The author(s) shown below used Federal funds provided by the U.S.
Department of Justice and prepared the following final report:

Document Title: 	

SMOOTHING THE PATH FROM PRISON TO
HOME

Author(s): 	

Brenner Brown; Robin Campbell; James A.
Wilson; Yury Cheryachukin; Robert C. Davis;
Jean Dauphinee; Robert Hope; Kajal Gehi

Document No.: 	

213714

Date Received: 	

April 2006

Award Number: 	

2002-RT-BX-1001

This report has not been published by the U.S. Department of Justice.
To provide better customer service, NCJRS has made this Federallyfunded grant final report available electronically in addition to
traditional paper copies.

Opinions or points of view expressed are those
of the author(s) and do not necessarily reflect
the official position or policies of the U.S.
Department of Justice.

SMOOTHING THE PATH
FROM PRISON TO HOME
A Roundtable Discussion on the
Lessons of Project Greenlight

Edited by
Brenner Brown
Robin Campbell
Vera Institute of Justice
December 2005

© 2005 Vera Institute of Justice. All rights reserved.
Additional copies can be obtained from the communications department of the Vera
Institute of Justice, 233 Broadway, 12th Floor, New York, New York, 10279, (212) 3341300. An electronic version of this report is available for download on Vera’s web site,
www.vera.org.
Requests for additional information about this project should be directed to Robin
Campbell at the above address or to rcampbell@vera.org.

Prison reentry is an important and extraordinarily complex national issue. This
supplementary document provides context and analysis of research findings about
Project Greenlight, an ambitious prison-based reentry demonstration project that the
Vera Institute of Justice conducted at the Queensboro Correctional Facility in Queens,
New York, from February 2002 to February 2003.
Drawing upon research literature and demonstrated best practices, Greenlight sought to
reduce recidivism among soon-to-be-released men by working with corrections and
parole staff to address a spectrum of reentry issues during the last 60 days in prison. In
evaluating the pilot, however, Vera researchers found that arrest rates among
Greenlight’s 348 participants were significantly higher than those of two different
comparison groups. After an independent peer review conducted by the U.S. Justice
Department’s National Institute of Justice found nothing in the research design or
execution to account for this disappointing finding, Vera staff had to consider the
possibility that the program itself was responsible for the negative outcomes.
In order to better understand what happened and to preempt any premature conclusions
about the field of reentry
The invited participants were:
based upon this single
• Eddie Ellis, chair, NuLeadership Policy Group,
enterprise, in April 2005
Medgar Evers College, City University of New
Vera hosted a roundtable
York
discussion of prominent
• Martin Horn, commissioner, New York City
researchers, expert
Department of Correction
practitioners, and former
Greenlight and select Vera
• Doris MacKenzie, professor of criminology and
staff. [See box at right.]
criminal justice at the University of Maryland,
Their conversation about the
College Park
program and the evaluation
• Orlando Rodriguez, professor and chair of the
outcomes yielded three
Department of Sociology and Anthropology at
dominant concerns that could
Fordham University
have factored into the
• James Wilson, professor of sociology and
Greenlight participants’
anthropology, Fordham University, and lead
higher rate of re-arrest: the
investigator on the Greenlight evaluation.
program’s design and
implementation may have
In addition to Michael Jacobson, director of Vera, the
lacked sufficient cultural
following Vera staff participated:
sensitivity, particularly in
• Mike Bobbitt, former Greenlight family counselor
terms of race; Greenlight
• Megan Golden, director of planning, Vera
participants may have
• Marta Nelson, former director, Greenlight
perceived some injustice in
• Tim Ross, director of research, Vera
being assigned to the
• Dan Wilhelm, director, State Sentencing and
program, leading to low
Corrections program, Vera
levels of commitment; and
participants’ heightened
expectations may have been unmet upon release, leading to unanticipated negative
behaviors.

Building on these themes, the roundtable participants identified several lessons. These
include: best practices should be adapted to be culturally competent and race-specific;
pre-release programs require more than two months’ time; participants need continuous
support and some single institution must be accountable for the transition process;
community-based service providers must be reliably able to meet released individuals’
basic needs, including housing, employment, and substance abuse treatment; and there is
a continuing need for research to better understand what works.
The text that follows, edited for ease of reading, represents highlights of the roundtable
discussion. Michael Jacobson, Vera’s director, moderated.
Michael Jacobson: We wanted to convene this group to coincide with the release of the
Greenlight research report. Specifically, I wanted to do something to provide some
context. I was concerned that the report itself, released in a vacuum, would be too easy
for some people to very narrowly interpret the results. I at least wanted to release a
discussion of the results for either future research or programming around reentry.
Marta Nelson: We started planning Greenlight in 1999, when “reentry” used to mean
when a rocket came back to earth. We began by doing the study The First Month Out,
where we followed a group of men and women leaving prison and some New York City
jail folks in their first month after release to see their experiences.1 At the time,
preparation for release was extremely haphazard and there was a sense from the people
we talked to that they would have liked to have been better prepared. Also, in New York
City we have a relatively rich community of post-release services for people coming out,
so we didn’t see a need to create another program to do what they were already doing.
Parole got involved because Parole was looking for a new role for its institutional parole
officers. We really wanted to hit all the highlights of good pre-release programming and
create a real menu of services for the folks to go through. And we did that.
Jim Wilson: The basic research program covered those elements which, either
anecdotally or empirically, we have some basis for belief that they would improve
offender integration into the community. In terms of looking at recidivism, we were
looking at the four big measures: all arrests, serious felony arrest only, parole
revocations, and re-incarcerations. Ultimately, Greenlight participants performed worse
on all recidivism measures than both the TSP and the Upstate groups.2 The question is,
Why? We feel that our research design was fairly solid, so what we’re left with is the
potential for a negative impact of the intervention.
There are several issues about the intervention itself. The cognitive behavioral
skills training—the Reasoning and Rehabilitation Program—clearly has a lot of support
1

The First Month Out can be downloaded from Vera’s web site, http://www.vera.org/firstmonthout.
Greenlight participants’ outcomes were compared to those of two other groups. The Upstate comparison
group was composed of individuals who remained in prison facilities in upstate New York until their
release, presumably with no pre-release services. The second group, a control group, participated in the
Transitional Services Program (TSP), a limited pre-release program run by the New York State Department
of Corrections, also at the Queensboro Correctional Facility.
2

Vera Institute of Justice

2

in terms of the literature.3 But in this case, the program was restructured to a significant
degree. It’s typically offered over a four- to six-month period, but it was condensed so
that Greenlight participants were involved on a daily basis over an eight-week period, and
class sizes were increased significantly.
We’re also posed with the notion of program-offender mismatch. Porporino and
Fabiano indicate in the [Reasoning and Rehabilitation] program manual that during the
initial periods of delivery, especially with high-risk offenders, you can engender a lot of
resistance if the program is too short. You release individuals to the street who may be
angry or frustrated and never get to the point where you reach that therapeutic effect.
The last potentially relevant concern is that there may have been implementation
issues. For example, we used individual case managers in the program as controls, and
what we see is that they appear to be largely responsible for the negative program effect:
Those participants who were associated with particular case managers appear to have
much worse outcomes than all the study participants that we followed.
Martin Horn: You followed the people who came through each case manager?
Tim Ross: Yes. But the two that Marta ranked in the middle for quality of service
delivery, the outcomes of their case loads were similar to TSP’s. While the managers
Marta rated the best and the worst had by far the worst outcomes. We had hoped, from a
research standpoint, that the two Marta considered weaker would have the worst
outcomes. Then we could say, “When you have good case managers, they do better;
when you have bad case managers, they do worse.” Not the case.
Eddie Ellis: All the case managers were DOCS [New York State Department of
Corrections] personnel?
Marta: No, there were two DOCS and two Parole officers.
Eddie: Did they have any prior training?
Marta: Yes, we trained them all in the programming, including the Reasoning and
Rehabilitation Program. The people who created that program also changed it for us, by
the way. They wanted to test what the intervention would be like if you did it in a shorter
period of time with this group of people. So it wasn’t like we took it and said, “Let’s
change it.” We changed it with them.
Michael: The really interesting piece, at least for me—and a lot of lay people I would
think—is that the results are not only not as good as TSP, but they’re worse than no
intervention, the Upstate sample.

3

The Reasoning and Rehabilitation program is a multifaceted cognitive-behavior program designed to
teach juvenile and adult offenders cognitive skills and values. It was developed by Dr. Robert Ross of the
University of Toronto, and by Canadian criminal justice practitioners Elizabeth Fabiano and Frank
Porporino. It is widely used throughout the Canadian correctional system, as well as in a number of states
in the United States.

Vera Institute of Justice

3

Doris MacKenzie: Is there anything in the program that would cause participants to feel
like they were required to do more things they didn’t want to do, so they didn’t see it as
fair? Did they see the program as, “This is to my benefit,” or as, “This is a lot; you really
make me work”?
Jim: In the post-release interviews they expressed more positive perceptions about the
program. In general, what they really appeared to like was the employment part. They
expressed frustration with having to participate in the cognitive classes, and there was
some expression of frustration with having to go to drug treatment classes when they
didn’t have a history of drug use. Other than that, I don’t recall anything that really stood
out. But there was some frustration expressed, yes, because they really were not given
much of a choice about participating in the program.
Doris: Or enough choice.
Jim: If they refused participation in the program, they could presumably lose their good
time and be sent back upstate to finish their sentences.
Marta: One person did that. No one else really thought of that as a choice. At the
beginning of the program there was a huge amount of resistance because 50 people were
shipped altogether and plopped down in Greenlight. It was supposed to be a trickle in,
little by little. And so, you know, they were mad.
Marty: And that was part of the study sample?
Marta: Yes. The first two cohorts were extra experimental because they were just
dropped in and we were still figuring the program out. That had probably more of a
climate of injustice to it. But then it got to be business as usual.
Marty: I’m just reminded that in 1985 we tried to do a pre-release drug treatment
program and a three-part evaluation. DOCS has this enormous difficulty getting programs
like this up and running and moving inmates there in an orderly fashion that does not
engender antagonism on the part of the individuals who are selected, or “ripped out” of
wherever they are—Attica, Clinton, Great Meadow, wherever. They were cozy where
they were: they had their own cell, they were living in the dorm, they had their little gig
going—now they don’t. If you go back and you read the reports from that project, you’re
going to find very similar results.
Eddie: As I read the study, I noticed that there was only one mention of the question of
race—although 94 percent of the people involved were either Latino or AfricanAmerican. The other thing that struck me as interesting is that it seems to me that
programs based on best practices are very general, one-size-fits-all. I was wondering
whether or not in the many parts of this program race or cultural competence were
factored in?
Marta: In what way? In terms of talking about it, in terms of staff, or in terms of…?

Vera Institute of Justice

4

Eddie: All of the above. In terms of talking about it, in terms of staff, in terms of how it
was presented, in terms of the literature, in terms of the symbols, in terms of the pictures
if there were any pictures.
Marta: Well, in terms of staff, the case managers were one Caribbean-American, two
African Americans, and one Latino. For the Greenlight folks, the whites were in the
minority. Race did come up in terms of how people ran the classes—because you’re
talking real life, it didn’t get shied away from. But it wasn’t on the curriculum as in,
“We’re going to talk about race today.”
Mike Bobbitt: There was a tension in the earlier cycles about the recognition that
although the staff overall was kind of diverse, it was white people who were really
running this program. Those were the occasions to talk about using the methodology or
the cognitive skills to address issues of race. But there was nothing woven into the
curriculum per se to create those opportunities or to unpack, “What does that mean?”
Marta: And you think that changed?
Mike: I can’t recall hearing it as much in the later cycles, but I do remember hearing it,
and I remember it from the early going.
Orlando Rodriguez: The impression that I get reading the program evaluation is that
this is a de facto evaluation of a pilot program because you had this early termination of
Greenlight after only one year of operation.4 The next thing you expect is that the
researchers would get together with the planners and say, “This is what we find. Is this
real or are there some program flaws?” You know? But it didn’t happen.
Jim: That’s absolutely correct. We had hoped to give the program approximately six
months or so to settle down and get into a rhythm before we really started looking at
individuals. We started tracking people early in hopes that we might be able to give some
feedback to the program planners, but ultimately they became part of our sample because
of the early termination of the program.
Orlando: And you don’t have that much process data that could help you.
Jim: No, clearly we would like to have videotaped group sessions, we would have liked
to have talked to the staff at the program on a regular basis, but we just did not have the
resources. We did some limited process work in that we had individuals attend class
sessions and talk to the staff to some extent, but there’s just not a lot of information there
that we can use to make any conclusions in terms of the program.
Marty: Can we go back to this whole issue of expectations? So you did a lot of referrals
to community-based services. But we don’t know what the Greenlight participants’
4

Project Greenlight was designed as a three-year demonstration program, but budget constraints led to the
program being funded in its original form for only one year.

Vera Institute of Justice

5

experiences were like when they connected with them after release. If you said, “We’re
going to send you to Project ABC and they’re going to help you find a place to live,” and
then when they get there someone from Project ABC says, “Look, we haven’t got any
places for you. Come back tomorrow”—we don’t really know what the impact of that
was. I personally think it’s a critical variable.
Marta: Yes, that’s quite true.
Jim: We did know from some of our interviews that their contacts with housing services
were not positive.
Marty: I think that’s pretty important. If I come out and I’m all brushed up and dusted
off, and I get a “Oh yeah? Come back tomorrow and we’ll see what we can do,” Boy,
that’s not what I expected.
Doris: Particularly when you’ve been through the more difficult program.
Eddie: How competent were the community-based groups?
Marta: It varied. You had the groups that come and do this all the time—you had the
Fortune Society and the Osborne Association—and then you had some more specialized
groups. Some of the presenters were excellent, some were not.
Marty: But the answer to Eddie’s question has to be that we really don’t know—there’s
no way of measuring.
Marta: Right.
Eddie: And in a lot of cases the empirical evidence is that they’re not doing such a great
job, right? The reports they turn in monthly look good. But if you talk to the people on
the ground, they say, “I didn’t get any help.”
Marty: Or, “They referred me to somebody else.”
Marta/Eddie: Right.
Michael: Can I just follow up on two quick points? First, the basis for so many of these
cognitive therapy programs comes from Canadian researchers—I’m sure there must be
someone in this country who also does this programming, I just don’t know who they are.
Does that programming get translated from the Canadian to the specifically American
urban contexts?
Doris: That was my question too. Because this curriculum has been used in many other
countries, but there’s probably a similarity between Canada and many European countries
that aren’t like our urban inner-city.

Vera Institute of Justice

6

Michael: That stuff is used all over the U.S., isn’t it?
Marty: But no one’s evaluated it.
Jim: There have been recommendations that, as you implement the program, you have
to assess how well it fits with your population. But I have no idea how extensively that’s
been done.
Michael: My second question is, Let’s assume people were mad—whatever that
means—either from being dumped from wherever they were into this program or the
program itself, or it raised their expectations and then post-release lowered them. Does it
make sense that that anger and resentment shows itself in the form of increased felony
arrests?
Eddie: I think it goes beyond that. The question of cultural competence is more than a
question of sensitivity. Cultural competence comes when you have people who are
culturally competent to deliver the material in a way that the student will receive it
properly and act on it accordingly. It doesn’t really matter if the instructors are African
American or not—you can have African-American instructors who culturally are not
competent. From reading the material, what I got was that we tried to do this on the
cheap. We didn’t really have a good initial training on that for people who were going to
be involved in the program.
Michael: And can you see that if all that stuff is true, it could lead to not just more
revocations but more felony arrests?
Eddie: If your expectation is that as a result of being involved in this project you’re
going to get some tools that are going to help you once you get out, and you get what you
assume to be those tools and you go out and the tools don’t work, your only recourse is to
do what it is that you do best, which is what got you in in the first place. And that seems
to be what happened.
Marty: Years ago, I read a study that concluded that parole violators were basically
organizational rejects: Whether or not your parole officer cited you for a violation could
be a function of whether or not you pissed him off. I just wonder whether teaching people
to be assertive, thoughtful, persistent—things that we normally think are good traits to
have in negotiating bureaucracies—is valuable if parole officers aren’t ready to receive it.
Eddie: I know people who have had negative interactions with parole officers to the
degree that they felt if they make one more report, they wouldn’t walk out of the parole
office. And those people decided that they would never report again: “They just catch me
when they catch me.” But that was a direct response to the interaction between them and
the parole officer.
Marty: I think that the same behavior—especially on a misdemeanor level—that causes
you to piss off your P.O. might also cause you to piss off a cop. Because now you’ve

Vera Institute of Justice

7

decided, I’m gonna live in this world, I have rights, and I have learned new skills and
how to deal with people in authority who I think are being unreasonable.
Marta: I think that’s an interesting theory in terms of the skills that you’re teaching
clashing with the culture of parole and police officers. But there was another element to
the Greenlight guys in that they had this release plan. There was a person at each parole
office to whom we faxed the release plan. But you then have a parole officer who has this
release plan that he didn’t help create but that their bosses told them they’re supposed to
follow. Even if the parolee was incredibly well behaved, that parole officer is still kind of
annoyed that he has to do this. And you might have had a case where the parolee wanted
to talk about it—because some of them really got into the release plans.
Marty: I’m still not clear how that necessarily affects the felony or misdemeanor arrests,
but it would certainly affect the revocation outcomes, and it would certainly affect the
violations of parole initiated by the parole officer absent an arrest. On the other hand, if
I’ve got my release plan and I’m planning to do good, and I get to a P.O. who basically
brushes me off, as Eddie said, I might choose to abscond. And having absconded, I may
just blow off the rest, and now I’m living on the street, and now I’m getting high. And
once I get high, it’s only a matter of time until I get arrested, right?
Doris: The research that I’ve seen that really supports that scenario is research about
procedural justice. You’re being required to do all this stuff and—whether it’s because of
a cultural thing or the demands placed on you that other people didn’t have, or because
the probation or parole agent wasn’t fair to you—you come out saying, “The procedures
aren’t fair. I wasn’t treated fairly.” There is some research—limited—showing that this
really does have an impact on behavior.
Marty: I think the interaction between parolee and parole officer after release is an
important variable that’s going to affect outcomes.
Dan Wilhelm: Marty, you’ve talked a couple of times now about the bureaucratic, or
cultural, response that the parolee encounters. As I’ve been listening to the conversation,
I’ve been trying to think, What’s the value of this evaluation for people who are trying to
either affect policy or influence specific programs?
Michael: Yes. One of the things we wanted to get here is some lessons learned.
Marty: I don’t think Greenlight was a failure in any way. I think it really begs a whole
host of new questions. You started this in what year, 1999?
Marta: Started planning it then. The project officially began in 2002.
Marty: We’ve learned a lot about reentry since then, right? Helping these individuals
change the way in which they lead their lives is very hard. And to think you can swoop in
and, in eight weeks with some sort of cognitive behavioral approach, change that is naïve.

Vera Institute of Justice

8

That’s the first thing. Secondly, I believe that the real work has to be done after release.
And someone has to own the responsibility for helping Person A affect their transition.
One of the things that I would like to know more about is the interrelationship
between sobriety and failure. I believe that all this stuff about criminogenic needs is
interesting, but why should we expect that ex-inmates are going to deal with their
criminogenic needs if they don’t have a place to live, if they’re dying to get high, and
they can’t afford to live because they don’t have a job? So we have to attend to those
basic human needs.
I think that another important lesson is that in-prison programs, without some
very, very strong follow-up, are bound to fail. Making referrals is not enough. The
program that they go to has to be expecting them— “I’m waiting for ‘Joe’ to show up”—
and has to know what programming ‘Joe’ had beforehand, right? And it has to run with
it. I think there were too many hands in the pot in Greenlight. There have to be a limited
number of hands and the people all have to be singing from the same hymnal. And the
other thing I would say is, All of the effective programming literature talks about the
need to appropriately match needs with programming and risk. Therefore, low-risk
people, we ought to just kiss ’em goodbye and give ’em a voucher for services after their
release.
Eddie: I’ve got a question. I haven’t read most of the literature, but I get the sense that
this project was set up more as a study than as a service.
Marta: No, absolutely not. I mean, it may have been naïve, but we really did think that
if you do more intensive preparation for release and you bring in this wealth of
community groups that are out there, what’s missing is they didn’t know about them.
Because that was something we would often see when we talked to people.
Orlando: The nagging thing for me is that Upstate group that did so well. Either there’s
something wonderful going on in the upstate correctional facilities or else they’re just
doing the normal thing. But there’s something about what happened in Queensboro with
these two programs, Greenlight and TSP, that created some worse effect.
Eddie: One of the things that I’ve heard about Queensboro is that it’s not a place that
people want to go to. I don’t remember all of the reasons. But not withstanding the fact
that it’s closer to the city, if you’ve got six months or more to do on your sentence, you
don’t want to go to Queensboro.
Marty: It’s oddly placed.
Eddie: And I understand the recreation facilities are not very good—they don’t have a
yard, for example. And all that kind of stuff is important. But I want to pick up on the
point about the constraints of the project, particularly the financial constraints. One of the
things I got from this evaluation is that if we’re going to have effective post-release
intervention it may require a greater level of commitment than what was spent in this
particular instance. The other thing is that in addition to post-release continuity we need
to have community-based operations that are viable and that can deliver on the

Vera Institute of Justice

9

expectations that people have coming out of prison. I think also that pre-release programs
need to have greater longevity. I think six months would be a minimum period of time,
and optimally you would want to have at least a year working with someone before they
got ready to go. Because by the time a person gets two months to the street, they’ve
already decided in their mind what they’re going to do. And Marty’s point about
providing the very bare minimum—the essential prerequisites for success—if we can’t
provide that in the initial stages then we’re looking at some lost causes.
Doris: This is kind of a different topic, but is it maybe that the targets were wrong? A
lot of our research points to the role of work and family bonds in reducing recidivism.
But when I’ve evaluated correctional programs, what has ended up being effective are the
individual change programs: education, vocational education, cognitive skills, drug
treatment. So one of the problems could be that Reasoning and Rehabilitation didn’t work
with this group, so we really didn’t see a change at the individual levels. Prior to getting a
job you really have to get an education, or change your attitudes towards work. So the
focus may have been on building those bonds and ties but the focus needs to be maybe on
individual-level change.
Eddie: Yeah, yeah.
Doris: That, of course, isn’t an eight-week program.
Eddie: Picking up on that, most of the programs that I see have been what we call
“deficit model” programs. That is to say, they start with the deficits and ask, How do we
tinker with those to modify them or whatever? But it seems to me that there is another
approach that could really focus on building on the strengths that people have. In terms of
program design, that should be uppermost in our minds. How do we capitalize on the
individuals’ strengths and begin to build on those as a kind of a segue to everything else?
Marta: I think we did some of that. Certainly the family program was meant to celebrate
the fact that there were family members here, and have the families talk together about
how they want to handle this transition. But we probably could have done a lot more.
Michael: One of the reasons we’re having this discussion is because we did a fairly
deep—and at the end of the day, probably pretty expensive—piece of research. We
happen to know an unbelievable amount about what happened: the outcomes, the intrameasures, the design of the service delivery itself in terms of the randomness of who was
assigned to the program. So what’s at issue is not just good programming, but the
research that you need to find out whether you’re doing good programming. It’s one
thing to ask legislators or governors for a few million dollars to do an intelligent reentry
program. But on top of that you really need to come up with a bunch of money to do this
research.
Jim: Evaluations of this type of programming have traditionally been “black box”
evaluations. You run a program and you look at the end to see whether or not people do
better. Then you say, “Great, the program worked or it didn’t work.” But you have no

Vera Institute of Justice 10

idea why. Anytime you’re going to propose a new research project you really want to be
able to tell people why a program worked. What we said at the very beginning of
Greenlight was, “Okay, Marta and her staff are doing all these things. At the end, will we
be able to say something like, ‘You can dump this housing component because it doesn’t
matter. You can dump this cognitive behavioral stuff because it doesn’t matter. It’s all in
the employment. Find everyone jobs and they’re going to do fine.’” What’s important
when you’re thinking about evaluations now is, What can you tell people about why a
particular program works the way it does?
Michael: This has been a really interesting discussion. We wanted to accompany this
report with some sort of illuminating and critical discussion to wrap this project and its
results in context. I think we’ve done that. So thanks a lot, everyone, for coming.
For more information about Project Greenlight, visit www.vera.org/greenlight, or email
tross@vera.org.

Vera Institute of Justice 11

SMOOTHING THE PATH
FROM PRISON TO HOME
An Evaluation of the Project Greenlight
Transitional Services Demonstration
Program
NIJ Grant #2002-RT-BX-1001

James A. Wilson
Yury Cheryachukin
Robert C. Davis
Jean Dauphinee
Robert Hope
Kajal Gehi
Vera Institute of Justice
December 2005

This project was prepared by the Vera Institute of Justice and supported by Grant No. 2002RT-BX-1001 awarded by the National Institute of Justice, Office of Justice Programs, U.S.
Department of Justice. Points of view in this document are those of the authors and do not
necessarily represent the official position of the U.S. Department of Justice.
Additional copies can be obtained from the communications department of the Vera Institute of
Justice, 233 Broadway, 12th Floor, New York, New York, 10279, (212) 334-1300. An electronic
version of this report is available for download on Vera’s web site, www.vera.org.
Requests for additional information about the research described in this report should be directed
to Robin Campbell at the above address or to rcampbell@vera.org.

Executive Summary
In partnership with the New York State Department of Correctional Services (DOCS) and the
New York State Division of Parole, the Vera Institute of Justice developed a 60-day, in-prison,
reentry program for men called Project Greenlight (Greenlight), which was delivered at the
Queensboro Correctional Facility in Queens, New York. Greenlight was designed to be as
affordable and as comprehensive as possible. Planners created the program design based on the
literature on “what works” and common sense about what substantive areas need to be covered
in a pre-release program. The program included job preparation training and counseling,
substance abuse relapse prevention, practical skills, cognitive skills, family reunification, and
homelessness prevention. It also produced a release plan for each participant that was jointly
designed by inmate and program staff.
To evaluate Greenlight, Vera researchers compared three groups of inmates totaling 739
study participants. Two of the groups included people randomly assigned to either Greenlight
(348 participants) or to a less ambitious program operated in the same facility by DOCS (278
participants). The third group (“Upstate”) consisted of 113 people released directly from upstate
prisons to New York City without specialized reentry programming. The three groups did not
have any statistically significant differences in criminal history, education, or demographics.
Contrary to the expectations of program planners and research staff alike, Greenlight
participants recidivated at higher rates than either of the comparison groups after one year postrelease. Most of these differences were statistically significant and held across multiple measures
of recidivism, including new arrests, new felony arrests, and revocation. Assignment to the
Greenlight group remained significant in multivariate models. The Upstate group, which
received no pre-release reentry programming, recidivated at the lowest rate.
These research results are as puzzling as they are unexpected. None of the non-programmatic
reasons Vera researchers examined, such as variation in parole supervision and borough of
residence, explain the differences in recidivism. Vera researchers identified several possible
programmatic explanations, such as program design and implementation and the impact of the
program on parolee expectations, but did not have enough information to explain the results
definitively. It is always possible that factors that we did not measure, such as motivation to
avoid criminal activity, might explain the results, but these factors are often correlated with the
variables we did use.
This study raises strong cautions about the desirability of intensive short-term reentry
programming. Other studies of longer and more expensive reentry programs with tighter links to
community-based services show that reducing recidivism among people leaving prison is
possible. The current findings suggest, however, that trying to save money by compacting
reentry programming into a shorter time frame may be counterproductive. Like mentoring
programs, a small dose of reentry programming may be worse than no programming at all.
To learn more about how the program produced these unanticipated results, the Vera Institute
convened a roundtable of experts from corrections, academia, and the community. Highlights
from their discussion are appended to our summary and technical reports.

Acknowledgments
We received assistance on this project from many quarters, beginning with the JEHT
Foundation, whose initial investment helped us develop the detailed proposal that was accepted
by the National Institute of Justice (NIJ). We also want to acknowledge the support of Thom
Feucht, assistant director for research and evaluation at NIJ, and our NIJ program officer,
Andrew Goldberg.
Katie Lapp and the New York State Division of Criminal Justice Services, as well as the
Department of Correctional Services and the Division of Parole, also provided crucial assistance
and support. Chauncey Parker, Roger Jeffries, Bruce Frederick, Glenn Goord, Paul Korotkin,
Martin Ciricione, Teri Salo, Michael Buckman, and Mary Ellen Flynn and their staff all gave
generously of their time and resources.
Our work benefited substantially, too, from several researchers who reviewed our work,
especially Caroline Friendship and the anonymous peer reviewers arranged through NIJ. We are
also grateful for the insight and support of Eddie Ellis, Medgar Evers College, City University of
New York; Martin Horn, commissioner, New York City Department of Correction; Doris
MacKenzie, University of Maryland, College Park; and Orlando Rodriguez, Fordham University.
To the Greenlight program staff who assisted us in understanding how the program worked,
reviewed our report, and showed grace and resilience in the face of findings that were
disappointing, we would like to extend special thanks. In particular, we would like to recognize
Marta Nelson, Michael Bobbitt, Ali Knight, and Nino Rodriguez. Numerous other people at Vera
also worked on this project, commented on drafts, edited our writing, and supported our work.
They include Chris Stone, Molly Armstrong, Robin Campbell, and Janet Mandelstam of the
Communications Department.
Finally, we want to thank the Greenlight Oversight Board members: Robert Lee Guy, Harold
Clarke, Carl Reynolds, Dora Schriro, Devon Brown, Bill Bales, Jeremy Travis, and Julio
Medina.
We are grateful for all this help and advice. However, any errors or omissions are the
responsibility of the authors.

