The Insurance Institute for Highway Safety spent all of yesterday touting its report on the effectiveness on red light cameras as if it were the most “comprehensive” study available on the topic. The group emphasized the effects cameras would have had on one hundred of America’s largest cities, but the report itself only looked at accident numbers in fourteen out of the 500 jurisdictions that have active photo ticketing programs. Industry researchers did not even know how many red light cameras were in use in the locations studied.
“Attempts were made to obtain historical information on the number of red light cameras in the study cities, but information on the scope of red light programs could not be obtained for many of the cities,” the IIHS study explained (page 9).
By comparison, a 2005 study by the Federal Highway Administration (FHWA) gathered data from fifteen cities that used automated enforcement. The study’s final results excluded eight of these locations because it was not possible to obtain complete accident reports, signal timing data, traffic flow numbers and other variables — twenty-eight in total. While the FHWA study suffered from its own methodological flaws, the government agency considered reporting results without detailed data to be unscientific.
“Exposure is the major determinant of intersection crashes,” the FHWA study explained. “Therefore, it is important to account for any changes between the before and after period, particularly if these changes are triggered by the measure. All studies reviewed have failed to do this accounting, conveniently assuming that red light cameras will not change exposure.”
IIHS did not bother gathering data regarding any of the factors FHWA considered essential, aside from looking up 1990 and 2000 population figures. In fact, the insurance industry relied upon the eight-year gap between the “before” and “after” periods to obtain the desired result. In locations like Chandler, Arizona the community went through significant changes — including the building of the Loop 101 and Loop 202 freeways — during this time. These new routes drew traffic away from intersections during the “after” period despite the increase in population. Without accounting for the change in traffic volumes, the figures would be misleading. Chandler accounted for the greatest decrease in citywide accidents in the IIHS report. IIHS not knowing which locations in the city had cameras could not check whether there was a difference between camera and non-camera locations.
A professor at the University of Illinois at Chicago did check and determined last year that there was no statistical difference before and after the cameras were installed in the Windy City. The data refuted the IIHS assumption that there is a so-called “spillover” or “halo” effect that spreads good driving habits throughout the photo-enforced jurisdiction. Between 2001 and 2008, use of cameras had no effect on the percentage of accidents that took place at intersections — the figure remained steady at about 25 percent (view report). IIHS claimed big accident reductions in Chicago, and in Baltimore, Maryland. The latter city last month reported inconclusive results from its photo ticketing program.
“Six of the 15 locations had more traffic accidents in 2010 compared to prior years, and eight of the 15 locations had fewer traffic accidents in 2010 compared to prior years,” the Baltimore County report explained. “One location had the same number of traffic accidents.”
As mentioned in Part One of this analysis of the IIHS report, the Insurance Institute is primarily the public relations arm of the automobile insurers that rely upon ticketing programs to increase annual premiums. Anne T. McCartt, senior vice president for research, was one of the three researchers who coauthored the report. Her doctorate is not in engineering but from the Rockefeller School of Public Affairs, which offers degrees only in public administration or politics — the perfect preparation for a career in manipulating the media to advance public policy.
A copy of the IIHS report is available in a 500k PDF file at the source link below.
Source: 
Effect of Red Light Camera Enforcement (Insurance Institute for Highway Safety, 2/1/2011)
[Courtesy:Thenewspaper.com]
Two thoughts:
1. What is “significant?” I know they found a reduction to be statistically significant, but there is a lot of room for play there.
2. If you compare a city in the 1990s vs a city now with red light cameras, aren’t the fatality rates going to be lower because of better passive safety?
+1. If something is statistically significant it means that it is unlikely to have been a result of chance, however it says nothing about whether it is a big difference or if the change was worth it if the costs are external. Fatalities could have gone from 10 per 100k passenger miles (or whatever) down to 9.999999 per 100kpm, and the difference could be statistically significant, but so what?
@ snabster:
If you do statistical analysis on a regular basis, you would compare the averages of the two samples (before/after) using a 2 sample-T test. With the null hypothesis being there is no difference in them and using the standard confidence interval of 95%, then if the statistical value for “p-value” in the analysis is less than 0.05, then there is a statistically valid difference in the two populations. Otherwise, no difference. Of course, you would initially need to perform analysis on the input data to determine its validity and/or any skewing, i.e. Anderson Darling Normality Test.
As to your second question, that would be an interesting study. If the data is available to separate the accidents w/wo passive systems at locations w/wo the red light cameras, then that analysis could be done as well.
Of course, figures don’t lie …… but liars figure (messing with the input data, as in adjustments to climate data and providing only the adjusted figures w/o allowing access to the raw numbers. But that’s a different topic……)
Having followed IIHS “studies” since the 55 NMSL era, every single one of them is an exercise in cherry picking the stats that support your contention.
I cannot tell you how many times, as a lobbyist and in the media, I had to counter IIHS claims with engineering studies or stats taken from state DOT, which were scientifically gathered and reported.
IIHS wants to normalize red light cameras so that they can get the real prize, speed cams. They’d love to see the USA look like Britain, or Australia, with a cam or four on every corner.
The “lives saved” claim is cut from whole cloth. The best was the blooper reel IIHS gave to the media, showing out of control drivers blow through intersections at 60 mph. Cams clearly stopped that activity, and how many of those drivers were DWI or being pursued.
