As homicide rates climbed around the country in the 1980s and 1990s, dozens of police departments began to upend the way they patrolled inner-city neighborhoods. Officers started getting out of their cars and making the rounds on foot, talking to residents and looking for opportunities to prevent crime — not just respond to it after the fact. Called “community policing,” the efforts coincided with drops in crime in some cities, making the approach seem like a great success.
But there is no evidence that lower crime rates had anything to do with community policing. An analysis based on 65 before-and-after assessments of community-oriented policing initiatives by a small group of criminologists found there was no statistically significant reduction in crime, or even in citizens’ fear of crime, as a result of these programs.
The research by these criminologists, led by David Weisburd of George Mason University’s Center for Evidence-Based Crime Policy, has refined the way that policing strategies get evaluated, yielding a new perspective on the state of the art — and sometimes directly contradicting conventional wisdom.
Weisburd is lead editor of a new book, What Works in Crime Prevention and Rehabilitation, which is the most comprehensive examination yet of policing strategies deployed by departments over the past two decades. Too often, the criminologists’ research shows, policing initiatives are driven by shifts in local crime rates or by theories that seem to make intuitive sense — like putting cops in closer contact with the communities they serve — rather than by hard empirical evidence. As a result, American cities tend to alternate between a breathless search for silver bullets to stop violent crime, and a narrative of failure in which new strategies themselves are blamed for perpetuating urban bloodshed — even when those strategies haven’t been given enough time to show results.
The good news, Weisburd says, is that the criminologists identified successful strategies to combat violent crime, especially gun crime, when properly deployed. By using more rigorous methods of evaluating programs, he explains, “we ended up with a strong body of evidence that crime prevention can work. And that’s an important message.”
One of the most effective policing methods, Weisburd and other researchers found, is pinpointed targeting by police of the most dangerous people and places in communities. It is also helpful, they found, when police or communities work to try to solve underlying social problems that can’t be erased by heavy-handed policing. These approaches are the core of one of the most consistently successful programs, called focused deterrence, in which cops and social services groups sit down with gang members to steer them away from lives of violence, rather than simply put them in handcuffs. Started in Boston in 1995, the approach has since been used in dozens of cities, including current programs in Chicago, Los Angeles, and New York.
Other policing strategies have shown promise, but have been discredited after spikes in urban gun violence pressure law enforcement officials and politicians to seek quick fixes. In Chicago, for instance, years of experimentation with aggressive seizures of illegal street guns and varieties of violence intervention under former Police Superintendent Garry McCarthy have been called into question by record-setting homicide totals and racial outrage over ugly police killings.
At the same time, the seemingly successful results of a crime-reduction policy can often lead to overly broad assumptions. Critiques of the Los Angeles Police Department’s cleanup of MacArthur Park, for instance, have credited trimming shrubs and picking up trash with a drop in muggings and other crimes.
Weisburd’s book and other major studies like it employ analytical tools called meta-analysis and systematic reviews — in essence, studies of studies that can compare results across thousands of individual evaluations to find significant impacts on crime rates. By taking a wide-angle view, researchers can discover sometimes-hidden patterns that aren’t changed by one study or two. They are also insulated from the politically motivated interpretations that sometimes drive the conclusions taken from smaller, individual inquiries.
The impulse to declare a crime prevention strategy a success or failure based on a narrow examination can lead to bad public policy decisions, says Daniel Webster of the Johns Hopkins Center for Gun Policy and Research. Take gun seizures. When they’re carried out as stop-and-frisk street sweeps, they can leave law-abiding citizens feeling as though the police are an occupying force. But that doesn’t change what the research shows about the approach when it is carried out in the prescribed way: using data to target the specific blocks and even specific people to search. “Focusing on gun offenders and their possession of guns is one of the most important things you can do for public safety in many communities,” says Webster. “We’ve largely forgotten that.”
Other prominent researchers have echoed Weisburd’s findings about crime prevention programs. Among the most notable big-picture analyses are papers by Rutgers’ Anthony Braga on policing, Harvard’s Thomas Abt on gun-violence prevention, and Anthony Petrosino of the WestEd Justice & Prevention Research Center on violence prevention.
The overall picture can still look messy, with many research gaps and unanswered questions remaining. Gun control and mass shootings, for example, get no attention in Weisburd’s book or in the other major studies mentioned here because there have not yet been enough smaller studies to form the basis for a larger analysis. Likewise, many specific programs and strategies — such as gunfire-detection systems like ShotSpotter — at this point remain unassessed by big-picture studies.