Table of Contents
Introduction……………………………………………………………………………... 1
Reentry Initiatives Nationwide…………………………………………..………. 3
Empirical Evidence Supporting Correctional Interventions…………………….. 5
Project Greenlight Program Overview………………………..……………........ 8
Key Hypotheses…………………………………………………………………..10
Data and Methods…………………………………………………………………….. 12
Overview of the Study Design………………………………………………….. 12
Data Sources……………………………………………………………………..17
Findings………………………………………………………………………………… 19
Data Analyses…………………………………………………………………… 19
Description of the Study Participants………………………………………….. 20
Interviews with Greenlight and TSP Parolees………………………………….. 22
Interviews with Parole Officers…………………………………………………. 25
Analysis of Recidivism Outcomes………………………………………………. 27
Multivariate Analysis: Cox Proportional Hazards………………………………. 34

Discussion and Conclusion……………………………………………………………. 43
Appendices ..................................................................................................... 48 

Appendix A: Project Greenlight Program Description………………………….. 48
Appendix B: Methodology and Data…………………………………………….. 70
Appendix C: Supplementary Tables…………………………………………….. 90
Appendix D: Instruments and Forms………………………………………….. 116

Introduction
This study reports findings of an evaluation of a prisoner reentry program operated by the Vera
Institute of Justice. Planning for reentry has become a national issue: the number of people
incarcerated in the United States has increased sixfold in the last three decades.1 By the end of
2002, more than two million people were incarcerated. Based on present rates of incarceration,
about one in eleven U.S. men will go to prison in their lifetimes, and that figure is far higher for
men of color. At present rates, one in every three African American males will spend time in
prison.2 Virtually all people who enter prison eventually leave.3
In 2001, more than 630,000 people exited state or federal prison—more than the population
of Boston, Denver, or New Orleans.4 And millions more are affected—when people are released
from prison, their neighborhoods, families and communities must adjust to their return. Some
neighborhoods, especially in disadvantaged, minority communities in the inner cities, have a
disproportionate number of people entering and returning from prison.5 Families in these
neighborhoods often suffer from high unemployment rates, a lack of appropriate housing, limited
access to quality health care and other problems that make re-integrating people returning from
prison that much harder.
According to recent statistics, five states (California, Florida, Illinois, New York, and Texas)
accounted for almost half of all releases nationwide in 2001; 16 states accounted for 75 percent
of those released from state prisons.6 Most data indicate that the majority of offenders released
nationwide return to “core counties” (counties that include the central city of a metropolitan
area), and that the releases appear to be concentrated within certain neighborhoods in those
counties.7 Sixty-six percent of all people leaving prison return to inner-city metropolitan areas;
and a disproportionate number return to a small number of inner city neighborhoods.8
1

“Facts about prisons and prisoners.” The Sentencing Project, November 2004. Data aggregated from the Bureau of
Justice Statistics. See http://www.sentencingproject.org/pdfs/5022.pdf.
2
If recent incarceration rates remain unchanged, an estimated 9 percent of men (or one in eleven) will serve in
prison in their lifetime. Bureau of Justice Statistics, www.ojp.usdoj.gov/bjs/crimoff.htm, accessed 12/4/2002.
3
Estimates indicate that somewhere between 90 and 95 percent of those who are incarcerated will eventually return
to the community. Travis, Jeremy, Amy Solomon, and Michelle Waul, From Prison to Home: The Dimensions and
Consequences of Prisoner Reentry (Washington, DC: Urban Institute, 2001); Petersilia, Joan, When Prisoners Come
Home: Parole and Prisoner Reentry (New York: Oxford University Press, 2003).
4
Harrison, Paige M., and Jennifer C. Karberg, Prison and Jail Inmates at Midyear 2002, BJS Bulletin, Washington,
DC: USDOJ, NIJ, April 2003, NCJ 198877. Harrison, Paige M., and Allen J. Beck. Prisoners in 2002, BJS Bulletin,
Washington, DC: USDOJ, NIJ, July 2003, NCJ 2002248.
5
See, for example, Lynch, James, and William Sabol, Prisoner Reentry in Perspective, Crime Policy Report, vol. 3.
Washington, DC: Urban Institute, 2001. Petersilia, Joan, and Jeremy Travis (eds.), From Prison to Society:
Managing the Challenges of Prisoner Reentry, Crime and Delinquency, Special Issue 47 (3) (July 2001); Travis,
Jeremy, Amy Solomon and Michelle Waul, From Prison to Home: The Dimensions and Consequences of Prisoner
Reentry, Washington, DC: Urban Institute, June 2001
6
Harrison and Karberg, Prison and Jail Inmates, 2003:6. Also see, Bureau of Justice Statistics, Correctional
Populations in the United States, 1998, BJS Internet Report, NCJ 192929. Retrieved as *.pdf file on March 3, 2003
from http://www.ojp.usdoj.gov/bjs/pubalp2.htm#cpus; Lynch and Sabol, Prisoner Reentry in Perspective, 15.
7
Lynch and Sabol, Prisoner Reentry. 2001:15. Also see, Gottfredson, Stephen D., and Ralph B. Taylor.
“Community Contexts and Criminal Offenders.” In Communities and Crime Reduction, T. Hope and M. Shaw

Vera Institute of Justice 1

The number of ex-offenders who are re-arrested and re-incarcerated shows the need for more
effective reentry practices. The largest national study to date found that 67.5 percent of offenders
released in 1994 are arrested for a new felony or serious misdemeanor within three years; a slight
increase from the 62.5 percent found by a similar study of prisoners released in 1983.9 Half of all
released felons returned to prison within three years—a quarter due to new convictions, and
another quarter due to parole violations. Nationally, the proportion of people who return to
prison for parole violations has been rising consistently since 1980 and parole violators are an
increasing proportion of all prison admissions; parole violators accounted for more than onethird of all new admissions to state prison in 1999.10
Many people exiting prison come from marginalized segments of the population; they often
lack resources such as stable family members and law-abiding friends who can support them
socially or financially as they reenter society. Finding a safe and affordable place to live is often
a major hurdle.11 Further, ex-offenders seldom have the educational skills or employment
background that would help to secure a job and rebuild their lives. They are often stigmatized by
family members, friends, prospective employers, and others because they are ex-offenders. In
addition, many people have issues with substance abuse that contribute to their struggles after
release. Because roughly 80 percent of people leaving prisons are under parole supervision
(either discretionary or mandatory), with its curfews, reporting requirements, travel limitations,
and mandated programs, even one violation of any of these conditions can land a person back in
prison.12 On top of the numerous personal obstacles, there are also clearly defined institutional
barriers encountered when they reenter society.13 Successfully reintegrating ex-offenders into
communities is a substantive task, and there are significant and far-reaching social, economic,
and policy implications associated with the large-scale reentry of offenders that are not yet well
understood.14 Given the array of barriers to successful reintegration, and the fact that government
(eds.). London, England: Her Majesty’s Stationary Office, 1985:62-82; Rose, Dina R., Todd R. Clear, and Judith
Ryder, Drugs, Incarceration and Neighborhood Life: The Impact of Reintegrating Offenders into the Community.
Prepared under NIJ Grant #1999-CE-VX-008;Travis, Solomon and Waul 2001.
8
Lynch and Sabol, Prisoner Reentry in Perspective, 2001:15; Rose, Dina R., Todd R. Clear, and Judith A. Ryder,
Drugs, Incarceration and Neighborhood Life: The Impact of Reintegrating Offenders into the Community,
Washington, DC: USDOJ, 1999.
9
Langan, P. A., and D. J. Levin. Recidivism of Prisoners Released in 1994. BJS Special Report, NCJ 193427.
Washington, D.C.: USDOJ, 2002. Beck, A. J., and B. E. Shipley. Recidivism of Prisoners Released in 1983. BJS
Special Report, NCJ 116261. Washington, D.C: USDOJ, 1989.
10
Between 1990 and 1998 the proportion of parole violators among new prison inmates increased 54 percent. It
should be noted that in all assessments of trends in state parole, California skews the national data substantially. For
more detail, see Hughes, T. A., D. J. Wilson, and A. J. Beck. Trends in State Parole, 1990-2000. Bureau of Justice
Statistics Special Report, NCJ 184735. Washington, DC: USDOJ, 2001.
11
See Nino Rodriguez and Brenner Brown, Preventing Homelessness Among People Leaving Prison, New York:
Vera Institute Justice, 2003.
12
Hughes, et al., Trends in State Parole, 2001:1.
13
For a discussion of institutional barriers, see Hirsch, Amy E., Sharon Dietrich, Rue Landau, Peter D. Schneider,
Irv Ackelsberg, Judith Bernstein-Baker, Joseph Hohenstein, Every Door Closed: Barriers Facing Parents With
Criminal Records. Retrieved March 3, 2003 http://www.clasp.org.
14
See, for example, Hagan, John, and Ronit Dinovitzer, Collateral Consequences of Imprisonment for Children,
Communities, and Prisoners, Crime and Justice: A Review of Research, vol. 26, Chicago: University of Chicago

2 Vera Institute of Justice

and society pay an enormous cost when reintegration fails and people return to prison,
policymakers, criminal justice agencies, and community-based organizations are focused today
on the need to prepare prisoners for release and to help them successfully return to their
communities.
The time immediately prior to and following an offender’s release from incarceration is
critical to successful reintegration, and focused transitional services programs may provide an
opportunity to smooth the transition and significantly improve post-release outcomes.15 A recent
survey of state correctional systems indicates that 39 of 41 states have some sort of pre-release,
reintegration, or transitional services programming.16 These programs vary with respect to their
duration, class hours and content, type of staffing, and degree of community participation. It is
likely that many programs were developed without a research-based understanding of the
transitional issues offenders face. Further, since two-thirds of the programs are voluntary and
many have caps on enrollment. No one knows how many offenders actually receive these
services, but the available evidence suggests that for most prison programs participation is
relatively low.17 Finally, there is little rigorous empirically-based research validating the efficacy
of transitional services or other pre-release programs.18
In sum, the large numbers of people leaving prison are seriously challenged in their efforts to
become part of a local community again, and their return entails serious consequences for both
their families and neighborhoods.
Reentry Initiatives Nationwide

Reentry initiatives are not new; it is more that the changing policy implications associated with
what have become historically high levels of incarceration have forced officials to focus on this
issue. A handful of state and local jurisdictions around the country offer programs to ease
transitions to the community from prison or jail. These programs take many forms, but only a
Press, 1999: 121-162; Petersilia, Joan, and Jeremy Travis (eds.), From Prison to Society: Managing the Challenges
of Prisoner Reentry, Crime and Delinquency, Special Issue 47 (3) (July 2001); Travis, Jeremy, Amy Solomon and
Michelle Waul, From Prison to Home: The Dimensions and Consequences of Prisoner Reentry, Washington, DC:
Urban Institute, June 2001.
15
See Nelson, Marta, Perry Deess and Charlotte Allen, First Month Out: Post-Incarceration Experiences in New
York City, New York: Vera Institute, 1999; Petersilia and Travis, From Prison to Society; Travis, et al., From Prison
to Home, esp. pp. 18-19.
16
American Correctional Association, “Pre-Release/Reintegration: A Survey Summary,” Corrections Compendium,
August (2000): 7-15. Non-responding states were Alabama, Georgia, Hawaii, Iowa, Kentucky, Michigan,
Oklahoma, Utah, and West Virginia. In addition, no response was received from the District of Columbia or the
Federal Bureau of Prisons. California and Nebraska indicate no programming, although California does have a reentry program that is not run by the Department of Corrections.
17
Austin, James, “Prisoner Reentry: Current Trends, Practices and Issues,” Crime and Delinquency 47 (3) (July
2001): 314-34; Camp, Camille, and George M. Camp (eds.), The Corrections Yearbook, 1998: Adult Corrections,
Middletown, CT: Criminal Justice Institute, 1999.
18
This is true of correctional interventions in general. See Sherman, Lawrence W., Denise Gottfredson, Doris
MacKenzie, John Eck, Peter Reuter, and Shawn Bushaway, “Preventing Crime: What Works, What Doesn’t, What’s
Promising?”, Final Report for NIJ, grant number 96-MU-MU-0019, Washington, DC: USDOJ, 1997, NCJ 165366.;
Gaes, Gerald, Timothy J. Flanagan, Laurence L. Motiuk, and Lynn Stewart, “Adult Correctional Treatment,” Crime
and Justice: A Review of Research, vol. 26, Chicago: University of Chicago Press, 1999: 361-426.

Vera Institute of Justice 3

few have a long track record. Some are prison-based, some community-based, others make use
of cooperative arrangements between departments of correction, parole, and probation, and
private community-based services and initiatives to provide a continuum of services between
prison and the community. In addition, the populations they are designed to serve are often quite
diverse in terms of their background characteristics, offense history, and overall level of risk.
Some programs are designed to address a single issue—focusing, for example, on substance
abuse, vocational issues, educational deficits, or family issues, while others offer a combination
of services that are often more comprehensive in nature. Single-focus programs can make an
important difference, but assessments of these types of programs suggests that efforts to address
multiple offender issues simultaneously are likely to have a greater impact.19
In Texas, for example, the departments of labor and correction jointly started a vocational
development program in 1985 that is open to state inmates two years from release. Project RIO
(Reintegration of Offenders) starts serving offenders while they are still in prison and offers job
training and work experience. A program evaluation found that RIO’s participants had higher
rates of employment after release and lower rates of returning to prison when compared to
prisoners who did not participate.20 Chicago’s Safer Foundation, a community-based provider of
employment services for ex-offenders established in 1972, offers a multitude of programs such
as basic education, life skills, and job placement.21 Although emphasizing employment services,
Safer works with “juvenile and adult probationers and parolees, community corrections residents
and persons in the county jail” with the philosophy of providing “avenues for them to let go of
the criminal life and buy into the mainstream; getting and keeping a job is a means to that end.”22
Some jurisdictions take a longer range approach to transitional services by planning for
reentry as soon as a person enters prison. Missouri takes this approach—the institutional regimen
is designed to mirror the outside world as much as possible.23 Prison work, recreation, and
treatment schedules mimic a workday. Inmates interview for jobs in prison and, if they do well,
advance to better paying jobs and other privileges.
New York State’s Department of Correctional Services operates several programs that
transfer inmates from across the state closer to their home community before release. Several
work release facilities located in New York City help inmates in their transition back to the
community. There are also interagency efforts to make sure that inmates have the necessary
identification and paperwork completed before release to smooth their transition. For example,
through a cooperative arrangement with New York State’s Department of Correctional Services
(DOCS) and Division of Parole, New York City’s Human Resources Administration (HRA)
19

Gaes, et al., 1999. Also see National Institute on Drug Abuse. Principles of Drug Addiction Treatment: A 

Research-Based Guide. Washington, DC: National Institutes of Health, 1999. 

20
Finn, Peter. Texas’ Project RIO: Reintegration of Offenders, NIJ Program Focus, NCJ 168637, Washington, D.C.: 

National Institute of Justice, 1998. 

21
Finn, Peter, Chicago’s Safer Foundation: A Road Back for Ex-Offenders, NIJ Program Focus, NCJ 167575, 

1998:91-94.

22
Finn, Chicago’s Safer Foundation, 1998:3-6. 

23
Schriro, D., and T. Clement. “Missouri’s Parallel Universe: A Blueprint for Effective Prison Management.” 

Corrections Today 63, no. 2 (2001): 140-143, 152. 


4 Vera Institute of Justice

works with inmates in their last months at a local correctional facility (Queensboro) to apply for
Medicaid benefits. HRA also screens inmates to assess their need for substance abuse treatment
and provides referrals to services as appropriate. The agency also works to pre-qualify inmates
for employment services before release.
New York is also in the process of implementing a multiphase transitional services program
that, similar to Missouri’s, begins with admission to prison. The first phase in New York’s
transitional services program, a two-week introductory and planning period, identifies an
inmate’s goals and uses them to create an individual transitional program plan. In this phase,
inmates apply for copies of their birth certificates and social security cards so that those copies
will be waiting for them at release. The second phase consists of 320 hours of programming
delivered by institutional staff that covers personal responsibility, maintaining ties with family
and friends, life skills such as time management, and coping with prison life.
As part of the third and final phase, inmates bound for New York City are transferred to a
prison in the borough of Queens. There, the Transitional Services Program (TSP) seeks connect
prisoners to family members, potential employers, and community organizations that can offer
them support.24
Empirical Evidence Supporting Correctional Interventions

In contrast to the “nothing works” philosophy of rehabilitation that dominated principles of
correctional practice during the 1970s and 1980s, the current consensus is that rehabilitative
efforts can and do have a positive impact in reducing offender recidivism.25 Some intervention
approaches, however, are better than others and certain treatment interventions may work better
with different offender populations.26 But few treatment-focused programs are rigorously
evaluated and the impact evaluations of existing programs need to make use of strong inferential
designs.27 Meta-analytic reviews have quantitatively synthesized diverse literatures on
correctional interventions. This work has challenged the “nothing works” philosophy and

24

DOCS. Transitional Services Program Manual. Albany, NY: DOCS, 2000. 

The “nothing works” conclusion was drawn from the work of Martinson (1974) and Lipton, Martinson and Wilks 

(1975). There is an increasing consensus that rehabilitative efforts can be effective—e.g., see Andrews, D. A. and J. 

Bonta, The Psychology of Criminal Conduct, Second Ed., Cincinnati, OH: Anderson, 1998; 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(3):369-404, 1990; Cullen, F. T. and P. 

Gendreau, “Assessing Correctional Rehabiliation: Policy, Practice and Prospects”, Pp. 109-176 in Policies,

Processes and Decisions of the Criminal Justice System, J. Horney (ed.), NCJ 182410, Washington, DC: USDOJ, 

2000; MacKenzie, D. L., “Criminal Justice and Crime Prevention”, Preventing Crime: What Works, What Doesn’t, 

What’s Promising. Pp. 9-1 – 9-76, NCJ 165366, Washington, DC: USDOJ, 1997. 

26
MacKenzie, Criminal Justice, 1997; Palmer, T., The Re-Emergence of Correctional Intervention, Newbury Park, 

CA: Sage, 1992. 

27
Gaes, G. G., T. J. Flanagan, L. L. Motiuk, and L. Stewart. “Adult Correctional Treatment”, Pp. 361-426 in Crime

and Justice: A Review of Research, Vol. 26, M. Tonry and J. Petersilia (eds.), Chicago, IL: University of Chicago 

Press, 1999; Sherman, L., D. Gottfredson, D. MacKenzie, J. Eck, P. Reuter, and S. Bushaway, Preventing Crime: 

What Works, What Doesn’t, What’s Promising, NCJ 165366, Washington, DC: USDOJ, 1997. 

25

Vera Institute of Justice 5

provides an empirical basis for understanding the positive impact of appropriate correctional
treatment.28
The treatment literature identifies a group of theoretically sound and empirically-based
principles that typify effective correctional treatment programs and which “represents the most
coherent approach to treatment now available.”29 These principles can be generally categorized
as follows: (1) treatment should address dynamic risk factors; (2) programs should employ
cognitive-behavioral, skills-oriented, or multi-modal treatment approaches; (3) the intervention
should focus on the needs of the participant with higher-risk offenders receiving more intensive
services; and (4) the treatment intervention should be implemented appropriately and in a wellsupported manner.30
Psychologically-oriented treatment approaches emphasize that intervention efforts must be
focused on changeable characteristics that are associated engaging in delinquent or criminal
behavior. Although static factors such as age, gender, ethnicity and criminal history are most
commonly associated with recidivism, there are dynamic risk factors that are strongly correlated
with recidivism and that can be targeted in treatment.31 Dynamic risk factors are often identified
as “criminogenic needs” and include pro-criminal or antisocial attitudes, cognitions, and
behaviors, pro-criminal associates, weak problem-solving and self-regulation skills, poor
socialization, family functioning and structure, unstable employment and weak educational
skills. More successful programs attempt to change antisocial attitudes and feelings, reduce procriminal associations, promote family relationships, increase self-control and problem-solving
skills, and reduce chemical dependencies.32 Meta-analyses on treatment approaches that target
dynamic risk factors, along with narrative reviews lend substantive support to the idea of
targeting treatment to such risk factors.33
28

Cullen and Gendreau, Assessing Rehabilitation, 2000; Gaes, et al., Adult Correctional Treatment, 1999.
See esp. Andrews, D. A., “The Psychology of Criminal Conduct and Effective Treatment”, Pp. 35-62 in What
Works: Reducing Reoffending, J. McGuire (ed.), New York: Wiley, 1995. Also see, Gaes, et al., Adult Correctional
Treatment, 1999:363; also see Andrews, et al., Does Correctional Treatment Work, 1990; Cullen and Gendreau,
Assessing Rehabilitation, 2000; Levrant, S, F. T. Cullen, B. Fulton, and J. F. Wozniak, “Reconsidering Restorative
Justice: The Corruption of Benevolence Revisited?” Crime and Delinquency 45(1):3-27, 1999; MacKenzie,
Criminal Justice, 1997.
30
Multi-modal, in this instance, refers to treatment regimens that are structured to address multiple deficits.
Attempting to address only substance abuse issues, for example, may not be as effective as addressing substance
abuse issues in combination with deficits in educational skills and job readiness skills (see Gaes, et al., Adult
Correctional Treatment, 1999:363-364; Andrews, et al., Does Correctional Treatment Work, 1990; MacKenzie,
Criminal Justice, 1997.
31
Gendreau, P., T. Little and C. Goggin, “A Meta-Analysis of the Predictors of Adult Offender Recidivism: What
Works!” Criminology 34(4):575-607, 1996. Andrews and Bonta, Psychology, 1998; Gendreau, Little and Goggin,
Meta-Analysis of the Predictors, 1996.
32
Andrews and Bonta, Psychology, 1998:357, Table 13.3.
33
Andrews, et al., Does Correctional Treatment Work, 1990; Lipsey, M. W., “What Do We Learn from 400
Research Studies on the Effectiveness of Treatment with Juvenile Delinquents?” Pp. 63-78 in What Works:
Reducing Reoffending, J. McGuire (ed.), New York: Wiley, 1995; Pearson, F. S., and D. S. Lipton, “A MetaAnalytic Review of the Effectiveness of Corrections-Based Treatments for Drug Abuse.” The Prison Journal
79(4):384-410, 1999; Cullen and Gendreau, Assessing Rehabilitation, 2000; Gaes, et al., Adult Correctional
Treatment, 1999; Gendreau, P., and C. Goggin, “Correctional Treatment: Accomplishments and Realities.” Pp. 27129

6 Vera Institute of Justice

Meta-analytic research and narrative reviews also provide a substantive basis for programs
that incorporate cognitive-behavioral, skill-oriented, multi-modal, or familial interventions.34 In
general, however, cognitive-behavioral treatment approaches have been identified as the
interventions most often associated with reducing offender recidivism.35 A review of the studies
evaluating cognitive skills programs concludes that modest support exists for recidivism
reductions in general offender populations, but that evidence exists for stronger positive results
among certain offender subgroups.36 Surprisingly few outcome studies exist.37 Much of the metaanalytic (and evaluation) research focuses on juveniles, making evaluations of adult correctional
treatment programs even more important.38
Although a number of studies show that cognitive-behavioral programming can be an
effective part of an intervention, there are limitations to this research. A recent meta-analysis
suggests that the strongest effects are achieved in demonstration programs, while much more
modest (though still positive) effects are found in correctional intervention programs.39 In
addition, these studies show that juveniles are more responsive than adults--who have the
smallest effects. Finally, there is no clear indication of which program elements are most
important in reducing offender recidivism when such programs are implemented.40
The risk principle suggests that higher risk individuals should receive more intensive levels
of treatment.41 To implement this idea requires that participants be assessed for their risk for a
specific outcome (e.g., recidivism or drug use). Empirical evidence generally supports the notion
that higher-risk individuals benefit more from targeted treatment than lower-risk individuals and
therefore emphasizes the necessity of appropriately identifying individuals who are higher-risk.42
Some evidence, however, suggests that medium-risk offenders may be more suitable targets for
programming than what are traditionally considered the “highest-risk” offenders.43
Treatment implementation and integrity should be consistent with the program’s design. A
program cannot be expected to be successful if it is poorly implemented, under-funded, or
279 in Correctional Counseling and Rehabilitation, P. Van Vourhis, M. Braswell, and D. Lester (eds.), Cincinnati,
OH: Anderson, 1997; MacKenzie, Criminal Justice, 1997.
34
Lipsey, What Do We Learn, 1995; Palmer, Re-Emergence, 1992; Gaes, G. G., “Correctional Treatment.” Pp. 712738 in The Handbook of Crime and Punishment, M. Tonry (ed.), New York, NY: Oxford University Press, 1998.
35
For a review, see Gaes, et al., Adult Correctional Treatment, 1999; also see Andrews, et al., Does Correctional
Treatment Work, 1990; Pearson and Lipton, A Meta-Analytic Review, 1999; Wilson, David B., Leana C. Allen, and
Doris L. MacKenzie. “A Quantitative Review of Structured, Group-Oriented, Cognitive-Behavioral Programs for
Offenders.” 2004, forthcoming.
36
Gaes, et al., Adult Correctional Treatment, 1999:375.
37
Gaes, et al., Adult Correctional Treatment, 1999; also see Sherman et al., Preventing Crime, 1997.
38
For exceptions, see Andrews, et al., Does Correctional Treatment Work, 1990; Pearson and Lipton, A MetaAnalytic Review, 1999; Lipsey, M. W., G. Chapman, and N. A. Landenberger, “Cognitive-Behavioral Programs for
Offenders”, The Annals of the American Academy of Political and Social Sciences 578(November 2001):144-157.
39
Lipsey, Chapman and Nanaberger, Cognitive-Behavioral, 2001:154-155.
40
Wilson, Allen and MacKenzie, 2004, forthcoming.
41
Andrews, et al., Does Correctional Treatment Work, 1990; Gaes, et al., Adult Correctional Treatment, 1999;
MacKenzie, Criminal Justice, 1997.
42
Andrews, et al., Does Correctional Treatment Work, 1990; Gaes, et al., Adult Correctional Treatment, 1999;
MacKenzie, Criminal Justice, 1997.
43
Porporino and Fabiano, Reasoning and Rehabilitation Program Manual, 2000.

Vera Institute of Justice 7

delivered by untrained, uncommitted, or unqualified staff.44 Researcher involvement in the
design and development of programs appears to contribute to larger treatment effects, primarily
through enhancing program integrity.45 An important caveat to note is that programs have often
been shown to show better results in the demonstration phase than when implemented to scale.46
In addition to the above general principles, several other key facets are associated with
effective treatment regimens.47 Occupying a large percentage of an offender’s time,
incorporating relapse prevention techniques and training, and linking offenders to capable
community service providers. There is also growing evidence that longer exposure to treatment
is associated with better post-release outcomes.48
Project Greenlight Program Overview

The New York State Department of Correctional Services (DOCS) and the New York State
Division of Parole, in conjunction with the Vera Institute of Justice, operated an institution-based
transitional services demonstration program that was piloted in New York State’s Queensboro
correctional facility (hereafter referred to as Project Greenlight) and is the focus of this report.49
The project was a Vera-developed demonstration program, and was largely implemented and
operated by Vera demonstration staff.
Planners designed Project Greenlight to provide participants with intensive transitional
services programming in the two prior to their release. The program drew extensively from the
literature on “what works” in criminal justice interventions and attempted to improve postrelease outcomes by (1) incorporating a multi-modal treatment regimen during incarceration and
(2) providing links to community-based service providers and parole officers after release.
Though research studies associated better outcomes, with longer exposure to programming, most
jurisdictions have not found the resources to implement such programs. Vera and correction
officials wanted to develop a short, intensive intervention that had the potential to serve a greater
number of people at a lower cost. Since participants are admitted to the program every week,
most program modules were delivered in a non-sequential manner, making it easy for new
participants to join the ongoing curriculum.
The Greenlight program focused on the key issues offenders face as they prepared to
transition from prison to the community.50 First, the intervention included a cognitive-behavioral
44

Gaes, et al., Adult Correctional Treatment, 1999; MacKenzie, Criminal Justice, 1997.
Lipsey, What Do We Learn, 1995
46
Lipsey, Chapman and Landenberger, “Cognitive-Behavioral Programs”, 2001.
47
Levrant, et al., Reconsidering Restorative Justice, 1999; also see Gendreau, P., “The Principles of Effective
Intervention with Offenders.” Pp. 117-130 in Choosing Correctional Options That Work, A. T. Harland (ed.),
Thousand Oaks, CA: Sage, 1996.
48
NIDA, Principles of Drug Addiction Treatment: A Research-Based Guide, Washington, DC: NIH, 1999; NIDA,
Therapeutic Community, NIDA Research Report, Inciardi, J. A., “A Corrections-Based Continuum of Effective
Drug Abuse Treatment,” NIJ Research Preview, Washington, DC: USDOJ, 1996.
49
The program began in February 2002; budget constraints led to its cancellation in January 2003.
50
See Appendix A and Charts A-1 and A-2 for a detailed description of the program components and participants’
daily and weekly schedules.
45

8 Vera Institute of Justice

component that focused on changing antisocial behaviors and thinking. Program developers
worked closely with the developers of the Reasoning and Rehabilitation Program to adapt the
intervention to a 60-day period.51 This type of programming has consistently been shown to be
most often associated with reductions in recidivism.52
A substantial portion of the intervention focused on addressing key measures such as
housing, employment, drug relapse prevention and substance abuse awareness, linking inmates
to community-based services, and facilitating relationships with field parole officers before
release.53 Inspired by research that demonstrated positive employment outcomes, the program
also offered job readiness training, which included preparing for and conducting interviews, and
guidance in workplace behavior.54 Participants worked with the project’s job developer, who
helped them in their job search and lines up interviews. For participants with a history of
substance abuse, the program provided relapse prevention and drug treatment readiness sessions.
Program staff also worked with staff from New York City’s Department of Homeless Services to
divert offenders away from the shelter system at release.
Inmates also participated in sessions that focused on practical living skills. These sessions
covered how to use public transportation, budgeting, time-management strategies, how to set up
and use a bank account, and where to get emergency cash or non-cash assistance when money is
scarce. This class had particular promise for those with lengthy periods of incarceration, and had
little exposure to ATMs or Metrocards (used to access the public transportation system).
The program also aimed to help participants build social supports available upon release.55
First, the project employed a community coordinator to build a network of community-service
providers and connect participants with providers before release. A full-time family counselor
and a part-time family specialist worked with offenders and their families to address potentially
volatile issues before release. The program also introduced participants to parole officers and

51

Ross, Robert, and Elizabeth Fabiano, Time to Think: A Cognitive Model of Delinquency Prevention and Offender
Rehabilitation, Johnson City, TN: Institute of Social Sciences and Arts, 1985. Gaes, et al, Adult Correctional
Treatment, indicate that the Cognitive Thinking Skills Program, the precursor of the Reasoning and Rehabilitation
program, is the most widely adopted cognitive program, although it is still not widely evaluated.
52
Gendreau, P., and R. R. Ross. “Effectiveness of Correctional Treatment: Bibliotherapy for Cynics.” Crime and
Delinquency 25 (1979):463-489; Gaes, et al., Adult Correctional Treatment, 1999; Lipsey, What do We Learn,
1995.
53
See Appendix A for a complete description of the program and its operational guidelines as implemented in the
Queensboro Correctional facility. Appendix A also includes two exhibits, the first illustrating the weekly schedule
for the eight week intervention and the second illustrates a typical daily schedule.
54
Piehl, A. M. “Economic Conditions, Work, and Crime.” In The Handbook of Crime and Punishment, edited by M.
Tonry. New York: Oxford University Press, 1998. Fagan, J., and R. B. Freeman. “Crime and Work.” In Crime and
Justice: A Review of Research, vol. 25, edited by M. Tonry. Chicago: University of Chicago Press, 1999; Fagan, J.
“Legal and Illegal Work: Crime, Work, and Unemployment.” In The Urban Crisis: Linking Research to Action,
edited by B. A. Weisbrod and J. C. Worthy. Evanston, IL: Northwestern University Press, 1997; Finn, Texas’
Project RIO, 1998. Finn, P. “Job Placement for Offenders in Relation to Recidivism.” Journal of Offender
Rehabilitation 28 ½ (1998):89-106.
55
Finn, Job Placement, 1998; Hanlon, T. E., R. W. Bateman, and K. E. O’Grady. “The Relative Effects of Three
Approaches to the Parole Supervision of Narcotic Addicts and Cocaine Abusers.” Prison Journal 79, no. 2 (1999):
163-181.

Vera Institute of Justice 9

parole supervision requirements to give them a better understanding of how they could avoid
parole-related problems.
Finally, program designers aimed to enhance coordination between DOCS and the Division
of Parole by hiring institutional parole officers and correction counselors to create release plans
that met the needs of both agencies. Each inmate’s work around employment, education,
community resources, family, and substance abuse was captured in the detailed release plan—a
step-by-step outline of what the participant planned to do in each area and the community
organizations he aimed work with after release. Inmates worked on these plans with their case
manager, who is also a correction counselor or an institutional parole officer. That case manager
in turn can discuss areas or programs in the proposed release plan with the field parole officer
who will supervise the person in the community. Case managers transmitted these plans to the
field parole officers. Monitoring a parolee’s progress on the release plan provided a framework
for supervision—but not conditions of parole. The field parole officer could make changes as
needed depending on changes in each person’s circumstances after release.
The Greenlight program sought to improve a number of interim outcomes that would,
hypothetically, influence subsequent recidivism. Program designers believed that the Greenlight
program’s emphasis on employment, for example, would lead to higher rates of employment and
increased job stability after release. Diverting offenders from the shelter system would lead to
more stable living situations. The program’s focus on drug relapse and connecting participants to
substance abuse treatment and counseling was expected to result in reduced drug use. The
detailed release plan and the network of community supports participants established prior to
release were expected to result in increased use of community services. The work on release
planning hoped to reduce the number of parole revocations.
Project Greenlight was originally designed as a three-year demonstration program, but
budget constraints led to the program being funded in its original form for only one year. In
2003, DOCS incorporated aspects of Greenlight into its own transitional services program. Our
research covers only the first year of the program’s evaluation, with all the advantages and
drawbacks that entails.
Key Hypotheses

We have two key hypotheses regarding outcomes for program participants which we examine
here.56 The program is specifically focused on addressing those issues most commonly identified
as impacting a participant’s general well-being and quality of life after release. Helping
participants locate stable housing, providing connections to either jobs, or employment services
after release, helping re-establish relationships with their families or friends, connecting them to
a variety of community-based services including self-help organizations, substance abuse
services, mental or general health providers, and providing them with the basic skills to navigate
56

There are multiple ancillary hypotheses dealing with intermediate outcomes as well; more detail can be found in
the Appendices.