The real deal is cameras go up. Once folks figure out where they are, cash drops. Yellows magically get shorter to catch more “out”. At some point they are used to strictly enforce right turn on red laws, when the take drops further. Meanwhile, “hiit in rear” accidents increase as drivers under a yellow light stop short and get hit.
Letting those pesky highway engineers set yellow intervals and approaches is the best way to minimize crashes. If your goal, on the other hand, is to tax motorists for a trivial out, then maybe not.
No camera every caught a drunk, fugitive or unlicensed driver.
When the 55 NMSL was repealed, IIHS claimed 6500 folks would die as a result. The fatals per vehicle mile traveled, the only relevant stat, dropped. The IIHS = Tobacco Institute
The IIHS and local govts would support a referendum that ended all red light cameras….if the measure was bundled with a new tax of $500 per driver and indexed to grow each year.
They would applaud such a ‘sensible step in the public interest’.
@ Speedlaw:
Having followed IIHS “studies” since the 55 NMSL era, every single one of them is an exercise in cherry picking the stats that support your contention.
Not true: how do explain how these studies are accepted by peer-reviewed journals, such as Accident Analysis and Prevention?
When the 55 NMSL was repealed, IIHS claimed 6500 folks would die as a result. The fatals per mile traveled, the only relevant stat, dropped. The IIHS = Tobacco Institute
No, the Institute did not make that claim — it was Ralph Nader, Joan Claybrook, Clarance Ditlow, and if I recall correctly, Advocates for Auto and Highway Safety. I challenge you to find the IIHS making such a claim. Tobacco Institute? Nice ad hominem attack.
I drew a similar comparison yesterday in a blog. When a self-serving industry pressure group are disseminating data that can improve their bottom line, the comparison is inevitable. And valid.
Which is one of the great ways IIHS stuff spreads as “authoritative”. They end up in various journals, and peer reviewed does not equal unbiased. I’ve run across a lot of IIHS stuff in attorney defense bar oriented law reviews and such…they bootstrap the credibility to that particular audience.
IIHS has a public “good guy” image for those who don’t look too hard in the car crash reels they release every six months or so, where TV news can run that new Mercedes SUV into a wall. The crash testing is probably the only pro-motorist thing they do, and even that is “accidental”. NHTSA is too in thrall to the manufacturers to be able to get any improvements, so IIHS there does a service by embarrassing the car makers into fixing things-this goodwill is used to then forward the “political” agenda, and since the media has already used the crash vids and reported “safest and least safe cars”, they go with the political claims in most cases.
Actually, the IIHS best 55 claim was that in 65 zones, 80 or so percent were “over 70 mph”, but this really meant they were mostly at 72 mph on open highways. Sounded really shrill though.
I recall Chuck Hurley saying the 6500 number on TV back in the day.
“As mentioned in Part One of this analysis of the IIHS report, the Insurance Institute is primarily the public relations arm of the automobile insurers that rely upon ticketing programs to increase annual premiums. Anne T. McCartt, senior vice president for research, was one of the three researchers who coauthored the report. Her doctorate is not in engineering but from the Rockefeller School of Public Affairs, which offers degrees only in public administration or politics — the perfect preparation for a career in manipulating the media to advance public policy.”
Yeah, if true, then these things really do undermine or at least call into question the soundness and trustworthiness of IIHS’s interpretation of the data (particularly as relates to the process and criteria of data selection). This is where an independent review of the data by a non-interested party would be extremely helpful as a check against the IIHS’s conclusions. I have no idea if such a review has been done, but you would think that if it had then the IIHS would reference it to support the credibility of their claims.
The study may as well have been partly underwritten by Redflex and ATS.
DO THE MATH
The following table lists the yellow light times in seconds per the FHWA MUTCD and standard kinematic formula with 10fpsps and 8fpsps braking.
20mph 2.5/5.9/6.7
25mph 2.8/6.7/7.6
30mph 3.2/7.4/8.5
35mph 3.6/8.1/9.4
40mph 3.9/8.9/10.3
45mph 4.3/9.6/11.3
50mph 4.7/10.3/12.2
55mph 5.0/11.1/13.1
60mph 5.4/11.8/14.0
65mph 5.8/12.5/14.9
Does this help explain why cities following the MUTCD have a red light running problem?
Does this help explain why people are screaming LENGTHEN THE DAMN LIGHTS!!!?
Does this help explain why hundreds if not thousands of Certified Traffic Engineers across the United States have chucked the MUTCD and went to using standard kinematic mathematical equations that date back to the day an apple beaned Sr. Isaac Newton?
This is not rocket science.
Tell the engineer how hard you want to brake and he or she will tell you the time and distance it will take.
Add 2-3s for reaction time and 0.5s for controller delays and you have it. Now go out and set the lights properly…
If the intersection has a traffic camera, add 0.25 seconds to both the controller and camera systems for coordination errors and a 1.0 second delay to the camera system before recording an infraction. We don’t want to appear abusive, and exploitative, do we?
It was either Georgia or South Carolina where yellows were mandated to be longer at camera locations, and the cams came down because they were no longer profitable.
It was Georgia and possibly Ohio.
HB1823 just introduced in the State of Washington Legislature seeks to make it 16 states camera-free.
The more you dig into the cameras, the more you grasp the depth and breadth of the corrosiveness of having them around. They are absolutely evil. Cities that love the revenue are crushing their populace. Twenty four reasons how and why your life goes down the toilet when they are installed.