By combing through the latest crop of research, The Trace has devised the following report card on policing strategies aimed at reducing gun violence. That very specific language — policing strategies to curb gun violence — is key to understanding this score sheet, as we have excluded programs and strategies focusing on other kinds of crime or carried out by players other than the police. (One exception is the community-based violence intervention approach called Cure Violence, which we included because of how closely related it is to a similar crime reduction strategy.)
As the examples below show, some of the most well-known uses of certain programs can lead to ambiguous conclusions about whether a strategy works. But that is the point of the big-picture research: to go beyond single experiments to find the larger trends.
Here’s what the new mega-studies can teach policymakers, law enforcement leaders, and the journalists who cover urban crime.
Table of Contents
The strategy: Law enforcement agencies identify the most violent gang members and other likely shooters in a city or neighborhood and summon them to a sit-down. At these call-ins, police, federal agents, and prosecutors threaten immediate crackdowns if the shooting doesn’t stop. In return, community groups and social services organizations offer pathways to new lives for gang members, in the form of counseling, treatment, training, and other assistance. The approach is formally known as focused deterrence, or “pulling levers policing” (for the tactic of pulling all the levers necessary to achieve the desired response). The strategy is most closely associated with the National Network for Safe Communities’ Operation Ceasefire programs. It was used most famously in a pilot program in the mid-1990s: the so-called Boston Miracle that reduced the city’s homicide total by half.
One place that tried it: Starting in 2003, homicides began increasing in New Haven, Connecticut, bucking a national decrease in violent crime. By 2011, the city’s murder rate exceeded that of Chicago and Washington, D.C. The following year, New Haven launched its version of Ceasefire. After identifying the most violent gangs responsible for the city’s surge in shootings, police warned gang members of an immediate crackdown if the gunfire continued, while offering to help them with housing, drug counseling, and job training. In the program’s first three years, monthly shootings dropped on average by nearly 73 percent — a result that a team of Yale University researchers determined could not be explained by factors other than the use of the program.
What the research shows: Numerous studies consistently rate the method as effective. One major paper, for instance, found “significant reductions in targeted crime problems, particularly gang homicide.” Focused deterrence is also among the highest-ranked programs in Weisburd’s What Works. But sometimes it fails. The strategy’s chief architect, David Kennedy of the John Jay College of Criminal Justice, documented in his memoir the eventual collapse of multiple Operation Ceasefire programs, including even the celebrated Boston initiative. Kennedy and the researchers who champion focused deterrence say that blame for its spotty record rests with police departments, which at times have failed to sustain the program or carry out its elements faithfully.
HOT SPOTS POLICING
The strategy: Advanced crime-mapping software identifies specific geographic areas with the highest risk of violence. These “hot spots” — sometimes as small as one block or even one building — are then targeted with intense patrols and other attention from police. The most effective hot-spots approach focuses exclusively on the places and people most likely to turn violent — not broad, heavy-handed sweeps. This approach also uses a tactic called problem-oriented policing, which encourages officers to address underlying causes of crime — what sparked a particular turf battle between gangs, for instance — rather than rely on methods that can lead to racial targeting, such as stop-and-frisk.
One place that tried it: A data analysis in 2009 identified 22,000 Philadelphia street intersections as the most dangerous in the city. Police added new foot patrols to the top 1 percent of those locations citywide, where 120 pairs of officers per day walked beats, talked to residents, and stopped known troublemakers in cars and on sidewalks. A National Institute of Justice report on the experiment found a reduction of 23 percent in reports of violent crime in the program’s first three months.
What the research shows: This strategy has repeatedly lowered crime rates in violent areas, according to several studies cited in Weisburd’s book, and even helped reduce crime in adjacent neighborhoods — disproving suspicions that such programs merely displace rather than prevent crime. Other studies found the program has a more modest impact, meaning that while this method can help prevent violence, it might not be as effective as other strategies.