10 Vera Institute of Justice

in the community are all expected to facilitate their reintegration into the community. These are
all interim outcomes that we examined in our post-release interviews with participants and to
some extent, parole officers.
Improvements in these interim outcomes were expected to lead to lower levels of recidivism:
rearrests, reconvictions, reincarcerations, and parole revocations. Our first hypothesis is:
H1:
A structured multi-modal transitional services program, located in or near
the community to which offenders will return, with links to community services
and field parole supervision, will result in lower rates of arrests (total and
felony).57
These same program elements were expected to reductions in parole revocations. In addition,
the program’s emphasis on connecting individuals to field parole officers before they are
released, structuring the working relationships between parole and DOCS personnel, and
providing release plans to parole officers were expected to improve participant perceptions of the
process of parole supervision and to reduce the rate of revocations. Our second hypothesis is:
H2:
A structured multi-modal transitional services program, located in or near
the community to which offenders will return, with links to community services
and field parole supervision, will result in lower rates of parole revocations.
We examine a number of the interim outcomes in the findings section and discuss the
implications for our measures of recidivism.

57

We use arrest as our primary outcome measure for several reasons. Arrest has been argued to be the most valid
outcome measure of recidivism (e.g., see Maltz, Michael. Recidivism. Orlando, FL: Academic Press, 1984). Second,
the short one-year time frame for our study makes the use of other measures (reconviction) problematic; the
processing time from arrest to conviction is often lengthy and is likely to be less useful as an outcome as a result.

Vera Institute of Justice 11

Data and Methods
Overview of the Study Design

58

Study design is crucial to any evaluation in that design strength is crucial to the conclusions that
one can draw at the end. Thus, we describe in significant detail the strengths and relative
weaknesses of our research design. It is also inherent in the research process that research
projects encounter known or unknown obstacles along the way—this often requires adjusting the
design in order to better maintain the integrity of the study design. We therefore also note
difficulties we encountered and how we adjusted the design to attempt to deal with those
difficulties.
Potential study participants were identified by the DOCS MIS/Classification section each
month based on specific eligibility criteria. Although inmates were not required to participate in
the research, they were required to participate in mandated programming per DOCS policy.
Refusal to participate could result in sanctions including loss of good-time release. Eligible
inmates were defined as conditional releases (CR) who were 75 to 105 days from release. The
time to release is critical because inmates must have at least 60 days to participate in the program
plus the time it takes to transfer from one facility to another. CR inmates were chosen because
they could be identified with sufficient time remaining to participate in the eight-week program;
CR inmates include individuals who had previously been denied release by the Board of Parole,
and inmates determinately sentenced under the 1995 and 1998 New York State Sentencing
Reform who were not eligible for discretionary release. Because Queensboro is a minimumsecurity institution, safety and security issues require that inmates not be assigned to the facility
if any of the following conditions apply: conviction for a sex offense, vicious or callous violence,
high institutional risk score, current disciplinary sanction, medical reasons, existing warrants, or
ineligible current assignment (e.g., work release, shock incarceration). In addition, Queensboro is
not structured as a male/female facility, and security and safety concerns dictate that all
participants will be male. Finally, study participants were selected only if they had a conviction
in one of New York City’s five boroughs to increase the chances that the inmate would make use
of NYC-based service providers that were brought into the facility.
Our evaluation is best described as using a "mixed" design, combining quasi-experimental
and experimental elements. Although our intent was to follow only people who were eventually
randomly assigned, early termination of the program led us to reconsider this decision. To
include the quasi-experimental haphazardly assigned individuals during the first five months
meant greater statistical power; by excluding them, the obvious trade-off was a loss of statistical
power for a purely randomized design. In discussing this, we considered the haphazard
assignment sufficiently strong so as to warrant the inclusion of those participants. We discuss
these issues in more detail in the following paragraphs.
58

See Appendix B for more detail on study design, issues encountered, and data sources.

12 Vera Institute of Justice

Our study participants are separated into three distinct groups and are defined in the
following manner (See Exhibit 1 for program flow). “Potentially” eligible participants were to be
transferred to the Queensboro facility based on DOCS transportation schedules and their time
remaining until release. All individuals identified by DOCS as “potentially” eligible did not
become part of our final sample until such time as they were assigned to a NYC parole office
after release—those not assigned to a NYC parole office were removed from the final analyses.
Those who were identified as eligible by DOCS, were not transferred to the pilot facility due to
space limitations, and were ultimately released from an Upstate (not in New York City)
institution constitute the first comparison group (hereafter referred to as Upstate—for Upstate
releases).59 To our knowledge, the Upstate group did not participate in any transitional services
programming before release. Inmates transferred to Queensboro were assigned to either the
intervention (hereafter referred to as GL or Greenlight) or the existing DOCS transitional
services program (hereafter referred to as TSP).
All participants identified by DOCS and transferred to Queensboro between February and
June 2002 were sequentially assigned to program slots on approximately a two-to-one basis,
Greenlight to TSP with the following caveat. During the first few weeks, the intervention beds
were filled with the first 52 participants. During the next two months while the initial
intervention participants were in the eight-week program, all assignments were to the TSP
program. About 100 participants were assigned to the TSP program during this period (a flow of
about 50 potential participants per month). Once the first GL participants began leaving prison,
we began the 2-to-1 assignment process on a daily basis. Thus, this assignment procedure to the
two programs at Queensboro essentially constitutes a “haphazard assignment” procedure, albeit
in a sequential manner.60 Although not randomized, “haphazard assignment procedures…can be
a source of stronger quasi-experiments” than other quasi-experimental designs.61
In terms of the haphazard assignment protocols, the institutional locator officer at
Queensboro received a list of inmates each day who were to be transferred to Queensboro on the
following day. Based on order of appearance on that list, individuals were sequentially assigned
such that the first two with a “conditional release” designation were assigned to the intervention,
the third to the TSP program, the fourth and fifth to the intervention, and so on as long as bed
space existed to accommodate the assignments. A running list was maintained so that each
subsequent day continued the assignment from the prior day—i.e., if the last assignment had
only one person assigned to the intervention, the first person the next day was designated GL, the
second person TSP, the third and fourth persons GL, and so on. Once individuals arrived at
Queensboro, only those identified during intake as returning to New York City were assigned to
59

As a result, individual characteristics did not play a role in whether these individuals were transferred to
Queensboro—their selection was based entirely on DOCS transportation schedules. It is possible however, that
some of these had their transfers cancelled due to misbehavior although DOCS cannot identify for us whether this
occurred for specific individuals.
60
Shadish, William R., Thomas D. Cook, and Donald T. Campbell. Experimental and Quasi-Experimental Designs
for Generalized Causal Inference. Boston: Houghton-Mifflin, 2002:303.
61
Shadish, Cook and Campbell, Experimental and Quasi-Experimental Designs, 2002:302-303.

Vera Institute of Justice 13

the Greenlight program. Individuals who indicated during the intake interview that they would
not be returning to New York City were then assigned to the TSP program. Thus, if an individual
was assigned to a program bed and during the program orientation it was discovered that this
person was not returning to New York, they were removed from the program and placed in the
TSP group. These removals were not “backfilled” from the existing TSP group, but new
assignments were made as they entered the institution. As noted before, all participants identified
as not returning to New York City were eventually removed from the analyses.
In July 2002, as the program went to its full capacity of 104, a trickle process random
assignment procedure was instituted. In the six-month period between July and December, there
were minor deviations (15 percent of assignments) from the assignment protocol. Although these
deviations were higher than we would have liked, the overall integrity of the randomization
procedure between July and December remained intact.62 Our “mixed” study design therefore
combines a haphazard sequential assignment process with a randomized assignment.
Although DOCS identified all potential participants, the final determination for study
inclusion depended on assignment to a NYC parole office—information we did not receive until
after release. Thus, a total of 677 potential participants were transferred to the pilot facility: 349
were assigned to the GL intervention, 328 were assigned to TSP (See Exhibit 1; also see
Appendix B, Table B-3). A total of 128 participants were identified as part of the Upstate group
(total number of potential participants=805).63 However, we anticipated that we would lose a
small proportion of individuals because although they may have been convicted in one of New
York City’s five boroughs, not all would return to the City. Eight percent of our overall sample
did not return to the city (64 of 805). As a result of the program requirement to serve only NYCbound participants where possible (and because program staff transferred at intake any inmates
identified as not returning to New York to the TSP program), a greater proportion of those
identified as TSP participants were in fact not coming back to NYC. In addition, there have been
2 known deaths and four individuals initially identified as returning to NYC who later transferred
out-of-state—we were unable to acquire dates of death or transfer for any of the cases in order to
censor them in the analyses. Therefore, after removing those 64 cases that did not return to NYC
and deleting the six cases with unknown dates, our final sample size for each group was as
follows: Greenlight N=344, TSP N=278, Upstate N=113 (Total N=735). Of the GL and TSP
62

Boruch, R. F. Randomized Experiments for Planning and Evaluation: A Practical Guide. Applied Social Research
Methods Series, Vol. 44. Thousand Oaks, CA: Sage, 1997:106. Our trickle-process design assigned individuals on
approximately a 2-for-1 basis of intervention to control. Deviations appear to occur most often when institutional
personnel ignored the random assignment instructions and disproportionately filled empty beds on the intervention
side when too many participants were released at once. There were also cases in which the institution did not notify
us that the intervention side was full and intervention assignments were shifted to the control group. Thus, the
availability of bed space (or lack of it) is entirely responsible for deviations from the assignment protocol. For
considerably more detail on the assignment procedures and other aspects of the methodology, see Appendix B.
Boruch suggests that up to a 15% deviation from treatment assignments may warrant further investigation—in this
case, we have good data on why deviations occurred.
63
We note again that identified participants are only potentially eligible until they have been identified as returning
to New York City. Once they have been released to a NYC parole office they officially become part of our final
sample.

14 Vera Institute of Justice

cases, 325 were assigned under the haphazard assignment process (186 GL; 139 TSP) and 297
were assigned under the trickle-process random assignment procedure (158 GL; 139 TSP).

Vera Institute of Justice 15


16 Vera Institute of Justice

TSP (N=135)

GL (N=183)

TSP (N=118)

TSP (N=40)

GL (N=17)

GL (N=143)

TSP (N=111)

TSP (N=28)

GL (N=17)

GL (N=141)

NYC Returns

TSP (N=139)

TSP (N=170)

Institutional Assignment

GL (N=186)

GL (N=189)

Haphazard Assignment
Initial Assignment

Upstate (N=113)

NYC Returns

Upstate (N=128)

Haphazard Assignment
Initial Assignment

Exhibit 1. Program Flow—Project Greenlight Evaluation

N=139

TSP

N=158

GL
TSP
N=278

GL
N=344

Upstate
N=113

Final Sample

Data Sources

Data for this evaluation come from diverse sources. We obtained initial background information
on participants from DOCS. The demographic and institutional data included the New York
State criminal justice identification number (NYSIDS); dates of incarceration and release; intake
scores on substance abuse; reading and math ability scores; arrest dates for the incarcerating
offense; original commitment offense; date of birth; educational attainment; and race/ethnicity.
In our analyses, we based commitment offense on data provided by DOCS. The New York State
Division of Criminal Justice Services (DCJS) provided detailed information on criminal history
in New York State. These data go back to 1970 and include arrest, conviction, and sentence
disposition for each recorded offense. The DCJS data for each offense also include county of
conviction, whether the offense involved a child victim, drugs, or a firearm, and multiple offense
descriptors.
Information on recidivism outcomes included arrests and parole violations after returning to
the community. Information on new offenses committed within New York State for a minimum
of 12 months post-release came from a comprehensive statewide database kept by DCJS—this
includes any new arrests or convictions in all of New York State, not just New York City. Each
new offense contained the same information listed above for criminal history information. The
Division of Parole provided violation information and data on drug tests conducted during parole
supervision.
Interviews conducted with parolees after release gathered information on offender
background characteristics and outcomes.64 Research staff contacted GL and TSP participants
during their last few weeks at Queensboro to gain their consent to participate in the research. The
voluntary interviews were conducted face-to-face two to three months after release either at the
participants’ parole appointments or at a mutually convenient time and place. The interview
instrument, derived from an instrument currently in use by the Urban Institute, focused on
housing, employment, income, use of community-based services, perceptions of parole officers,
and adherence to parole supervision requirements. Other questions focused on emotional and
psychological states, family relationships, past and present drug use, neighborhood perceptions,
and satisfaction with and perceived usefulness of transitional services programming. We
augmented the interviews post-release interview summaries written by the interviewers and data
collected from supervising parole officers from hardcopy case folders.
We used three approaches to locate this highly transient population. While administering
consent forms at Queensboro, research staff asked participants to provide an extensive list of
contact information. In addition, we arranged with the Division of Parole to conduct interviews
at parole offices on individual reporting days for those participants we were unable to contact.
Finally, for reincarcerated participants, the NYC Department of Correction allowed us to conduct
interviews in their facilities—this included people rearrested or detained for parole violations.
64

Our interview instrument is attached as part of Appendix D.

Vera Institute of Justice 17

This is especially important because in many studies, interviews are conducted with participants
who remain in the community, but are unable to access those who have been reincarcerated. In
all, 109 Greenlight and 48 TSP participants were interviewed after they were released. Appendix
B (and Table B-4) provides more detail on methods used to contact parolees for interviews,
response rates and the difficulties involved.
We also interviewed parole officers of participants we could not locate, and surveyed 19 of
the 26 field officers assigned to supervise the Greenlight participants. In addition to their
perceptions of the program, parole officers were also asked a short series of questions about
employment, housing, and parole violations for a randomly selected group of GL and TSP
participants.65

65

More detail from parole officer interviews is provided in the Appendices; program perceptions are summarized in
Appendix A, methodological considerations in Appendix B, and summaries of the information on study participants
in Appendix C. Instruments are included in Appendix D.

18 Vera Institute of Justice

Findings
Data Analyses

In this section, we examine the evidence and available measures indicating the impact of the
intervention on a variety of outcomes. We begin with demographic information and offense
histories to identify any pre-existing differences between the three study groups that might
account for differences in post-release outcomes. Next, we present information gained from
interviews with parole officers about the study participants’ performance on parole. Adherence
to parole supervision requirements was assessed through parole officer ratings of participant
performance on a seven-point scale where we asked the officer to rate adherence to (1) reporting
requirements and (2) general requirements. We also examine parolee performance based on the
number of times participants were cited for violating the conditions of parole in the first three
months after release.
The following section discusses information gained from interviews with Greenlight and TSP
participants. These interviews included participants’ assessments of the usefulness of
information contained in pre-release programs and proximate outcomes such as post-release
contact with social services, employment, drug use, and family relations.
The last section reports arrest and revocation outcomes. Here we examine differences
between the groups in new arrests, new felony arrests, and revocations for a minimum of one
year after release from custody. Survival analysis is used to examine post-release failure and
control variables are introduced through Cox proportional hazard models.66 Life tables are
generated for each group examining time-to-failure, survival curves are plotted, and tests of
significance are shown for the survival trends using the Kaplan-Meier method.
Studies should typically examine multiple measures of recidivism—alternative measures of
recidivism to arrest and parole revocations are usually new convictions and reimprisonment. In
this study, parole revocations essentially constitute a measure of return to prison. While all
parole revocations in New York State do not end in reincarceration, the Division of Parole is
reporting only those cases that are returned to prison.67 There were no reincarcerations of
individuals who successfully completed parole supervision during our study period (that is, new
convictions), although several individuals experienced new convictions while still under parole
supervision. In short, all incidences of reimprisonment in our data were also associated with a
parole revocation. Finally, we do not examine reconvictions in these data; in a short follow-up
period such as we have here for one year after release, a process leading to a new conviction is

66

Luke, Douglas A., and Sharon M. Homan. “Time and Change: Using Survival Analysis in Clinical Assessment
and Treatment Evaluation.” Psychological Assessment 10(4): 360-378, 1998. Yamaguchi, Kazuo. Event History
Analysis. Applied Social Research Methods Series, Vol. 28. Newbury Park, CA: Sage Publications, 1991.
67
Other outcomes are possible but not reported to us. For example, a parolee may be revoked by the administrative
law judge and then returned to an increased level of parole supervision.

Vera Institute of Justice 19

time consuming. In our study, there were only 15 new convictions in the follow-up period for all 

groups.68
In our examination of arrest survival probabilities, we censor cases (as non-failure terminal
events) that are revoked because those individuals are no longer at risk for arrest. In the
examination of revocation survival probabilities, we censor cases (as non-failure terminal events)
that successfully complete parole supervision because those individuals are no longer at risk for
revocation. We do not censor arrests because arrests do not typically remove the individuals
from the community for significant periods of time—that is, they are often still at risk for
revocation (or rearrest).
Description of the Study Participants

Table 1 shows basic demographics (age, education, and race) for the 735 study participants
broken down by study group. ANOVA indicates that Greenlight, TSP, and Upstate participants
have very similar characteristics across the board. Trivial differences in mean age at release are
noted across the three groups, as is true of educational levels and racial characteristics. Most
people in the study group have less than a high school education and a majority are of AfricanAmerican descent or Hispanic ethnicity.

Table 1. Group comparisons on basic demographic variables (ANOVA analysis).
Variables
GL
TSP
Upstate
Total
(N=735)
(N=113)
(N=278)
(N=344)
Age at release
33.4
32.5
33.4
33.6
Mean
9.4
9.1
9.5
9.5
Std. Dev
Education
8 years or less
9-11 years
12 or more
Missing

Sig.
ns

(N=658)
(N=100)
(N=240)
(N=318)
11.6%
11.7%
11.0%
11.6%
49.5%
50.0%
48.0%
49.7%
38.9%
38.3%
41.0%
38.7%
N=26 (7.6%) N=38 (13.7%) N=13 (11.5%) N=77 (10.5%)

ns

(N=735)
6.3%
56.2%
37.6%

ns

Race/Ethnicity*
N-H White/Other
N-H Black
Hispanic

(N=344)
5.2%
57.6%
37.2%

(N=278)
7.6%
55.8%
36.7%

(N=113)
6.2%
53.1%
40.7%

*Racial/ethnic categories are non-Hispanic White/Other, non-Hispanic Black and Hispanic. Other is combined with
White due to the very small percentage of cases in both categories.

68

New arrests and convictions are provided by the Division of Criminal Justice Services. Revocations are provided
by the Division of Parole. Reimprisonment is confirmed by the Department of Correction.

20 Vera Institute of Justice

Table 2 provides an overview of criminal histories of the study participants obtained from
official records maintained by DCJS and DOCS.69 Primary commitment offense categorizations
are based on Uniform Crime Report codes provided by DOCS. DOCS data was used in this case
because there was no missing information and we assumed that the commitment offense would
be accurate based on DOCS records. We separated robbery as an offense because of its
significance as a property and violent offense and the numbers of individuals in the category.
Property offenses include burglary, larceny, motor vehicle theft, forgery, stolen property, fraud,
and embezzlement. Violent offenses include homicide, rape, and assault.
As is shown, the three groups were closely matched on all variables—none of the group
differences in the table rise to the .05 level of statistical significance. On average, Greenlight
participants have slightly more prior arrests and convictions than TSP participants, who in turn
have more arrests and convictions than Upstate participants.70 In terms of the primary offense
(commitment offense), the Upstate group is comprised of fewer drug offenders and more robbery
offenders. In addition, data provided by DOCS shows that the Upstate group had more people
who self-reported alcohol-only abuse and fewer individuals who reported abusing alcohol and
drugs in combination.

Table 2. Criminal history and substance use indicators by program group (ANOVA
analysis).*
Variables
Total Arrests
Mean
Std. Deviation
Total Convictions
Mean
Std. Deviation
Felony Arrests
Mean
Std. Deviation
Felony Convictions
Mean
Std. Deviation
Misdemeanor Arrests
Mean
Std. Deviation
Misdemeanor Convictions
Mean
Std. Deviation
Age at first arrest

GL
(N=339)

TSP
(N=274)

Upstate
(N=112)

Total
(N=725)

Sig.
ns

8.65
9.1

7.81
8.1

6.67
9.7

8.02
8.8

5.94
7.6

5.31
6.8

4.33
8.5

5.45
7.5

4.57
4.7

4.12
3.9

3.57
4.3

4.25
4.4

1.76
1.6

1.72
1.6

1.53
1.6

1.71
1.6

3.11
5.7

2.63
5.2

2.09
7.4

2.77
5.8

4.18
6.8

3.59
6.1

2.80
7.6

3.75
6.7

ns

ns

ns

ns

ns

ns

69

Criminal history information obtained from DCJS for 10 of the 735 participants is missing. 

Table C-1 presents a correlation matrix of key criminal history and demographic variables. As one would expect, 

there are relatively strong correlations between our criminal history measures. Tables C-2 and C-3 show 

proportional distribution of felonies and misdemeanors among study groups rather than simple means. 


70

Vera Institute of Justice 21

Mean
19.7
19.9
19.8
19.8
Std. Deviation
5.0
5.3
4.6
5.1
Primary Offense
N=344
N=278
N=113
N=735
ns
Robbery
29.1%
34.9%
37.2%
32.5%
Violent
20.9%
15.5%
21.2%
18.9%
Drug
22.7%
21.2%
13.3%
20.7%
Property
15.1%
17.3%
15.0%
15.9%
Other
12.2%
11.2%
13.3%
12.0%
Substance abuse
N=324
N=269
N=108
N=701
ns
None
35.2%
37.5%
31.5%
35.5%
Alcohol
8.6%
7.4%
15.7%
9.3%
Drugs
33.3%
30.9%
34.3%
32.5%
Alcohol and Drugs
22.8%
24.2%
18.5%
22.7%
Missing
N=20 (5.8%)
N=9 (3.2%)
N=5 (4.4%) N=34 (4.6%)
*Note: For all arrest and conviction information and age at first arrest, there is missing criminal history data for 5
GL, 4 TSP, and 1 Upstate participants.

Interviews with Greenlight and TSP Parolees

We interviewed Greenlight and TSP participants after release in order to assess the impact of the
program on different outcomes (more detailed information on the number of interviews
completed is shown in Appendix Table B-4—the instrument itself is included in Appendix D).
These interviews included information on perceptions of the programs. We first examined
Greenlight and TSP responses on how helpful they found different aspects of their pre-release
program. The results are displayed in Table 3. The table indicates that of 11 types of information
asked about their pre-release programs, Greenlight participants were more positive than TSP
participants in nine of the areas., Due in part to the small number of responses, these findings
reached statistical significance on just two of the measures (information about gaining child
custody and attending Alcoholics Anonymous and other support groups).71

Table 3. Proportion of Greenlight and TSP respondents who found aspects of pre-release 

programming helpful.

Interview response
Greenlight
TSP
Significance 

(n=85)
(n=16)
Life skills
Finding a place to live
Finding a job
Continuing education
Gaining child custody
71

86%
58%
75%
81%
88%

88%
50%
73%
64%
43%

ns
ns
ns
ns
.03

Although 109 GL and 48 TSP participants were interviewed, participants were not asked to answer this question if
they responded “NO” to a prior question on whether they received pre-release programming at Queensboro. In
addition, several respondents simply refused to answer the questions. It is significant in and of itself that such a
small proportion of the TSP participants interviewed recognized that they were actually involved in pre-release
programming while at Queensboro.

22 Vera Institute of Justice

Drug/alcohol treatment
AA & other support groups
Health care
Counseling
Financial assistance
Obtaining identification

75%
77%
77%
73%
78%
85%

65%
50%
77%
64%
60%
67%

ns
.07
ns
ns
ns
ns

We hypothesized that if Greenlight had made a significant impact, we would see its effects
on a series of proximate outcomes. These included referrals to social service agencies, postrelease contacts with social services, better and more stable employment and housing outcomes,
drug use, family relations, participation in community activities, and knowledge of parole
conditions. We believed that if Greenlight were to have an impact on recidivism it would occur
through the positive impact on practical outcomes such as employment and “pro-social”
behaviors such as less drug use.
These outcomes and behaviors were also assessed during interviews with Greenlight and TSP
participants. (Upstate participants were more difficult to locate after release and only three
interviews were conducted—thus, we do not report on those data here.) However, the proportion
of total parolees interviewed was low; 32 percent for the Greenlight group and 17 percent for the
TSP group (See Appendix B). The success rates were better for Greenlight participants because
we had more information on their post-release residential plans as a consequence of the greater
ease in contacting them while still incarcerated. We also had greater ease in contacting them
through their parole officers. Because the proportion of parolees interviewed was low, we were
concerned that information collected on the interviews would not be representative of the entire
Greenlight and TSP samples. We were particularly concerned that we would have less success
trying to interview parolees who remained engaged in criminal careers. Therefore, we compared
interviewed parolees and those not interviewed within the Greenlight and TSP groups on
demographics, criminal history, and recidivism; the results are presented in Appendix Table C-4.
The table does not suggest any significant differences between interviewed and not interviewed
participants on any of the variables on which they were compared. This may be due in part to our
efforts to contact individuals in the community as well as those rearrested or revoked and held in
local facilities.72
As noted, we asked parolees multiple questions to assess social service referrals, post-release
contacts with social services, drug use, family relations, knowledge of parole conditions, and
opinions about being on parole. Survey items in each of these areas were condensed into additive
scales which simply summed the number of items answered positively. Statistics on these scales
72

Although social scientists commonly assume that high response rates increase the validity of their findings, an
emerging view that brings this assumption into question holds that low response rates can yield valid information
and that under some conditions increased response rates can in fact lead to problematic results. See Feinberg, Barry.
“Choosing a Method”, and Schulman, Mark, “Fielding the Study”, presented at AAPOR meetings, New York, NY,
6/22/2004.

Vera Institute of Justice 23

(means, standard deviations, and internal consistency) are presented in Appendix Table C-5. The
interview also queried parolees about time spent working, time involved in community activities,
and whether their first night out was spent on the street or in a shelter. Differences between
Greenlight and TSP parolees are displayed in Table 4.

Table 4. Comparisons between Greenlight and TSP interviewees on post-release measures.
Value
Greenlight
TSP
Measures
Significance
Range
(N=116)
(N=48)
# Service referrals 

# Service provider contacts 

Knowledge of parole conditions 

Opinions about parole supervision 

Family relations 

Times using drugs in past month 

Weeks employed 

Community/self-help activities/mth 

Spent first night on street/in shelter 


0–9
0–9
0–9
0–1
0–1
0-90
0-48
0-55
0 - 100%

2.20
1.23
8.31
0.78
0.80
1.77
2.72
11.44
7%

1.01
0.51
8.16
0.65
0.83
6.14
2.10
10.98
4%

.001
.001
.03
ns
ns
.02
ns
ns
ns

The table shows that Greenlight participants self-reported receiving significantly more 

referrals than TSP participants to social services, including substance abuse treatment, 

employment interviews or training, housing, education, anger management, and family 

counseling. Greenlight participants reported an average of 2.20 service referrals compared to 

1.01 for TSP participants.73 These referrals resulted in significantly different reported rates of 

contact with service providers upon release from prison; Greenlight participants reported an 

average of 1.23 post-release contacts with providers compared to 0.51 for TSP participants.74

Table 4 also shows that Greenlight respondents differed from TSP respondents in their
approach to parole supervision. The table indicates slight differences between Greenlight and
TSP respondents in knowledge of conditions of parole. Greenlight respondents were aware of an
average of 8.31 conditions asked about—including regular reporting; notifying parole of changes
in residence or employment or if arrested; not associating with criminals; obeying all laws; not
possessing firearms or drugs; observing a regular curfew; and seeking employment. This was
significantly more than TSP respondents, who were aware of 8.16 conditions of parole.75 GL
respondents also differed from TSP participants in opinions about parole. Greenlight respondents
were significantly more positive according to this scale that included questions about the value of
parole and opinions about the parole officer.76
73

t(158) = -3.84, p < .001. 

t(158) = -3.39, p < .001. 

75
t(158) = -2.24, p < .03. 

76
t(158) = -2.98, p < .01.
74

24 Vera Institute of Justice

Finally, Table 4 indicates that Greenlight respondents self-reported less drug use than their
TSP counterparts. TSP respondents reported an average of 6.14 drug experiences in the month
prior to the interview compared to just 1.77 for Greenlight participants.77 Table 4 does not show
any significant differences between Greenlight and TSP interviewees in the following measures:
•	 Family relations scale, which assesses the extent to which parolees relied on family
members for advice, assistance, and support;
•	 Number of weeks worked since release;
•	 Number of days involved in various community and self-help activities (church, 

recreational club, organized sports, ex-offender groups, and self-help groups); 

•	 Whether the respondent spent his first night out of prison in a homeless shelter.
In sum, Greenlight participants express slightly more positive views of the intervention and
the program appears to have had an impact in several areas. But along several significant
dimensions, we found no difference between the two groups on either the program perceptions or
reported outcomes after release.
Interviews with Parole Officers

We completed fewer post-release interviews than we anticipated, so we asked parole officers a
series of very basic questions about the performance and outcomes of parolees within the first
three months after release.78 The parole officer interviews enhanced our understanding of
participant outcomes—especially for housing, employment, and adherence to parole supervision,
but we expected parole officers to provide a more independent measure of these outcomes. To
identify the parolees we wished to collect information about from the parole officers, we
randomly sampled 45 percent of the GL (for a total N of 157) and TSP (for a total N of 107)
study participants, choosing our sample size on the number of interviews it seemed reasonable to
complete and simultaneously minimizing the burdens on the field parole officers we would be
talking to. We completed interviews with parole officers for 66 percent of the GL participants
(N=103) and 59 percent of the TSP participants (N=63).79 We asked parole officers to assess
how well the participants adhered to reporting and other parole requirements, whether they had
lived in a shelter during the first three months after release, whether they had been employed
during the same period, and to indicate the participants’ current living situation.
We asked the first two questions in order to get a sense of whether the GL participants
differed from the TSP participants from the parole officers’ perspective. We expected that GL
participants would receive more positive ratings because of the program’s efforts to familiarize
77

F[1,127] = 5.38, p < .02. Although we do not present the data here, data obtained from the Division of Parole on
random drug tests conducted after release and while still under parole supervision showed no inter-group differences
on drug use.
78
See the parole officer interview in Appendix E for more detail on the exact questions and response categories.
79
We note that we do not have a large enough N to ensure adequate statistical power for these data.

