Until recently, Broken Windows critics could dismiss New York’s success in fighting crime as anecdotal, suggesting correlation without causation. But in recent years, at least three randomized experiments, published in refereed journals, attest to Broken Windows’ impact on crime. Rutgers criminologist Anthony Braga and his colleagues conducted two field experiments: the first in Jersey City, New Jersey; and the second in Lowell, Massachusetts. In each case, multiple high-crime areas of the city were identified and randomly assigned to experimental and control conditions. Police in the control areas continued routine policing. In the experimental areas, police took a problem-solving approach that, in both cities, involved aggressive order maintenance. Crime declined in the experimental areas at greater rates than in the control areas and was not displaced to adjacent neighborhoods.
In the Netherlands, experimenters took a different approach. Their findings support the central social insight of the Broken Windows theory: that disorder breeds crime. University of Groningen social scientist Kees Keizer and his colleagues conducted six field experiments in which they artificially created opportunities for crime in both orderly and disorderly environments. In one case, they placed an envelope containing visible cash so that it was hanging out of a postbox. The baseline condition was a clean postbox; the experimental condition was a postbox covered with graffiti and surrounded by litter. In the baseline condition, 13 percent of those who passed the postbox stole the money. In the experimental condition, 27 percent stole the money. The other five experiments had similar outcomes.