فصل 11

Chapter Eleven

Case Study: The Kansas City Experiments

A century ago, a legendary figure in American law enforcement named O. W. Wilson came up with the idea of “preventive patrol.”1 Wilson believed that having police cars in constant, unpredictable motion throughout a city’s streets would deter crime. Any would-be criminal would always have to wonder if a police car was just around the corner.

But think about it. When you walk down the street of your neighborhood, do you feel like the police are just around the corner? Cities are vast, sprawling places. It’s not obvious that a police force—even a large police force—could ever create the feeling that they were everywhere.

This was the question facing the Kansas City Police Department in the early 1970s. The department was about to hire extra police officers, but it was divided over how to deploy them. Should they follow Wilson’s advice—and have them drive randomly around the city? Or assign them to specific locations—such as schools or difficult neighborhoods? To resolve the question, the city hired a criminologist named George Kelling.

“One group said riding around in cars doesn’t improve anything, it doesn’t do anything,” Kelling remembers. “Another group said it’s absolutely essential. That was the standoff. Then I was brought in.” Kelling’s idea was to select fifteen beats from the southern part of the city and divide them into three groups. It was a big area: thirty-two square miles, 150,000 people, good neighborhoods and bad neighborhoods, and even a little farmland on the fringes. One of the three groups would be the control group. Police work would continue there as it always had. In the second neighborhood, Kelling would have no preventive patrol at all; police officers would respond only when called. In the third neighborhood, he would double and in some cases triple the number of squad cars on the streets.

“Nothing like this had ever been done in policing,” Kelling remembers. “This was 1970. Nothing had been written about police tactics.…This was at a very primitive stage in policing.” People like O. W. Wilson had ideas and hunches. But police work was considered an art, not a science that could be evaluated like a new drug. Kelling says that many people told him his experiment would fail, “that the police simply weren’t ready for research. I wouldn’t be able to do it. They’d sabotage it.” But Kelling had the backing of the city’s police chief. The chief had spent the bulk of his career in the FBI, and he was shocked to learn how little police departments seemed to know about what they did. “Many of us in the department,” the chief would later admit, “had the feeling we were training, equipping, and deploying men to do a job neither we, nor anyone else, knew much about.” He told Kelling to go ahead.

Kelling ran the experiment for a year, meticulously collecting every statistic he could on crime in the three areas of the study. The result? Nothing. Burglaries were the same in all three neighborhoods. So were auto thefts, robberies, and vandalism. The citizens in the areas with beefed-up patrols didn’t feel any safer than the people in the areas with no patrols. They didn’t even seem to notice what had happened. “The results were all in one direction and that was, it doesn’t make any difference,” Kelling said. “It didn’t matter to citizen satisfaction, it didn’t matter to crime statistics, it just didn’t seem to matter.” Every police chief in the country read the results. Initially, there was disbelief. Some urban police departments were still committed Wilsonites. Kelling remembers the Los Angeles Police Chief standing up at one national law-enforcement conference and saying, “If those findings are true, every officer in Kansas City was asleep at the switch because I can assure you that’s not how it is in Los Angeles.” But slowly resistance gave way to resignation. The study came out as violent crime was beginning its long, hard, two-decade surge across the United States, and it fed into the growing feeling among people in law enforcement that the task before them was overwhelming. They had thought they could prevent crime with police patrols, but now the Kansas City PD had tested that assumption empirically, and patrols turned out to be a charade. And if patrols didn’t work, what did? Lee Brown, chief of the New York City Police Department, gave a famous interview in the middle of the crack epidemic in which he all but threw up his hands. “This country’s social problems are well beyond the ability of the police to deal with on their own,” Brown said. He had read George Kelling’s Kansas City report. It was hopeless. No matter how many police officers a city had, Brown said, “You could never have enough to use traditional policing techniques to deter crime.…If you don’t have a police officer to cover every part of the city all the time, the chance of an officer on patrol coming across a crime in progress is very small.” In 1990, President George H. W. Bush came to Kansas City. He spent the morning in one of the city’s poorest and most violent neighborhoods, then gave a speech to a group of local police officers. He tried to be upbeat. He failed. The homicide rate that year in Kansas City was three times the national average. It would go up again in 1991 and again in 1992, then once more in 1993. There wasn’t much to say. Halfway through his remarks, Bush was reduced to simply listing the terrible things happening on the city’s streets: A four-year-old boy shot dead in a suspected crack house; an eleven-year-old kid gunned down outside another drug den, allegedly at the hands of a fourteen-year-old guard; in a downtown bar, a mother sells her baby for crack; and a firebombing leaves three generations dead, including a grandmother and three little kids—the headlines are horrifying, sickening, outrageous.