Vera Institute of Justice 25

them with parole policies and procedures. We asked questions on employment and shelter
because these were outcome elements that the GL program was intended to impact. Moreover,
these are post-release situations that parole officers should be familiar with since parolees are
required both to seek and maintain employment as part of the general parole requirements and to
notify their supervising officer of changes in their living situation. We asked parole officers to
provide this information (and the current living situation) in the context of the first three months
so that we might combine the parole officer reports with self-reports (interviews completed in the
first three months) for more comprehensive outcome information.
Finally, we asked parole officers to provide information to us on the number of written
citations of eight specific parole reporting requirements. The Division of Parole has 12 specific
supervision requirements and a thirteenth for special conditions. All parolees are required to
adhere to the first 12 and may be required to adhere to special conditions of parole (e.g., such as
attending drug treatment). If a parolee does not adhere to specific requirements, they may be
cited by the supervising officer for not doing so. If the parolee continues to break parole’s
supervision rules or commits a serious enough offense, parole may issue a violation warrant
which results in a hearing in front of an administrative law judge. The judge may revoke parole
and return the person to prison to complete his/her sentence.80 In this case, we asked parole
officers to indicate whether the parolee had been cited for any of eight different supervision
requirements and if so, how many times. The results from the interviews with tests of mean
differences are in Table 5.
Table 5. Parole Officer Assessments of Parolee Performance (Means and T-tests).
Variable
Means
T-statistic
GL
TSP
(N=103) (N=63)
Outcomes
Met Reporting Requirements 

5.98
6.48
-1.949ns
5.04
5.32
- .828ns
Met General Requirements 

Lived in Shelter 

.10
.14
- .755ns
Employed After Release 

.33
.34
- .366ns
Parole Supervision
Contact PO at Release 

Make Office Reports 

Won’t Leave Assignment Area 

Notify PO of Changes 

Will Reply to PO Inquiries 

Will Notify PO of Arrest 

No Criminal Acquaintances 

Will Obey the Law 

80

.02
.33
.10
.06
.03
.06
.02
.03

.03
.13
.06
.02
.08
.11
.05
.11

- .490ns
1.386ns
.632ns
1.175ns
- .840ns
-1.101ns
- .845ns
-1.404ns

The evidence may be found insufficient and the parolee returned to supervision. It is also the case that the
parolee’s status may be revoked, but outcomes other than prison may ensue, e.g., a return to parole supervision with
more intensive supervision. We reiterate that in this study, revocation refers to reimprisonment.

26 Vera Institute of Justice

None of the differences between the GL and TSP participants were statistically significant. The
TSP group did slightly better in meeting reporting and general parole requirements Despite the
positive perceptions of the parole officers about the strength of the release plan for employment
services after release (see Appendix A), actual outcomes are similar to TSP participants with no
release plan. Greenlight did not make a significant impact in any of these key elements. Data
provided by the parole officers from case folders also showed no statistically significant
differences. Results from the parole officer interviews were consistent with the interviews with
program participants in finding no differences between the two groups.
Analysis of Recidivism Outcomes

Table 6 shows the raw group differences in outcomes (expressed as rates) a year after the last
study participant had been released. Greenlight participants had more total arrests (44 percent of
the intervention group experienced one or more misdemeanor or felony arrests) than the TSP (35
percent) or Upstate (32 percent) groups.81 The table also shows that 24 percent of the Greenlight
sample was rearrested on a felony charge compared to 19 percent of the TSP sample and 16
percent of the Upstate sample. Differences between groups were significant at the .05 level for
total arrests, but not for felony arrests. GL participants had their parole revoked at a significantly
higher rate than the two comparison groups and the TSP group at a higher rate than the Upstate
group.
Table 6: Arrest and revocation outcomes for entire study period as of April 1, 2004.
Outcomes
GL
TSP
Upstate
Total
Significance
Any new arrest
New felony arrests
Revocations

44%
24%
29%

35%
19%
25%

32%
16%
17%

39%
21%
25%

.02
ns
.05

Table 7 presents parole outcomes. Approximately 30 percent of both the GL and TSP groups
had been successfully discharged from parole supervision as of April 1, 2004. More than 50
percent of the Upstate group was still under parole supervision at study end. We think this is
because a block of Upstate participants were identified toward the end of the study’s recruitment
phase, and less time had elapsed for them to complete parole supervision at the end of the study.
Table 7. Parole status at the end of study by program group for NYC returns.
Parole Status
Greenlight
TSP
Upstate

81

In cases of multiple arrests, we use the date of the first arrest. In the analysis of felony arrests, if a study
participant experienced one or more misdemeanor arrests prior to a felony arrest, the date of the felony arrest is
used.

Vera Institute of Justice 27

Successful Discharge
Non-Reporting/Warrant Issued
Revoked
Still Under Supervision

109
26
102
107

(31.7%)
( 7.6%)
(29.7%)
(31.1%)

81
28
69
100

344

Total

(29.1%)
(10.1%)
(24.8%)
(36.0%)

278

27
7
20
59

(23.9%)
( 6.2%)
(17.0%)
(52.2%)

113

The raw rates presented in Table 6 do not account the time at risk. If the GL participants, for
example, had more time in the community than the other two groups, one might expect higher
rates of arrests or revocations. To compensate, we used survival analysis. Table 8 summarizes
the results of life-tables generated using this approaches.82
Table 8. Summary of life-table cumulative survival probabilities by primary outcome by
program group (N entering risk interval shown in parentheses).
Greenlight
TSP
Upstate
Total Arrests
6 months
12 months
18 months

.828
.659
.527

(N=280)
(N=211)
(N= 71)

.870
.758
.596

(N=235)
(N=196)
(N= 63)

.856
.732
.661

(N=93)
(N=77)
(N=34)

Felony Arrests
6 months
12 months
18 months

.917
.820
.727

(N=297)
(N=238)
(N= 89)

.934
.870
.780

(N=245)
(N=213)
(N= 70)

.928
.880
.861

(N=100)
(N= 88)
(N= 38)

Parole Revocations
6 months
12 months
18 months

.902
.749
.632

(N=280)
(N=215)
(N= 60)

.906
.790
.714

(N=235)
(N=190)
(N= 50)

.926
.868
.794

(N=97)
(N=84)
(N=27)

Table 8 shows the cumulative proportion surviving at six, 12, and 18 months after release for
each group. The cumulative proportion surviving “estimates the probability of surviving to [a
specific time interval], or longer, without occurrence of the [failure] event”, in this case, arrest or
revocation.83 The cell at 6 months for Greenlight, for example, indicates that 82.8 percent of
Greenlight participants had not been arrested after six months. The cumulative survival provides
comparable measures for our three groups when there may be different times at risk. Our primary
82

Complete life-tables with statistics are shown in Appendix C. 

See Luke and Homan, “Time and Change”, Psychological Assessment, 1998: 365. Cumulative proportion 

surviving is given as S`(t) = S`(t-1) ((ner – nte) / ner), t > 0, where ner is the number exposed to risk for time period t,

and nte is the number of terminal events during time t.


83

28 Vera Institute of Justice

indicator is one year after release (12 months), although we also show the values for 18 months
as well. In addition, we also include in the table the N for the number of cases entering the risk
interval to provide a sense of sample sizes and numbers upon which the survival probabilities are
based. Although we have data for a period of almost two years, the Ns beyond 18 months are too
small to provide reliable estimates.
In almost every instance, we found that people who received in-facility re-entry
programming recidivated more often than those that did not, and that the more intense the
programming, the greater the recidivism. Focusing on the 12 month interval, for total arrests, the
data show that 65.9 percent of the GL participants remain arrest free compared to 75.8 percent of
the TSP group and 73.2 percent of the Upstate group. Thus, 10 percent more of the Greenlight
participants experienced an arrest compared to the TSP group after 12 months. Only a portion of
all these arrests involved felony charges, but the pattern remained the same. More GL
participants were arrested on felony charges than the comparison groups, though the difference
was not statistically significant. GL participants also experienced more parole revocations than
the TSP group, and the TSP group more than the Upstate group.
Figures 1, 2, and 3 provide a more detailed visual presentation of the survival curves for the
first year (see Figures C-1 through C-7 for the entire follow-up period, including graphs of the
hazard rates). The three groups are roughly similar in total arrests through the first eight weeks,
but then the GL group starts to experience more arrests (see Figure 1). From weeks 12 to 32, the
gap between the groups is stable, but for the rest of the year follow up the gap between GL and
the comparison groups expands. At one year after release, the percentage not experiencing an
arrest is labeled as the last data point. Figure 2 shows a similar trend with respect to felony
arrests. The Upstate and TSP curves are almost identical—the difference in total arrests between
these two groups is the result of misdemeanor arrests.
Figure 3 presents life-table data on parole revocations. Despite the different mechanism
involved in a revocation, the pattern for revocations mirrors that for arrests. Overall trends begin
to show more pronounced group differences after about 28 weeks, with the TSP group ultimately
falling between Upstate and GL in overall success.

Vera Institute of Justice 29

0.5

0.6

0.7

0.8

0.9

1.0

0

4

8

12

UP Survival

TSP Survival

GL Survival

16

20

28
Weeks at Risk

24

32

36

40

Figure 1. Cumulative survival (life table) for total arrests one year after release.


30 Vera Institute of Justice

Proportion Surviving

44

48

0.66

52

0.73

0.76

Proportion Surviving

0.5

0.6

0.7

0.8

0.9

1.0

0

4

8

12

UP Survival

16

TSP Survival

GL Survival

20

28
Weeks at Risk

24

32

36

40

Figure 2. Cumulative survival (life table) for felony arrests one year after release.

44

52

Vera Institute of Justice 31

48

0.82

0.87

0.88

0

4

32 Vera Institute of Justice

0.5

0.6

0.7

0.8

0.9

1.0

8

12

UP Survival

TSP Survival

GL Survival

16

20

28
Weeks at Risk

24

32

36

40

Figure 3. Cumulative survival (life table estimates) for parole revocations (reimprisonment)
one-year after release.

Proportion Surviving

44

48

52

0.75

0.79

0.87

Although Table 8 and Figures 1-3 are informative in illustrating overall group differences, it
is important to assess whether these inter-group trends are substantively different.84 Table 9
provides the results of Kaplan-Meier analysis for the group comparisons.85 We show mean
survival times, standard errors, and 95 percent confidence intervals (CI) in order to more
meaningfully interpret the differences.86 We used two statistics generated by the Kaplan-Meier
method, the log-rank and the Breslow. These tests compare the number of terminal events to the
expected number for each time interval in the study. The primary difference is the manner in
which weights are applied to each interval. For the log-rank, each interval is equally weighted.
Because the Breslow weights each interval by the number at risk during that interval, earlier
observations tend to be weighted more. The Breslow that is based on the square root of the
number at risk in a given interval. It is considered the more conservative statistic., and given the
larger Ns in the risk intervals in the earlier part of the study, the more appropriate measure.
Table 9. Comparison of study group survival times, Kaplan-Meier method (GL is the
primary comparison for statistics).
Mean
Standard
95%
Log-Rank
Breslow
Survival
Error
Confidence
Statistic
Statistic
Time (in
Interval
and
and
weeks)
Significance Significance
Total Arrests
Greenlight
TSP
Upstate

70.62
79.28 

78.54 


1.98


2.28
3.43

66.74, 74.49
74.81, 83.75
71.81, 85.27

4.75, .029
4.28, .039

5.45, .019
2.44, .118

Felony Arrests
Greenlight
TSP
Upstate

84.55
90.93 

91.17 


1.69


1.97
2.64

81.25, 87.86
87.06, 94.80
85.99, 96.34

1.99, .158
4.15, .042

2.12, .146
2.63, .105

Parole Revocations
Greenlight
TSP
Upstate

79.51
85.83 

88.84 


1.87


2.23
2.91

75.85, 83.16
81.46, 90.20
83.14, 94.54

1.07, .301
6.24, .013

.79, .375
4.87, .027

84

We do not present assessments of the differences between our two comparison groups. We discuss these
differences at a later point in the paper.
85
See Luke and Homan, “Time and Change”, Psychological Assessment, 1998: 367, for more detail.
86
A number of authors have pointed out the weaknesses inherent in simply relying upon significance tests without
more fully understanding the data, e.g., Maltz, Michael. “Deviating from the Mean: The Declining Significance of
Significance.” Journal of Research in Crime and Deliquency 31(4): 434-463.

Vera Institute of Justice

33

The Breslow statistic indicates a statistically significant difference between Greenlight and
TSP in total arrests and statistically significant differences between Greenlight and Upstate in
parole revocations. Confidence intervals between GL and TSP for total arrests do not overlap,
though there is a slight overlap between GL and Upstate. Slight overlap of the intervals occurs
between GL and TSP for felony arrests, with substantively more correspondence between GL
and Upstate. The trend is reversed for parole revocations, with only the slightest overlap between
GL and Upstate. In this case, the confidence intervals are consistent with the Breslow tests of
significance and reinforces our confidence in these results.
These data indicate significant differences at one year after release—a relatively short
follow-up period. The log-rank statistics suggest that if the current trends continue their present
course, a longer follow-up period will result in significant results between the GL and Upstate
groups on total arrests and felony arrests. This does not preclude the possibility, however, of
other significant differences emerging over a longer follow-up period for other comparisons.
Multivariate Analysis: Cox Proportional Hazards

The Kaplan-Meier tests and survival data do not control for other factors that might explain the
results. Cox regression is a standard method for modeling time-to-event data in the presence of
censored cases, while controlling for variables that might influence outcomes such as age at
release, age at first arrest, education, ethnicity, prior arrests, drug use, and offense type. Though
we did not find significant demographic or criminal history differences between the groups (see
Tables 1 and 2), we conducted these analyses to increase the precision of the statistical tests. The
results of these analyses are shown in Tables 10 through 15.
Each table shows three separate models: the first with standard demographic and criminal
history covariates, the second includes a dummy for the intervention, and the third includes a
four-category proxy variable for the intervention. Covariates were chosen based on widely
recognized predictors of adult reoffending and available data.87 Each model presents the
unstandardized coefficients (b) and the exponentiated coefficient (Exp (B); also known as the
hazard ratio). The sign of the b coefficient indicates the direction of the effect on the hazard rate;
a positive value means an increased hazard rate for the outcome of interest (for example, arrest),
and a negative value means a reduction in the hazard and increased survival times. A
straightforward interpretation of the Exp (B), of 1.05 for prior arrests (see Table 10) indicates
that a one unit increase in prior arrests (each additional arrest) results in a five percent increase in
the hazard rate, or the rate of failure [(1.05 – 1.00)*100=5%].88

87

E.g., Gendreau, Paul, Tracy Little, and Claire Goggin. “A meta-analysis of the predictors of adult offender
recidivism: What Works!” Criminology 34(4): 575-607. Our variables, however, are primarily what Gendreau, et
al., and others, term static variables. We have no indicators, for example, of mutable characteristics such as offender
motivation to change (dynamic predictors) that might explain the differences.
88
See Luke and Homan, “Time and Change”, Psychological Assessment, 1998:369, and Pampel, Fred C. Logistic
Regression. Thousand Oaks, CA: Sage Publications, 2000:36-37, for more detail on the interpretation of the
coefficients.

34

Vera Institute of Justice

For all six tables, prior arrests are the most consistent predictor of an increased probability of
failure. We specified multiple models using a variety of criminal history indicators, settling on
prior arrests as our primary indicator. Age at release shows a consistent significant negative
relationship with the various outcomes, indicating a higher probability of failure among younger
study participants. Age at first arrest is seldom related to any of the recidivism measures (see
Table 12 for an exception). Although this seems counterintuitive as the empirical evidence
suggests that the earlier the involvement with the criminal justice system, the more likely one
will engage in later criminal behavior, it seems likely that this is due to the limitations of the data
we have available to us. DCJS primarily provides age at first adult arrest—i.e., those arrested at
age 16 and older. Since many (or most) of these individuals are likely to have come into contact
with the criminal justice at much earlier ages, perhaps as young as 10-12 years of age, and we are
missing the bulk of juvenile arrests, our measure of age at first arrest is unlikely to be as
influential as if it was truly a measure of first arrest.
In terms of a program effect, Model 2 in Table 10 shows that the addition of control variables
does not mediate the negative effect of the intervention. Model 2 indicates that the intervention
group shows a 41 percent increase in the hazard for arrest over the TSP comparison group.
Model 3 presents an alternative view of the intervention effect. There is sufficient evidence
within the literature that services delivered by poorly trained or even ‘incompetent’ staff can
have detrimental effects for those receiving services. In an analysis of recidivism outcomes for
juvenile offenders, for example, juveniles participating in sessions with therapists rated as ‘not
competent’ or ‘borderline’, had significantly higher failure rates.89 In this case, individual case
managers associated with the GL intervention delivered all programming (except substance
abuse awareness), including the cognitive-behavioral and practical skills, and worked one-on-one
with the participant to develop a release plan. In essence, each participant was guided by a single
case manager through the program. In Model 3, we treat each case manager as a proxy for the
intervention program. We see that although all of the case managers are associated with poorer
outcomes than the TSP group, two case managers in particular appear to be associated with a
significant negative impact on intervention outcomes. Tests of the models in Table 10 using the
log likelihood values indicate that the addition of the intervention variables in Models 2 and 3
result in significant model improvement.90
Table 10. Cox Regression of Demographic/Criminal History Variables on Total Arrests,
Greenlight and TSP.
Variables
Model 1
Model 2
Model 3
b
Exp (B)
b
Exp (B)
b
Exp (B)

89

Barnoski, Robert. 2004. “Outcome Evaluation of Washington State’s Research-Based Programs for Juvenile
Offenders. Olympia, WA: Washington State Institute for Public Policy. Available at http://nicic.org/Library/019447.
90
Differencing the -2LL values gives a chi-square value with an associated degrees of freedom that can be used in
combination with a chi-square table to assess model improvement. See Pampel, Logistic Regression, Sage, 2000:4548.

Vera Institute of Justice

35

Age at release
Education
Race/Ethnicity
NH White/other
NH Black
Hispanic
Prior arrests
Primary offense
Robbery
Violent
Drugs
Property
Other
Substance abuse
None
Alcohol only
Drugs
Alcohol and drugs
Age at first arrest
Study Group
Case Manager
TSP
GLCaseMgr1
GLCaseMgr2
GLCaseMgr3
GLCaseMgr4
Model Chi-Square
-2 LL
Df

-.04
-.02

.96***
.98

-.04
-.02

.96***
.98

-.04
-.02

.96***
.98

--.10
-.06
.05

--1.11
.94
1.05***

--.05
-.12
.05

--1.05
.89
1.05***

--.02
-.12
.05

--1.02
.89
1.05***

---.22
.51
.39
-.50

--.80
1.66**
1.47*
.61^

---.27
.49
.40
-.51

--.77
1.63**
1.49*
.60^

---.26
.50
.41
-.49

--.77
1.65**
1.50*
.61^

--.54
.14
.16
-.02

--1.71*
1.15
1.17
.98

--.52
.15
.15
-.02
.34

--1.69*
1.16
1.16
.98
1.41**

--.53
.15
.15
-.02

--1.70*
1.16
1.16
.98

--.21
.16
.45
.52

--1.23
1.18
1.57*
1.68**

99.92***
2822.30
13

105.74***
2815.48
14

^p<.10; *p<.05; **p<.01; ***p<.001; b is the unstandardized coefficient. 


36

Vera Institute of Justice

110.68***
2812.00
17

Table 11. Cox Regression of Demographic/Criminal History Variables on Felony Arrests,
Greenlight and TSP.
Variables
b

Model 1
Exp (B)

b

Model 2
Exp (B)

Model 3
Exp (B)

b

Age at release
Education
Race/Ethnicity
NH White/other
NH Black
Hispanic
Prior arrests
Primary offense
Robbery
Violent
Drugs
Property
Other
Substance abuse
None
Alcohol only
Drugs
Alcohol and drugs
Age at first arrest
Study Group
Case Manager
TSP
GLCaseMgr1
GLCaseMgr2
GLCaseMgr3
GLCaseMgr4

-.04
.03

.96**
1.03

-.04
.03

.96**
1.03

-.04
-.03

.96**
1.04

--.06
.02
.04

--1.06
1.02
1.04***

--.02
-.03
.04

--1.02
.98
1.04***

--.04
.03
.04

--1.04
1.03
1.04***

---.20
.45
.49
-.31

--.82
1.57^
1.63^
.74

---.23
.45
.50
-.32

--.80
1.58^
1.64^
.72

---.22
.47
.53
-.31

--.80
1.61*
1.69*
.73

--.31
.36
.09
-.02

--1.36
1.42
1.09
.99

--.30
.36
.08
-.02
.26

--1.35
1.44^
1.08
.98
1.30

--.33
.38
.09
-.02

--1.40
1.46^
1.09
.99

--.42
-.01
.28
.33

--1.53^
.99
1.32
1.39

Model Chi-Square
-2 LL
Df

39.95***
1584.38
13

41.72***
1582.18
14

43.69***
1580.27
17

^p<.10; *p<.05; **p<.01; ***p<.001; b is the unstandardized coefficient.

Vera Institute of Justice

37

Table 12. Cox Regression of Demographic/Criminal History Variables on Parole
Revocations, Greenlight and TSP.
Variables
b

Model 1
Exp (B)

b

Model 2
Exp (B)

b

Model 3
Exp (B)

Age at release
Education
Race/Ethnicity
NH White/other
NH Black
Hispanic
Prior arrests
Primary offense
Robbery
Violent
Drugs
Property
Other
Substance abuse
None
Alcohol only
Drugs
Alcohol and drugs
Age at first arrest
Study Group
Case Manager
TSP
GLCaseMgr1
GLCaseMgr2
GLCaseMgr3
GLCaseMgr4

-.02
-.03

.98*
.97

-.02
-.03

.98*
.97

-.02
-.02

.98*
.98

--.00
-.42
.04

--1.00
.66
1.04***

---.03
-.44
.04

--.98
.64
1.04***

---.12
-.53
.04

--.89
.59
1.04***

---.14
.24
.34
.35

--.87
1.27
1.41
1.42

---.17
.24
.35
.34

--.85
1.27
1.42
1.40

---.11
.22
.35
.39

--.89
1.24
1.43
1.48

--.17
.38
.32
-.01

--1.18
1.46*
1.38
.99

--.16
.39
.33
-.01
.21

--1.17
1.48*
1.39
.99
1.23

--.17
.41
.31
-.01

--1.18
1.51*
1.37
.99

---.18
.25
.03
.54

--.83
1.29
1.03
1.72**

Model Chi-Square
-2 LL
Df

48.81***
1978.51
13

50.08***
1976.78
14

^p<.10; *p<.05; **p<.01; ***p<.001; b is the unstandardized coefficient. 


38

Vera Institute of Justice

58.47***
1969.56
17

Table 13. Cox Regression of Demographic/Criminal History Variables on Total Arrests,
Greenlight and Upstate (N=451).
Variables
Model 1
Model 2
Model 3
b
Exp (B)
b
Exp (B)
b
Exp (B)
Age at release
Education
Race/Ethnicity
NH White/other
NH Black
Hispanic
Prior arrests
Primary offense
Robbery
Violent
Drugs
Property
Other
Substance abuse
None
Alcohol only
Drugs
Alcohol and drugs
Age at first arrest
Study Group
Case Manager
Upstate
GLCaseMgr1
GLCaseMgr2
GLCaseMgr3
GLCaseMgr4

-.03
-.03

.97*
.96

-.03
-.04

.97**
.96

-.03
-.04

.97**
.96

---.33
-.46
.05

--.72
.63
1.05***

---.35
-.48
.05

--.70
.62
1.05***

---.40
-.50
.05

--.67
.61
1.05***

---.31
.36
.42
-.15

--.73
1.43^
1.51^
.86

---.32
.32
.40
-.16

--.72
1.37
1.49^
.85

---.30
.35
.43
-.12

--.74
1.42
1.53^
.89

--.27
.11
.28
-.06

--1.30
1.11
1.32
.94**

--.34
.16
.29
-.06
.37

--1.40
1.17
1.33
.94**
1.45^

--.35
.16
.27
-.06

--1.42
1.17
1.31
.94**

--.22
.21
.46
.56

--1.25
1.23
1.59^
1.75*

Model Chi-Square
-2 LL
Df

106.11***
2003.35
13

107.36***
1999.59
14

111.94***
1996.01
17

^p<.10; *p<.05; **p<.01; ***p<.001; b is the unstandardized coefficient.

Vera Institute of Justice

39

Table 14. Cox Regression of Demographic/Criminal History Variables on Felony Arrests,
Greenlight and Upstate (N=451).
Variables
Model 1
Model 2
Model 3
b
Exp (B)
b
Exp (B)
b
Exp (B)
Age at release
Education
Race/Ethnicity
NH White/other
NH Black
Hispanic
Prior arrests
Primary offense
Robbery
Violent
Drugs
Property
Other
Substance abuse
None
Alcohol only
Drugs
Alcohol and drugs
Age at first arrest
Study Group
Case Manager
Upstate
GLCaseMgr1
GLCaseMgr2
GLCaseMgr3
GLCaseMgr4

-.01
-.04

.99
.96

-.02
-.03

.98
.97

-.02
-.03

.98
.97

---.48
-.68
.02

--.62
.51^
1.02*

---.54
-.74
.02

--.58
.48^
1.02**

---.53
-.69
.02

--.59
.50^
1.03**

---.24
.35
.63
.04

--.79
1.43
1.88*
1.04

---.25
.32
.60
.06

--.78
1.37
1.82^
1.06

---.23
.34
.62
.09

--.79
1.40
1.85*
1.10

--.14
.41
.27
-.05

--1.15
1.50
1.31
.95^

--.23
.47
.27
-.05
.58

--1.26
1.61^
1.31
.95^
1.79*

--.27
.49
.27
-.05

--1.32
1.63^
1.31
.95^

--.71
.37
.59
.68

--2.04*
1.45
1.80^
1.98*

Model Chi-Square
-2 LL
Df

33.08**
1102.42
13

36.31***
1097.45
14

^p<.10; *p<.05; **p<.01; ***p<.001; b is the unstandardized coefficient. 


40

Vera Institute of Justice

37.75**
1096.04
17

Table 15. Cox Regression of Demographic/Criminal History Variables on Parole
Revocations, Greenlight and Upstate (N=448).
Variables
Model 1
Model 2
Model 3
b
Exp (B)
b
Exp (B)
b
Exp (B)
Age at release
Education
Race/Ethnicity
NH White/other
NH Black
Hispanic
Prior arrests
Primary offense
Robbery
Violent
Drugs
Property
Other
Substance abuse
None
Alcohol only
Drugs
Alcohol and drugs
Age at first arrest
Study Group
Case Manager
Upstate
GLCaseMgr1
GLCaseMgr2
GLCaseMgr3
GLCaseMgr4

-.03
-.09

.97*
.92

-.04
-.08

.97**
.93

-.03
-.06

.97*
.94

--.24
-.24
.04

--1.27
.79
1.04***

--.13
-.35
.04

--1.14
.71
1.04***

---.11
-.47
.04

--.99
.62
1.04***

---.18
.15
.52
.24

--.83
1.17
1.69^
1.28

---.15
.14
.53
.30

--.86
1.15
1.70^
1.35

---.07
.13
.55
.38

--.93
1.14
1.73^
1.46

---.20
.15
.26
-.003

--.82
1.16
1.30
1.00

---.08
.25
.31
-.00
.67

--.95
1.28
1.37
1.00
1.96**

---.06
.28
.29
-.00

--.94
1.32
1.34
1.00

---.27
.66
.43
.96

--1.31
1.94*
1.53
2.62***

Model Chi-Square
-2 LL
Df

49.17***
1323.21
13

53.22***
1315.41
14

60.92***
1309.78
17

^p<.10; *p<.05; **p<.01; ***p<.001; b is the unstandardized coefficient.

In addition, in the analysis of parole revocations we also tested several models (data not
presented here) that included a time-dependent covariate for arrest after release.91 In other words,
we tested the assumption that time to arrest was predictive of time to parole revocation. Using a
91

For more detail on the use of time-dependent covariates, see Allison, Paul D. Event History Analysis. Thousand
Oaks, CA: Sage Publications, 1984; Luke and Homan, “Time and Change”, Psychological Assessment, 1998;
Yamaguchi, Event History Analysis, 1991. for more detail on the use of time-dependent covariates.

Vera Institute of Justice

41

simple dichotomous variable for arrest is inappropriate because arrest is time-dependent. That is,
the dichotomous variable does not reflect the differing times at risk—those at risk longer have
higher cumulative probabilities of an arrest. These analyses indicate that parole revocations are
strongly related to arrests in the models testing the Greenlight and TSP comparisons and only
slightly significant in the models for the Greenlight and Upstate comparisons. In general, the
addition of the arrest variable as an additional control does not impact our findings other than to
suggest its impact on revocations. We also modeled several analyses of parole revocations
without an associated arrest (data not presented).92 In these analyses, our dependent outcome was
parole revocation which did not have an associated arrest. In all models tested, there were no
significant differences between the Greenlight and two comparison groups.
There are at several other important points to make. Other factors might contribute to our
finding that the program is associated with an increased probability of failure. One example is
the expanding literature on the impact of neighborhood of residence for involvement in criminal
offending.93 Although we attempted to gather information on post-release residence, too few
post-release interviews were conducted to have sufficient information on residence after release,
and we were unable to collect the data from the Division of Parole’s paper records. The 26 parole
field office boundaries in New York metro area offer a very rough proxy, but including this
variable in the Cox Regression models for total arrests, felony arrests, and parole revocations
added little to the overall model fit (the X2) and did not mediate the influence of study group in
any way. It is possible that neighborhood effects might account for our findings, but by the same
token, it is possible that including neighborhood effects might make group differences starker.
Finally, we also conducted several analyses examining those haphazardly versus those
randomly assigned for the GL and TSP groups. The results of those analyses are shown in
Appendix C, Figures C-8 and C-9 for total arrests (we do not show Cox Regressions or other
outcomes). In sum, GL performs worse than TSP regardless of the assignment protocol. For
those who were haphazardly assigned, failure rates after one year were slightly better than for
those who were randomly assigned after the program had been operating for almost five months
[Cumulative survival probabilities—life-table estimates, one-year outcomes: Haphazard
Assignment (GL=70%; TSP=78%); Random Assignment (GL=61%; TSP=74%)]. It is both
puzzling and significant that survival probabilities decreased for those who participated in the
intervention program after it had already been operating for a period of time.

92

See Appendix C, Figure C-7, and Tables C-15 through C-17, for base comparisons. 

See, for example, Bursik, Robert J., and Harold G. Grasmick. Neighborhoods and Crime: The Dimensions of

Effective Community Control. San Francisco, CA: Lexington Books. 


93

42

Vera Institute of Justice

Discussion and Conclusion
The results we obtained were unexpected and puzzling. We have evidence that Greenlight
participants found some aspects of their pre-release programming more helpful than TSP
participants and that Greenlight participants had more referrals, made more service contacts, and
were more knowledgeable about conditions of parole than TSP participants. Yet, on both
measures of new arrests and revocations, the pattern was the same: Greenlight participants did
worse than TSP participants who, in turn, did worse than Upstate parolees. The positive
intermediate impacts of the program did not translate to reductions in aggregate recidivism. We
offer several theories that might explain these results, and discuss them in light of the limited
data we have.
Flaws in assignment processes often explain results. It is possible that some form of
systematic bias was introduced in the assignment for which we do not have adequate measures.
If a larger proportion of offenders assigned to the TSP group, for example, were more motivated
to change their behavior, this might account for our results. One of the weaknesses of our
evaluation is the lack of measures for what have been termed "dynamic" risk factors, i.e., those
individual-level variables that are specific targets of change. Intervention staff discontinued the
LSI-R risk assessment instrument midway through the program and we were unable to
administer the same instrument in a timely fashion to members of the TSP control group. This
constrained our ability to assess inter-group measures on dynamic variables.
Based on what we know about the haphazard assignment process, we have no evidence of
systematic bias from this process. There are no statistically significant differences between the
groups in demographics or criminal histories.94 About half of the Greenlight and TSP and all
Upstate participants were assigned to treatments haphazardly, not randomly, but haphazardly
assigned GL participants were arrested more frequently. If haphazard assignment created a bias,
it was for Greenlight, not against.
Participants entered Greenlight or TSP based only on their processing sequence, and DOCS
records indicate that this was the case. Assignment to the Upstate group was supposed to be
based solely on transportation availability (or more specifically, lack of availability). However,
we do not have documentation that would verify that transportation availability was the sole
criterion used. If, for example, some potential parolees were sent to Queensboro because a
DOCS official believed they were especially in need of the pre-release programming, then the
Greenlight and TSP samples may have contained higher proportions of high-risk participants. A
more likely possibility is that individuals may not have been transferred to Queensboro due to
behavior problems—that is, they were not transferred because of institutional misconduct before
being moved to the pilot facility and were removed from the transfer list at the institution—this

94

We tested for several potential differences in criminal histories, including charge level and type of crime, but did
not detect any statistically significant differences.

