Red light running causes about 800 crash deaths per year, and about half of the people who are killed aren't the signal violators. They're pedestrians and people in vehicles that are struck by motorists committing the violations. Another 165,000 people are estimated to be injured in red-light running crashes each year. High red-light violation rates increase during peak travel.

Evaluating the effectiveness of red-light cameras at two intersections along Philadelphia's busy Roosevelt Boulevard, researchers from the Insurance Institute for Highway Safety separated camera effects from the effects of extending yellow lights to give approaching motorists more warning that signals were about to turn red.

Sometimes these two measures have been introduced simultaneously, which has caused confusion about their relative benefits. The new study reveals that both measures reduce signal violations, but it's the cameras that make by far the biggest difference. They all but eliminated the signal violations that remained after yellow lights were lengthened at the Roosevelt Boulevard intersections.

"Violations virtually disappeared at the six approaches to the two intersections we studied," said Richard Retting, the Institute's senior transportation engineer and lead author of the Institute's new red-light camera study. "This decrease in violations is all the more remarkable because the intersections were such high-crash locations. In fact, they had been identified as having some of the highest crash rates in the nation."

Researchers tallied signal violation rates at intersections before and after extension of yellow lights and again after red-light camera enforcement had been in effect for about a year. The first step reduced signal violations by 36 percent. The cameras reduced the remaining violations by 96 percent. At the same time, violations didn't change much at intersections without cameras in Atlantic County, New Jersey, about 50 miles away.

How They Work and Why They're Needed
Cameras help by deterring violators and, thus, preventing collisions. The cameras, connected to signal lights and sensors that monitor traffic, automatically photograph vehicles driven into an intersection after the light has turned red -- not just as the light changes but a specified amount of time after.

The 2004 legislation authorizing camera use in Philadelphia requires photos of the rear license plates of vehicles in violation but not images of the motorists. Owners of the identified vehicles are subject to $100 fines.

In Philadelphia and elsewhere with camera enforcement, conspicuous signs warn motorists as they approach camera-equipped intersections. The signs posted along Roosevelt Boulevard include images of traffic signals and the words, "PHOTO ENFORCED."

"This policy flies in the face of red-light camera critics who claim the cameras are all about catching people, writing lots of tickets, and raising money," Retting points out. "The true purpose of cameras is to reduce crashes by getting motorists to stop at red lights, so the most successful programs don't produce any revenue at all."

Results of the Philadelphia study also rebut camera opponents who insist that lengthening yellow signal intervals is all that's needed to reduce intersection crashes. It isn't. Appropriate yellow intervals are important, but cameras make a much bigger difference, said the Institute in a release.

Violations at the six approaches to the two Philadelphia intersections ranged from 8 to 251 per 10,000 vehicles before any changes were introduced. After yellow signal timing was lengthened, violation rates declined by 20 to 63 percent, depending on the location. After cameras had been operating for about a year, the rates declined an additional 87 to 100 percent.

At the intersection approach with the highest violation rate (251 per 10,000 vehicles), the first step of extending yellow lights produced a decline to 198 violations per 10,000 vehicles. Then with red-light cameras, the rate dropped to 1.8 per 10,000. Signal violation rates at four of the six intersection approaches plummeted to fewer than 2 per 10,000.

Before this research, the most widely cited evaluation of cameras in a U.S. community was