Broken Windows

Stop me if you've heard this before. A fired-up principal with a revolutionary educational philosophy takes over a failing inner-city public school, and turns it around so impressively that before long it's the the school that every parent wants their kid to go to. We all know how this story ends. The principal encourages everybody else to follow the revolutionary system, but somehow when people try to put it into practice elsewhere, it never works quite as well.

As everybody knows, the dedication and enthusiasm of the principal and his staff is usually much more important than whatever system they're using. He might believe wholeheartedly that it's the system which is proving itself, but there's a world of difference between a dedicated group of teachers giving their all to prove the success of a radical new pedagogic philosophy, and a worn-down group of teachers being told to stop doing it this way, start doing it that way.

New York City knows the story very well, of course, which is why it has its system of charter schools. New York is full of wonderful unique schools, but no one's trying to duplicate them.

And so to Broken Windows. Here's the story in a nutshell: Rudy Giuliani and Bill Bratton, full of crimefighting zeal, take over the NYPD with a revolutionary philosophy. Crime comes down. So now Bratton is a huge advocate of said philosophy, the Broken Windows theory. Critics, on the other hand, say that correlation is not causation. Yes, crime came down in New York City, but it came down everywhere. Or it came down because of the end of the crack-crime epidemic. Or it came down because of the rise in abortions in the early 1970s. Or it came down because New York City put more cops on the beat in tough neighborhoods. Or some combination of all of the above.

Meanwhile, academics are studying the Broken Windows theory, doing things like literally counting broken windows, and then taking polls which seem to show that there's no correlation between the number of broken windows and how "disordered" people think a neighborhood is. Some of them are coming to the conclusion that Broken Windows isn't empirically rigorous.

The debate about Broken Windows is silly on both sides. Counting broken windows doesn't prove anything: the point about the theory is not so much that broken windows get fixed, and much more that the police care that windows are being broken in the first place. Meanwhile, Bratton resorts to ad hominem attacks on his critics:

Many social scientists are wedded to the idea that crime is caused by the structural features of a capitalist society — especially economic injustice, racism, and poverty. They assume that true crime reduction can come only as the result of economic reform, redistribution of wealth, and elimination of poverty and racism.

It's time, I think, to tone down the rhetoric and get much more pragmatic. Crime fighting isn't science, as Bratton comes close to admitting:

What particularly galls police about these critiques is that ivory-tower academics — many of whom have never sat in a patrol car, walked or bicycled a beat, lived in or visited regularly troubled violent neighborhoods, or collected any relevant data of their own "on the ground" — cloak themselves in the mantle of an empirical "scientist" and produce "findings" indicating that broken windows has been disproved. Worse, they allege that police have had little to do with the declines in crime. Police don't have time for these virtual-reality theories; they do their work in the real world.

Which is why it's ironic that Bratton himself clings to the idea that Broken Windows has the status of empirical, scientific fact.

The way I see it, Broken Windows is just like the pedagogical theories in charter schools. If you have a police chief who believes in it, and who can energise his police force, and is backed up by the mayor, then it's very likely that crime will come down. But ultimately, that's probably more a function of the police chief and the mayor and the energised police force than it is a function of the universal applicability of Broken Windows.

Posted by Felix at 13:52 EST

Comments

I don't see how Bratton's attack on the "economic injustice, racism, and poverty" basis of crime is in the least ad hominem. But whatever.

More to the poinjt, my own take on why the "broken windows" approach works refers back to Jane Jacobs' "eyes on the street" approach. Fixing broken windows is a sign that people are watching and care and will call the cops.

Posted by: David Sucher at 16:29 EST, March 05, 2006

David, it is an ad hominem attack, since Bratton is going after the social scientists, and what they're "wedded to", rather than the actual substance of their papers.

And before we can work out why Broken Windows works, we need to ascertain whether Broken Windows works. That's the debate I'm talking about, and the debate which I ultimately conclude is rather silly. Really, it's Bratton's policing which works, rather than a theory.