But in the early 1990s, twenty years after the first Kansas City experiment, Kansas City decided to try again. They hired another brilliant young criminologist named Lawrence Sherman. As they had with George Kelling, they gave him free rein. It was time for Kansas City Experiment Number Two. Why not? Nothing else was working.

2.

Lawrence Sherman thought the focus ought to be on guns. He believed the sheer number of guns in the city was what fueled its epidemic of violence. His plan was to try a number of ideas in sequence, rigorously evaluate their effectiveness—as Kelling had done—and pick a winner. He called a planning meeting with a group of the city’s senior police officers. They chose as their testing ground Patrol District 144: a small, 0.64-square-mile neighborhood of modest single-family homes, bounded to the south by 39th Street and to the west by Highway 71. District 144 was as bad as Kansas City got in the early 1990s. The homicide rate there was twenty times the national average. The area averaged one violent felony a day and twenty-four drive-by shootings a year. A third of the lots were vacant. Just a few months before, an officer had been on patrol through 144 when he saw some kids playing basketball in the street. He stopped, got out, and asked them to move. One of the players threw the basketball at his head, then two others jumped him. It was that kind of place.

Sherman’s first idea was for two-man teams to knock on every door in the neighborhood over a three-month period. The officers would introduce themselves, talk about gun violence, and give the residents a flyer with an 800 number on it: if they heard anything about guns, they were encouraged to call in an anonymous tip. The plan went off without a hitch. In many of the visits, the officers were trailed by a graduate student in criminology, James Shaw, whose job was to evaluate the program’s effectiveness. Sometimes the officers stayed for as long as twenty minutes, chatting with people who had never had a police officer come to their door other than to make an arrest. In his subsequent write-up, Shaw was effusive: The police went to every residence in that community, some more than once, and talked to residents in a friendly, non-threatening manner. In response, people were very receptive and glad to see the police going door to door. People frequently responded with comments like “God bless you all, we shoulda’ had a program like this before,” or “Thank God! I didn’t think you all would ever come.” In the end, 88 percent of the people visited said that they would use the hotline if they saw any guns. So how many calls came in—after 858 door-to-door visits over three months? Two. Both were about guns in another neighborhood.

The problem, everyone soon realized, was not that the residents of District 144 didn’t want to help. They did. It was that they never left their houses. “It’s starting to sound like Beirut around here,” one homeowner told Shaw, and if you’re so scared that you never leave your house, how on earth do you know who has guns or not? Shaw wrote: Not unlike residents in many other inner-city neighborhoods, these people have become like caged animals in their own homes; bars on the windows are the norm. One is not surprised even to see bars on second-story windows. More dismal however is the fact that in house after house the blinds are drawn and drapes closed up tightly, blocking out any trace of the outside world. These elderly people lock themselves up and shut themselves in. They hear the world outside, and it sometimes sounds like a battle zone. But they can’t see anything.

The group’s next idea was to train officers in the subtle art of spotting concealed weapons. The impetus came from a New York City police officer named Robert T. Gallagher, who in eighteen years on the force had disarmed an astonishing 1,200 people. Gallagher had elaborate theories, worked out over many years: street criminals overwhelmingly put their guns in their waistbands (on the left side, in the case of a right-hander), causing a subtle but discernible hitch in their stride. The leg on the gun side takes a shorter step than the leg on the nongun side, and the corresponding arm follows a similarly constrained trajectory. When stepping off curbs or getting out of a car, Gallagher believed, gun carriers invariably glance toward their weapons or unconsciously adjust them.

Gallagher flew to Kansas City, with great fanfare, the month after the failed hotline experiment. He gave seminars. He made videos. The officers took notes. The television program 20/20 sent a camera crew to record the technique in action on the streets of Kansas City. Nobody spotted anything. 20/20 came back again. The same thing happened—nothing. Whatever magical skills Robert T. Gallagher possessed were not, apparently, transferable to the beat cops of Kansas City. Two of the team’s best ideas for curbing gun violence had failed. They had one left.