Vera Institute of Justice

43

would presumably lead to a higher-risk comparison group, an outcome at odds with the data we
present.
Half of the Greenlight and TSP samples were randomly assigned, but there were minor
deviations (15 percent) from the assignment protocol. These deviations were primarily due to
breakdowns in communication about available bed space, leading to the assignment of
individuals to the intervention when beds were not available, or too many available beds on the
intervention side that the institution was compelled to fill. While not the laboratory ideal, this
assignment process is common in criminal justice field tests.
Neighborhood effects might also explain inter-group differences. Greenlight participants
might have returned to more criminogenic neighborhoods than TSP or Upstate parolees, and
either through greater disadvantage, more criminal opportunities or a greater law enforcement
presence, this influenced recidivism rates. But we have no evidence of differential assignment
and our analysis of parole offices as proxies for neighborhood did not show a bias.
Other explanations of inter-group differences are based on the hypothesis that something
about the treatments caused the observed results. For example, the “treatment” for Greenlight
participants involved not only pre-release programming, but also assignment to one of a small
group of parole officers chosen for the study. Parolees in the other groups were assigned to a
broader pool of parole officers. If the Greenlight parole officers were more likely to find out
about and report to the police crimes committed by their parolees, that might explain the
observed differences in arrests. We think this explanation unlikely. If the Greenlight parole
officers were “more strict,” then the Greenlight participants should have more technical parole
violations (independent of arrest). There are, however, no differences between the groups on this
measure.
The last explanation is the most alarming: the pre-release programming itself resulted in
higher rates of new crimes. This outcome would be contrary to the current consensus that
cognitive-behavioral programming leads to at least a modest reduction in recidivism. While the
bulk of the research literature supports this view of cognitive-behavioral programs, there have
been reports of some negative outcomes, and such effects are not unheard of in the literature on
crime prevention programs that intercede to try to prevent delinquent behavior among youth or
intimate partner abuse among adults.95
Given the intensive nature of the program over a relatively short duration, the intervention
may have provided a bonding opportunity among the participants that led to higher recidivism
rates. Living together and isolated from other inmates and attending classes and other forms of
programming all day long every day for eight weeks may provide the participants an opportunity
to interact and share experiences that ultimately lead to greater criminal offending. This is a

95

See, for example, Pullen, Suzanne. Evaluation of the Reasoning and Rehabilitation Cognitive Skills Development
Program as Implemented in Juvenile ISP in Colorado. Denver, CO: Colorado Division of Criminal Justice, 1996;
see Porporino and Fabiano, Reasoning and Rehabilitation, T3 Associates, esp. Appendix D; McCord, Joan. “Cures
That Harm: Unanticipated Outcomes of Crime Prevention Programs.” Annals of the American Academy of Political
and Social Science 587(May):16-30, 2003; also see Davis and Medina, 2000; Harrell, 1996.

44

Vera Institute of Justice

weak explanation: there are other intensive programs (for example, boot camps) and isolated
prison groups (for example, therapeutic communities) that do not produce similar outcomes.
Alternatively, the Greenlight program may have raised expectations to an unrealistic level
about what life would be like on the outside. If Greenlight participants came to believe that
finding jobs and housing would be relatively easy, the hard reality of returning to life on the
outside might have created frustration and anger. These responses could lead to more criminal
behavior. We have no data to support or refute this theory, but it does not fit well with
Greenlight’s programming—which sought to prepare people by examining the many challenges
people leaving prison face. For example, the family counseling program explicitly sought to
destroy the common fantasy that returning to family life would be an easy and fulfilling.
A more likely explanation may exist in the way that Greenlight implemented the cognitive
behavioral program model. The demonstration program included the Reasoning and
Rehabilitation cognitive-behavioral program, but structured the program in a fundamentally
different way than in prior implementations and evaluations. Although developed with the
program designers, the overall structure of the cognitive-behavioral program departed in many
ways from the “optimal” program structure.96 For example, developers of the program
recommend class sizes for the cognitive skills component of eight to 13; Greenlight classes
generally held 26 participants, double to triple what is generally considered ideal.97 Given that
the cognitive-behavioral component is dependent on an active and engaged learning environment
with small class sizes, increasing the class size may be problematic for program delivery. The
program was also restructured to fit the sequence into a two-month rather than four- to six-month
design. These changes may have led to different program effects.
The issue may be less a matter of program design than a mismatch between the program and
the targeted offender population. Although we do not have standardized measures of recidivism
risk for the people in our study, based on criminal histories, this population fits a "high-risk"
profile. A “conditional release” status indicates either poor behavior that led to denial of parole
release or a level of offense seriousness that justified sentencing under New York State’s
determinate sentencing statutes. In either case, this type of intervention may have different
effects for a high-risk group than for a low- or moderate-risk population
Finally, implementation issues might explain the results. Our resources allowed only for an
outcome evaluation; we conducted relatively little process evaluation research. Furthermore,
program staff did not report any glaring implementation problems.98 We have little ability to
comment on program integrity, or the actual quality of implementation. Our finding that specific
case managers are associated with many of the negative program results is a concern. Whether
more or better training might have resulted in better program delivery and subsequently better
outcomes remains unknown. Although two of the case managers were primarily associated with
96

See Fabiano and Ross, Reasoning and Rehabilitation Program Manual.

See Appendix A, Cognitive-Behavioral Sessions, for more detail on how the cognitive-behavioral model 

implemented at Queensboro differed from what Fabiano and Ross describe as the “optimal” program structure. 

98
These issues are addressed in more detail in Appendix A.
97

Vera Institute of Justice

45

significant negative outcomes compared to the TSP group, results associated with the other case
managers were still poor in comparison. It may also mean that the characteristics of case
managers (Reasoning and Rehabilitation “tutors”) as suggested by the program designers (for
example, above average verbal and interpersonal skills, humility, enthusiasm, sensitivity to
group dynamics, etc.) are extremely important as part of program implementation and program
fidelity. It is also worth noting that although we did not find any significant differences between
groups, there may have been some sort of assignment bias within the program under our
assignment protocols, that led to some case managers being assigned more difficult or
criminogenic participants, and this would account for the differences in recidivism by case
manager.
Implementation issues could be the result what might be termed “start-up” problems. As
noted previously, the program was not given a chance to iron out the problems that are inevitably
associated with the start-up of any new program. These specifically might operate through
confusion (among program staff and participants) concerning objectives and procedures,
resistance to change, staff being less competent in delivering a new program, staff
communicating their dissatisfaction (deliberately or inadvertently) to participants, or stress on
resources associated with “novice implementation” among others.”99 We found, however, that
intervention participants who went through the program after it had been operating for five
months (those randomly assigned vs. haphazardly assigned) had poorer one-year arrest
outcomes.
We also want to point out (as one reviewer reminded us) that although we have indicated the
relevance of certain omitted variables as limitations of our analysis (e.g., dynamic risk factors,
neighborhood effects), the explanatory power of these variables decreases in relation to the
degree of statistical association with the variables we have included in our analyses. We know
that dynamic risk factors are independent of static variables to some degree, but it is unclear how
much. Those who may have had more extensive criminal histories, for example, may have been
more likely to come from neighborhoods of greater disadvantage, or to express more anti-social
attitudes and beliefs.
We do not have definitive evidence to explain our results. Given the available evidence, we
believe that the causes are more likely found in the program design or implementation rather
than in the assignment process: Program participants with the most extensive pre-release
programming (Greenlight) had worse outcomes than people who had some pre-release
programming (TSP) who, in turn had worse outcomes than people who had no pre-release
programming. In drug research, this is called a dosage-response relationship. When the response
increases monotonically with higher levels of a drug, it is considered evidence that it is indeed a
drug and not other causes that is having the observed effect on outcomes. .

99

Frederick, Bruce. Personal communication. Frederick’s point is well-taken about conducting impact evaluations
during the initial stages, and worth re-emphasizing.

46

Vera Institute of Justice

We have given a great deal of thought to the implications of our findings. Although there are
numerous lessons to take away from both our evaluation and the program we examined, we
focus our remaining discussion on those lessons that seem most important.
First, our study should not be considered an indictment of rehabilitative programming. In
contrast to the famed “nothing works” philosophy that dominated corrections during the 1970s
and 1980s, recent research findings are robust in their conclusion that rehabilitation can and
should be a cornerstone of sound correctional policy. Much of the current work on rehabilitation
continues to help define the parameters of effective treatment. Although we reiterate the point
that there is no single program that appears to be appropriate for all offenders, there is substantial
evidence that appropriate treatment does work when targeted to offenders with specific needs.
Second, the lack of a thorough implementation (or process) evaluation constrains our ability
to identify programmatic issues that might be tied to the negative findings we report. We are
certainly not the first to argue the importance of implementation evaluations to understand the
results of outcome evaluations, but the limitations of the work reported here highlight again the
necessity of including an implementation study with any impact evaluation.
At the outset of this evaluation, no one suspected that the results would turn out as they did.
At this point, all we can do is speculate on the reason for the pattern of results that we obtained.
But at a time when dollars are being poured into reentry programs, we believe that it is important
that these findings be followed up with other research that can support or refute them and, if they
are supported, offer more detailed reasons for the puzzling results. As noted criminologist Joan
McCord stated, potential adverse consequences of interventions in criminal justice ought to be
treated as seriously as they are in the pharmaceutical field: “Social programs deserve to be
treated as serious attempts at intervention, with possibly toxic effects, so that a science of
intervention can prosper.”100

100

McCord, “Cures That Harm”, 2003:17.

Vera Institute of Justice

47

Appendix A. Project Greenlight Program Description
Introduction

The following program description is largely drawn directly from the operational guidelines for
Project Greenlight and reflects changes in the program after the first six months of operation.
The original operational guidelines for the program were jointly developed by the Vera Institute,
the New York State Division of Parole, and the New York State Department of Correctional
Services. All three agencies approved the guidelines to implement the program. The following
description was provided by the Greenlight operations director and modified but not developed
by the research staff. Its inclusion is intended to provide a descriptive overview of the project’s
structure and operation. This should be of significant interest to others interested in
implementing similar programs and the difficulties inherent to that process.
Our evaluation is primarily an outcome evaluation although we made every effort to observe
the program while in operation. Research staff visited the program at different intervals,
observed different classes, and spoke with program staff. Limited funding prevented a full-blown
process evaluation, but we spoke with DOCS personnel and Parole officers, program staff,
contracted vendors, and community service providers and observed both the DOCS Transitional
Services Program and the Greenlight Program in operation when time permitted. In some cases,
the discussions with program staff and external agencies working with the program took the form
of more formal interviews.
In addition, the Greenlight Program maintained a database on site (the PMA—Program
Management Application) that collected information on participation in different types of
classes, how often participants attended classes or other sessions, and other types of information
such as counselor ratings of participant engagement or performance. Wherever possible, we use
this data and our observations to assess program implementation in a strict sense (i.e., whether
classes were implemented as stated in Exhibit A-1 and A-2) and whether or not there was
compliance on the part of the participants.
In general, our observations and interviews suggest that the description of the program
structure is generally consistent with actual implementation. There were, however, some
indications that in the short time frame that the program operated, a number of obstacles were
encountered. Some of these were successfully negotiated, but it is unclear that all were resolved.
Given that the project was terminated after only 10 months of operation (rather than the full three
years it was scheduled to operate), it is not surprising that program staff were unable to
successfully resolve all program implementation issues. There are, however, other issues
inherent to program implementation independent of program structure. Perhaps the most
important of these is the issue of program integrity. The recognition that quality of program
delivery (program integrity or fidelity) may be a significant source of variation in program
outcomes has become prominent. Thus, it is important to note that our limited process work does
not allow us to adequately assess program integrity. Our research observations of program
implementation are addressed as part of the following program description.

48

Vera Institute of Justice

As noted above, these limited observations do not allow us to accurately assess how well the
program was delivered in terms of its fundamental integrity. Although we can often indicate
whether or not the program structure was implemented as planned (e.g., whether participants
attended classes, whether or not classes were held as intended), we were unable to complete the
detailed process work to allow us to assess how well the program was delivered. Here, however,
is an example of at least one prior evaluation of the same cognitive program applied in a
different program context (and structured differently):
“A review of video-taped sessions of program delivery revealed that the program
delivered by JISP (Juvenile Intensive Supervision Probation) coaches barely met
the standards of R&R [Reasoning and Rehabilitation] program developers.
Findings from this review indicate that, while the content of the program was
delivered, the process of actually imparting knowledge and skills to the offenders
barely occurred. Several shortcomings were noted by the video tape evaluator,
such as not linking crucial information, lack of lesson preparation, inability to
explain concepts or explaining concepts incorrectly, inappropriate combination of
program sessions, and failure to make the program relevant to adolescents.”101
In the evaluation of the juvenile ISP, two scales were used to measure pre- and post-intervention
attitude and cognitive change. One scale revealed small, generally positive improvements in
eight of nine skill areas for intervention participants and three of nine areas for controls. In a
somewhat more comprehensive battery, respondents registered a radically different pattern of
change; intervention participants scored worse on 14 of 14 composite measures (statistically
significant in six of 14) and grew worse for control participants in 12 of 14 for controls
(statistically significant for 10 of the composite scores).102 We highlight this issue to illustrate
that although we deal with implementation on a very basic level, it is by no means more than
that. We do not have the detailed information to make qualitative assessments of program
integrity.
Program Staff

The staff was to operate as a multidisciplinary team to monitor inmates’ progress and share
expertise, information, and ideas between team members. Staffing for the project and each team
member’s specific duties were as follows:
ƒ	 Correction Counselors (2 positions): DOCS counselors ran the cognitive-behavioral
group sessions, facilitated groups on practical skills and job readiness, held individual

101

Pullen, Suzanne. Evaluation of the Reasoning and Rehabilitation Cognitive Skills Development Program as
Implemented in Juvenile ISP in Colorado. NIJ Grant No. 93-IJ-CX-K017. Denver, CO: Colorado Division of
Criminal Justice, 1996. The video tape evaluator in this case was one of the developers of the R&R program.
102
We note that these data, like many other evaluations, are based on small sample sizes.

Vera Institute of Justice

49

ƒ

ƒ

ƒ

ƒ
ƒ
ƒ

ƒ

ƒ

50

case management meetings every two weeks with a set caseload of 26 inmates each, and
participated in team meetings.
Institutional Parole Officers (IPOs—2 positions): in the course of performing their other
duties, IPOs ran cognitive-behavioral sessions, facilitated groups on practical skills and
job readiness, held individual case management meetings every two weeks with the same
size caseload as DOCS counselors and participated in weekly team meetings.
1 Alcohol and Substance Abuse Treatment (ASAT) team (1 counselor and 1 program
assistant—part time, approximately eight hours a week): both members of the ASAT
team were Queensboro correctional staff. The counselor was a Certified Alcohol and
Substance Abuse Counselor (CASAC). Both ran substance abuse groups.
DOCS Correction Officers assigned to the housing unit: the officers were responsible for
maintaining a secure environment, contributing to team meetings based on their
observations and interactions with inmates, and working with other team members to
forge the most appropriate policies and responses to inmate behavior.
A Family Re-integration Specialist (part time): the specialist, an MSW, facilitated groups
for inmates and participating families or other support providers.
A Family Counselor: the counselor recruited participants and their family members to
participate in the family reintegration program and co-facilitated the sessions.
A Community Coordinator: the coordinator was to build a network of community groups
to serve the variety of Greenlight participants’ needs. This included preparing the groups’
representatives to come into Queensboro to discuss their services with participants and
monitoring those groups’ performance. The difference between the community
coordinator and existing DOCS and parole positions of the same nature was that the
coordinator had the time, the flexible hours, and the mobility to go into the community to
meet with groups, assess their operations, and report back. The coordinator’s work was
front-loaded. She worked full-time at the beginning of the demonstration to build a
network of reliable providers that would come into the facility, and was to finish this task
before the end of the demonstration when the the network would be turned over to DOCS
and parole.
A Job Developer: the job developer was contracted from a local employment agency that
worked with difficult-to-place populations. The job developer met individually with
inmates, worked towards getting them interviews and jobs, ran the job resource room,
and participated in team meetings.
An operations director: this position was to exist only during the demonstration project
phase. The operations director was responsible for the day-to-day operation of the
program, including scheduling classes, assigning participants to caseloads and classes,
representing Project Greenlight at program committees where inmates were assigned to
the program, and acting as the point of contact for day-to-day operational matters with
Queensboro’s deputy superintendent for programs, senior counselors, and the senior
parole officer. The operations director also worked with parole and DOCS staff to revise

Vera Institute of Justice

curricula and program delivery throughout the demonstration to arrive at a version ready
to go to scale. Over the course of the demonstration, the operations director was to pass
on these functions to existing DOCS and parole supervisory staff.
ƒ	 A Project Director: this position was to exist only during the demonstration phase. The
project director was responsible for the overall operation of the program. The director
acted as a liaison to DOCS and Parole senior management, and was responsible for
funding and contractual issues. The director monitored, analyzed, and evaluated the
achievement of program objectives and standards. The director acted as primary liaison
to Vera research staff.
ƒ	 A Project Assistant: this position was to exist only during the demonstration phase. The
project assistant assisted the operations director and project director with their tasks, and
filled in for absent facilitators in groups.
ƒ	 Field Parole Officers who supervise the inmates once they were released: each parole
bureau in the five boroughs of New York City was to have a parole officer (26 field
offices and 26 field officers) to take all Project Greenlight participants in addition to
members of the two control groups (as well as other parolees to make up a standard size
caseload). They were to communicate with Greenlight parole officers and correction
counselors about the participants and their release plans before and after release and to
report on participants’ progress.
Staff Development. Vera staff, together with DOCS and parole training staff, presented the job
readiness and job development, relapse prevention, substance abuse awareness, and practical
skills curricula to the correction counselors and parole officers who facilitated these groups. The
developer of the cognitive-behavioral curriculum, the Cognitive Thinking Skills Program, trained
project staff in delivering the curriculum and was to provide ongoing support and follow-up
training. Part of this training included a train-the-trainer component so that project staff could
train other agency staff in the curriculum and techniques. Staff had ongoing sessions with the
community coordinator to learn about new community resources. The correction officers were
introduced to the goals and philosophy of the project. Refresher sessions on the project’s
curriculum for existing staff members and new sessions for new staff occurred throughout the
demonstration.
Staff Supervision. Correction officers, counselors, and institutional parole officers, who
delivered much of the programming, already functioned within an established management
structure in their agencies. While they worked with and were supported by the operations
director and project director, who were Vera employees, they were not supervised by program
staff. They continued to report to their senior correction counselor and senior parole officer at
Queensboro.
Given this, DOCS, Parole and Vera developed a process to deal with operational and
performance issues affecting the project. The program facilitator and Vera’s project director

Vera Institute of Justice

51

raised any concerns and suggestions as they arose through the DOCS and Parole chain of
command. At the same time, Parole and DOCS line staff were encouraged to bring any concerns
through the same chain of command. In addition, the Queensboro deputy superintendent for
programs and a Parole designate met frequently with the operations director and/or the project
director to discuss project operations and resolve any problems.

Program Operations and Elements
Program Participation

Mandating Program Participation. Inmates selected for the program were required to
participate to keep their conditional release date. This was in keeping with DOCS policy
regarding conditional release: inmates are required to participate in any assigned programming to
earn such release.103 Refusal and/or unsatisfactory participation or disciplinary sanctions led to a
review by the DOCS Time Allowance Committee, in accordance with existing DOCS rules.
Transfer to Queensboro, Orientation, and Assessment. Intervention participants were transferred
to the north side of the sixth floor of the Queensboro Correctional Facility in Long Island City
(Queens), New York, arriving approximately 60 to 90 days before their scheduled conditional
release date. Inmates arrived Monday through Friday and began participating in a one-week
Greenlight orientation on the following Monday. The orientation introduced participants to each
of the program elements and to the staff.
Assessments and the Release Plan. Program staff long maintained that one of the key
differentiating aspects of the Greenlight program was the construction of the detailed release
plan which provided a step-by-step outline of what participants needed to do after they were
released. During the first weeks of the project each inmate was individually assessed. The initial
protocol called for the correction counselor or facility parole officer to whom the inmate is
assigned (the “case manager”) to administer a Level of Service Inventory – Revised (LSI-R)
assessment.104 Program staff began the program using this instrument along with an internally
developed intake instrument and the Socrates 8D (a substance abuse assessment instrument).
After the first five months, program staff stopped administering the LSI-R because it was
103

It is worth noting here that although DOCS policy requires inmates to participate in the program, they are not
required to participate in the research component. This is discussed in more detail in the section on methodology.
104
The LSI-R is a well-validated risk and needs assessment instrument (see Andrews, Don, and James Bonta. 2000.
Level of Service Inventory Revised—User’s Manual. Toronto, Ontario, Canada: Multi-Health Systems, for a review
of the relevant research on the LSI-R). Traditional risk assessment scales based solely on demographic information
and data from official records are often weakly to moderately correlated with recidivism measures. Scales that
incorporate psycho-social, social functioning, or other instrumental characteristics appear to have more validity as
risk assessment instruments. See, for example, Klein, Stephen P., and Michael N. Caggiano. The Prevalence,
Predictability, and Policy Implications of Recidivism. NIJ Grant # 83-BJ-CX-0011. Santa Monica, CA: Rand, 1986;
Andrews and Bonta, LSI-R: User’s Manual, 2000:1-3.

52

Vera Institute of Justice

deemed too time-consuming to administer. Program staff also felt that the LSI-R was redundant
because the information they needed to construct inmate release plans was being collected in the
other instruments and in materials in each inmate’s correction guidance folder, such as
certificates from completed programming, an employment profile, and notes from previous
counselors and institutional parole officers.
The failure to follow through with the LSI-R is a significant drawback to our study and the
demonstration project. Based on our conversations with state officials, one of the key interests in
designing the reentry project was the general utility of a well-validated assessment instrument in
both addressing offender needs and risk after release, especially given the strong evidence on the
relationship between criminogenic factors and criminal recidivism. There is substantial evidence
that the variables tapped by the LSI-R (“criminogenic” risk factors) are significantly related to
offender recidivism.105 Although the information for needs assessment was collected through
other intake measures, the LSI-R is a well validated needs/risk assessment instrument which
differentiates individuals not only on their need for services, but also on their risk for recidivism.
In order to use the LSI-R assessment information collected by program staff, research staff had
conducted initial discussions with DOCS to administer the LSI-R to control group inmates at
Queensboro when administering research consent forms. This pursuit was dropped when it was
discovered that program staff had dropped the LSI-R. It is unclear how well the intake
instruments substituted by program staff fulfill either of the risk or need assessment function.
And although data from the program-developed intake instruments were available to assess
intervention group risk in our quantitative analyses, we had no comparable measures for our
control groups.

Daily Skill-Building
The project’s programming had three principal elements: (1) providing daily skill-building in the
areas of employment, substance abuse prevention, and practical and cognitive skills; (2)
cultivating a network of community-based support; and (3) involving government agencies
beyond the criminal justice system. Exhibit A-1 (at the end of Appendix A) lays out how these
three elements (tracked along the left side of the exhibit) unfold over an inmate’s stay and also
provides a visual reference for the remaining discussion.
Every day and in structured settings, project staff offered inmates the information and tools
they needed to find and hold a job, avoid drug and alcohol relapse, and make good decisions in
all areas of their life after release. Over time, project staff adjusted the content and the approach
to providing daily programming. Because inmates arrived at and were discharged from
Queensboro every day, the project was structured to take rolling admissions. Programming was
designed to be non-sequential and classes were designed to accommodate people working at
different levels.
105

See Andrews and Bonta, Psychology of Criminal Conduct, 1998.

Vera Institute of Justice

53

Job Preparation and Development

The job preparation and development aspect of the program was delivered in two phases. First,
inmates completed a four-week, non-sequential course on job readiness, covering such areas as
job searching, preparing for and going on interviews, and workplace behavior. A new topic was
covered each week in a way that did not depend on prior information but that incorporated ideas
that some inmates had learned in previous weeks and others would be introduced to later.
Participants could move through the topics in any order.
After four weeks of group sessions, inmates entered the second phase of the employment
program: working individually with the project’s job developer to find work. The job developer
staffed a job resource room to which inmates reported in groups during the second half of their
stay at Queensboro. Over the course of four weeks in the room, every inmate was supposed to
have had several opportunities to speak individually with the job developer. The job developer
guided individuals in their own job search, as well as lined up interviews for participants
considered “job ready.” The job developer also invited employers to come into the facility to
present information about their company and the kind of people they were looking to hire. Vera
contracted with a job development organization in New York City that had experience
successfully working with ex-offenders. By contracting with the organization, Vera gained the
expertise of the job developer as well as the commitment, expertise, and supervision of the
contracted agency.
Research staff interviewed program staff and the job developer who was contracted to
provide employment help to the Greenlight program. As noted, the program contracted with an
outside for-profit vendor designed to provide no-cost employment services to ex-offenders,
substance abusers, and people with physical disabilities. The agency had experience doing
assessment and testing to help people make informed career choices, and they have relationships
with over one thousand mainstream companies. They largely recruited for food service,
maintenance, security, customer service, driving, retail, and clerical positions.
The contracting company was very interested in the Greenlight program because it was an
opportunity for them to expand their services and client base. As part of the developmental work,
the contractor further developed relationships with outside companies with whom the contractor
had worked previously. Inmate needs were also assessed in the employment-job development
process; in some cases, participants were referred to various job placement/training agencies,
such as Employment Initiative, Grant Associates, and/or the Osborne Association.106 Referrals
were generally based on the perceived degree of assistance needed by the participant. Someone
who needed more help and case management would be sent to Strive, an agency that provided
more of a classroom setting. If the participant had difficulty reading, he might have been referred
to VESID (New York State Office of Vocational and Educational Services for Individuals with
Disabilities), which provides vocational and educational training.
106

54

See, for example, http://www.osborneny.org/;

Vera Institute of Justice

Participants met with the job developer four times during their time in the Greenlight
Program. There were, therefore, four milestones to the program. First, when the inmates were
being oriented into Project Greenlight in the first week, they were given an introduction to the
job development program. It was presented as an optional part of the program, but the benefits of
participating in the employment component were explained in detail. Individuals who wished to
participate could volunteer and sign up to be contacted later.
The second meeting was a 15 to 30 minute individual assessment with each inmate. This
provided an opportunity to assess their needs and strengths and weaknesses in terms of skills.
The third session was scheduled in a group setting with anywhere from three to six inmates for a
one-hour session. During this time, the central focus was on the structure of job interviews; all
the participants engaged in mock interviews and role playing as if in a real interview. As part of
the process, the participants would critique themselves and each other and provide feedback on
their strengths and weaknesses.
A “resource room” was set up at the Queensboro facility for the final meeting. If the inmates
had skills with computers, for example, they would type up resumes. During this time, a
workshop was also conducted on how to use the newspaper to find a job and on how to keep a
job. As their exit date neared, participants met one-on-one with the job developer to go over their
resume, help them focus their resumes based on particular types of jobs, provide get referrals to
various job placement agencies, and try to set up actual interviews with various companies
before the inmate was released. Finally, if the inmate needed appropriate clothing for a job
interview, they were given referrals to places that give out free clothing, such as churches.
There were however, multiple issues with implementing the employment component. A
representative started working with Greenlight participants in March 2002, putting in 17 hours a
week in the prison every week from Monday through Thursday. Because there were 102
Greenlight program participants at any given time (once the program was at capacity), one of the
biggest initial obstacles was how to work with those 102 participants in only 17 hours per week.
This presented a logistical problem which was only partially resolved over time as the contract
representative began working with study participants.
Although the job developer had access to the participants for only four weeks, in many cases
he would only be able to meet with them in the two weeks before they were released due to
conflicts in their Greenlight class schedules. The general perception was that there was
insufficient time as it was and in the cases in which some participants did not get to the
employment component until two weeks before release, it was difficult to complete the needs
assessment and subsequent sessions in a useful manner. Although they attempted to work around
this particular problem, it continued as a major obstacle throughout the course of the program.
The job developer indicated that if he could have changed one thing about his role within
Project Greenlight, he would have done it full-time rather than part-time, because he feels that he
did not get enough time with the participants. He also indicated that available resources to teach
the inmates how to use the Internet would have been a big help (Internet access is prohibited to
inmates in institutional settings). Finally, although speakers from different referral agencies did

Vera Institute of Justice

55

come into the institution, he felt he was never able to offer them any substantive human
resources exposure.
The job developer did indicate, however, that the inmates were very receptive to him. He
thought this was primarily because he spoke truthfully to them, and they liked that he didn’t
work for Vera or DOCS; according to him, inmates expressed a lot of resentment towards both
organizations. He also felt that inmates were a little more respectful of him than the rest of the
Greenlight staff because they knew that they needed jobs but did not feel that sitting in a class
and learning life skills and cognitive-behavior would get them where they wanted. The developer
indicated that he worked hard to make inmates realize that what they were learning in class
would also be helpful. He put a lot of effort into developing a rapport with them, so as to build
inmates’ trust in him.
Working relationships within the institution, especially with Greenlight staff were generally
described as excellent. The job developer met with the Greenlight staff on a weekly basis to
discuss individual cases with them. This helped Greenlight staff determine which other kinds of
referrals to give to the inmates. Drawing primarily upon our interview with the job developer, the
main strength of the job development program and Greenlight overall was that all the people
involved had strong working relationships. The coordination between the family counselor, the
community coordinator, and the job developer was extremely functional in helping the inmates
have practical plans and expectations after release. These services allowed the inmates to deal
with the realities of securing a place to live, finding work, and establishing good ties with family
members.
Substance Abuse Readiness and Relapse Prevention

For inmates with a history of substance abuse—the vast majority—the project offered one of two
approaches depending on their score on the SOCRATES (an instrument that measures a person’s
willingness to acknowledge their substance abuse and their readiness to change). Those who did
not believe that they have a problem with drugs or alcohol took part in groups designed to
educate them about the effects of substance use and to prepare them to enter a treatment program
after release. Those who acknowledged their abuse and who expressed some willingness to
consider other options went to a relapse prevention group in which each person designed a
relapse prevention plan, and prepared to enter treatment upon release.107
Data from the Department of Correctional Services indicates that approximately 70 percent
of all study participants had some form of substance abuse problem upon admission to prison
(using the MAST I and II—Michigan Alcohol Screening Test), a figure consistent with national
averages. Post-release interviews with study participants suggest that inmates without a history
of substance use were required to participate in one of the two types of sessions. Study
participants often expressed a great deal of frustration at being required to sit through drug
107

Both programs are drawn from Terence Gorski et al., Relapse Prevention and the Substance Abusing Criminal
Offender. Washington, DC: Dept. of Health and Human Services, 1993.

