results "It should be noted that these results are based on small sample sizes and observed trends in the means and therefore require further research to validate". The conclusions of this study are supported by the data supplied.