Posted on October 3, 2005  Comments (3)

D.C. Red-Light Cameras Fail to Curb Accidents by Del Quentin Wilber and Derek Willis.

The explanation of the data presented in the article does not support claim made in the headline.

I’m not sure why accidents should go down at intersections with red light cameras. First what percentage of accidents are caused by red light running? Second, is sending tickets somehow preventive to a specific location?

It would seem the assumption behind their conclusion is people who run red lights are so careful that they notice a camera and chose not to run the red light that they would have otherwise run. That doesn’t make much sense to me. If red light cameras work I would think they work because people learn if they run red lights they will be ticketed and therefore stop running all red lights therefore decreasing red light running at all lights. Or perhaps they don’t and lose their driver’s license due to too many violations.

I can imagine that some people who choose to run red lights figure out that a couple of specific locations that they frequently use have camera and therefore they choose not to run those red lights but continue to run other red lights but this seems unlikely to be of such an impact as to decrease red light running significantly. I would think either red light running everywhere decreases or it does not decrease significantly not that people learn where they can violate the law and where then cannot. Though that is merely conjecture on my part.

I am not sure why they think placing a camera is going to get people to stop running that red light. And I am not sure why that would be what you would target anyway. Wouldn’t you want people to stop running all red lights?

Again the article seems to be muddling the analysis of the data. What is the goal of the red lights cameras (to reduce red light running at those intersections or everywhere?). What percentage of accidents are due to red light running?

Are the conditions (other than the red light cameras) identical to the previous years? If there is more traffic and if more traffic means more accidents then it could be the equivalent of saying that accidents increase as HDTVs were introduced to the marketplace. Why do accidents keep increasing the more that people use HDTV’s. Somehow I think more people are driving with cell phones today than in previous years. Are more accidents being caused by drivers with cell phones?

It may be that the actual data has been analyzed sensibly and the article just doesn’t explain it well, but based on the data from the article the data doesn’t seem to say much of anything of value and doesn’t seem to support the conclusions stated in the article.

You might think this means the cameras are ineffective. However that would only be the case if drivers were so selective with red light running that they chose to run some red lights and not others. And second if the intersections with cameras experienced the same changes as other intersections (same increases in traffic…) and if those intersections were not at some tipping point which meant they would have actually increased by some percentage in excess of the average intersection absent those cameras.

It could be you have 10 really bad intersections and the traffic exceeds the safe capacity and therefore as traffic increases the level of accidents increases at a much greater rate. Then if you took sensible and effective measures at those intersections but those reductions to the accident level were not enough to overcome the deteriorating other conditions (say increased traffic) they could have worse results compared to the average intersection. That data would not support the conclusion that the measures taken were ineffective however, careless analysis of the data could lead some to believe that is what the data said.