56

Vera Institute of Justice

education when they had no history of substance use. This was a concern for research staff in
that the question was raised as to whether forcing individuals to participate in programming for
which they have no need (as opposed to coercive programming for which they do have a need)
can engender resistance to programming and lead to counterproductive outcomes.
Cognitive-Behavioral Sessions

An inmate’s daily life and work in the project was anchored by a mandatory program of
cognitive-behavioral sessions, using the Reasoning and Rehabilitation curriculum.108 An
evaluation of the program (under its title in Canada, the Cognitive Skills Program), the largest
study of any cognitive-behavioral treatment of offenders to date, suggests a reduction in
recidivism (a statistically significant six percent), depending strongly on the original
commitment offense. The program was found to have a stronger effect on violent, sex, and drug
offenders than on robbery and non-violent property offenders.109 In general, the evidence
suggests that cognitive skills programs have been the interventions most consistently associated
with reductions in recidivism overall.110
The Reasoning and Rehabilitation sessions, similar to other cognitive skills programs, teach
pro-social cognitive skills like effective problem solving, the ability to consider consequences,
and accepting delayed gratification.111 They also teach a number of social skills such as asking
for help, expressing complaints in a non-aggressive manner, and responding to failure. The
curriculum, led by a correction counselor or parole officer as facilitator, was meant to be very
interactive and involve exercises and role-plays designed to get participants to learn new skills
together.
The cognitive-behavioral component was one of the most interesting aspects of the
Greenlight model in that its implementation was radically different from the recommended, or
“optimal,” program structure that is typically evaluated.112 As noted, cognitive-behavioral skills
training has substantial empirical support in the literature for achieving small to moderate, but
robust, reductions in post-release recidivism measures. Suffice it to say that there is now general
agreement that such cognitive-behavioral programming, when effectively implemented, can
produce reductions in recidivism.

108

Ross, Robert R., and Elizabeth. Fabiano. Time to Think: A Cognitive Model of Delinquency Prevention and 

Offender Rehabilitation. Johnson City, TN: Institute of Social Sciences and Arts, 1985. 

109
Gaes, Gerald G., Timothy J. Flanagan, Laurence L. Motiuk, and Lynn Stewart. “Adult Correctional Treatment.” 

Pp. 361-426 in Crime and Justice: A Review of Research, Vol. 26., M. Tonry and J. Petersilia (eds.). Chicago, IL: 

University of Chicago Press, 1999:379. 

110
See, for example, Andrews 1995; Gaes, et al., 1999; Cullen and Gendreau, 2000; Palmer 1996; Gendreau 1996;
Losel 1995.
111
Other cognitive skills programs are readily available that may require less modification for the largely urban
populations imprisoned in the U.S.—see for example, the “Thinking for a Change” curricula, endorsed by the
National Institute of Corrections, and available in the public domain at http://www.nicic.org.
112
Although we review some of the basic elements here, see Porporino and Fabiano, Reasoning and Rehabilitation,
2000, for more detail on “optimal” program structure.

Vera Institute of Justice

57

The issue of program structure may have important implications for our findings in that we
are unaware of evaluations of the program in which it was implemented in a manner that
diverged greatly from that generally recommended by program developers. Virtually all
evaluations which show positive (or no) intervention effects are based on the recommended
implementation model. This is important in that it is generally conceded that we know little
about how well cognitive skills programs function to reduce recidivism when implemented other
than typically suggested.113 Given the importance of this issue, we briefly review how the
Greenlight model diverged from the “optimal” program model.
Reasoning and Rehabilitation Program Structure. Porporino and Fabiano (2000) devote several
sections of their program overview to program structure and implementation issues (Chapters 6,
7, and 8), focusing on program delivery, how training sessions are to be conducted and other
standard implementation issues such as sustaining program integrity, training of tutors, offender
selection, etc.. While it is important to note that they strongly emphasize that the program is
NOT “therapy,” much of their focus on delivery and implementation is consistent with
therapeutic techniques and depends on effective staff delivery and engagement. A quick
overview of these issues is warranted because implementation appears to be a fundamental, but
often ignored, key to program effectiveness, and may constitute the best reason for why we are
seeing the results we are seeing.114
1.	 Group Composition: Porporino and Fabiano indicate that “the ideal group size for
cognitive training is from six to about 10 participants” (p. 40) with eight as the optimal
number. We take it as axiomatic that program efficacy is based on an engaged group; it is
difficult to imagine, therefore, that the Greenlight program, which had 26 participants per
class, could keep all the group members engaged to the point that they would benefit
from the sessions. Small group sizes allow for the group members to interact and for the
instructor to keep those participants engaged in the process. Larger groups are likely to
dilute those effects. In addition, it is a standard understanding in social psychology that
group size can significantly impact group dynamics.
2.	 Program Delivery: Porporino and Fabiano indicate that the 38 sessions can be delivered
in varying sessions, but that the optimal delivery is 2-3 times per week, and that
delivering more frequently (as was the case with Greenlight, see Exhibit A-1), “…could
be expected to interfere with application or practice of skills as they are learned and could
create confusion, lapses of attention, or content overload for offenders (p. 40).”
3.	 Cognitive Skills “Tutors”: Crucial program implementation is tied to effective tutors
delivering the program. Although they emphasize the notion that the program is meant
for delivery by line staff and require no special technical qualifications or credentials,
they do note that staff should have specific abilities and personal characteristics (e.g.,
empathy, sensitivity, strong interpersonal skills, successful experience in managing
113
114

See Porporino and Fabiano, Reasoning and Rehabilitation, 2000.
Porporino and Fabiano, Reasoning and Rehabilitation, 2000; Gendreau, Goggin and Smith, 1999.

58	 Vera Institute of Justice

groups of poorly motivated individuals, humility, and a thorough understanding of the
social-cognitive model).
4.	 Training of “Tutors”: The program manual notes that “Tutors need to complete an
intensive training workshop before they can appreciate the many subtleties for optimal
style and methods of delivery.” Tutor preparation should consist of at least two weeks of
formal “Tutor certification instruction” which includes practicing delivery techniques in a
role-play classroom setting, feedback on delivery style, and understanding of program
objectives. Following training, new trainees should be monitored through videotaping of
some of their program methods.” Greenlight staff went through a far shorter 3-day “trainthe-trainer” session. This would raise questions about how effective the staff could be,
given the emphasis of the program developers on this point.
5.	 Targeting Services to Offenders: This is a significant issue from the research literature—
broad-based intervention programs have less of an impact than programs that target
specific services to offender needs. In this instance, the Greenlight program can be
viewed as a broad-based program since all program participants were delivered the same
services. Targeting services to the offenders who need them seems to produce the best
results. This is one of the key rationales underlying an assessment instrument in the early
stages. Offenders who are identified as “low-risk” are not likely to benefit from
services—they are already at low risk of reoffending and there seems to be little that can
be done to further reduce their risk. Although much of the literature has speculated over
the last decade that “high-risk” offenders may benefit most, the evidence is somewhat
mixed; there are some indications that “medium-risk” offenders may benefit most—
higher-risk offenders have been shown in some cases to be more resistant and need
longer programs. There is certainly evidence, however, that “high-risk” offenders can be
targets for programming; it is simply that short-term programs such as Greenlight may
not be of sufficient length to make a real impact for more serious offenders.115
Although Greenlight program staff worked with the developers of the Reasoning and
Rehabilitation program to adapt the intervention to the 8-week time-frame, the key point is that
there are significant deviations from the optimal program model that raises serious questions
about the efficacy of the program.
Practical Skills

Inmates participated in a series of discussions, role-plays, and other exercises designed to help
them learn basic skills needed to negotiate living in New York City—skills that were not
necessary or reinforced in prison.116 The sessions covered how to use public transportation,
115

There is very little evidence that short-term interventions can produce significant results to begin with; most of
the evidence with criminal justice populations (especially drug offenders) suggests that longer is better, but that
there may be declining returns after about 12 months (e.g., Lipsey, 1990 (cited in Andrews, 1995).
116
Skills-oriented programs, along with cognitive-behavioral and other multi-modal approaches, have been shown to
produce the best effects in reducing recidivism (Loesel, Friedrich. “The Efficacy of Correctional Treatment: A
Review and Synthesis of Meta-Evaluations.” Pp. 79-114 in J. McGuire (ed.), What Works: Reducing Reoffending.
New York, NY: John Wiley and Sons.

Vera Institute of Justice

59

particularly Metrocards, which will be new to some people; tips for financial planning and
budget shopping; time-management strategies; the benefits and procedures for setting up a bank
account and how to use an Automated Teller Machine (ATM) (for those who, based on their
crime of commitment, are not barred from doing so); and where to get legitimate sources of
emergency cash or non-cash assistance. While the information will not be new to every inmate, it
is important to remind everyone that mastery of such skills can help them manage difficult
situations that otherwise could lead to a crisis and possibly re-incarceration.

Cultivating a Network of Community Support
Taking advantage of Queensboro’s relative proximity to the neighborhoods where people will
live after release, the project was designed to help inmates connect with the individuals best able
to guide and support them in the community: parole officers, family members, and
representatives of appropriate service agencies. Establishing these relationships in advance eases
several challenges after release.
Getting to Know Parole Officers

The institutional parole officers who were part of the program’s staff gave talks to groups of
inmates about community supervision, explaining what officers do and what offenders can do to
succeed on parole and get the most out of the experience.
The Division of Parole worked with Greenlight program staff to familiarize supervising field
parole officers with the program and its goals as well as some of the more complicating elements
of the program, especially the individualized release plan developed by program staff. On April
9th, 2003, representatives from the Vera Institute of Justice and the NYS Division of Parole and
Department of Correctional Services introduced the Greenlight program to approximately 25
New York State parole officers from the five boroughs of New York City. The officers assigned
to supervise study participants participated in a half-day training to familiarize them with the
rationale underlying the specific assignment, the expectations regarding their supervision, and
what the program was trying to achieve.117 This session was intended to provide the officers with
some general information about Greenlight, and more specifically, to discuss the role they would
play in the program. The session began with an overview of the program, noting DOCS’
development of transitional services programming, how the final phase of this transitional
services programming is structured by Greenlight, and the role that parole officers were
117

It is important to note that although all study participants were originally to be assigned to a set of specific parole
officers upon reentry (26 field officers), difficulties in identifying and assigning study participants were never fully
resolved and as a result, a much larger number of program participants were assigned to officers unfamiliar with the
program than originally intended. In addition, inter and intra-office transfers of cases after release led to the
involvement of an even greater number of field parole officers. As a consequence, we are unable to control for the
variation in officer supervision as originally intended.

60

Vera Institute of Justice

anticipated to play in the continued evolution of transitional services. Most importantly, they
were familiarized with the individual aspects of the treatment plans assigned to parolees. This is
particularly important because if the individualized release plans were treated as conditions of
parole, this might have placed an increased burden on Greenlight participants that may not be
experienced by comparison group participants who do not have detailed release plans in place as
part of their supervision. In practice, this means that parole officers might have potentially cited
Greenlight participants for a greater range of violations than controls based on the existence of
the release plan—in essence, leading to a greater number of parole violations for the intervention
group. Parole officers, who oversaw the release plan after participants exit prison, were
recognized as key to the continued development of the release plan. Not only were they expected
to supply feedback about the efficacy of the release plan, but they were also given the authority
to modify the plan according to the individual needs of each parolee.
Project staff held regular meetings thereafter with select groups of the participating parole
officers. These meetings initially occurred every month and a half and were eventually
conducted on a quarterly basis after the first six months. The primary intent was to flush out how
the release plan would be used by field officers, to make minor adjustments to the Release Plan
to make it more relevant to the officers and parolees, to establish some consistent communication
between the institution and the field and to develop policies on the development of the release
plan and how it would be shared with the field. Research staff conducted short qualitative
interviews with field parole officers who were supervising the program participants about the
program and their knowledge and perceptions of it.118
In these brief qualitative interviews, we primarily addressed such issues as the officers'
knowledge of the program, perceptions of the release plan, interaction with institutional parole
officers, perceived obstacles in how participants transition from the program to the community,
and their views about differences between the program and control group participants. Our
purpose was to keep the interviews as short as possible to minimize the burden on field parole,
while surveying the field officers who were most responsible for implementing the post-release
transition phase of the program. We focused on the 26 field officers originally assigned to the
program as they handled the bulk of the cases despite the problems in post-release assignments.
We interviewed 19 of the 26 officers (one was on medical leave and we were unable to contact
the other six).
Officers were generally familiar with the Greenlight program because most had attended
informational sessions with program staff on various occasions. These informational sessions
were generally perceived as useful in providing an overview and orientation regarding the
program and its purpose. Views of the program were mixed—some officers indicated a positive
view, while others indicated that their views of the program became less positive over time. Field
officers generally also expressed positive perceptions of their interactions with both institutional
118

A short quantitative interview about program participants was also administered to field parole officers to
supplement the information obtained from post-release interviews. The information from these interviews about
program participants is discussed more fully in Appendix F.

Vera Institute of Justice

61

parole and project staff. There was no indication of regular contact between field and
institutional parole; rather, the contact as irregular and accompanied problems as program
participants transitioned to the community. In such cases, institutional parole officers seemed
able to help field parole resolve most issues that were encountered.
In terms of specific program elements, we focused our questions on two particular
components: the release plan and Medicaid cards. Given that we wanted to keep the interviews
short, and because both were central pieces of the in-prison component, these two components
seemed suitable as key proxy measures for how the program functioned overall. Parole officers
gave the release plan mixed reviews. Almost all indicated they received the plans in a timely
manner and many perceived the release plan as useful—especially when they viewed the
program’s post-release referrals as appropriate. The release plan was largely viewed as positive
because parolees had a “roadmap” before they were released and parole officers did not have to
start the process from scratch. However, a substantial number felt like the release plans created
more work because of inappropriate referrals and indications that the process began to break
down over time. Based on these interviews, inappropriate referrals included post-release referrals
to programs that were too far away, inappropriate programming assignments after release, and as
time passed, fewer participants receiving the appropriate referrals they needed. Some officers
indicated that although referrals seemed consistent and appropriate in the beginning, that there
were few or no program referrals at all for program participants assigned to parole caseloads, and
that the problem seemed to worsen over time. In addition, some of the officers seemed to
indicate that given the difficulties associated with the release plan, they would resort to their own
resources and make program assignments based on their own experience.
The largest number of officers indicated, however, that the lack of Medicaid cards was a
significant problem. Almost all stated that participants on their caseloads did not have cards
waiting when they were released even though they had completed the paperwork during their
stay at the facility. They would often start the paperwork over which would result in a 6-week
delay before the letter of coverage would arrive. In some cases, additional delays would arise
when parolees would start the paperwork over and find that they already had an open file, but no
Medicaid card or letter assigning coverage. There were two key problems with this: first,
program participants were essentially “on hold” for services such as substance abuse treatment or
mental health programs that Medicaid traditionally covers. Second, parole officers indicated an
increase in frustration among program participants at not having the cards available when they
were expecting them.
The field officers also often expressed frustration at the assignment of caseloads. While all
26 were designated to receive all program participants from Greenlight and DOCS,
misassignments occurred regularly and parole officers found themselves supervising individuals
in precincts not typically under their jurisdiction. Further, there was more work for these officers
in trying to locate participants who were supposed to be on their caseload but were not. Although
this was a topic of conversation at the meetings between program staff and the parole officers,

62

Vera Institute of Justice

this situation was never resolved and continued to frustrate field officers throughout the duration
of the project.
Re-establishing Family Relationships

The project ran family reintegration groups during the evening led by staff members with
expertise in family counseling. Each inmate was asked to identify a family member or, where
appropriate, close friend or mentor, who was willing to attend these sessions. Once the person or
persons were identified, the family counselor contacted them and invited them to join a family
reintegration group for four sessions.
There were three types of family reintegration groups: those for couples, those for “family of
origin”(parents, siblings, extended family) and those for co-parents (parents who were not
currently in a relationship, but who had children together). The groups discussed absence,
expectations, communication, trust, reintegration into the family and reintegration into the local
community. Family members had occasional opportunities to meet alone with the group leader,
who offered brief counseling and referrals to appropriate agencies in the community. Family
members were also encouraged to visit inmates at other times when they could spend time
together in a less formal environment. Each group ran for four one-hour sessions.
The purpose of the program was to “ease the transition of men returning to their families
from prison by discussing the opportunities and challenges awaiting them and their families in an
atmosphere that was non-judgmental and which provided support and creative problem solving.”
This component of Greenlight was led by a family reintegration specialist and a family
counselor. When the inmates were being oriented into Greenlight, the family counselor would
present this program to the inmates as an opportunity to work through some issues with family
members in a “loosely structured setting,” explaining in detail his role as family counselor and
the purpose and benefits of the program. Participation in this program was completely voluntary.
If an inmate chose to participate, he had to choose which type of group to partake in: the familyof-origin, the co-parent, or the couples group. The family counselor determined which group to
place the inmate in by simply asking him which types of issues he wished to focus on.
When an inmate volunteered to participate, the counselors would call the family member(s),
explain the process with them, and schedule a session for them to come in. If the counselor could
not reach the family member(s) by telephone, he would send them a letter. If this also proved
ineffective, he would make a house visit and explain the program in person. The Family
Reintegration Program was structured as a four-week program in which the inmate and his
family member(s) met once each week. Each of the four sessions had a different theme: 1)
absence, 2) expectations, 3) readjustment, and 4) reintegration. Often, there were multiple
families and/or couples present at sessions.
The family reintegration component encountered some initial difficulties in getting inmate
participation and eventually engaged in an outreach effort to include more individuals. Ninetyfive inmates volunteered to take part in the Family Reintegration Program. Of these, 38 actually
participated in one or more session(s). The family counselors did 12 home visits; four of these

Vera Institute of Justice

63

resulted in family participation in the sessions. There were difficulties in contacting families due
to wrong or disconnected phone numbers, as well as incorrect addresses and frequent relocating
on the part of the family member(s). In addition, the family member(s) would initially appear
eager to take part, and then fail to attend the sessions. It seems likely that some of these problems
in contacting people could have been resolved had the program operated for a longer period of
time or had been longer than two months in length. In other words, if the counselors had access
to the inmates for 90 days rather than 60, it may have been easier for them to contact and
persuade family members to attend the sessions.
The ex-offenders interviewed about their experiences with the Family Reintegration Program
generally spoke highly of it. For example, one person stated that the counselor “taught us how to
hold a family together; how to speak to one another without arguing.” Another person said that
“these family sessions are necessary for many inmates and their families because it allows the
individuals to know where they stand with one another, and this is important…. Family members
tend to feel completely abandoned; my wife definitely felt that way.... The counselors would tell
you what you need to hear, not what you want to hear.”
The Family Reintegration Program essentially encouraged prisoners to re-establish and
maintain ties with family members who they believed would support their efforts to successfully
reintegrate back into society. The sessions seemed to have generated positive perceptions among
the participants, based on the interviews conducted with them. Program recruitment strategies,
however, prohibit any conclusions from being drawn about the effectiveness of this particular
Greenlight program element in promoting positive outcomes.
Making Contact with Service Providers in the Community

The project was designed to connect inmates with community-based service providers before
release. Each week three to four groups came into the facility to make presentations to program
participants in groups, and in some instances individually. Making these connections ideally
helped shift some of the burden of supporting people after release from parole to private agencies
with client rosters to fill.

Expanding Government Involvement in Release Planning
Many public agencies influence what happens to people after they leave prison. Some city
agencies, like the Human Resources Administration, the Department of AIDS Services, and the
Administration for Children’s Services, have former inmates on their caseload, or deal with the
children of former inmates. Other agencies, like the state Department of Labor and the city
Board of Education, have no caseload relationship with this population but could help this group
succeed after release. Bringing these agencies into the process of preparing inmates for release—
and bringing their staff physically to Queensboro—gave inmates an opportunity to learn about
the public programs they could draw on and, in some cases, a chance to enroll in them before
release. These connections were expected to reduce both the likelihood that people would go

64

Vera Institute of Justice

without needed services after release and the burden parole faced in brokering relationships,
giving field officers more time to devote to supervision.
Understanding and Applying for Public Benefits

To receive substance abuse treatment and medical care immediately after release, inmates who
are eligible for Medicaid applied for Medicaid as soon as they entered Queensboro. Their
applications were processed and approved by Medicaid officials at the New York City Human
Resources Administration during their stay at the facility and their benefits were supposed to be
activated by the time they exited the institution. However, interviews with supervising field
parole officers and program participants after release suggest that benefits were often not
activated and parolees had to start the application process over again.
Locating Housing for the Homeless

Greenlight was committed to doing everything possible to avoid placing project graduates in
homeless shelters. As part of creating a release plan, staff linked inmates to private groups and
public agencies that provided housing—both permanent and transitional. The project had an
agreement with the New York City Department of Homeless Services to share information and
assessments about participants who, as a last resort, were headed towards the shelter system in
order to reduce their stay in the city’s intake shelters and move them more swiftly to program
shelters.119
Facilitating Child Custody and Visitation

New York City’s Administration for Children’s Services (ACS) oversees the placement and care
of children in foster care and the collection of child support from non-custodial parents,
regardless of whether their children are in care. Recently, the agency has been focusing on the
intersection between foster children and parents in the criminal justice system. ACS is interested
in facilitating visits between children and their incarcerated parents and developing other
services that help preserve and strengthen these family bonds and improve the chances of
reuniting children with their biological parents. ACS representatives visited Queensboro to help
inmates with children in foster care understand the issues and regulations surrounding custody,
visitation, and child support, and the services ACS offers to parents who are temporarily unable
to care for their children at home.
Improving Education and Literacy

The New York City Board of Education runs a number of free programs—offered at locations
across the city and at different times—for adults seeking a G.E.D. Similarly, the New York City
119

Counselors first received the inmate’s consent to share assessment information with DHS and to have his location
in publicly supported transitional housing disclosed to service providers.

Vera Institute of Justice

65

Public Library runs free literacy programs, available in branch libraries in all five the boroughs.
The project tried to bring representatives from these organizations to Queensboro to discuss
these services with participants. The City University of New York runs college-level programs
for ex-offenders free of charge or at reduced tuition; outreach workers come to Queensboro to
meet with inmates individually and offer workshops for groups.

Daily Schedule
Exhibit A-2 lays out the structure of a typical weekday and details inmate movement. The day
begins with two 55 minute periods of programming devoted to orientation, practical skills, job
readiness and job development, or drug treatment readiness, depending on where each inmate is
in his eight-week stay (see Exhibit A-1). Community groups also come in to meet with inmates
during these periods. Lunch is next, followed by a two-hour cognitive-behavioral group session
(except for Monday, the visiting day). Dinner is next, followed by family sessions and
workshops with community service providers.

66

Vera Institute of Justice

DRUGS
COG.
SKILLS

Week One

Week Two

Week Three

Week Four

Job Readiness

Job Readiness

Job Readiness

Job Readiness

Substance Abuse
Awareness-Relapse
Prevention

Substance Abuse
Awareness-Relapse
Prevention

Substance Abuse
Awareness-Relapse
Prevention

Substance Abuse
Awareness-Relapse
Prevention

Reasoning and
Rehabilitation

Reasoning and
Rehabilitation

Reasoning and
Rehabilitation

Reasoning and
Rehabilitation

Community
Linkages

Community
Linkages

Community
Linkages

Family
Reintegration
Session

Family
Reintegration
Session

Family
Reintegration
Session

Workshop with
DHS on housing

Workshop with the
Board of Education
to learn about adult
education and
alternative
programs
Individual meeting
with case manager
to work on release
plan

Workshop with
ACS on custody
and related issues

RELEASE
PLANS

OTHER
AGENCIES

PAROLE
SERVICES

Introduction to
Parole Supervision

Community
Linkages

FAMILY

COMMUNITY LINKS

PRACT.
SKILLS

DAILY CLASSES

JOBS

Exhibit A-1. Project Greenlight Weekly Schedule

Workshop with
HRA to apply for
Medicaid and other
benefits

Individual
assessment with
case manager

Vera Institute of Justice

67

Week Five

Week Six

Week Seven

Week Eight

Work with Job
Developer

Work with Job
Developer

Work with Job
Developer

Work with Job
Developer

Reasoning and
Rehabilitation

Reasoning and
Rehabilitation

Reasoning and
Rehabilitation

Reasoning and
Rehabilitation

Practical Skills

Practical Skills

Practical Skills

Practical Skills

Transmitting Release
Plan to Field Parole
Officer
Community
Linkages

68

Community
Linkages

Community
Linkages

Community
Linkages

Family Reintegration
Session

Workshop with
CUNY about college
level programs

RELEASE
PLANS

OTHER
AGENCIES

FAMILY

SERVICES

COMMUNITY LINKS

PAROLE

SKILLS

COG.
SKILLS

DAILY CLASSES

DRUGS

JOBS

Exhibit A-1 (continued).

Vera Institute of Justice

Individual meeting
with case manager

Individual meeting
with case manager

Exhibit A-2: Project Greenlight Daily Schedule 

ACTIVITY / LOCATION / INMATES
7 a.m.
Mess for breakfast52 inmates
Mess for breakfast52 inmates

MOVEMENT / ISSUES
7:2052 inmates to mess for breakfast.
7:4052 inmates to mess for breakfast.
7:40Smoke break.
7:50104 inmates to gym for unit meeting.
8:45Smoke break.
9:0052 inmates go to Program Space for programs. 26
inmates go to Program Space for electives. 26 inmates work
in job resource room.
9:45All inmates switch classrooms.

8 a.m.
9 a.m.

Program 1 room 13 inmates120
Program 1 room 26 inmates

Program 1 room 26 inmates
Job Resource Room 26 inmates
10 a.m.
Program 1 room 13 inmates
Program 1 room 26 inmates

Program 1 room 26 inmates

11:0052 inmates to mess for lunch.
11:2052 inmates to mess for lunch.

Job Resource Room 26 inmates
11 a.m.

12:30104 inmates to Program Space for cognitivebehavioral group sessions

Mess for lunch52 inmates
Mess for lunch52 inmates
12 p.m.

1 p.m.

Cognitive Skills Sessions
(except Mondays)
4 classrooms
26 each

2 p.m. 104 inmates
Recreation,
Dayroom or
Call Out
3 p.m. Case management
12 inmates

3:0012 inmates from Program Space to Case Management.
Other inmates to recreation or housing.

4 p.m.
Mess for dinner—52 inmates
Mess for dinner—52 inmates

4:4552 inmates to mess for dinner.
5:0552 inmates to mess for dinner.

5 p.m.
Family Session (Mon-Thurs) (Mess Hall) 5 inmates

5:305 inmates to Processing Center.
Other inmates to Program Space.

6 p.m.
Family Session (Mon-Thurs) (Mess Hall) 5 inmates
7 p.m.

6:555 inmates from Program Space switch with 5 inmates
from Processing Center

120

The substance abuse program was split into two groups that will take place in separate rooms: Drug Treatment
Readiness and Relapse Prevention.

Vera Institute of Justice

69

Appendix B. Methodology and Data
Study Overview

Our original research proposal to NIJ was based on a three-year evaluation of the transitional
services pilot program using a design that employed both randomized and quasi-experimental
elements. We anticipated approximately 96 eligible participants each month to be assigned to
one of three study groups (N=1,152/year; N=3,456 for the three-year pilot duration). DOCS
released approximately 5,000 conditional release (CR) inmates in 2002; thus, based on these
trends, our eligible study population was approximately one-fifth of all conditional release
inmates each year. However, since we are focused on individuals returning to New York City,
we point out that approximately 60 percent of all inmates incarcerated in a New York State
prison are from NYC; thus if the 60 percent figure holds true for CR inmates, then approximately
3000 would return to NYC each year. As a consequence, our study population was expected to
constitute almost 40 percent of all CR inmates returning to NYC. We had proposed to collect
data on participants who passed through the program during the first 21 months (N=2,016). We
chose the 21-month timeline in order to be able to conduct follow-up interviews in a timely
manner, and to allow for at least one year of post-release time in the community for all
participants before the end of the three-year period.
There were multiple criteria for study eligibility as shown in Exhibit B-1; these are the exact
criteria as defined by the DOCS MIS/Classification section in their programming algorithms.
Conditional release inmates who were between 75 and 105 days from their anticipated release
date were identified every month by the DOCS MIS/Classification division for assignment to the
study. Conditional release inmates were chosen for this program because they were the group
who could be most easily identified and transferred to Queensboro at least 75 days prior to
release.121 Conditional release inmates include felons who have previously been denied release
by the Board of Parole, and inmates determinately sentenced under both the 1995 and 1998 New
York State Sentencing Reform who were not eligible for discretionary release. The 75 to 105day criterion is based on the two-month (60-day) length of the transitional services
demonstration program combined with the issues inherent to relocating eligible inmates to the
pilot facility in a timely manner. The logistics of transporting inmates to the demonstration
facility means that there will be some variation in the amount of time inmates spend in the
program (e.g., those identified as 105 days from release and transported quickly may reside in
the program several weeks longer than those with shorter times to release—there was not any
more than a four-week difference in time spent at Queensboro for any of the participants). Thus,
some participants assigned to the program stayed at Queensboro little more than two months and
others stayed up to three months.

121

Inmates eligible for discretionary release (parole) appear before the Board of Parole approximately 60 days
before their parole eligibility date.

70

Vera Institute of Justice

Because Queensboro is a minimum-security institution, safety and security issues required
that inmates not be assigned to the facility if any of the following conditions apply: conviction
for a sex offense, vicious or callous violence, high institutional risk score, current disciplinary
sanction, medical reasons, existing warrants, or ineligible current assignment (e.g., work release,
shock incarceration, lockdown). In addition, as Queensboro is not structured as a male/female
facility, all participants were male. Finally, the program was focused on inmates returning to
New York City who could benefit most from the NYC-based service providers that were part of
the program. The program therefore used a proxy variable to determine NYC eligibility based on
a conviction in one of New York City’s five boroughs (Brooklyn, Bronx, Manhattan, Queens,
and Staten Island). Thus, similar to prior work, we anticipated losing a small percentage to
attrition because they will not return to New York City despite having been convicted in one of
those counties.122
Table B-1. Eligibility criteria used by NYS DOCS MIS/Classification section for
determining study eligibility.
1.	 Only under-custody male inmates with a TAC type of APPR and an Earliest Release Type of CR as of Sunday.
2.	 EXCLUDE inmates whose CR date is NOT within the specified number of days in the future. Regular run is 75
to 105 days in the future.
3.	 EXCLUDE inmates who have not had a date comp done since admission.
4.	 EXCLUDE inmates on outcount.
5.	 EXCLUDE inmates owned by Work Release, Shock, ASAT, CASAT, CASAT Feeder, Willard, Transit,
SHU200.
6.	 EXCLUDE inmates owned by Queensboro General.
7.	 EXCLUDE inmates owned by “Special Needs”, Southport, Upstate SHU.
8.	 EXCLUDE inmates owned by Max facilities.
9.	 EXCLUDE in-transit inmates.
10. EXCLUDE inmates housed in Infirmary beds.
11. INCLUDE indictment counties: New York City (5 boroughs) only.
12. EXCLUDE inmates with warrants.
13. EXCLUDE inmates with Outstanding Approval or T.O. for reason 24 (Work Release) or 29 (Open Date) OR
Outstanding Referral, Approval, or T.O. to Queensboro General.
14. EXCLUDE inmates with Initial Guideline Escape Score of 12 or greater.
15. EXCLUDE inmates with Current Guideline Institutional Risk Score of 6 or greater.
16. EXCLUDE inmates with Initial Guideline Other Security Characteristics 04 (Pattern of Callous Violence), 05
(Vicious Violence), 09 (Sex Crimes), 13 (Aggressive Homosexual), or 27 (Mariel Cuban).
17. INCLUDE inmates with Temp. Release Application Status of Denied or Cancelled (EXCLUDE inmates with
Temp. Release Application in progress).
18. EXCLUDE inmates with pending DCP Tier 2 or 3 Tickets
19. EXCLUDE inmates currently serving SHU, Keeplock or Cube Confinement
20. EXCLUDE inmates with OMH Level of 1, 2 or 3. EXCLUDE inmates with MED Level of 1.
21. Referral reason = 31 (Queensboro Release Program).

