Alan Hevesi vows not to “take a step backward in the fight against crime.” Peter Vallone opines that the shooting of Amadou Diallo was a tragedy “not just for Amadou Diallo, [but] for the four officers involved.” Mark Green is even courting former city Police Commissioner William Bratton, the original architect of Giuliani’s crime-control strategy, promising he’ll “put more cops on the street.”

Even though Diallo and Dorismond made New Yorkers of all colors uneasy about quality-of-life policing, mayoral candidates are loath to distance themselves from Giuliani’s infamous “zero tolerance” tactics. Most people credit Giuliani for the city’s drop in crime, and New Yorkers seem willing to pay a steep price for their newfound sense of safety. But what if everything that popular opinion thinks about the city’s declining crime rate–right down to the theory it’s supposedly based on–turns out to be wrong?

Giuliani’s policing strategy, developed under Bratton, drew on the popular “broken windows” theory: that little things–public drinking, turnstile jumpers, broken windows–add up to a looming sense of disorder that keeps ordinary, law-abiding people off the streets, leaving the public sphere to criminals. But when it comes to policing–especially policing neighborhoods–it turns out that citizen support for the police is one of the “little things” that matter.

The idea behind “broken windows” is to make public space hospitable for everyone by eliminating incivilities that keep people off the streets. It’s designed with policing communities, not people, in mind. Zero tolerance policing, by contrast, aggressively targets people–with results that even broken windows proponents agree can be disastrous. George Kelling, one of broken windows’ originators, is a vigorous defender of New York’s policing. But even he points out that the theory as he originally conceived it was intended not to justify aggressive police stops, but to reinforce neighborhood cohesion. “What’s going on is the restoration of public spaces,” says Kelling. “In some places, it’s done wonderfully, and in some places, it’s done terribly.” So how did the broken windows theory turn into zero-tolerance policing–and can it be salvaged?

regardless of the tactics used to implement it, broken windows itself has hit hard times. Kelling and James Q. Wilson made criminological history in 1982, when they argued in Atlantic Monthly that public disorder–panhandling, graffiti, groups of unsupervised youth–signals would-be criminals that no one is watching. Disorder makes citizens withdraw from public places, they wrote, ceding those areas to disorderly people who feel little restraint to commit crimes. Wilson and Kelling believed disorder and its message of tolerance for crime was contagious, spreading from citizen to citizen and neighborhood to neighborhood–a kind of domestic domino theory.

There’s one major problem: a significant lack of evidence that disorder encourages crime. Several recent studies have blown apart “disorder-policing” research from the early 1990s, showing instead that the contribution of disorder to crime rates has been small at best. University of Chicago sociologist Robert Sampson found that many of the factors that produce disorder in neighborhoods–poverty, inequality, poor housing–also produce crime. In Baltimore, Temple University criminologist Ralph Taylor found that disorder and crime are loosely connected, but without a causal relationship. And in 1998, University of Arizona law professor Bernard Harcourt reanalyzed data from six cities, joining the chorus that found no evidence of a solid connection between disorder and crime.

But the influential theory survives in the public imagination as proven fact–and as the justification for zero-tolerance policing. When Giuliani hired Bratton as police commissioner, he applied the broken windows theory to a range of minor crimes: jaywalking, panhandling, open bottles. Bratton even worked with Kelling, as head of New York’s transit police, to target turnstile jumpers–whom he would later call “the biggest broken window in the transit system.” Later dubbed “zero tolerance,” Bratton’s method of policing swept the nation, earning praise and emulation from departments nationwide.

Kelling, a professor of criminal justice at Rutgers, disowns the term “zero tolerance,” claiming that it’s often used to describe tactics far afield from his work. But the argument most often used to justify zero tolerance policing–frequently cited by Bratton himself–derives from his interpretation of Kelling and Wilson’s theory: that when you crack down on minor crimes, you catch more serious offenders and take guns off the streets, leading to a drop in homicides.

Recent studies show these claims to be overblown at best, flatly wrong at worst. Looking at New York’s experience, John Jay College professor Andrew Karmen analyzed the role of aggressive policing in New York City’s declining murder rates. In the resulting book, New York Murder Mystery, Karmen documented that the crime reduction fruits of these stop-anyone-suspicious tactics were meager, and simply could not explain the decline in homicide. After all, other cities had similar declines in crime without New York’s tactics.

But zero tolerance wasn’t just an experiment that didn’t pan out. Instead of building community confidence in public space, such policing undermined it by aggressively policing poor people in poor places. Studies by the state attorney general, the Civilian Complaint Review Board, and newspaper accounts all tally with what citizens already knew: that police stops were often accompanied by verbal and physical abuse. In the city’s poorest neighborhoods, and especially among African-American and Latino citizens, zero tolerance left a bitter aftertaste of citizen distrust of the police.

A growing body of evidence shows that citizens are more likely to comply with the law and cooperate with law enforcement when they feel they have been treated fairly by the police and the courts. New York University social psychologist Tom Tyler showed in his 1990 book Why People Obey the Law that fair treatment is a critical factor that shapes how citizens behave toward law and the criminal justice system. Closer to home, a recent study by Rob Davis of the Vera Institute of Justice showed much the same in two precincts in the Bronx: Fair treatment, or “procedural justice,” is a far more effective way to solve persistent crime problems plaguing neighborhoods.

And here, even Kelling agrees with the critics. Police-citizen partnerships at the street level are the best version of his theory, he says. “I have always defined it as negotiating a consensus about neighborhood standards of behavior,” Kelling explains. “Police action reinforces that standard. The problem with zero tolerance is that it gives the impression that police set the standard.”

Todd R. Clear is a professor at John Jay College of Criminal Justice. Jeffrey Fagan is a professor of law and public health at Columbia University.