“BROKEN WINDOWS” POLICING
The strategy: This approach, which is actually a variety of programs that fall under the title “disorder policing,” strives to prevent violence and other serious crime by first addressing minor signs of social and physical disorder. (The theory holds that when cars with broken windows sit parked on a street for days, for instance, the message to the neighborhood is that law and order has broken down.) Adopted in various forms by hundreds of cities by the late 1990s, the strategy quickly became infamous for its more strident iterations — such as “order maintenance” and “zero tolerance” policing, in which police blanket neighborhoods and arrest people for low-level offenses. When misused, these methods tend to antagonize the very neighborhoods they were meant to protect. In the program’s more subtle forms, police collaborate with community members and other agencies to address underlying social problems that cause crime.
One place that tried it: The strategy gained national attention when it was used in New York City starting in the 1990s. Police targeted a wide range of disorderly and illegal behaviors: open-air drug markets, street-walkers, “squeegee men” who harassed drivers, panhandlers, public drinking, and turnstile jumpers in the subways, among others. These crackdowns proved popular with a public fed up with crime, and coincided with historic drops in the city’s violence rates. But they also caused rifts with the city’s black and Latino residents, who were disproportionately arrested by police.
What the research shows: A 2015 study by Braga and colleagues — which combined results from reports in multiple cities — gives broken-windows policing high marks for reducing crime. But the strategy’s success depends on its implementation. The paper concluded that the approach works when it’s used as a “community problem-solving” tool, but found that the sort of “aggressive order maintenance” methods used in New York City have little impact on crime. (There’s no scientific consensus on which combination of strategies deserves credit for the city’s dramatic drop in crime rates since the 1980s.) Braga’s own research documents one success story, in Lowell, Massachusetts. A 2005 randomized experiment showed that in the city’s 34 violent-crime hot spots, the use of broken-windows policing — cleaning vacant lots, razing abandoned buildings, boosting building-code inspections, and the like — caused a 20 percent drop in crime overall, with even larger reductions in robberies and assaults.
The strategy: As part of a broader hot-spots policing program, police teams focus exclusively on getting guns off the streets of violent neighborhoods. Trained to identify people likely to be carrying illegal weapons, and supplied with intelligence identifying known gun offenders, these guns squads frequently stop vehicles and pedestrians to search for weapons.
One place that tried it: In Kansas City, Missouri, in the 1990s, police deployed the strategy in an impoverished neighborhood with a homicide rate 20 times the national average. Before beginning the gun patrols, police went door to door to spread word of the coming crackdown and to ask citizens to report people carrying illegal weapons. Gun squads then spent 200 nights working the streets. Over 29 weeks, gun seizures increased by 65 percent and gun crimes fell by 49 percent. Evidence showed that the program not only reduced homicides and drive-by shootings, but also that it did so without pushing gun violence into surrounding neighborhoods.
What the research shows: The largest study to date of hot-spots patrols focused on seizing guns found they effectively reduce gun-related crime. Six out of seven additional major studies on the program’s use in the U.S. and Colombia revealed that gun crimes declined in targeted neighborhoods between 10 and 71 percent, with the American efforts producing the largest drops in violence.
PROMISING BUT UNPROVEN
The strategy: The gist: Treat gun violence as a contagious disease, and it can be cured with the help of “interrupters” — usually former gang members — who intervene to prevent retaliatory violence. These interventions can include mediating disputes and offering social services to gang members and other likely shooters. The key to this strategy is a tight focus on small groups of people who face the greatest risk of gunshot injuries. The work of Yale’s Andrew Papachristos, for instance, shows how violence can spread within these networks like an epidemic. “As an individual’s exposure to gunshot victims increases, so, too, do that individual’s odds of victimization,” he writes. Founded in Chicago in 1999 as CeaseFire, confusingly adopting the name of a similar focused-deterrence approach, the strategy now often goes by the name Cure Violence. More than two dozen other cities have implemented the program.
One place that tried it: After Baltimore’s version of the strategy, Safe Streets, showed dramatic results in one violent neighborhood in 2007 — homicides stopped completely in the first 22 months of the program — it was rolled out in three other areas of the city. According to a study overseen by Johns Hopkins’s Webster, the results varied widely, from big drops in shootings in some neighborhoods to an increase in homicides in another. Webster says that, on paper, Baltimore’s uneven results might suggest the entire approach is bunk. But a closer look revealed that whether the strategy succeeded or failed depended on how faithfully it was implemented. For instance, crime rates were unchanged where interrupters lacked the skills and drive to do the work effectively. But when properly motivated, trained, and supervised, the workers succeeded at preventing further violence. “Investing in the right guys to do this kind of work with good supervision clearly will pay off,” Webster says.