Once they were defined as “potentially” eligible under the criteria established, participants
were assigned to one of three study groups. 123 Space limitations at Queensboro precluded the
122

Gottfredson, S. D., and R. B. Taylor. “Community contexts and criminal offenders.” Pp. 62-82 in Communities
and Crime Reduction, T. Hope and M. Shaw (eds.). London: Her Majesty’s Stationary Office, 1988.
123
Of those defined as “potentially” eligible, only those who eventually returned to New York City constituted the
final sample.

Vera Institute of Justice

71

transfer of all potential participants identified by DOCS to the demonstration facility. Thus, those
who did not get transferred in a timely manner and whose time to release dropped below 60 days
had their transfer orders cancelled. These individuals constitute the first comparison group; those
who are released directly from institutions outside of New York City and who received no
transitional services programming (hereafter referred to as Upstate releases). As a result, we
expected no individual selection biases based on these determination criteria because assignment
to this group was entirely dependent on a combination of (1) DOCS transportation schedules, (2)
space limitations at Queensboro, and (3) time remaining to release.124
Those who were transferred to Queensboro were assigned to one of two primary study
groups; the Greenlight intervention (hereafter Greenlight or GL) and a comparison group which
participated in the existing DOCS Transitional Service Programming (hereafter TSP). Our
original research protocol called for the study to begin tracking research participants in July 2002
and to have a fully randomized design in place at that time. We began following participants
from program inception (February 2002) in order to provide feedback to the program where
possible as the program started up and addressed initial implementation issues.125 In February
2002, as the eligible participants were identified, the first 52 participants were immediately
assigned to the intervention until the program was filled; after the program was filled to the 52bed capacity, remaining assignments were to the TSP program.126 After the initial bed space was
filled, haphazard sequential assignment to one of the two groups at Queensboro on
approximately a two-to-one basis (intervention to comparison) was the primary protocol between
March and mid-July.127
In terms of the haphazard assignment process, the Queensboro Institutional Locator Officer
received a faxed list each day showing the transfers to take place the next day. For the first five
months after the program was filled and the first program participants began exiting the program,
she would assign individuals to the intervention or the control groups on a two-to-one basis in
the order of the appearance of the names on the faxed list. If on a specific day, the last name
assigned was to the control, she would start the next day assigning two participants to the
intervention; if on the prior day the assignment ended with one person assigned to the
intervention, the next day would assign the first person on the list to the intervention group, the
next person to the control, the next two people to the intervention, and so on. Records kept by the
locator officer indicate that this procedure was consistently followed during this time. Because
124

We recognize the possibility however, that DOCS transportation schedules may not move individuals from
certain institutions at the same rate and that those not transferred may potentially differ on institutional
characteristics, prior programming, etc.
125
Due to early program termination, however, we now include those early program participants as part of our
study.
126
Although the pilot program was originally designed to accommodate 104 participants, facility construction
limited program capacity at 52 until July 2002 when construction was completed.
127
Haphazard assignment protocols are often considered to approximate randomized designs reasonably well and
are among the stronger quasi-experimental designs (Shadish, William R., Thomas D. Cook and Donald T. Campbell.
Experimental and Quasi-Experimental Designs for Generalized Causal Inference. New York: Houghton-Mifflin.
2002).

72

Vera Institute of Justice

potential program participants (as well as controls) were not to be transferred to Queensboro
unless bed space was available on the intervention wing, deviations from the assignment
procedure were not an issue. A trickle-process random assignment was instituted in July 2002
(discussed below) and remained in place through the program’s termination at the end of
December 2002.
It is not uncommon for studies to encounter a “pipeline” effect and find that the target
population is smaller in size or harder to locate than originally expected in the planning stage.128
Clearly, the total number of expected participants was reduced given the early termination of the
program. Instead of 2,016 total study participants, our final count was 805 potentially eligible
participants. Our original estimate of 96 eligible participants per month was slightly reduced; on
average, we found that DOCS was able to identify slightly fewer eligible participants than was
originally estimated although DOCS did indicate to us at an early stage that they were having
difficulty identifying and tracking individuals who did not transfer to Queensboro. Our first
group,therefore, is slightly smaller than originally projected even given the early termination.
Thus, in the 10-month interval between February and December 2002, a total of 805 “potential”
study participants (Greenlight N=349; TSP N=328; Upstate N=128) were identified for an
average of 81 per month, 16 less per month than originally expected; this number is consistent
however, with the later estimations provided by DOCS.
As we have noted, because identification of these individuals was based on county of
conviction, we assumed from the start that some number of individuals would not be returning to
New York City (NYC) even if identified as eligible. Our final determination as to whether
individuals returned to New York City was based on parole office assignments made by the
Division of Parole. This could not be determined from official record data before release
although some participants knew at the time of assignment whether they would be returning to
the city. Program requirements dictated that if individuals were determined to be returning to
some location other than NYC within the first few days of being assigned to the intervention at
Queensboro, they were removed from the program and placed in the TSP program. This was
done so that the intervention could make the most use of the NYC community-based service
providers for those who were returning to NYC. Since placement in either group at the
demonstration facility was random (or haphazard) for non-NYC parolees, we assumed that this
would not disparately impact our design because we eventually removed all participants from
analysis who were not returning to New York.
Group Assignments and Achieving a Randomized Design

Although initial participants were assigned to the program in a haphazard manner and haphazard
designs may have the advantage of being stronger than other quasi-experimental designs, they do

128

Boruch, Robert F. Randomized Experiments for Planning and Evaluation: A Practical Guide. Vol. 44, Applied
Social Science Research Methods Series. Thousand Oaks, CA: Sage, 1997.

Vera Institute of Justice

73

not replace the desirability of a true randomized design.129 Being able to achieve a random
assignment protocol was one of the most challenging aspects of the study; the difficulties in
achieving randomized designs are certainly not specific to criminal justice research, but the
difficulties are often amplified due to the nature of the setting.130 Once inmates were identified as
eligible for the program and were transferred to the pilot facility, there were a number of
constraints inherent to this particular project that complicated the assignment of individuals to
treatment and comparison conditions in a random manner. Program capacity at the facility was
initially designated as 104 beds for treatment participants with approximately 50 beds available
for assignment to the comparison group.131 Eligible inmates were transferred to the facility on a
daily basis Monday through Friday. Facility staff received a list of the inmates to be transferred
the day before the inmates arrived, bed assignments were made upon arrival and once assigned,
inmates remained in that bed until release. The unequal group sizes complicated the assignment
process, and day-to-day variability in the number of participants and the number of available
beds for either condition further constrained the method that could be used to achieve a
randomized design. Thus, simple allocation procedures were not well-suited to achieve a
randomized assignment protocol.
The flow of participants into the pilot facility at Queensboro required a trickle-process
randomization procedure.132 There are numerous issues inherent to trickle randomization
including the ways in which these methods of assignment can be corrupted or subverted; thus we
were especially attentive to these issues in designing the randomization procedure.133 As a result,
there were several key criteria we kept in mind. First, the primary criterion for the randomization
procedure was the ability to document the assignment process to ensure the integrity of the
procedure. Simple allocation procedures, for example, may allow individuals to be assigned to
one condition or another based on the order of their appearance in a sequence, but may not allow
researchers to know whether particular individuals were assigned based on order of
appearance.134 Second, we could not use simple random assignment because there may not have
129

Thus, although we had no pre-existing measures of such characteristics as prior DOCS programming or
individual motivations and attitudes, randomized designs are generally considered the best guard against nonrandom assignment of unknown characteristics that might vary systematically under other assignment protocols.
130
See Boruch, Randomized Experiments,1997. Boruch elaborates many of the difficulties inherent to randomized
field designs. Numerous other sources also describe the issues inherent to implementing a randomized design, see,
for example, Shadish, William, R., Thomas D. Cook and Donald T. Campbell. Experimental and QuasiExperimental Designs for Generalized Causal Inference. Boston, MA: Houghton Mifflin Company, 2002.
131
There were 52 intervention beds available between February and June; an additional 52 beds were added in July
2002 when a renovation project was completed at the facility. Over the study period, the number of beds available to
control participants was variable whereas the number of intervention beds was fixed.
132
See, for example, Braucht, G. Nicholas, and Charles S. Reichardt. 1993. “A Computerized Approach to TrickleProcess, Random Assignment.” Evaluation Review 17(1):79-90; also see, Shadish, Cook and Campbell,
Experimental and Quasi-Experimental Designs, 2002. Further, we are grateful for the comments of Dr. Mark Lipsey
at Vanderbilt University and Dr. Charles Reichardt at the University of Denver regarding trickle randomization
procedures and the issues inherent to implementing those procedures. Their insights and comments were invaluable
in developing the current random assignment design—any remaining limitations are strictly our own.
133
Braucht and Reichardt, “A Computerized Approach”, 1993:79-90.
134
Boruch, Randomized Experiments, 1997.

74

Vera Institute of Justice

been sufficient bed space to accommodate them. Although the TSP program was more flexible in
the numbers of participants it could accommodate (these individuals were simply assigned bed
space in the institution upon arrival), the GL program participants were housed in a separate
wing from the rest of the institutional population. Simple random assignment of inmates could
have resulted in more assignments than bed space available. In designing the assignment process,
researchers consulted with DOCS institutional staff to construct a procedure that minimized any
additional burdens on those staff. In the end, we achieved a randomization procedure that led to a
true randomized design, minimized burdens on existing institutional staff, and allowed us to keep
an audit trail for the assignment protocol.
When the list of the next day’s incoming inmates arrived at the facility each day, the
institution’s locator officer faxed the list to researchers at Vera. Vera researchers conducted the
randomization procedure (described below) and faxed the assignment list back to the locator
officer the same day. The trickle-process randomization procedure, approved by the DOCS
Program Planning, Research and Evaluation unit, allowed the DOCS institutional intake officer
to randomly assign eligible inmates to the intervention or control group as they entered the
facility at Queensboro.
Once research staff received the list of incoming inmates from DOCS, we listed all CR
inmates in a spreadsheet in the order in which they appeared on the incoming list as shown in
Column 4 of Exhibit B-2. After entering the inmate identifier into the spreadsheet, each available
bed was numbered in a list (columns 1 and 2) and a random numbers table was used to generate
the order of the bed assignments.135 We listed the beds in column two in a ratio of their
availability—in Exhibit B-2, they are shown in the 2-to-1 ratio we assigned them.136 The random
numbers table determined the order of the assignment in Column 3. Thus, the first inmate
number in Column 4 (31X5019) was assigned the 3rd bed in the list (Column 1) because of its
appearance first in the random numbers table (Column 3). We therefore achieved a randomized
method of assignment by randomizing the available beds and assigning them to the inmates
based on the order of their appearance on the incoming transfer list.
Table B-2. Randomized Bed Assignment Procedure.
Column 1
Column 2
Column 3
Column 4
Bed
Treatment
Assignment
Inmate
Number
Condition
Order
Number
1
GL
3
31X5019
2
GL
6
49X5025

Column 5
Inmate
Assignment
TSP
TSP

135

Boruch, Randomized Experiments, 1997. We used the Rand Corporation’s “A Million Random Digits.”
http://www.rand.org/publications/classics/randomdigits/randomdata.html. Santa Monica, CA: Rand, March 2002.
Boruch points out that this type of numbering is a means of creating an audit trail to ensure how individuals are
being identified and randomly assigned. We downloaded the text file of random numbers in March 2002 and printed
out several hundred pages of random digits. Each time a new assignment was to be made, we used the next page in
the sequence of printed pages.
136
If we were notified that bed space availability on the GL side was restricted, we varied the assignment ratio.

Vera Institute of Justice

75

3
4
5
6
7

TSP
GL
GL
TSP
GL

7
1
2
5
4

49X5534
31X5657
41X5064
31X5088
40X5794

GL
GL
GL
GL
GL

There were multiple concerns given the difficulties with random assignment procedures.
First, it was possible that all beds in the intervention were filled on some days requiring that all
assignments be to either the control condition—this was more likely to involve intervention beds
rather than comparison beds because treatment beds were restricted to two wings on a single
floor at Queensboro. Transfers to Queensboro were only supposed to occur on an available bed
capacity basis. Although research staff made every attempt to ascertain available bed space, it
was possible that bed space counts were inaccurate leading to forced deviations from the
assignment process. We were also concerned about large imbalances in the assignment process
on a day to day basis although this never became a problem with our study.137
Table B-3. Assignment Protocol and Efficacy of the Randomized Design

Initial Assignment

Final Group Assignment
Greenlight
TSP
Upstate

Total

All Assignments
Haphazard Assignment
GL
TSP
UPST*
Random Assignment
GL
TSP

189
---

-170
--

--128

189
170
128

143
17

40
118

---

183
135

Total

349

328

128

805

186
---

-139
--

--113

186
139
113

141
17

28
111

---

169
128

NYC Returns—Final Sample
Haphazard Assignment
GL
TSP
UPST*
Random Assignment
GL
TSP

137

Reichardt (personal comment, 4/26/02) has indicated that large imbalances in the assignment process from day to
day (e.g., too many days where the assignment process might be 4:1 or 5:1 rather than 2:1) can also lead to bias.
However, this is also something that can be easily identified and if it occurs too often, data can be analyzed by the
relative variations or strata in assignment to examine for potential differences.

76

Vera Institute of Justice

Total

344

278

113

735

*Upstate assignments occurred throughout the entire period of 2/02 through 12/02. In addition to the 64 that did not
return to NYC upon release, another 6 individuals (4 TSP, 2 Upstate) were removed in the final sample due to two
known deaths and transfers out-of-state after being released to NYC.

Exhibit B-3 shows the assignment of potentially eligible participants to the three study
groups during the assignment period between February and December, and the assignment of
those who were subsequently defined as eligible (i.e., our final study group). Final eligibility was
ultimately determined by research staff based on post-release parole assignment locations. All
study participants who were under parole supervision in one of NYC’s five boroughs were
included in the final analysis. Those who were under parole supervision outside of NYC were
not included in the final analysis as not meeting one of the defining eligibility criteria.
Of particular note in Exhibit B-3 is that although the number of GL participants haphazardly
assigned drops by three from 189 to 186, the comparable TSP group drops by 27 from 170 to
143. This is primarily due to identification of non-NYC parolees being identified at program
intake and being routed away from the GL program. One of the issues of central focus however,
is adherence to the randomization protocol established and we focus on the final sample of 741
in the second half of the exhibit. Of the 169 participants that research staff assigned to the GL
intervention, 28 were eventually assigned by the institution to the TSP program (a 16.6 percent
deviation)—this primarily occurred when research staff were unaware of space limitations at the
facility on the intervention side. Of the 128 assigned to the TSP group, 17 were eventually
assigned at the institution to the GL program (a 13.3 percent deviation). This was the most
surprising aspect of the deviation from the assignment process in that research participants were
not supposed to backfill open slots in the program once assigned to the TSP condition. We
anticipated some deviation because of space problems on the intervention side (i.e., assignments
to TSP after researchers had assigned them to the intervention) but to minimize deviation from
the design, intervention assignments were to only have been filled with new institutional
assignments. Nevertheless, for the approximately five and a half months that the randomization
procedure was in place, only 45 of 297 research participants were reassigned contrary to the
random assignment protocol (15.5 percent total deviation). As with other elements of the
program, we did not completely resolve these issues before the program was terminated.
A Note about Issues of Selection Bias

Evaluations of the type proposed constantly deal with selection bias issues and the present study
is no exception.138 We attempt, however, to deal with these issues in several ways. We have
indicated that the first five months of assignments to intervention and control groups were
haphazard, and that for the last five and a half months, we maintained a randomized design in
our evaluation. Thus assignments to our primary groups involve a relatively strong quasi138

E.g., see Gaes, Gerald G. “Correctional Treatment.” Pp. 712-738 in The Handbook of Crime and Punishment, M.
Tonry (ed.). New York, NY: Oxford University Press, 1998.

Vera Institute of Justice

77

experimental method initially followed by a randomized design later. Although inmates are not
required to participate in the research, they were required through DOCS regulations to
participate in the programs. As a consequence, self-selection was not an issue in which program
was assigned. In addition, we identified our participants and followed them from approximately
10 weeks before release. For our primary outcome measures (arrest, incarceration, parole
violation) and demographic information from publicly available official records, we have
information for all identified participants.
There is always a loss of subjects during the program phase, which typically compounds
selection bias issues. Indications from the present study suggest that DOCS and GL program
staff actively worked to keep participants in the program and that there are very few issues
associated with program dropout. Over the course of conducting program activities, fourteen
inmates were removed from the GL program for disciplinary, physical health, or other issues—
some of these individuals were restored to the program. We include all of these however, as
intervention program participants. Although we did not receive research consent for post-release
interviews from all participants (discussed in the following sections), we can indicate how many
declined to participate in those interviews. Thus, selection bias does not appear to be a
problematic issue in our study.
Gaining Research Consent

Study participants were required by DOCS policy to participate in the Greenlight program as
they are required to participate in any DOCS administered programming; refusal to participate
carries potential consequences including loss of good-time release. Program participants,
however, were not required to participate in the project’s research component. During the first
few months, research staff contacted study participants upon the first interview date after release
and requested them to participate in the research. Signed consent forms were obtained along with
contact information. Within several months of the program’s commencement, we had reached an
agreement with DOCS to seek research consent from study participants while they were still
incarcerated at Queensboro. This served a dual purpose. First, although we were not able to
obtain consent in this manner from the Upstate group because we were not able to reach them,
we did contact significant numbers of the participants while they were still incarcerated to ask
them to participate in the evaluation of the program. In addition, we also asked for post-release
contact information. Although we did get contact information for a substantial proportion of the
individuals who consented to participate in the research, we found that much of the information
was not useful and that most of our contacts occurred through our arrangements with the
Division of Parole. Exhibit B-4 provides the relevant information on the numbers of individuals
who refused program participation either before or after assignment to the program, the numbers
removed from the program and finally the numbers of individuals who were asked to participate
in the research and how many agreed and declined. These numbers are based on the total number
of participants returning to New York City (N=741).

78

Vera Institute of Justice

Table B-4. Research Participation
Greenlight
(N=344)
Agreed to participate
Did not agree to participate
Unable to Contact
Interview Rates
Total Group
Of Those Contacted

173
23
148

(50.3%)
( 6.7%)
(43.0%)

31.7% [109 / 344]
55.6% [109 / 196]

TSP
(N=278)
101
18
159

Upstate
(N=113)

(36.3%)
( 6.5%)
(57.2%)

-------

17.3% [48 / 278]
40.3% [48 / 119]

-----

Further Distinguishing the Three Study Groups

Our research design is based on tracking three different groups of offenders who are identified as
potentially eligible for the program: the Greenlight intervention participants who were relocated
to the pilot facility approximately 8 to12 weeks before they were released, the TSP comparison
group transferred to the Queensboro correctional facility in a similar timeframe, and a
comparison group released directly from New York State institutions well away from their home
community of New York City. It is seldom the case in criminal justice research that control or
comparison groups receive no treatment, and our study is no exception.139 We have already
described the intervention program in some detail in Appendix A along with basic
implementation issues. For the two comparison groups, the TSP group received the existing
transitional services programming provided by DOCS. To our knowledge, the Upstate group
received no pre-release services of any kind.
In terms of differences between the DOCS comparison group housed at the pilot facility and
the DOCS comparison group released from Upstate institutions, another critical difference
between the two groups is the proximity to their home communities. An intuitive assessment
suggests that providing services close to one’s home community where participants can be more
easily connected to existing community-based service providers and other mechanisms of
support (e.g., family, community organizations) will be more effective. The available evidence
suggests that relocating individuals closer to their home communities before release can have a
positive impact on outcomes.140 If all participants in a transitional services program are returning
139

Weisburd, David, Anthony Petrosino, and Gail Mason. “Design Sensitivity in Criminal Justice Experiments.”
Pp. 337-379 in M. Tonry (ed.) Crime and Justice: A Review of Research, vol. 17. Chicago, IL: University of
Chicago Press, 1993.
140
See, for example, McGuire, James, and Philip Priestley. “Reviewing ‘What Works’: Past, Present and Future.”
Pp. 3-34 in J. McGuire (ed.), What Works: Reducing Reoffending—Guidelines from Research and Practice.
Chichester, England: John Wiley and Sons, 1995.

Vera Institute of Justice

79

to the same community, it is logistically much easier to tailor the program to those individuals
than it would be for a program whose participants are being released to multiple communities
and often to communities at great distances from where the program is being offered and the
institution is located.
However, because TSP is our primary comparison group, it is essential that some description
of the program is necessary in order to understand what the Greenlight intervention is being
compared to. Anecdotal and empirical evidence generally suggests that most institutional
programs are often not well-conceptualized, not delivered in an effective manner, and reach very
few participants.141 In addition, it is also not uncommon for a competing program to take on the
characteristics of, or incorporate elements of, an intervention program, especially if it is operated
in proximity to the intervention.142 Our original proposal stated that the comparison group
participants were to “receive DOCS’ current Phase IV programming, a still-evolving intervention
[delivered] in an eight to twelve week period.” DOCS Phase IV programming “is primarily
based on the individual needs of each inmate and will include:
I. Assessment
A. Institutional Summary—accomplishments during incarceration
B. Profile of Concerns—inmates presenting problems
C. Document Review
II. Transitional Services
A. Family Counseling
B. Job Fairs
C. Community Preparation
D. Referrals”143
DOCS has been in the process of designing and implementing a comprehensive transitional
services program for several years. The original schema called for a four-phase program that
begins transitional services programming at admission. The fourth and final phase was structured
so that “inmates in Phase Four will participate in programming that will make the greatest impact
on their reintegration needs.”144 The Greenlight program was a demonstration model of
structuring and delivering transitional services to inmates returning to the community. The
DOCS Transitional Services Program, developed by DOCS staff, was in a state of evolution
during our study period from February 2002 through March 2003. We conducted interviews with
DOCS staff at the Queensboro facility, and research staff attended some of the TSP classes being
held at the facility in order to get at least a fundamental understanding of the programming being
141

See, for example, Austin, James. “Prisoner Reentry: Current Trends, Practices, and Issues.” Crime and
Delinquency 47(3):314-334, 2001, for an overview of prison programs.
142
Boruch, Randomized Experiments, 1997.
143
DOCS. Transitional Services Program Manual. Albany, NY: DOCS, 2001:76.
144
DOCS. Transitional Services Program Manual, 2001:75. The “four-phase” program has been replaced with a
three-phase programming sequence although it has not yet been fully implemented.

80

Vera Institute of Justice

offered. These interviews and personal observations suggest the following depiction of DOCS
Transitional Services Programming offered at Queensboro during our study period. As with the
GL program, this depiction is primarily descriptive of the program and is not intended as an
assessment of program fidelity.
Interviews with facility staff responsible for administering the program suggest that a basic
transitional services program was offered beginning in late March of 2002. This basic
introductory program had several components including money management (e.g., budgeting,
how to open a checking account, get a bank loan or mortgage), general classes, and a cognitivebehavioral component that focused on coping with changing circumstances and decision-making
processes. In addition, DOCS, Parole and the New York City Human Resources Administration
worked in a joint effort to qualify eligible inmates for Medicaid coverage before they exited
prison. By early 2003 (after our study had ended), the program had added several additional
elements including a class conducted with the inmates that dealt with family integration, sessions
dealing with time management and leisure time, journal writing sessions designed to help
participants engage in self-examination through writing or drawing, and “portfolio” sessions that
dealt with more pragmatic issues such as the federal bonding act, employer tax credits, resume
writing along with sample resumes and cover letters, and forms to apply for identification such
as birth certificates and social security cards.
Upon arrival at Queensboro, a short assessment was conducted by intake staff. During this
assessment, determinations are made about inmates’ capability for work outside of the facility
until such time as they are released, their post-release living arrangements, emergency contact
information, and their general mental status and whether or not they are suicidal or have a history
of suicide attempts. Whereas Greenlight participants had no facility work obligations, TSP
participants did have work assignments while at the facility and these typically took precedence
over participation in programming. During our study period, TSP classes were conducted four
days a week (Monday through Thursday) from 8:15 to 10:30 am and from 12:30 to 2:30 pm by
seven DOCS counselors. The length of the TSP programming is approximately five to six weeks.
Although inmates were supposed to receive a disciplinary “ticket” if they did not report to
classes when not at work, our observations and conversations with staff suggest that this seldom
occurred, if ever. In addition to the above sessions, a number of community-based organizations
made presentations to TSP participants every two or three months about the availability of
services after release including housing, employment, and substance abuse and other healthrelated services (including HIV/AIDS appointments). There was also a program emphasis on job
readiness in which inmate skills were matched to Department of Labor apprenticeship programs.
It is not clear how often participants went to classes on a regular basis. In-class observations
suggest that participants attended at their discretion, and class rolls were not kept to document
attendance. Researchers also had a great deal of difficulty in reaching TSP participants in
Queensboro to administer the research consent form. When scheduling times to administer
consent forms, institutional staff had much more difficulty in locating TSP than GL participants
for “call-outs” to meet with research staff. This suggested that there was no internal

Vera Institute of Justice

81

documentation of which participants were supposed to be in which classes at which time—and
therefore, that no regularly-enforced schedule for attending classes existed. This problem is
reflected in the lower numbers of TSP participants we were unable to contact for research
consent forms as shown in Exhibit B-4 in the last section.
Although the TSP program appears to have developed significantly more detail over the
course of time, there are real questions about the degree of its implementation and use. Similar
program elements to those used by Greenlight were incorporated into the program although the
form of those programs appear to be substantively different (e.g., although a family component
was said to exist in the TSP program, it involved only the inmate and not the inmate’s family as
the Greenlight component did). There are substantive questions about the rigor of
implementation; many inmates did not appear to attend or be in classes when they were supposed
to be, interviews with institutional staff indicated that many inmates did not spend much time in
the program, especially given institutional work requirements, and classes were limited to four
days a week and four hours per day. We were unable to obtain documentation on the class
lessons used in the program. In addition, interviews with program staff also suggested that the
cognitive component was less useful because some inmates were simply involved for too short a
period of time. In sum, it is clear that DOCS has an operating program although it is not clear
how intensive that program is, what role assessment plays in inmate classes, or how well
implemented the program is in general. Given that DOCS was still in the developmental stages
of Phase IV programming when the Greenlight project launched, these issues are neither
surprising nor unexpected; newly structured programs often encounter implementation issues
which take time to resolve in such a way that the program operates smoothly.145 It is without
question that some of these same issues are inherent to the Greenlight program as previously
discussed in Appendix A.
Designated Parole Officers and Differential Enforcement Issues

There is clearly a potential for bias arising from parole supervision; differential enforcement
practices across parole offices and parole officers could impact rates of parole violations and
therefore our findings and results. Few studies recognize, much less are able to, account for
patterns of differential enforcement by parole offices or officers. Although it is not possible to
resolve all potential for bias arising from this particular arena, we attempted to address these
issues in our study with the cooperation of the Division of Parole.146 In the five boroughs of New
York City, there are 28 parole field offices staffed by more than 600 parole officers. The
Division of Parole designated 26 specific field officers in New York City to supervise
participants in this evaluation as part of their regular caseload. These 26 officers were to
145

See, for example, Harrison, Lana D. and Steven S. Martin. Residential Substance Abuse Treatment for State
Prisoners: Implementation Lessons Learned. NIJ Special Report, NCJ 195738. Washington, DC: USDOJ, 2003
(April). Web-only document downloaded May 10th, 2003 from http://www.ncjrs.org/pdffiles1/nij/195738.pdf.
146
We note that Parole raised this suggestion with us and offered the assistance of their field officers in trying to
cope with this particular problem.

82

Vera Institute of Justice

supervise all study participants (intervention and controls) for the duration of their supervision
period in the community.
Individual level participant data collected from the Division of Parole as part of the research
design therefore included identifying information on the assigned borough and parole office in
that borough and the supervising parole officer. Thus, although the potential for differences
between parole officers in their individual practices cannot be eliminated, reducing the number
of supervising officers to a select would have allowed for an examination in greater detail of the
outcomes under their supervision. This provided research staff the opportunity to examine
differences within and between parole officers in the distribution of cases assigned to them and
the handling of those cases. Further, we have also collected the written violation summaries for
those study participants who were revoked in order to examine in greater detail the behaviors
leading to revocation.
As noted in a prior section, although there was an attempt to assign all research participants
to the caseload of these 26 designated field officers, misassignment of cases at release was
common and a frustrating experience for the parole officers. In addition, reassignment after
release also proved problematic when these cases were transferred to other offices or boroughs
and were again not transferred to the caseload of the designated officer.
Data: Post-release and Parole Officer Interviews

In terms of the timing of the interviews, both anecdotal and prior evidence strongly suggests that
the choices made and issues faced by released offenders in the period immediate following
release can have a profound impact on both their short and long-term outcomes.147 As a
consequence, the original study design included two post-release interviews with study
participants: these interviews were to occur one month after release (Time 1) and then again six
months later (Time 2). Interviews were originally intended to be conducted with randomly
selected participants after release from prison. Criminal justice populations are notoriously
difficult to follow and despite a multiple strategy for contacting individuals in the community,
we continued to have difficulty. As a result, we quickly made the decision to contact all
participants that we could locate. We had three primary strategies for locating and contacting
individuals after they were released.
First, when we contacted potential participants while they were incarcerated and asked them
to participate in the research project, we also asked for post-release participant contact
information. We asked them to provide similar information for family members and friends who
were likely to be in contact with the participant after release. We also made a decision to contact
each participant at least once after release even if they had declined to participate while still
incarcerated. Although a number of individuals still declined to participate, a sufficient number
147

See, for example, Fifield, Adam. “Life on Freedom Street.” New York Times (New York, NY), 14(1): 2001:1;
Nelson, Marta, Perry Deess and Charlotte Allen. The First Month Out: Post-Incarceration Experiences in New York
City. New York, NY: Vera Institute, 1999.

Vera Institute of Justice

83

agreed to participate after release (after initially declining while incarcerated) to make the effort
worthwhile.
Our second approach involved a cooperative arrangement with the Division of Parole. We
provided parole with the names of individuals under parole supervision whom we were trying to
locate. Parole, in turn, would provide us with reporting dates and the field parole office the
parolees were reporting to. We then contacted the parole officer to arrange an interview in a
private setting at the field parole office on a day that the individual was scheduled to meet with
his parole officer. Research staff typically contacted parole officers one to two weeks before an
interview was due, and would meet the participant at the parole office during a regularly
scheduled visit. Although we gained most of our interviews in this way, it still proved to be
labor-intensive because parolees did not always keep their appointments or make their reporting
dates.
Finally, because it is common for interviews not to be conducted with those who have been
reincarcerated, we hoped to address this in part in the following manner. Because our focus was
on those returning to New York City, individuals who have been rearrested or reincarcerated for
parole violations are held in city institutions. We contacted the New York City Department of
Correction to conduct interviews with those that we could identify as being held locally. We
would search the city database for the NYSIDS (New York State Identification number) of study
participants and once located, we contacted city officials for permission to interview those
individuals if they had been returned to custody. Thus, we were able to conduct a substantial
number of interviews with study participants who had already “failed” through the cooperation
of the city DOC, by conducting interviews in local jails in each borough and at Riker’s Island.
In practice, our first interview usually occurred between six and 12 weeks after release. Early
termination of the program coupled with insufficient funding led us to the conclusion that we
would not have sufficient time to complete a reasonable number of second interviews, and that
our efforts would be better focused elsewhere. Because we had only just begun conducting Time
2 interviews (3 completed), we discontinued them when the program was officially cancelled.
We therefore focused on increasing the number of Time 1 interviews and conducting a
limited set of interviews with parole officers. Our original Memorandum of Understanding with
the Division of Parole allowed for review of hard copy file folders at field parole offices, but the
shortened time frame for the study precluded a more detailed assessment of those data. The
compromise solution was to ask field parole officers to provide answers to a set of basic
questions regarding participants’ outcomes (most specifically, housing and employment) and
performance under parole supervision during the first three months after release using case
folders for reference.148 We used SPSS to generate a list of randomly selected parolees stratified
by study group for data collection purposes.
We also note that evaluations of this type commonly use pre- and post-intervention measures
of study participants and controls. Although we considered this approach, especially given the
148

84

We also asked parole officers a set of qualitative questions at the same time—see Appendix A.

