Austin red-light camera results minimal, overhyped

Whenever you see someone claiming dramatic decreases in accidents due to red light cameras - a result which contradicts the experience of most jurisdictions - it's always worth examining the data closely. In Austin, reports the Statesman:

cameras were phased in at nine intersections — with more than one camera at some — beginning in May 2008. They have brought in more than $100,000 in citations.

Police Lt. Brian Gruetzner , who oversees vehicular homicide and the red-light cameras, says the 30 percent decrease in wrecks " is pretty substantial for anyone."

"We really haven't seen any increases (at those intersections)," Gruetzner said, "and I see it as a very positive thing."

That sounds pretty good, huh? Too bad it's a misrepresentation.

Gruetzner says the city hasn't witnessed any increases in wrecks at red-light camera intersections, but that's false. In three of the nine intersections, according to a chart provided by the Statesman (see below), wrecks at intersections increased the year after they were installed. At the intersection of I-35 and 15th, which I pass through frequently, the number of accidents increased from six to sixteen the year after cameras went in!

So how can Officer Gruetzner say "We really haven't seen any increases (at those intersections)"? It's just not true. The Statesman published the graphs so they had the data. Why let the police spin it that way?

At three other intersections, accidents declined slightly the year after cameras went in but remained higher than they were a couple of years ago. At intersections 5, 6, and 7, wrecks declined the year after cameras were installed, but were still at a higher level than in 2006. That indicates reductions likely result from routine fluctuations, not the initiation of cameras.

In all, at six of the nine intersections accidents increased or are still higher than in the past. That's evidence of failure, not success! If this is really about safety, not revenue, cameras at all six of those intersections should be immediately removed as unnecessary.

What's more, at the intersection with the biggest drop in accidents, it seems questionable to attribute the decline to cameras. Intersection #1 on the list made up most of the 30% decline, going from 15 wrecks in 2007 to 2 wrecks in '08, the year they were installed, and 2 in '09. However, the cameras weren't installed until October '08, so they don't explain the decline in accidents during the first two-thirds of the year.

Which brings us to the point that in all instances we're talking about very small numbers with large margins of error - including at intersections that increased, decreased, or basically stayed the same. It'd be premature to draw any conclusions from this data, but concluding the cameras "worked" is not just premature, it contradicts the intersection-by-intersection details.

What a phony baloney use of data to support what's clearly little more than a money grab!