Not long ago, I facilitated a discussion among policymakers, criminal justice professionals, and representatives from community organizations about the large number of incarcerated individuals, the recidivism rate after release, and the effect of both on resources, especially local jails.
As we discussed what we know about effective rehabilitative programming, one attendee could not contain his ire. He strongly asserted that the individuals in his jails had been in program after program after program until they had been programmed nearly to death, and it had not made a whit of difference.
I believe he took offense when I asked him what evidence he had that they actually were good programs and that they worked. “We know they’re good programs—and they don’t work,” he responded.
His response is emblematic of the continuing nationwide debate on rehabilitation and correctional programs. The perceived failure of prison to deter criminal behavior—as evidenced by high recidivism rates and the substantial costs associated with an increasing number of ex-prisoners who unsuccessfully return to the community—has renewed interest in promising rehabilitative approaches. Nothing has fueled this renewed interest like the recent discussions on Project Greenlight.
Project Greenlight was a short-term, prison-based reentry demonstration program. It was jointly operated by the New York State Department of Correctional Services and the New York State Division of Parole and administered by program developers from the Vera Institute of Justice. Here, I offer a basic overview of the program and, most importantly, discuss the somewhat controversial findings from an evaluation sponsored by the National Institute of Justice.
What Did Project Greenlight Offer?
Offenders tend to leave prison much as they enter: lacking practical and interpersonal skills and possessing few economic and social resources. They tend to encounter significant barriers, both formal and informal, when they return to the community. In an effort to help offenders meet some of these challenges, Project Greenlight was designed as an intensive, prison-based reentry program to be delivered during the 8 weeks immediately preceding an inmate’s release from prison.
The developers of the Project Greenlight program drew extensively from the literature on correctional interventions and from anecdotal evidence about the services that offenders need to succeed when they return home. The key elements of Project Greenlight were:
- Cognitive-behavioral skills training. The foundation of the Project Greenlight program was cognitive-behavioral skills training because the research indicates that this type of program shows the most consistent results in reducing offender recidivism. Cognitive-behavioral programming is based on the theory that if offenders commit crime due to poor socialization, they can be resocialized toward more prosocial thinking and behavior.
- Employment. Project Greenlight employed a job counselor to work with program participants on how to write a résumé and improve their interview skills. If inmates were perceived to be job-ready, the counselor matched them with employment opportunities that might lead to stable work upon release.
- Housing. Because homeless shelters generally do not provide good living situations, the program worked with the New York City Department of Homeless Services to find short- and long-term housing for inmates who did not have a place to go upon release.
- Drug education and awareness. Participants were required to attend drug education or relapse prevention classes to help them deal with addictive behaviors.
- Family counseling. When a person returns home after a long absence, the adjustment can be difficult for the entire family. A counselor worked in the evenings with some Project Greenlight participants and their families to help them prepare for the inevitable strains that arise when an absent family member returns home.
- Practical skills training. Classes in practical skills offered guidance to Project Greenlight participants on a wide variety of tasks—some straightforward, such as how to use a subway card; some complex, such as how to open and manage a bank account, access emergency sources of food or cash, and regain voting rights. The program also helped participants obtain proper identification documents and Medicaid coverage before leaving prison.
- Community-based networks. Project Greenlight developed a network of community-based organizations to provide participants with social support after they were released.
- Familiarity with parole. Participants were introduced to parole officers and familiarized with the parole process to promote greater adherence to the conditions of parole.
- Individualized release plan. Project Greenlight staff worked one-on-one with participants to develop an individualized release plan. At its most basic level, this plan was akin to a “day planner,” reminding offenders what they planned to do upon release and when they would do it. The plan also attempted to provide a degree of structure to the participants’ postrelease activities, helping them add order to what was likely to be a very disorienting time. The release plan was given to the participants’ parole officers to make them aware of the goals and tasks established by parolees before their release.
The Greenlight Study
In the Project Greenlight Study, 735 inmates were divided into three groups and followed for at least 1 year (some for 2 years) after release. The intervention group of 334 inmates received the Project Greenlight programming. One comparison group (referred to as the UPS group) comprised 113 inmates who were released directly from prisons in upstate New York without any pre-release services. The second comparison group comprised 278 inmates who participated in the transitional services program (TSP) already in existence at the facility (in the same prison as the Greenlight participants).
Project Greenlight was designed to emphasize specific services that would improve certain interim quality-of-life outcomes and, as a result, would affect subsequent criminal behavior. The developers believed, for example, that helping parolees (who would otherwise end up in a homeless shelter) find stable housing would reduce criminal behavior. The program also had a job counselor to help participants develop their interview skills and connect with potential employers, with the goal of better employment, gained more quickly, for a longer duration.
Interim Quality-of-Life Outcomes
Data from evaluation surveys of participants and parole officers indicated:
- Employment, family relationships, and use of homeless shelter. There were no differences between the Project Greenlight group and the control groups.
- Parole knowledge and adherence. Although Project Greenlight participants demonstrated significantly more familiarity with parole conditions and were more positive about parole, there was no difference in adherence to parole conditions between the Project Greenlight group and the control groups.