3.

The winning entry in the Kansas City gun experiment was deceptively simple. It was based on a quirk in the American legal system.

The Fourth Amendment to the U.S. Constitution protects citizens from “unreasonable searches and seizures.” That’s why the police cannot search your home without a warrant. On the street, similarly, a police officer must have a good reason—“reasonable suspicion”—to frisk you.2 But if you’re in your car, that standard is not at all hard for a police officer to meet. Traffic codes in the U.S. (and in fact in most countries) give police officers literally hundreds of reasons to stop a motorist.

“There are moving violations: speeding, running a red light. There are equipment violations: a light that doesn’t work, a tire not quite right,” legal scholar David Harris writes.

And then there are catch-all provisions: rules that allow police to stop drivers for conduct that complies with all the rules on the books, but that officers consider “imprudent” or “unreasonable” under the circumstances, or that describe the offense in language so broad as to make a violation virtually coextensive with the officer’s unreviewable personal judgment.

There was even a Supreme Court case in which a police officer in North Carolina stopped what he thought was a suspicious driver, using the pretext that one of the car’s brake lights was out. As it turns out, it’s perfectly permissible in North Carolina to drive with one brake light out, so long as the other one works. So what happened after the driver of the car sued, claiming he had been stopped illegally? The Supreme Court ruled in favor of the officer. It was enough that he thought driving with only one brake light seemed like an infraction. In other words, police officers in the United States not only have at their disposal a virtually limitless list of legal reasons to stop a motorist; they are also free to add any other reasons they might dream up, as long as they seem reasonable. And once they’ve stopped a motorist, police officers are allowed, under the law, to search the car, so long as they have reason to believe the motorist might be armed or dangerous.

Kansas City decided to take advantage of this latitude. Sherman’s proposal was for the police department to detail four officers, in two squad cars. Their beat would be District 144. They were told not to stray outside the area’s 0.64 square miles. They were freed from all other law-enforcement obligations. They didn’t have to answer radio calls or rush to accident scenes. Their instructions were clear: watch out for what you think are suspicious-looking drivers. Use whatever pretext you can find in the traffic code to pull them over. If you’re still suspicious, search the car and confiscate any weapon you find. The officers worked every night from 7 p.m. to 1 a.m., seven days a week, for 200 consecutive days. And what happened? Outside District 144, where police business was conducted as usual, crime remained as bad as ever. But inside 144? All of the new focused police work cut gun crimes—shootings, murders, woundings—in half.

Remember, the police had all but given up by that point. Hotline? Nobody calls it. Concealed-weapons detection? A crew from 20/20 comes down and twice goes home empty-handed. Lee Brown, up in New York City, was mourning the powerlessness of the police to do anything serious about violent crime. Everyone remembered the previous Kansas City experiment, which had plunged the law-enforcement community into twenty years of despair. But now the same city had come back, and this time they were declaring victory. “I don’t know why it didn’t occur to us to really focus on guns,” the Kansas City police chief said after the results came in. He was as stunned as everyone else at what just two extra patrol cars had accomplished. “We usually focus on getting the bad guys after a crime. Maybe going after guns was too simplistic for us.” The first Kansas City experiment said that preventive patrol was useless, that having more police cars driving around made no difference. The second Kansas City experiment amended that position. Actually, extra patrol cars did make a difference—so long as officers took the initiative and stopped anyone they thought suspicious, got out of their cars as much as possible, and went out of their way to look for weapons. Patrol worked if the officers were busy. The statistics from the final report on the experiment were eye-opening. Over the seven months, each patrol car issued an average of 5.45 traffic citations per shift. They averaged 2.23 arrests per night. In just 200 days, the four officers had done more “policing” than most officers of that era did in their entire careers: 1,090 traffic citations, 948 vehicle stops, 616 arrests, 532 pedestrian checks, and 29 guns seized. That’s one police intervention every forty minutes. On a given night in the tiny 0.64 square miles of 144, each squad car drove about twenty-seven miles. The officers weren’t parked on a street corner, eating doughnuts. They were in constant motion.

Police officers are no different from the rest of us. They want to feel that their efforts are important, that what they do matters, that their hard work will be rewarded. What happened in District 144 provided exactly what the profession of law enforcement had been searching for: validation.

