Here's a link to the full study pdf. (As an aside - why can't newspaper stories link to their sources just like bloggers do?)
Red-light cameras apparently reduced overall collisions at dozens of monitored intersections across Texas, according to a state transportation study.
The report, released today by the Texas Department of Transportation, concludes that crashes declined overall by 30 percent at a sampling intersections, many of them in Houston.
"While these results cannot conclusively determine that red light cameras are responsible for the overall reduction in crashes," the report reads, "the presence of the treatment provided some effect on the frequency of crashes at the selected intersections for the limited time period of this analysis."
The study examined crashes from July 1, 2007 to June 30, 2008 at select intersections at 12 cities required to report crashes under a new state law.
The data show that right-angle collisions reduced by 43 percent, while rear-end collisions increased by 5 percent, mirroring some other studies across the nation.
These are surprising data that contradict many past studies on the topic. For example, in Lubbock red light cameras were discontinued after accidents overall increased 52% at intersections with cameras. Similarly, the state of Virginia eliminated their use after studies in every city using the devices found the number of accidents increased. In other jurisdictions, studies have found reductions in right-angle accidents but nearly equal increases in rear-end collisions, including in injury accidents.
So how did this study come up with such radically different results? The short answer may be that much of their data is incomplete and speculative. For camera operators who began before 2008:
there is no requirement for the local authority to provide a report to the Texas Department of Transportation concerning the 18 months of pre-installation crash data even if the system remains active in 2008. ...Similarly, agencies' post-installation data was often truncated, particularly for jurisdictions that only began their programs since the beginning of the year. So the researchers had to take limited datasets and analyze them projecting forward.
This presents a problem in reporting since some local authorities reported preinstallation crash data while others did not. This made the process of analyzing the effectiveness of the red light camera system difficult to perform since no base line data was present for some local authorities. In short, there was no metric to determine the rise, fall or static percent difference in crash rates at some of the reported treatment intersections.
Only 12 of 26 cities provided pre-installation crash data; two of those 12 did not provide any post-installation crash data. For that matter, "Of the 24 cities that provided post-installation intersection crash data, 14 failed to provide pre-installation crash data." It seems likely that pre-installation crash data may be under-reported, especially for rear-enders, since many people don't call the police over a minor fender bender but the camera will pick up everything.
So researchers ended up with an extremely small dataset from intersections chosen because other data was flawed or non-existent, not because it was representative of the whole: "Ultimately, there were 10 local authorities that provided pre and post-installation intersection crash data. The information provided represented 56 different intersections within these 10 reporting communities."
In addition, we know in many cases agencies don't report all their crime data to the state on many more serious types of offenses. So one wonders whether reporting on traffic accidents contains similar gaps and errors.
Finally, while other factors may contribute to accidents, researchers only looked at a single variable so it was:
difficult to determine the impact that red light cameras had as a safety countermeasure because other crash variables could have produced a biasing effect on the number of red light running collisions that occurred. As such this analysis provided only a limited descriptive investigation of the self-reported local authority red light camera data that was provided to the Texas Department of Transportation.These pre- and post-installation data shortcomings, along with an outcome so radically different from other longitudinal studies, make this report highly suspect, IMO. But then, I'm biased toward using engineering solutions to reduce red light running instead of reacting to every problem by mulcting the public with tickets and fines.
Going forward, at least for agencies whose cameras came online after the new law took effect, it's possible that we'll have reliable data for analysis in the future. But there really doesn't appear to be enough data here for researchers to come to any real conclusions.
BLOGVERSATION: More from Blog Houston and Lose an Eye, It's a Sport.