- Service referrals and contacts. Project Greenlight participants received more service referrals and reported more contacts with community services after release.
Project Greenlight participants showed worse outcomes for every type of recidivism at 6 and 12 months after release. The chart “Percent of Participants Who Recidivated at 6 and 12 Months” shows the percentage of each group that experienced any kind of arrest (misdemeanor or felony), felony arrest only, and parole revocation. It is especially noteworthy—because it is statistically significant—that the overall arrest rate for the Project Greenlight group was 10 percent higher than that for the TSP group at 12 months postrelease (34 percent versus 24 percent). Also statistically significant is the 12 percent more parole revocations experienced by the Project Greenlight group than the UPS group at 12 months postrelease (25 percent versus 13 percent).
Several findings of the evaluation were at odds with program expectations. Most notably, Project Greenlight participants’ postrelease outcomes were significantly worse than those of the TSP and UPS groups. The evaluation found that the Project Greenlight program had no effect on the interim outcomes that it was designed to address—including housing, employment, and parole—and that Project Greenlight participants fared significantly worse than the two control groups in rearrest and parole revocation rates at the 1-year mark. In addition, although Project Greenlight participants displayed greater knowledge of parole conditions, showed more positive attitudes toward parole, received more service referrals, and reported greater contact with service providers after release, none of these translated into better outcomes.
Why Did Project Greenlight Participants Do Worse?
Project Greenlight had been viewed positively by many people: program developers and staff, participants, corrections officials, policymakers, and community advocates. Why, therefore, were the results so different from the perceptions? Why did the Project Greenlight intervention fail to reduce recidivism? Indeed, why did participants show substantially worse outcomes than both of the control groups?
Although selection bias is always a potential concern—did more crime-prone individuals end up in the Project Greenlight group than in the control groups?—the strength of the evaluation (both design and methodology) suggests that selection bias was not responsible for the negative outcomes. A more likely explanation is that something associated with the program or its implementation contributed to the negative findings. There are several potential explanations.
Obviously, Project Greenlight’s curricula had the potential to yield positive outcomes. It also had the potential to result in no difference among the three groups, but it is difficult to imagine that the program’s practical-skills or cognitive-behavioral training, for example, were somehow inherently criminogenic. The same curricula have been used extensively elsewhere, under a variety of conditions with a diversity of populations, with positive outcomes. It is therefore highly unlikely that the program’s content was responsible for the negative results.
It seems equally unlikely that referrals to community organizations, housing providers, and other community services would lead the Project Greenlight group to be rearrested at higher rates. In short, the program curricula seem relatively innocuous in their potential for creating negative outcomes.
There are reasons to suspect, however, that program implementation, including program design, might have resulted in the negative outcomes.
First, the standard cognitive-behavioral program that, in the past, has produced robust results in reducing offender recidivism was radically restructured in the Project Greenlight program. The recommended class size for cognitive-skills training is 10 to 13 participants; the Project Greenlight class size was 26. Given that many incarcerated people have limited interpersonal skills and education and are likely to be impulsive, a small class size is considered crucial in helping them maintain attention and helping instructors deliver material.
The cognitive-behavior model upon which Project Greenlight was based typically delivers services twice weekly for 4-6 months. The Project Greenlight program compressed the delivery of services, however, into daily classes for 8 weeks. These and other changes to the standard cognitive-behavior program model raise questions about how effective Project Greenlight could have been considering the deviations from what has long been considered the optimal program. In addition, participants in the Project Greenlight group were transferred from one prison to another—and were required to participate—suggesting the possibility that they could have been overwhelmed and perhaps even frustrated and angry about their participation.
The relatively short nature of the program might not have given participants enough time to get past any negative emotions and resistance generated by coerced participation.
Although the developers of Project Greenlight drew elements from the literature on correctional interventions, there were some key failures—most notably, ignoring the treatment principles that form the foundation of effective programming. There is general agreement that interventions should be directed toward high-risk participants and that assessing risk and needs should be a part of any intervention protocol. Project Greenlight staff found, however, that the assessment tool was too cumbersome and time-consuming to administer and therefore dropped it.
Another basic treatment principle is that interventions should target participants’ specific needs. Project Greenlight was a broad-based intervention in which everyone in the group was exposed to the same program elements. Postrelease interviews indicated that some participants felt significant frustration and anger about being forced to attend drug education sessions when they had no history of substance use. It should also be noted that an emerging body of evidence suggests that the delivery of intensive services to low-risk individuals may be counterproductive.
In addition to program design problems, Project Greenlight could have been poorly implemented. As a general proposition, implementation has clearly been identified as one of the most significant obstacles to an effective intervention. The evaluation found a correlation between Project Greenlight participants who worked with specific case managers and the program’s negative outcomes. Additionally, some participants in the Greenlight group were observed to be disengaged and appeared uninterested.
Project Greenlight attempted to create a comprehensive intervention by pulling together diverse program elements to address the multiple needs of participants. The program was clearly attractive to policymakers and corrections officials because of its short duration and the large number of individuals who could receive the programming. Based on the evaluation, however, one can seriously question whether Project Greenlight was a “hodgepodge of unproven and unstandardized clinical interventions” all lumped together. Although this may seem to be a harsh characterization, it might be an accurate portrayal of the program that was finally implemented.
