A much-praised Pittsburgh anti-gang program gets a critical review from RAND, but it's not the last word on the subject.
Seven years ago, when Pittsburgh experienced an upward spiral in youth homicides, community leaders went to a former gang leader for advice on how to stop them. A year later, in 2004, the Allegheny County Department of Human Services established a violence prevention program called One Vision One Life, which deployed ex-prisoners? including gang members–as streetwise mediators to quell gang disputes and connect troubled youths in targeted neighborhoods to social services that could get them off the streets.
It surprised no one in Pittsburgh that the program would be run by the man who gave the advice. Richard Garland, a Philadelphia native whose gang-banger days were far behind him, had already developed a citywide reputation, as director of a Pittsburgh youth job training program, for successfully steering young people away from violence and crime.
One Vision, supported by $1.2 million in federal and local government funding and foundation grants, seemed to fulfill the community's hopes. It won praise from local clergy; and as late as last year, Pittsburgh Police Chief Nate Harper singled it out, in an interview with the Pittsburgh Tribune-Review, as the prime factor responsible for breaking up a conflict between two rival gangs.
Yet despite the praise, and a host of similarly compelling anecdotes, Garland's program received a blow last month when a RAND Corporation evaluation found aggravated assaults over the previous decade had actually increased in the neighborhoods where One Vision had been working: the crime-challenged Pittsburgh areas of Northside, Hill District and Southside. Moreover, the RAND study, which compared One Vision neighborhoods with similar Pittsburgh communities, found it had no effect on homicide rates.
Initial media reports focused misleadingly on the report's assertion that there was an “association” between the increased violence and One Vision's presence in those communities, but RAND researchers later made clear that this was merely a statistically significant correlation. –and did not imply that the organization contributed to the violence in any way.
RAND researcher Jeremy M. Wilson, an associate professor at Michigan State University's School of Criminal Justice who was one of the authors of the report, admits the evaluators were puzzled by the Pittsburgh data, which seemed to suggest that the program did not appear to have its intended effects. “It's a head scratcher,” he says, pointing out that despite the correlation the study can't actually explain what caused the bump in assaults, and there is no evidence that the nonprofit's activities were responsible for them.
Garland has since adopted a number of the RAND report's recommendations, and One Vision is now a partner in a larger citywide violence prevention strategy launched this month, called the Pittsburgh Initiative to Reduce Crime. Adopting a gang-prevention approach that has been shown to work in cities across the country, the new initiative brings together law enforcement, social services and community members to tell gang members that violence and homicides will no longer be tolerated.
But the questions raised by the RAND study also underline some fundamental questions about violence prevention in America's troubled urban neighborhoods. What kinds of strategies can stanch violence when so many factors, like geography and gang dynamics, are at play? Do some intervention methods do more harm than good? Are there limitations to the kind of street outreach and mediation that One Vision performs?
RAND's Wilson has some theories about why the numbers may have turned out the way that they did. Some other factor that the researchers had not considered–such as a change in public policy or demographics–could have impacted the final results, he notes. Additionally, studies of Los Angeles, Boston and Chicago have shown that, in some cases, gang prevention tactics can increase cohesion among gang members or heighten tensions between gangs, which can lead to an increase in violence. Or, since One Vision was working in the most violent neighborhoods to start with, it's not unreasonable to see a continued increase in assaults in the short term.
Therefore, RAND evaluators are not ready to dismiss the Pittsburgh model. “There is no denying that One Vision has made a difference in the lives of individuals,” Wilson says. “They are having an impact at the individual level in terms of affecting people's lives.”
However, adds Wilson, the approach still requires more thorough testing. He observes that the “dosage” –the number of people One Vision was reaching–”isn't high enough where it can be detected at a neighborhood level or beyond.”
RAND also found that documentation was not systematically collected, even though the program supposedly relies on data collection to guide its work. In addition, only 15 percent of One Vision's clients identified themselves as gang members. Researchers concluded that the organization tended to “focus on those who were most in need, as opposed to those at risk for violence,” Wilson says.
One Vision's Garland disagrees. His group, he says, works with “the kids no one wants.” “I think they asked the wrong questions,” he added in an interview with The Crime Report. “They were looking at the wrong things, and because programs like this are so hard to evaluate, they still don't know what to look for.”
Although Garland concedes that the study has helped him identify areas of improvement, such as the need for greater documentation and accountability, and increased pay and training for his street outreach workers, he is adamant that assault and homicide data should not be used as the critical measure of success or failure of his approach.
“You can't document that we potentially stopped four or five murders,” Garland told The Crime Report. “We have guys working with different cliques to make sure tensions stay down, and you can't quantify that.”
“One Vision One Life is very effective in small ways and in individual circumstances, but to think that they could affect homicides all over the city is expecting more than they can do,” adds Jay Gilmer, coordinator of the Pittsburgh Initiative to Reduce Crime. “We are very happy to have them as partner.”
Impact of Street Outreach
David Kennedy, director of the Center on Crime Prevention and Control at John Jay College of Criminal Justice, is serving as a consultant to the new Pittsburgh initiative, and he agrees that the issues underlying violence prevention are complex.
Kennedy, who has helped pioneer successful anti-violence and anti-gang programs across the country, says the RAND results are not surprising, because street outreach programs by themselves don't generally produce large reductions in violence.
“The experience historically and presently with that kind of gang intervention work is that it doesn't produce results on its own,” says Kennedy. “Everyone would like this to be different.”
But, Kennedy adds, “That's not to say it produces no results.”
Kennedy advocates for programs with “more robust architecture” than One Vision, which usually involve a partnership among law enforcement, social services, and community leaders. Street outreach is often a component; but these programs, which now operate under a national umbrella group, co-chaired by Kennedy, called the National Network for Safe Communities typically involve explicitly communicating to gang members a direct rejection of violent behavior. And this model has been shown to reduce violence and homicides by one-third to one-half in communities such as Los Angeles, Indianapolis and Boston.
These more elaborately organized projects involve strong participation by community leaders who make clear that “although they are not rejecting the gang members, they are rejecting (their) violent behavior,” says Kennedy.
Law enforcement should also provide a clear message that the entire gang, and not just individuals, will be held responsible for any potential violence, he adds. “It's a deliberate attempt to turn peer pressure against violence, to affect how gang members police themselves.”
Building on improvements suggested by RAND, One Vision has become a part of this more complex approach through the new Pittsburgh initiative, which is in fact affiliated with Kennedy's National Network.
“Long before the RAND report, there was an agreement that One Vision One Life workers would be a central partner in that larger strategy,” says Kennedy, who has been working with Pittsburgh officials since 2008.
One Vision will receive an additional $75,000 a year in city funding as part of the initiative, and Garland says he continues to be committed to his work over the long term. “It's a moral thing for me,” he says. “We had a role in what happens to this community, and we have a responsibility to build it back up.”
Evaluations like the RAND report, which provide sometimes uncomfortable results, are critical to criminal justice reform, according to Greg Berman, the director of the New York City-based Center for Court Innovation and the co-author of a newly published collection of case studies, Trial & Error in Criminal Justice Reform.
“There’s lots of reasons why replications of a successful model can go wrong,” Berman says. “It often comes down to context: every place is different, with its own politics and personalities. In general, I worry that there is an enormous pressure to make quick judgments about new programs without giving them a chance to learn from their mistakes and without a nuanced look at their results.”
Bernice Yeung is a San Francisco-based freelance journalist.