Tribune study: Chicago's safety claims for camera enforcement are dubious

The seeming disparity in the amount of injury accidents is accounted for in a couple of different ways. What it comes down to is how you count the injury accidents. Not to say there aren't a lot of injuries resulting from rear-enders, but the question is of severity. Just counting accidents with injuries will give you one set of numbers while taking into account the severity of the injuries will yeild still a different number.

T-Bone crashes with injuries usually result in hospitalization, rear-enders rarely do even though someone may be transported, they are often released without checking into the hospital. So, it comes down to counting the medical costs of the injuries and the length of time a person was in care rather than just the raw number.

Here's the report the researchers submitted to the Chicago Tribune which goes into a fair amount of statistical and modeling detail about how they analyzed the data.
http://www.chicagotribune.com/news/watchdog/redlight/chi-red...
I think this report is not behind a paywall, but I can't tell for sure.

One thing I can tell from the Tribune paywall article that leads the researchers and Tribune to conclude that the city overestimated the safety effect is that the city was counting reduction in crashes (and went back farther in time than the cameras were in place at some corners for the baseline accident rate), and that during the study period, injury accidents were declining in Chicago anyway. The city was attributing all of the injury accident reduction in this period to RLCs when the study data says only a much smaller portion of it can be attributed to RLCs.

In the interest of fairness, and tippo' the hat to jgermann for that, since I oppose the use of these cameras, nonetheless, one thing that could be argued to suggest that the Tribune study underestimates possible improvements in safety due to RLCs in Chicago is that the Tribune study control group consisted of other intersections in Chicago that did not have RLCs. But if RLCs were influencing drivers to make safer choices (such as by not gunning the accelerator to beat the red light) city-wide because there are so many cameras, then one could say that the cameras have helped promote safer driving than the study suggests. To really control for the influence of RLCs in traffic safety, it might have been better for the researchers to use random intersections from a different large midwestern city where these cameras are not used.

However the increase in rear-end injury accidents which opponents have long cited as a potential safety negative with these cameras-- some supporters of RLCs here have said, prove it--where's the data really showing this??--did in fact show up in the data. RLC opponents will argue that the increase in rear-end injury accidents attributed to RLCs more than compensated for any decrease in T-bone and right-angle injury accidents due to RLCs in the Chicago Tribune data.

And Box Car I get your point about severity, but they likely don't have data on that. Was an ambulance rolled? or was there a fatality? would be in the city accident data. but data on the medical cost, pain, and suffering to the victims and their friends and families are of course not there. I would not be too cavalier though about possible severity of a rear-end accident. I agree that on average T-bones are the worst, but whiplash is not just in the imagination of plaintiff attorneys; it does very negatively impact some peoples' lives long-term. Also rear-end accidents are fatal less often than T-bones but still can be fatal especially when there is a mismatch in vehicle size or when there is a multi-vehicle accident stopping motion of a vehicle in the middle of a line, and both these scenarios certainly do happen.

I didn't mean to discount any injuries that may happen in a rear-ender. It's when taken on a whole, the injuries are typically not as severe as those in a T-Bone. There are a lot of instances where there are no injuries in a T-Bone, just as there are many more incidents of non-injury in the rear-enders.

None of these can be classified as "accidents" as there was nothing accidental about them. Someone wasn't paying attention or decided to chance the situation and lost.

Just this morning my wife witnessed a rear-ender at an intersection caused by a car not moving when the light turned green. The car behind saw the light change and went forward. There was damage but no injury. (And no, it wasn't an intersection with a camera.)

Agreed on both counts.

@dagarmin, your comments were much appreciated. Our country needs more people looking critically at the facts and making their own judgments rather than just listening to their favorite "talking heads".

It has always been relatively easy to take any report/study and extract statements out of context to support one's position. thenewspaper.com has done this for years.

I appreciate your posting the link and taking the time to read the report.

I always like to look at the "abstract" and/or the Summary and Conclusions portions of reports to get a sense of what the authors concluded.

[emphasis added]

I have always been a proponent of Automated Traffic Enforcement (ATE) not because ATE significantly improves safety (although on balance I think it does) but because it produces revenue badly needed by municipalities. AND, this revenue is paid by law breakers rather than the general population.

And that's a good point. Chicago's finances are in terrible shape.