What the research shows: A review of the research last year found the strategy lowers some types of crime in some places, but not consistently enough to brand it an unqualified success. When it does work, however, the results can be impressive. One study, for example, analyzed 16 years of data that compared areas in Chicago that used the Cure Violence approach with those that didn’t. In four of seven Cure Violence sites, shootings declined between 16 to 34 percent because of the program, researchers concluded. In another study, Massachusetts’ Safe and Successful Youth Initiative was credited by American Institute for Research and WestEd researchers with preventing nearly 1,000 violent crimes between 2011 and 2013, and with reducing — by more than half — the odds that program participants would end up in jail. The program’s tactics included protecting kids walking to school, alternatives to school suspensions, and, most critically, interrupting potential acts of violence by confronting would-be offenders with warnings and offers of social services.
The strategy: Running parallel with the debate over how police should respond to street violence is a discussion over how they should respond to violence committed by individuals within their own ranks. The effort to mend relations between police and residents of America’s most violence-prone neighborhoods is the focus of two approaches: promoting legitimacy and procedural justice. Though not aimed strictly at reducing violence, these programs try to foster healthy relationships between police and the communities they serve by retraining police to neutralize racial bias and to de-escalate confrontations with citizens. As officers are transformed from “warriors” to “guardians,” in the lingo of theorists, these efforts ultimately seek to curb violent crime.
One place that tried it: The Minneapolis police department has a history of racial tensions with local Somali refugees and a documented pattern of racially biased arrests. So the department was a natural choice for one of six pilot programs of the National Initiative for Building Community Trust & Justice, a federally funded initiative begun in 2014 to cultivate trust between police and residents of minority neighborhoods. The Minneapolis program trains the city’s 800 officers to recognize and avoid deeply ingrained racial attitudes. This training includes simulations in which officers’ reactions to black and white suspects, armed or unarmed, show how race alters their perception of a threat, even when the officers themselves are people of color. Another priority of the Minnesota program: changing policies to reduce the chances of lethal conflicts between police and citizens that further enflame racial tensions. Begun last year and still in progress, Minneapolis’s efforts have been frustrated, paradoxically enough, by racial tensions with local police. Widespread protests broke out in December after two white officers shot and killed a young black man. Meanwhile, gun violence in the city’s predominantly black neighborhoods is up sharply this year, accelerating a five-year increase in shootings.
What the research shows: Studies have found strong evidence that this approach promotes better relationships between residents and police. But its impact on crime rates remains uncertain. A leading report surveying the latest research concluded that there is at least a “marginally significant” effect on crime, but there are too few studies to reach more definitive conclusions.
WHAT DOESN’T WORK
The strategy: Community-oriented policing is one of the most widely used policing strategies, funded by billions in federal grants since the 1990s. The approach is driven by two primary goals: to move police beyond a strictly reactive role, and to involve community residents in reducing crime. At its most basic level, this approach deploys cops on foot patrols, and has them organize social events like picnics and basketball tournaments to get to know local residents. In its more substantive forms, community policing emphasizes crime-prevention efforts, designed in collaboration with local businesses and neighborhood groups, rather than traditional command-and-control policing that generally relies on reacting to crime instead of trying to prevent it.
One place that tried it: Lee Brown stressed community-oriented policing when he was chief of police in Houston in the 1980s. With homicides in the nation’s fourth-largest city topping 700 in 1981, the year before he took over, Brown’s department instituted a host of community-outreach activities in violent neighborhoods, including meetings, parades, and youth recreation programs, many of which remain in place. Homicides in Houston did eventually ebb, reaching a low point in 2011, long after Brown had moved on to run the New York Police Department. Despite this turnaround, researchers were unable to find any clear evidence that community policing was responsible for the drop in violent crime.
What the research shows: An analysis based on 65 before-and-after assessments of community-oriented policing programs turned up no evidence of significant reductions in crime. “Simply engaging the community doesn’t seem to translate directly into crime reduction gains,” Rutgers’ Braga wrote last year. Still, he said, community policing serves as a solid foundation for crime prevention, provided it includes “important inputs to help focus crime reduction strategies such as problem-oriented policing, hot spots policing, and focused deterrence approaches, which do seem to reduce crime.”
[Photo: Spencer Platt/Getty Images]