“Officers who recovered a firearm received favorable notoriety from their peers, almost to the point that recovery of a firearm came to be a measure of success,” Shaw wrote in his account of the program. “Officers could frequently be heard making statements such as ‘I’ve just got to get a gun tonight,’ or ‘I haven’t gotten a gun yet; tonight will be the night!’” In 1991 the New York Times ran a front-page story on the miracle in Kansas City. Larry Sherman says that over the next few days his phone rang off the hook: 300 police departments around the country bombarded him with requests for information on how he had done it. One by one, police departments around the country followed suit. To give one example, the North Carolina State Highway Patrol went from 400,000 to 800,000 traffic stops a year in the space of seven years.

The Drug Enforcement Agency used “Operation Pipeline” to teach tens of thousands of local police officers across the United States how to use Kansas City–style traffic stops to catch drug couriers. Immigration officials started using police stops to catch undocumented immigrants. Today, police officers in the United States make something like twenty million traffic stops a year. That’s 55,000 a day. All over the United States, law enforcement has tried to replicate the miracle in District 144. The key word in that sentence is tried. Because in the transition from Kansas City to the rest of the country, something crucial in Lawrence Sherman’s experiment was lost.

4.

The Lawrence Sherman who went to Kansas City is the same Larry Sherman who had worked with David Weisburd in Minneapolis a few years earlier, establishing the Law of Crime Concentration. They were friends. They taught together for a time at Rutgers, where their department chairman was none other than Ronald Clarke, who had done the pioneering work on suicide. Clarke, Weisburd, and Sherman—with their separate interests in English town gas, the crime map of Minneapolis, and guns in Kansas City—were all pursuing the same revolutionary idea of coupling.

And what was the principal implication of coupling? That law enforcement didn’t need to be bigger; it needed to be more focused. If criminals operated overwhelmingly in a few concentrated hot spots, those crucial parts of the city should be more heavily policed than anywhere else, and the kinds of crime-fighting strategies used by police in those areas ought to be very different from those used in the vast stretches of the city with virtually no crime at all.

“If crime is concentrated on a few percent of the city streets,” Weisburd asked, “why the hell are you wasting resources everywhere? If it’s coupled to those places and doesn’t move easily, even more so.” The coupling theorists believed they had solved the problem that had so confounded the earlier days of preventive patrol. How do you effectively patrol a vast urban area with a few hundred police officers? Not by hiring more police officers, or by turning the entire city into a surveillance state. You do it by zeroing in on those few specific places where all the crime is.

But think back to those statistics from North Carolina. If you go from 400,000 traffic stops in one year to 800,000 seven years later, does that sound like focused and concentrated policing? Or does that sound like the North Carolina State Highway Patrol hired a lot more police officers and told everyone, everywhere, to pull over a lot more motorists? The lesson the law-enforcement community took from Kansas City was that preventive patrol worked if it was more aggressive. But the part they missed was that aggressive patrol was supposed to be confined to places where crime was concentrated. Kansas City had been a coupling experiment.

Weisburd and Sherman say they have trotted out their maps and numbers, trying to convince their peers of the Law of Crime Concentration, to little effect. Back in the 72nd precinct in Brooklyn where he began his work, after a long day roaming the neighborhood, Weisburd would turn to the police officers he had been walking with and say, “Isn’t it strange how we’re returning again and again to the same blocks?” They would look at him blankly.

“I was in a meeting with the deputy commissioner [of police] in Israel,” Weisburd recalls.

Someone at the meeting said, “Well, David finds that crime doesn’t just move around the corner. And that would suggest that you ought to become more focused.” This guy turned around and he said, “My experience tells me that that’s just not true. I don’t believe that.” That was the end of that.3 Is something wrong with Israel’s deputy commissioner of police? Not at all. Because his reaction is no different from the behavior of the highway patrol in North Carolina, or the Golden Gate Bridge Authority, or the literary scholars who speak confidently of Sylvia Plath’s doomed genius. There is something about the idea of coupling—of the notion that a stranger’s behavior is tightly connected to place and context—that eludes us. It leads us to misunderstand some of our greatest poets, to be indifferent to the suicidal, and to send police officers on senseless errands.

So what happens when a police officer carries that fundamental misconception—and then you add to that the problems of default to truth and transparency?

You get Sandra Bland.