I didn't pay enough attention to that limitations paragraph in the report. I suspect that what the study authors meant by "spillover effect" is along the lines of what I was suggesting about drivers in Chicago possibly being more generally careful even at intersections where there aren't any RLCs. Because the city put so many cameras up, drivers may feel like they can't be sure where all of them are and be more cautious than they might have been if no cameras had ever been installed. This could be a confounding factor that could mean RLCs have more of a positive effect on driver safety than was controlled for in the data or suggested in the study conclusions.

More researchers looking at effects of RLCs on driver safety scientifically and with an open mind should be appreciated by all who don't have an axe to grind... a rarity in Chicago politics BTW.

Ars Tech article

arstechnica.com/tech-policy/2014/12/major-chicago-study-finds-red-light-cameras-not-safer-cause-more-rear-end-injuries/

Terrible is an understatement to describe our finances. You need a stronger word to describe a city that has $23 billion worth of unfunded liabilities. It was funny on the Sunday local news program watching the 2 mayoral challengers carefully dance around how they plan to come up with the money for the $1.2 billion pension payment due next year.

I was quite surprised that you submitted a link to an article that extracted statements out of context in order to mislead readers right after dagarmin and I had been discussing the entire report.

Let me requote the report's conclusions:

[emphasis added]

Let me note that the study applied only to "signalized 4-legged intersections" with the authors saying that their conclusions "should not be generalized to other intersection types and outside the study area."

Note particularly that the author's used the word "significantly reduce" when applied to angle and turning injury collisions but do not use any modifier when speaking of the increase in rear-end injury collisions.

Despite this, the article you linked to only noted " increase rear end injuries".

if the city of Chicago does anything it is because someone(political cronies) ends up with a bunch of money in their pockets, and to hell with everyone else!

Anything done by the politicians in "Crook" County Illinois is Dubious.

I need to correct a statement I made in that paragraph from post three in this thread. I was inaccurate in saying that injury accidents were declining in Chicago during the study period comparing before RLCs to after RLCs.

Right around the time RLCs were installed in Chicago, the city loosened the standards by which accidents were counted as injury accidents. Injury accidents declined statistically in the following years. I don't see in the reporting the reason the standards were changed, but the Tribune did not charge the city with rigging things by changing the way injury accidents were counted. I infer that change may have been done by other city workers who had nothing to do with RLCs; it may have been coincidental.

The Tribune's objection was that the mayor attributed all of the decline in injury accidents to RLCs when most of it was really due to the change in the way incidents were counted as injury accidents. It was misleading analysis and spin, and the mayor should have known that.

I need to correct a statement I made in that paragraph from post three in this thread. I was inaccurate in saying that injury accidents were declining in Chicago during the study period comparing before RLCs to after RLCs. The reporting standards for traffic incidents had changed during the study period, so it becomes difficult in one quick glance without in-depth statistical analysis to say whether injury accidents were really declining at that time.

Right around the time RLCs were installed in Chicago, the city loosened the standards by which accidents were counted as injury accidents. Injury accidents declined statistically in the following years. I don't see in the reporting the reason the standards were changed, but the Tribune did not charge the city with rigging things by changing the way injury accidents were counted. I infer that change may have been done by other people, perhaps in a different jurisdiction than the city, who had nothing to do with Chicago's RLCs; that change may have been coincidental.

The Tribune's objection was that the mayor attributed all of the decline in injury accidents to RLCs when most of it was really due to the change in the way incidents were counted as injury accidents. It was misleading analysis and spin, and the mayor should have known that.

This whole discussion points out why many of us can advocate having safer roads and, at the same time, question the use of enforcement cameras. When the purpose of law enforcement is to generate revenue in the promotion of public safety, we've got it backwards.

The City that rewards corruption, its amazing the Trib funded this study.

the city that pushed and promoted YOUR current president's elections. Can you see why the entire country is in such a poor economic state?

I have taken a lot of statistics and biostatistics courses, and a number of risk communication courses.

Most people know the saying attributed to Mark Twain about "lies, damn lies and statistics". But I prefer the statement "Never criticize the statistics, attack the assumptions that the statistics are based on."

As a result, I try not to be overly swayed by numbers and try to identify the underlying assumptions that a study is based on.

Studies of red light cameras seem to bring out the "confirmation bias" in readers. That is why I like to see people like jgermann and dagarmin attempt to objectively assess these types of studies, and challenge over-generalizations.

I may not always agre with their position, but at least some thought goes into their commentary.

Thanks to all and Merry Christmas!