The passing of of eminent political scientist James Q. Wilson has brought renewed attention and almost universal praise to his most famous work, Broken Windows Theory. While Wilson will certainly be missed and Broken Windows was a rare and valuable fresh addition to police doctrine, now that his theory has been put to the test for 30 years, it’s time to reevaluate its success and some of its assumptions just as Wilson began to do later in his life.
“Broken Windows: The Police and Neighborhood Safety” was a groundbreaking article by Wilson and Police Foundation researcher George Kelling published in the March 1982 edition of The Atlantic Monthly. The basic premise as implemented and remembered today was that small crimes like graffiti and vagrancy lead to larger crimes like burglary and murder by eroding order and a sense of control in communities. The name comes from an observation that if a broken window in a building is not fixed immediately, the other windows will likely be broken as well. Broken Windows gained many followers including New York Mayor Rudy Guiliani and NYPD Commissioner Bill Bratton. New York’s stunning decrease in crime was subsequently attributed to the aggressive strategy and police work was affected nationwide, shifting the focus from solving major crimes to responding to minor offenses. Even today, departments are continuing to switch to the Broken Windows strategy, such as Detroit, which plans to adopt the approach to cope with its infamous crime problems.
But what were Wilson and Kelling actually proposing? “Broken Windows” was much more nuanced and complex than simply targeting minor offenses and provided methods and rational that most would now find controversial. The article fondly recalled the rough and tumble early days of policing where patrolmen on foot might rough up young thugs to teach them a lesson without a warrant, arrest, or conviction. They also saw the key role of the police officer as enforcing order rather than the law. This was the central tenet of Broken Windows and the major rationale for focusing on minor or victim-less crimes. Order in this case is defined by the needs, tastes, and traditions of the community. Wilson and Kelling explained how even in poor neighborhoods, orderly communities typically have a set of upstanding citizens, accepted derelicts to be controlled and follow a set of rules such as not drinking in public or not lying on stoops, and strangers to be kept out and moving. The idea is that functional communities can handle their own but will not tolerate outsiders. The authors themselves recognize this to be potentially troublesome, and express fears that such a manner of policing would be used to impose racial segregation. Even if officers are vigilant to draw the line at enforcing segregation as Wilson and Kelling suggest, the original policy is intentionally extrajudicial and discriminatory, made up of ambiguities that ripe for abuse.
And on what did Kelling and Wilson base their ideas? The authors had a few major influences. The first was basic inference from their own daily lives about ordered and untended spaces, for example the broken windows of the title. They also drew on several studies that, though revolutionary and revealing, were tangential to the question at hand. Among them was a 1969 study by a Stanford psychologist, Philip Zimbardo, who left an abandoned car in the Bronx and an identical car in Palo Alto. Vandals attacked the Bronx car within 10 minutes, it was stripped in the first 24 hours, and then randomly damaged. In the more orderly Palo Alto, the car was untouched for over a week until Zimbardo hit it with a sledgehammer. Within a few hours the car was entirely destroyed and flipped upside down. The other study was conducted by Kelling and the Police Foundation on the effect of foot patrols across New Jersey in the 70s. Increased vehicle patrols had already been shown not to affect crime and, unsurprisingly, foot patrols failed to lower the crime rate compared to control groups. Despite this, however, there was a notable increase in community satisfaction with police service and sense of security.
From these anecdotes and research not directly related to crime prevention, Kelling and Wilson formed a hypothesis that we’ve improperly labelled a theory. While observation and conjecture are sufficient to form a hypothesis, a theory needs to be rigorously tested and confirmed. Broken Windows Theory sounds very reasonable from the limited data presented, but presuming it to be an accurate model of crime would be extrapolation. To illustrate the dangers of this practice, Mark Twain once calculated that, due to alterations in the path of the river for irrigation, the Lower Mississippi was shrinking at an average rate of a mile and a third a year. Using that information, he then projected a date in the future when it would disappear entirely and a date in the past when it hung over the Gulf of Mexico “like a fishing pole.”
As a hypotheses, however, Broken Windows was much more robust, and held substantial promise at a time when crime seemed to be an irreversible trend. To some extent, it has since been tested. The most famous example of the implementation of Broken Windows theory, as stated earlier, was in New York in the 90s, where it was attributed with a remarkable decrease in crime. Unfortunately, that popular conclusion doesn’t hold up under more scientific scrutiny. There was no control group for New York, but when compared to other cities that did not become tough on minor crimes and violations over the same period, it isn’t clear that New York fared any better due to “zero tolerance” policies. Broken Windows principles were only a small part of the changes New York was going through both with crime and policing, not the least of which was the decline in the crack epidemic.
