Experimental policy: when social science trumps politics

The "broken windows" theory of policing proposes that crime arises out of certain conditions and that tolerance of misdemeanor crimes can result in greater incidences of more serious crimes. If crime-ridden neighborhoods are cleaned up, if graffiti and jaywalking and loitering ordinances are enforced, the broken windows theory predicts violent and property crimes will abate in those neighborhoods.

As a theory, it sounds reasonable, and it has in fact been put into place as policy in many cities across the globe, including New York (where it is known as "zero-tolerance policing"), Boston, and yes, Milwaukee. Some of the cities saw crime rates drop dramatically, some saw smaller improvements, and some saw no change. So, as theories go, maybe it's less like a scientific theory, to be proven or disproven, and more like a philosophy, unable to be proved. (And of course, some of this variance in success is probably explained by the fact that not every police force is willing or able to implement all the elements of the practice and other neighborhood improvement efforts may have played a part, as well.)

Policymakers and public safety officers in Lowell, MA were adherents to the philosophy, but open to the possibility that, as a theory, it could be tested. So they worked with researchers from Harvard and Suffolk universities to implement a local policy experiment--something extremely rare in public policy research.

The experiment involved implementing "broken windows" policing practice in some high crime neighborhoods and policing as usual in others. Yes, you read that right, in some high crime neighborhoods they resisted making a policy change they believed in, so that they could measure the effect of the change in other neighborhoods. Policy effectiveness is most often measured after implementation and usually by third-party researchers such as academics. Local policymakers testing out a policy themselves, prior to adoption, is not the norm.

Why is this not done more often? Well, as you can imagine, no matter how the experiment turns out, testing a policy is quite a risk, politically. If the policy is proven not to work, local critics will not only argue time and money was wasted on it, all critics will now have data as ammunition against the policy. If the policy is proven to have a positive effect, there's an argument that time and money were wasted in waiting to fully implement it. For elected officials, research is not often worth the political risk. When it comes to crime, there are obviously other, more imperative risks as well.

Beyond the political risks, there are other reasons officials tend not to experiment with policy. At the local level, there is often neither the time nor money to give a new policy a test-run. At the state level, pilot projects are more common but, rather than being designed as empirical tests of effectiveness, they are usually seen as ways to phase-in implementation of new policy.

Kudos to Lowell for using scientific method to make policy decisions (the experiment confirmed their hypothesis and they have implemented the new policy) and for creating data that other policymakers can utilize. More municipalities should follow their lead. After all, in times of tough budgets, having reliable data can help when making the hard decisions.