What Have We Learned?
I considered beginning this article, as many discussions of corrections do, with the standard description of the U.S. social experiment in mass incarceration: the consequences to our society, communities, and families of having more than 2 million people incarcerated and nearly 700,000 admitted to and released from prison every year. I hope, however, that the experience I described in the opening of this article demonstrates the frustration of many criminal justice professionals. We do not really know about many of the programs currently being used, and some real lessons can be learned from the negative outcomes of a program like Project Greenlight.
First, whenever an intervention is contemplated and implemented, there is always an implicit assumption that “good” is going to come of it. Human behavior is complex, however, and we are still trying to understand it in a variety of ways, from the biological to the sociological to the philosophical. Perhaps we should also hold the assumption that an intervention program might do harm. Clearly, the implementation of every program should have precisely stated outcomes and a way to assess those outcomes on a regular basis.
Second, the “what works” literature on correctional interventions discusses programming that is known to work. Often, these discussions focus on the programs themselves without exploring why they work. The treatment principles that underlie effective programming were often ignored in Project Greenlight. This opened the program developers to the critique that they created a “kitchen sink” program —and one with negative outcomes at that.
Third, although Project Greenlight was labeled a reentry demonstration program, it had in fact no real reentry component. It was prison-based, with no structured followup in the community. Given what the reentry literature says about the need for postrelease services, it appears that an individualized release plan such as the one developed for Project Greenlight participants does not provide the necessary structured followup. Some States recognize the potential for structured postrelease assistance—for example, although still untested, Connecticut’s Building Bridges program allows parolees to work with a case manager for up to 1 year after release.
Finally, it is crucial to recognize that if Project Greenlight had not been evaluated, the program would be regarded as an unqualified success, based solely on the positive perceptions of those involved. Despite all the promise and positive perceptions, the program resulted in more harm than good. Could there be a clearer example of why program evaluations are needed?
I can understand the frustration expressed by the professional I mentioned in the opening of this article. We might continue to talk about the positives of rehabilitation, but when practitioners and the public see the constant churning of individuals through the criminal justice system, they see a failed system based on programs that do not work. If we continue to place offenders in programs that are positively perceived but that remain untested, we might continue to produce outcomes similar to Project Greenlight. Without effective evaluations of our programs, we run the risk of programming offenders nearly to death—and it still will not make one whit of difference.
Chart: Percent of Participants Who Recidivated at 6 and 12 Months
* Difference in the indicated pairs (by row) is statistically significant at p < .05.
For More Information
- Wilson, J.A., and R. Davis, “Hard Realities Meet Good Intentions: An Evaluation of the Project Greenlight Reentry Program,” Criminology and Public Policy 5 (2) (2006): 303-338.
- Brown, B., R. Campbell, J.A. Wilson, Y. Cheryachukin, R.C. Davis, J. Dauphinee, R. Hope, K. Gehi, Smoothing the Path From Prison to Home, final report submitted to the National Institute of Justice, Washington, DC: 2006 (NCJ 213714), available at www.ncjrs.gov/pdffiles1/nij/grants/213714.pdf.
About the Author
Dr. Wilson is an assistant professor of sociology at Fordham University in New York.
About This Article
This article appeared in NIJ Journal Issue 257, June 2007.
[note 1] The JEHT Foundation contributed the initial funding to begin the Project Greenlight evaluation.
[note 2] This literature has grown extensively. See Travis, J., A. Solomon, and M. Waul, From Prison to Home: The Dimensions and Consequences of Prisoner Reentry, Washington, DC: Urban Institute, 2001, available at www.urban.org/UploadedPDF/from_prison_to_home.pdf.
[note 3] More information on cognitive-behavior research is included in Andrews, D.A., I. Zinger, R.D. Hoge, J. Bonta, P. Gendreau, and F.T. Cullen, “Does Correctional Treatment Work? A Clinically Relevant and Psychologically Informed Meta-Analysis,” Criminology 28 (1990): 369-404.
[note 4] Frank Porporino, codeveloper of the Reasoning and Rehabilitation program—a multifaceted cognitive-behavior program used throughout Canada and the United States to teach juvenile and adult offenders cognitive skills and values—assisted me in clarifying some of my explanations of the negative findings.
[note 5] Lowenkamp, C.T., and E.J. Latessa, “Increasing the Effectiveness of Correctional Programming Through the Risk Principle: Identifying Offenders for Residential Placement,” Criminology and Public Policy 4 (2) (2005): 263-290.
[note 6] Rhine, E.E., T.L. Mawhorr, and E.C. Parks, “Implementation: The Bane of Effective Correctional Programs,” Criminology and Public Policy 5 (2) (2006): 347-358.
[note 7] Marlowe, D.B., “When ‘What Works’ Never Did: Dodging the ‘Scarlet M’ in Correctional Rehabilitation,” Criminology and Public Policy 5 (2) (2006): 342.