When judging the effectiveness of Broken Windows policing, it’s also important to remember that many agencies, departments, pundits, and politicians may be talking about different things when they say they based their policies on the Broken Windows Theory. Invariably this means going after minor offenses more aggressively, but this can be done in a number of ways, as evident by all the buzz words that has surrounded the approach such as “community policing” and “zero tolerance.” As Fred Leland notes, implementing a Broken Windows approach tends to be very selective. Kelling and Wilson explain, for example, that you can’t work with the community from a passing patrol car, but departments will continue to try. Implementing Kelling and Wilson’s suggestions completely has yet to be attempted, however, as that would be a major departure from current norms and even laws. American police officers still work to uphold the law before the will of the community they serve, and must work within it. They are also expected to answer all the calls for service they receive, making a return to foot patrol and almost complete officer discretion practically impossible except in limited situations. That isn’t what most people mean when they praise or call for Broken Windows policing, and it’s not clear whether Wilson and Kelling would recognize many of the law enforcement strategies attributed to their theory.
Some elements of Broken Windows Theory and the methods developed from it, however, have been put to the test, and we now have a better idea of what works and what doesn’t. The National Research Council recently reviewed studies done on the various elements of Broken Windows Theory to determine which have the strongest evidence in their favor. Among policing strategies that relied mostly on sanctions such as fines and arrests, unfocused efforts such as rapid response and increased patrol showed little to no evidence of effectiveness, while focused efforts such as “hot spots” policing had moderate to strong evidence of effectiveness. Strategies that apply a diverse array of approaches were more effective overall, with unfocused efforts such as increasing police contact with citizens having weak to moderate evidence of effectiveness, focused strategies like problem-oriented policing showing moderate evidence, and the most focused techniques, problem solving in hot spots, showing strong evidence of effectiveness. Any of these approaches can be seen as some part of Broken Windows, whether that’s the unproven method of making a community appear more orderly with highly visible patrols or the more supported problem-based hot-spot method of making the most high-crime area seem well tended by targeting their specific issues such as burglaries or drug corners.
Given the data, it seems that the “keeping nice neighborhoods nice” interpretation, with all of its overt and tacit implications, doesn’t hold up as an effective policing strategy. And while many of the benefits of Broken Windows are unproven, we know that it has its costs, both financial and societal, such as filling already overcrowded prisons with non-violent offenders. Those minor arrests also don’t come cheap. In Cop in the Hood, Peter Moskos notes how there are numerous street blocks in Baltimore that cost the city over a million dollars a year in processing, overtime, legal fees, etc. due to arrests, with many cases getting thrown out anyway due to an overloaded court system.
An unfocused, sanctions based approach, the traditional interpretation of Broken Windows, has shown little evidence of reducing crime and has numerous downsides Focusing on specific problems and implementing a variety of methods such as crime prevention, however, has been shown to work and may even save money. This isn’t a condemnation of Wilson or the departments that have seen success following his ideas, and in fact, given the evidence over the years, they approach the strategy with more nuance than politicians and journalists throwing the term “Broken Windows” around. When asked in interviews, including one on NPR after Wilson’s death, if he attributed more aggressive police sanctions to the dramatic reduction in crime in New York, Bratton, perhaps the most public supporter of Broken Windows Theory, is quick to point out that the NYPD implemented a mix of strategies, that Broken Windows and “zero tolerance” were only a few, and that other factors were also at play. In another interview, Wilson says the same, and openly states that Broken Windows is far from proven. Wilson even softened his stance on crime and punishment with the decline of the crack epidemic, such as life imprisonment without parole for juvenile offenders, which he first supported then opposed.
All of this nuance – The true meaning of Broken Windows Theory, the questionablelegality of a true community-based, Broken Windows approach, the lack of true evidence for a hypothesis that is popularly called a theory, and the need for a multifaceted strategy for the complex issue of crime – gets lost in the glowing praise and wide acceptance of Broken Windows by officials, politicians, pundits, and journalists, especially in recent days after James Q. Wilson’s passing. the results Too often the results are more questionably effective and questionably ethical policies like the ones which even Wilson grew to oppose. This is a disservice to Wilson’s memory, not only because it cheapens his ideas and legacy but because it does not adhere to his rigorous intellectual standards.