Posted by: Felix at 16:43 EST, March 05, 2006

And my main point is that it does work, at least in my own experience. "Fixing things" acts as a territorial marker. It says that this ground is claimed and a vandal will be challenged. To me it's obvious and commonsense.

But then again, this is not a subject I know about; so what other alternatives are there in police theory?

Posted by: David Sucher at 17:04 EST, March 05, 2006

But David, fixing broken things is an effect, not a cause. Heightened vigilance can be shown have a causal connection on decreased crime. But you can't prove that window repairs correlate to vigilance (the city could dispatch crews to repair windows without causing a change in personal vigilance on the part of residents).

There are better metrics, including the central component of Bratton's policy, COMPSTAT, which enabled a more dynamic response to patrolling (rapid escalation of patrols in areas with high levels of crime reporting -- it seems fairly obvious, but you have to understand what a sea change this is; a friend was mugged on the Williamsburg bridge in 1999, and had to go in person to the precinct for a follow up, since they couldn't manage to fax her a copy of the complaint -- because they didn't have a functioning fax machine.)

There was absolutely a measurable correlation between increased patrolling and reduction in crime. I;m sure there were some repaired windows as well.

Posted by: 99 at 18:20 EST, March 05, 2006

"...fixing broken things is an effect, not a cause."

Huh? I don't understand.

I guess I am saying --
Does the theory work? And why?
-- that "fixing things" is in fact very much a territorial marker and has a direct causal impact in deterring bad behavior.

Posted by: David Sucher at 18:46 EST, March 05, 2006

David, you are asserting that "fixing things" deters bad behaviour. The various empirical studies cited in the Boston Globe piece seem to show that, in fact, fixing things does not deter bad behaviour. Bill Bratton takes issue with those empirical studies. I think that concentrating on theory is silly, and that all that matters is how police chiefs and police officers behave in practice -- a weak police chief with no support from the mayor or from his officers might implement Broken Windows techniques as much as he liked, but no good would come of it.

You also seem to elide two different concepts: that of the police cracking down on minor vandalism to combat violent crime, on the one hand, and that of individual citizens "fixing things" like broken windows in order to get more support and sympathy from the police, on the other.

You say that you think Broken Windows works -- maybe you could elucidate for us exactly what Broken Windows is, in your view, and then explain whether your belief is simply a gut feeling, or whether it has an empirical basis.

Posted by: Felix at 19:12 EST, March 05, 2006

Broken windows theory was never any news to any half-decent community development worker, it's common sense. It's not saying 'fixing things deters bad behaviour;' it's saying that having an ordered environment encourages informal social control, which, together with other factors, can deter antisocial behaviour. Just fixing things, as an isolated policy, is unlikely to work. Which raises the question, what are the other factors? Well, for instance, what about the extent to which an ordered environment reduces the fear of reprisals, so that residents might be more willing to intervene; or the extent to which residents in mixed housing tenure are inclined to maintain their own ordered environment through subtle peer pressure; or the impact of family networks on informal social control within a neighbourhood... - these are links in the chain which need clarifying.

On the importance of the perception of police support as a component in informal social control, there was a really good paper by Rowland Atkinson and John Flint called 'Order born of chaos?' in Policy and politics back in 2004. Liz Richardson at LSE has recently come up with a similar finding.

kevin

Posted by: Keivin Harris at 10:07 EST, March 07, 2006

Kevin above says it nicely.
I was starting to form a response here and for my own blog but I keep coming back to "What other choice is there but to fix the broken wiondows?"

Posted by: David Sucher at 10:44 EST, March 07, 2006

Heightened vigilance can be shown have a causal connection on decreased crime. But you can't prove that window repairs correlate to vigilance (the city could dispatch crews to repair windows without causing a change in personal vigilance on the part of residents).

If the city (why the city by the way and not property owners generally?) can despatch a maintenance crew then someone is watching surely?

Posted by: ian at 6:52 EST, March 24, 2006