Vera Institute of Justice

focus of the cognitive behavioral component on changing anti-social thinking patterns, we
ultimately were unable to obtain pre-intervention measures. DOCS declined to allow us in-prison
access to interview study participants to administer such interviews, limiting our interactions
with study participants to administering research consent forms. We note however, that in a
personal conversation with one of the cognitive behavioral program designers, he indicated that
he held such measures as suspect in their ability to accurately assess change in criminal thinking
patterns. In sum, it was not clear that questions in an instrument of this type could adequately
measure changed attitudes and thinking.
Power of the Research Design

It is common for many treatment effectiveness studies to be underpowered, and this is equally
applicable, if not more so, to criminal justice research studies.149 In our initial design, it was clear
that there would be a sufficient N to maintain adequate statistical power to guard against Type II
errors (i.e., missing an effect due to a small sample size when one is actually present). We did,
however, choose the interview sample sizes to maintain sufficient power for more detailed
statistical analyses. This eventually became a non-issue when the program was terminated early
and it was clear that we were not going to be able to get sufficient interviews for to maintain
adequate power.
In our final sample, our study has more individuals assigned to the treatment condition than
the control conditions, and our major between-group comparisons are for the GL and TSP
groups. Unbalanced research designs result in a loss of power, but assignment to treatment and
control conditions were dictated by bed space allocation allowed by DOCS. Nevertheless, we
had a strong interest in examining within-group differences for the intervention group and the
larger treatment group worked in our favor. 150 We therefore have an unbalanced design that
allows us to make between-group comparisons that remain statistically powerful. In the selection
of an appropriate sample size for each of the post-release groups, it is important to specify the
assumptions and parameters used in understanding power calculations.
For most of our analyses, we set the significance level at .05, using a one-tailed test because
our hypotheses are specified in a certain direction (this also allows us to reduce our sample size
while maintaining power). No convention seems to exist for determining reasonable power, so
we set power at .80, a value that has been suggested for general use.151 The importance of effect
size in determinations of research designs is crucial; prior evidence from reviews and metaanalyses indicates a great deal of variation in expected effect sizes from criminal justice

149

Lipsey, Mark W. Design Sensitivity: Statistical Power for Experimental Research. Newbury Park, CA: Sage,
1990:73. Brown, Stephen E. “Statistical Power for Criminal Justice Research.” Journal of Criminal Justice
17(1989):115-122.
150
Kraemer, Helena C. and SueThiemann. How Many Subjects? Statistical Power Analysis in Research. Newbury
Park, CA: Sage, 1987; Lipsey, Design Sensitivity, 1990.
151
Lipsey, Design Sensitivity, 1990.

Vera Institute of Justice

85

research.152 Further, it has been noted that treatment administered in institutional settings may
dampen a program’s effect.153 Given this and the relative scarcity of anticipated effect sizes for
these types of interventions, we chose a sample size based on what appears to be a relatively
modest effect size of .25. Under these parameters, an average sample size of approximately 200
was necessary per each group.154
For our primary analyses then, using a GL sample of 349, basic computations under the
above assumptions indicate that control group sample sizes need to be at least an N of 140 each.
Our TSP group N is more than 200 although our Upstate group is not of sufficient size to allow
us to make statements with sufficient confidence without relaxing our assumptions regarding
power. Within-group analyses will also suffer this issue as well due to the smaller Ns.
Major Deviations from the Original Research Design—Summary

In sum, there have been several major deviations from the original research design that bear a
specific, summary discussion. First, the original design was based on the program’s continued
operation for three years; the program was essentially terminated due to lack of funding after
operating for 12 months. The major implication is that our study sample size is significantly
smaller than expected. Original estimations indicated slightly more than 2,000 potentially
eligible participants for the three study groups; the final sample of those who were potentially
eligible was 806 with a final sample of 741 study participants returning to NYC. As a
consequence, the total number of participants that were interviewed are significantly smaller as
well.
Time 1 interviews were originally scheduled to be conducted at random; it quickly became
clear that this was going to be too difficult to schedule. As a consequence, we decided to conduct
interviews with as many individuals as we could possibly find after release, either through our
contact information, parole reporting dates, or those who had been reincarcerated through new
arrests or parole violations. Although this means that our interviews are biased towards those
who are easiest to locate (as is true of all interviews), we conducted a large number of Time 1
interviews with participants who had already been reincarcerated in New York City.
We also decided to eliminate the second post-release interview. The sample size for the Time
1 interviews had been decided based on an N that would provide sufficient statistical power to
152

Lipsey, Design Sensitivity, 1990; Weisburd, David, Anthony Petrocino and Gail Mason. “Design Sensitivity in
Criminal Justice Experiments.” Pp. 337-79 in Crime and Justice: A Review of Research, M. Tonry (ed.), Vol. 17.
Chicago, IL: University of Chicago Press, 1993; Andrews, Don A., Ivan Zinger, Robert D. Hoge, James Bonta, Paul
Gendreau, and Francis T. Cullen. “Does Correctional Treatment Work? 1990; Gaes, et al., “Adult Correctional
Treatment” 1999; Lipsey, Design Sensitivity, 1990; MacKenzie, Doris L. “Criminal Justice and Crime Prevention.”
Pp. 9-1 through 9-76 in Preventing Crime: What Works, What Doesn’t, What’s Promising. NCJ 165366.
Washington, DC: USDOJ, 1997; Weisburd, Petrosino and Mason, “Design Sensitivity”, 1993; Lipton, Douglas S.
The Effectiveness of Treatment for Drug Abusers under Criminal Justice Supervision. NCJ 157642. Washington,
DC: USDOJ, 1995.
153
Andrews, et al., “Does Correctional Treatment Work?”, 1990.
154
Sample size determination is based on power charts presented in Lipsey, Design Sensitivity, 1990:90-96.

86

Vera Institute of Justice

address Type II errors including expected attrition by the second set of interviews. Given the
lower than expected number of initial interviews, we decided the research would be better served
by increasing our focus on the Time 1 interviews. In addition, we supplemented part of that
information by obtaining information from case folders maintained by the parole officers on
whether parolees had housing, employment and were adhering to parole supervision
requirements.
We also were unable to obtain information about risk assessment as originally planned. Our
original plan was to pilot our interview instrument for approximately three months, make needed
changes to the instrument, and incorporate the Level of Service Inventory-Revised (LSI-R)
questions into our instrument at that revision. Although initial interview rates were reasonable
(approximately 45 percent of eligible participants), we decided to seek permission from DOCS to
administer the LSI-R instrument to TSP participants while they were still incarcerated at
Queensboro. This would have served several purposes. The first is that since Greenlight program
staff were administering the LSI-R to intervention participants, we would have been able to
administer the instrument to TSP participants at a similar point in time before programming
(when we approached them to participate in our research project). In addition, we expected that
this would have resulted in an increase in the numbers of individuals for whom we collected this
piece of information. In this case, it would have been easier to contact study participants while
still incarcerated than to follow them after release. DOCS was receptive to this approach in
general, but as the program’s cancellation became increasingly likely and we discovered that the
program had stopped using the LSI-R, we ended our efforts to incorporate this aspect into the
research. It is worth noting at this point that Greenlight program staff found the LSI-R instrument
difficult and time-consuming to administer.
Key Hypotheses

I.

Recidivism Outcomes
H1:
A structured multi-modal transitional services program, located in or near
the community to which offenders will return, with links to community
services and field parole supervision, will result in lower rates of recidivism
(arrest, conviction, incarceration).
H2:
A structured multi-modal transitional services program, located in or near
the community to which offenders will return, with links to community
services and field parole supervision, will result in lower rates of technical
violations.

II.

Intermediate Outcomes

Vera Institute of Justice

87

H3: Greenlight participants will have lower rates of relapse and report greater use
of substance abuse treatment services after release from incarceration than
comparison group members.
H4: Greenlight participants will have higher rates of employment and report more
stable employment after release from incarceration than comparison group
members.
H5: Greenlight participants will report less use of shelter housing and have more
stable long-term housing arrangements after release from incarceration than
comparison group members.
H6: Greenlight participants will report more stable and better family relationships
after release from incarceration than comparison group members.
H7: Greenlight participants will report greater use of community services than
comparison group members after release from incarceration.
H8: Greenlight participants will report greater program satisfaction than 

comparison group members. 

Studying Impact and Implementation

As we have indicated, our report and analysis is focused on an impact evaluation of the Project
Greenlight Transitional Services Demonstration Program. However, we are aware that there is
considerable support for including an evaluation of the implementation and process of any
particular program when an impact evaluation is conducted.155 Rossi, Freeman and Lipsey
(1999), for example, state that “…it is generally not advisable to conduct an impact evaluation
without including at least a minimal process evaluation.”156 Because there is substantial evidence
that treatment integrity is a fundamentally important component of any treatment implementation
155

Gaes, “Correctional Treatment”, 1998; Gaes, et al., “Adult Correctional Treatment” 1999; Rossi, Peter H.,
Howard E. Freeman and Mark W. Lipsey. Evaluation: A Systematic Approach. Sixth Ed. Thousand Oaks, CA:
Sage, 1999; Scheier, Mary A. “Designing and Using Process Evaluation.” Pp. 40-68 in Handbook of Practical
Program Evaluation, J. S. Wholey, H. P. Hatry, and K. E. Newcomer (eds.). San Francisco, CA: Jossey-Bass, 1994.
156
Rossi, Freeman and Lipsey, Evaluation, 1999:199; Sherman, Lawrence, Denise Gottfredson, Doris MacKenzie,
John Eck, Peter Reuter, and Shawn Bushway (eds.). Preventing Crime: What Works, What Doesn’t, What’s
Promising. NCJ 165366. Washington, DC: USDOJ, 1997. Sherman, et al., in their report to Congress, made three
recommendations for evaluating crime prevention programs (a Statutory Evaluation Plan). One of those
recommendations suggested that federal research funds be limited to conducting impact assessments only. It is
important to note however, that they make this recommendation based on the limited availability of evaluation
research funds, not on the relative value of process evaluation research. This position stands in stark contrast
however, to Gaes’ (1998:717) statement as a response to the lack of process evaluations that “Perhaps funding
agencies should require that program evaluations include treatment process reports.”

88

Vera Institute of Justice

process, and since program delivery cannot be assumed to be controlled, inferences about
program effectiveness may be confounded with poor program implementation.157 In short, what
is perceived to be an ineffective program may in reality be a poorly implemented program. As a
result, we conducted qualitative interviews with program and institutional staff, field parole
officers, and program participants, and visited the program and observed classes of both the GL
and the TSP program on several occasions. Thus, although our evaluation was funded as an
impact evaluation, we included minimal process assessments when the opportunity arose to
bolster our understanding of program implementation. These issues were discussed in more
detail in Appendix A.

157

Gaes, “Correctional Treatment”, 1998; Gaes, et al., “Adult Correctional Treatment”, 1999; Lipsey, Mark W.
“What Do We Learn From 400 Research Studies on the Effectiveness of Treatment with Juvenile Delinquents?” Pp.
63-78 in What Works: Reducing Reoffending, J. McGuire (ed.). New York, NY: John Wiley, 1995; MacKenzie,
“Criminal Justice”, 1997; Pullen, S. Evaluation of the Reasoning and Rehabilitation Cognitive Skills Development
Program as Implemented in Juvenile ISP in Colorado. Prepared in part under NIJ Grant #93-IJ-CX-K017. Denver,
CO: CO Dept. of Public Safety, August, 1996. Scheier, “Process Evaluation”, 1994.

Vera Institute of Justice

89

Appendix C. Supplementary Tables
Table C-1. Bivariate correlations, aggregate criminal history measures (N=735).
Age at Last
Arrest
Total Arrests
Age at Last Arrest
Felony Arrests
Felony Convictions
Misd Arrests
Misd Convictions

.557***
1.000***
---------

Felony
Arrests
.820***
.561***
1.000***
-------

Felony
Convictions
.563***
.490***
.724***
1.000***
-----

Misd
Arrests
.903***
.431***
.505***
.323***
1.000***
---

Misd
Convictions
.949***
.462***
.686***
.387***
.935***
1.000***

*** p < .001 (2-tailed tests).

Table C-2. Prior felony arrests and felony convictions by study group.
GreenLight

Prior Felony Arrests
0
1-2
3-5
6-9
10+
Missing
Total
Prior Felony Convictions
0
1
2
3
4
5+
Missing
Total

90

Vera Institute of Justice

TSP

Upstate

Total

15.4%
29.4%
24.7%
15.4%
14.8%
0.3%

20.2%
25.5%
26.2%
16.7%
11.3%
0.0%

24.3%
33.0%
20.0%
13.0%
9.6%
0.0%

18.6%
28.5%
24.6%
15.5%
12.7%
0.1%

100.0%

100.0%

100.0%

100.0%

25.3%
27.6%
17.7%
14.5%
8.7%
5.8%
0.3%

31.9%
17.7%
22.0%
13.8%
9.9%
4.6%
0.0%

38.3%
20.0%
14.8%
13.0%
11.3%
2.6%
0.0%

29.8%
22.7%
18.9%
14.0%
9.6%
4.9%
0.1%

100.0%

100.0%

100.0%

100.0%

Table C-3. Misdemeanor arrests and convictions by study group. 

GreenLight

Prior Misdemeanor Arrests
0
1-2
3-5
6-9
10+
Missing
Total
Prior Misdemeanor Convictions
0
1
2
3
4
5+
Missing
Total

TSP

Upstate

Total

43.6%
26.2%
14.2%
5.8%
9.9%
0.3%

47.5%
26.2%
12.4%
6.7%
7.1%
0.0%

58.3%
22.6%
10.4%
6.1%
2.6%
0.0%

47.4%
25.6%
13.0%
6.2%
7.7%
0.1%

100.0%

100.0%

100.0%

100.0%

36.3%
16.3%
9.9%
6.7%
4.4%
26.2%
0.3%

41.8%
13.8%
8.5%
6.0%
7.4%
22.3%
0.0%

54.8%
14.8%
5.2%
9.6%
0.0%
15.7%
0.0%

41.3%
15.1%
8.6%
6.9%
4.9%
23.1%
0.1%

100.0%

100.0%

100.0%

100.0%

Vera Institute of Justice




91

Table C-4. Comparison of Interviewed and Not Interviewed Parolees. 

GL with
GL no
TSP
TSP no
interview
interview
interview
interview
(N=109)
(N=235)
(N=48)
(N=230)
Demographics
Age at release
Mean
Std. Deviation
Education
8 years or less
9-11 years
12 or more years
Race/Ethnicity
White/Other
Black
Hispanic

ns
33.79
9.16

33.49
9.62

33.73
9.27

33.37
9.59

9.3%
50.5%
40.2%

12.7%
49.3%
38.0%

12.5%
52.5%
35.0%

11.5%
49.5%
39.0%

5.5%
59.6%
34.9%

5.1%
56.6%
38.3%

10.4%
56.3%
33.3%

7.0%
55.7%
37.4%

60.41
46.18

55.01
43.48

55.51
43.70

50.16
39.85

9.38
10.10

8.32
8.6

7.96
7.91

7.78
8.15

6.48
8.48

5.70
7.14

5.29
6.99

5.32
6.80

ns

Criminal History Indicators
Time served (in months)
Mean
Std. Deviation
Total Arrests
Mean
Std. Deviation
Total Convictions
Mean
Std. Deviation
Age at first arrest
Mean
Std. Deviation
Substance abuse
No
Alcohol
Drugs
Alcohol and Drugs

19.18
4.47

19.90
5.25

20.13
4.88

19.89
5.37

38.8%
6.8%
31.1%
23.3%

33.5%
9.5%
34.4%
22.6%

46.7%
6.7%
31.1%
15.6%

35.7%
7.6%
30.8%
25.9%

Recidivism
Arrest within 1 yr
Felony arrest within 1 yr
Revocation within 1 yr

29.8%
16.4%
22.8%

37.8%
18.6%
26.3%

27.3%
13.6%
24.4%

24.1%
12.8%
21.2%

92

Vera Institute of Justice

Sig.

ns

ns

ns

ns

ns

ns

ns
ns
ns

Table C-5. Scale Statistics for Interview Items. 

Mean
Service referrals
0.20
Service contacts
0.11
Knowledge of parole conditions
0.95
Opinions about parole supervision
0.74
Family relations
0.81
Days using drugs in past month
2.99
Weeks employed
2.54
Days involved in community activities
11.30
Spent first night on street or in shelter
6%

Std Deviation
0.21
0.14
0.15
0.26
0.26
9.75
4.71
13.39
---

Alpha
0.72
0.50
0.96
0.87
0.92
NA
NA
NA
NA

Life-Tables

Life-table Description: Life tables are an accurate method of summarizing time-to-event data and
are fairly straightforward in their interpretation. For the purposes of this analysis, terminal (or
failure) events are defined as (1) arrests, (2) felony arrests, and (3) parole revocations. Time
intervals are based on weeks at risk. The Number Entering column denotes the number of cases
surviving at the beginning of the two-week interval. The Number Withdrawn refers to the
number of cases that are censored, or removed from risk, as non-failure terminal events during
that interval. With an analysis of arrests, for example, we censor cases for the following reasons:
1) a participant was still active as of April 1, 2004 when we completed our observations; or 2) a
participant’s parole supervision was revoked. The Number Exposed to Risk is a computation
based on the number starting the interval and those censored. Terminal Events are the number of
failure events (arrests, felony arrests, or revocations) during that time interval. The Proportion
Terminating is the number of Terminal Events divided by the Number Exposed to Risk and the
Proportion Surviving is simply equal to (1 – Proportion Terminating). The Cumulative
Proportion Surviving estimates the probability of surviving to the beginning of that interval
without experiencing a failure event. The Hazard Rate is the instantaneous rate of event
occurrence at that interval given that the event has not already occurred.158

158

See Luke and Homan, Psychological Assessment, 1998; or Yamaguchi, Event History Analysis, 1991; for more
detail.

Vera Institute of Justice

93

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100

94

344
340
335
331
328
318
309
303
299
298
292
287
284
280
277
275
271
264
255
248
243
234
231
226
222
215
211
202
193
184
178
168
157
142
125
115
110
103
89
71
58
53
50
48
44
33
20
19
19
19
12

0
0
0
0
1
1
0
0
0
0
3
1
1
1
1
2
3
2
1
0
4
0
2
0
2
1
3
7
6
4
6
8
13
15
8
5
3
8
18
13
5
3
1
3
10
12
0
0
0
6
11

Vera Institute of Justice

344.0
340.0
335.0
331.0
327.5
317.5
309.0
303.0
299.0
298.0
290.5
286.5
283.5
279.5
276.5
274.0
269.5
263.0
254.5
248.0
241.0
234.0
230.0
226.0
221.0
214.5
209.5
198.5
190.0
182.0
175.0
164.0
150.5
134.5
121.0
112.5
108.5
99.0
80.0
64.5
55.5
51.5
49.5
46.5
39.0
27.0
20.0
19.0
19.0
16.0
6.5

4
5
4
3
9
8
6
4
1
6
2
2
3
2
1
2
4
7
6
5
5
3
3
4
5
3
6
2
3
2
4
3
2
2
2
0
4
6
0
0
0
0
1
1
1
1
1
0
0
1
1

0.012
0.015
0.012
0.009
0.028
0.025
0.019
0.013
0.003
0.020
0.007
0.007
0.011
0.007
0.004
0.007
0.015
0.027
0.024
0.020
0.021
0.013
0.013
0.018
0.023
0.014
0.029
0.010
0.016
0.011
0.023
0.018
0.013
0.015
0.017
0.000
0.037
0.061
0.000
0.000
0.000
0.000
0.020
0.022
0.026
0.037
0.050
0.000
0.000
0.063
0.154

0.988
0.985
0.988
0.991
0.973
0.975
0.981
0.987
0.997
0.980
0.993
0.993
0.989
0.993
0.996
0.993
0.985
0.973
0.976
0.980
0.979
0.987
0.987
0.982
0.977
0.986
0.971
0.990
0.984
0.989
0.977
0.982
0.987
0.985
0.984
1.000
0.963
0.939
1.000
1.000
1.000
1.000
0.980
0.979
0.974
0.963
0.950
1.000
1.000
0.938
0.846

0.988
0.974
0.962
0.954
0.927
0.904
0.886
0.875
0.872
0.854
0.848
0.842
0.834
0.828
0.825
0.819
0.806
0.785
0.766
0.751
0.735
0.726
0.717
0.704
0.688
0.678
0.659
0.652
0.642
0.635
0.620
0.609
0.601
0.592
0.582
0.582
0.561
0.527
0.527
0.527
0.527
0.527
0.516
0.505
0.492
0.474
0.450
0.450
0.450
0.422
0.357

0.006
0.007
0.006
0.004
0.013
0.012
0.009
0.006
0.002
0.009
0.003
0.003
0.005
0.003
0.002
0.003
0.006
0.011
0.009
0.008
0.008
0.005
0.005
0.006
0.008
0.005
0.010
0.003
0.005
0.004
0.007
0.006
0.004
0.005
0.005
0.000
0.011
0.017
0.000
0.000
0.000
0.000
0.005
0.006
0.007
0.009
0.012
0.000
0.000
0.014
0.033

0.006
0.007
0.006
0.005
0.014
0.013
0.010
0.007
0.002
0.010
0.004
0.004
0.005
0.004
0.002
0.004
0.008
0.014
0.012
0.010
0.011
0.007
0.007
0.009
0.011
0.007
0.015
0.005
0.008
0.006
0.012
0.009
0.007
0.008
0.008
0.000
0.019
0.031
0.000
0.000
0.000
0.000
0.010
0.011
0.013
0.019
0.026
0.000
0.000
0.032
0.083

0.006
0.009
0.010
0.011
0.014
0.016
0.017
0.018
0.018
0.019
0.019
0.020
0.020
0.020
0.021
0.021
0.021
0.022
0.023
0.024
0.024
0.024
0.025
0.025
0.026
0.026
0.026
0.026
0.027
0.027
0.027
0.027
0.028
0.028
0.028
0.028
0.029
0.031
0.031
0.031
0.031
0.031
0.032
0.033
0.035
0.038
0.043
0.043
0.043
0.048
0.072

0.003
0.003
0.003
0.003
0.004
0.004
0.004
0.003
0.002
0.004
0.002
0.002
0.003
0.002
0.002
0.002
0.003
0.004
0.004
0.003
0.004
0.003
0.003
0.003
0.004
0.003
0.004
0.002
0.003
0.003
0.004
0.003
0.003
0.003
0.003
0.000
0.005
0.007
0.000
0.000
0.000
0.000
0.005
0.006
0.006
0.009
0.012
0.000
0.000
0.014
0.030

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-6. Life Table for Greenlight Total Arrests (Felonies and Misdemeanors).


0.003
0.003
0.003
0.003
0.005
0.005
0.004
0.003
0.002
0.004
0.002
0.003
0.003
0.003
0.002
0.003
0.004
0.005
0.005
0.005
0.005
0.004
0.004
0.005
0.005
0.004
0.006
0.004
0.005
0.004
0.006
0.005
0.005
0.005
0.006
0.000
0.009
0.013
0.000
0.000
0.000
0.000
0.010
0.011
0.013
0.019
0.026
0.000
0.000
0.032
0.083

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
102
104
106

278
276
271
268
267
264
262
258
255
251
247
245
242
235
233
229
223
222
221
218
217
214
211
209
204
201
196
187
181
173
156
141
125
108
103
99
94
83
71
63
43
39
37
35
33
32
24
14
14
13
13
9
2
2

0
0
0
0
0
0
1
0
2
0
0
1
4
1
1
2
0
0
1
0
0
2
1
2
1
3
4
2
4
13
13
15
15
4
4
4
11
9
8
16
3
1
1
0
1
7
10
0
1
0
4
7
0
2

278.0
276.0
271.0
268.0
267.0
264.0
261.5
258.0
254.0
251.0
247.0
244.5
240.0
234.5
232.5
228.0
223.0
222.0
220.5
218.0
217.0
213.0
210.5
208.0
203.5
199.5
194.0
186.0
179.0
166.5
149.5
133.5
117.5
106.0
101.0
97.0
88.5
78.5
67.0
55.0
41.5
38.5
36.5
35.0
32.5
28.5
19.0
14.0
13.5
13.0
11.0
5.5
2.0
1.0

2
5
3
1
3
2
3
3
2
4
2
2
3
1
3
4
1
1
2
1
3
1
1
3
2
2
5
4
4
4
2
1
2
1
0
1
0
3
0
4
1
1
1
2
0
1
0
0
0
0
0
0
0
0

0.007
0.018
0.011
0.004
0.011
0.008
0.012
0.012
0.008
0.016
0.008
0.008
0.013
0.004
0.013
0.018
0.005
0.005
0.009
0.005
0.014
0.005
0.005
0.014
0.010
0.010
0.026
0.022
0.022
0.024
0.013
0.008
0.017
0.009
0.000
0.010
0.000
0.038
0.000
0.073
0.024
0.026
0.027
0.057
0.000
0.035
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.993
0.982
0.989
0.996
0.989
0.992
0.989
0.988
0.992
0.984
0.992
0.992
0.988
0.996
0.987
0.983
0.996
0.996
0.991
0.995
0.986
0.995
0.995
0.986
0.990
0.990
0.974
0.979
0.978
0.976
0.987
0.993
0.983
0.991
1.000
0.990
1.000
0.962
1.000
0.927
0.976
0.974
0.973
0.943
1.000
0.965
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000

0.993
0.975
0.964
0.960
0.950
0.942
0.932
0.921
0.914
0.899
0.892
0.884
0.873
0.870
0.858
0.843
0.840
0.836
0.828
0.824
0.813
0.809
0.805
0.794
0.786
0.778
0.758
0.742
0.725
0.708
0.698
0.693
0.681
0.675
0.675
0.668
0.668
0.642
0.642
0.596
0.581
0.566
0.551
0.519
0.519
0.501
0.501
0.501
0.501
0.501
0.501
0.501
0.501
0.501

0.004
0.009
0.005
0.002
0.005
0.004
0.005
0.005
0.004
0.007
0.004
0.004
0.006
0.002
0.006
0.008
0.002
0.002
0.004
0.002
0.006
0.002
0.002
0.006
0.004
0.004
0.010
0.008
0.008
0.009
0.005
0.003
0.006
0.003
0.000
0.004
0.000
0.013
0.000
0.023
0.007
0.008
0.008
0.016
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.004
0.009
0.006
0.002
0.006
0.004
0.006
0.006
0.004
0.008
0.004
0.004
0.006
0.002
0.007
0.009
0.002
0.002
0.005
0.002
0.007
0.002
0.002
0.007
0.005
0.005
0.013
0.011
0.011
0.012
0.007
0.004
0.009
0.005
0.000
0.005
0.000
0.020
0.000
0.038
0.012
0.013
0.014
0.029
0.000
0.018
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.005
0.009
0.011
0.012
0.013
0.014
0.015
0.016
0.017
0.018
0.019
0.019
0.020
0.020
0.021
0.022
0.022
0.022
0.023
0.023
0.024
0.024
0.024
0.025
0.025
0.025
0.026
0.027
0.028
0.028
0.029
0.029
0.030
0.030
0.030
0.031
0.031
0.033
0.033
0.038
0.040
0.041
0.043
0.046
0.046
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.048

0.003
0.004
0.003
0.002
0.003
0.003
0.003
0.003
0.003
0.004
0.003
0.003
0.003
0.002
0.003
0.004
0.002
0.002
0.003
0.002
0.003
0.002
0.002
0.003
0.003
0.003
0.004
0.004
0.004
0.004
0.003
0.003
0.004
0.003
0.000
0.004
0.000
0.007
0.000
0.011
0.007
0.008
0.008
0.011
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

Vera Institute of Justice

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-7. Life Table for TSP Total Arrests (Felonies and Misdemeanors).


0.003
0.004
0.003
0.002
0.003
0.003
0.003
0.003
0.003
0.004
0.003
0.003
0.004
0.002
0.004
0.004
0.002
0.002
0.003
0.002
0.004
0.002
0.002
0.004
0.004
0.004
0.006
0.005
0.006
0.006
0.005
0.004
0.006
0.005
0.000
0.005
0.000
0.011
0.000
0.019
0.012
0.013
0.014
0.021
0.000
0.018
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

95

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
102

96

113
113
112
110
109
107
104
102
100
99
98
96
95
93
93
91
90
89
85
84
83
83
82
81
80
79
77
66
50
41
39
38
38
38
36
36
36
34
34
34
33
31
30
30
30
28
27
23
21
21
17
6

0
0
0
0
0
2
0
0
0
0
2
0
0
0
0
0
0
3
0
0
0
0
0
1
0
1
9
16
8
2
0
0
0
0
0
0
2
0
0
1
1
1
0
0
2
0
4
1
0
4
11
6

Vera Institute of Justice

113.0
113.0
112.0
110.0
109.0
106.0
104.0
102.0
100.0
99.0
97.0
96.0
95.0
93.0
93.0
91.0
90.0
87.5
85.0
84.0
83.0
83.0
82.0
80.5
80.0
78.5
72.5
58.0
46.0
40.0
39.0
38.0
38.0
38.0
36.0
36.0
35.0
34.0
34.0
33.5
32.5
30.5
30.0
30.0
29.0
28.0
25.0
22.5
21.0
19.0
11.5
3.0

0
1
2
1
2
1
2
2
1
1
0
1
2
0
2
1
1
1
1
1
0
1
1
0
1
1
2
0
1
0
1
0
0
2
0
0
0
0
0
0
1
0
0
0
0
1
0
1
0
0
0
0

0.000
0.009
0.018
0.009
0.018
0.009
0.019
0.020
0.010
0.010
0.000
0.010
0.021
0.000
0.022
0.011
0.011
0.011
0.012
0.012
0.000
0.012
0.012
0.000
0.013
0.013
0.028
0.000
0.022
0.000
0.026
0.000
0.000
0.053
0.000
0.000
0.000
0.000
0.000
0.000
0.031
0.000
0.000
0.000
0.000
0.036
0.000
0.044
0.000
0.000
0.000
0.000

1.000
0.991
0.982
0.991
0.982
0.991
0.981
0.980
0.990
0.990
1.000
0.990
0.979
1.000
0.979
0.989
0.989
0.989
0.988
0.988
1.000
0.988
0.988
1.000
0.988
0.987
0.972
1.000
0.978
1.000
0.974
1.000
1.000
0.947
1.000
1.000
1.000
1.000
1.000
1.000
0.969
1.000
1.000
1.000
1.000
0.964
1.000
0.956
1.000
1.000
1.000
1.000

1.000
0.991
0.974
0.965
0.947
0.938
0.920
0.902
0.893
0.884
0.884
0.875
0.856
0.856
0.838
0.829
0.819
0.810
0.801
0.791
0.791
0.782
0.772
0.772
0.762
0.753
0.732
0.732
0.716
0.716
0.698
0.698
0.698
0.661
0.661
0.661
0.661
0.661
0.661
0.661
0.641
0.641
0.641
0.641
0.641
0.618
0.618
0.590
0.590
0.590
0.590
0.590

0.000
0.004
0.009
0.004
0.009
0.005
0.009
0.009
0.005
0.005
0.000
0.005
0.009
0.000
0.009
0.005
0.005
0.005
0.005
0.005
0.000
0.005
0.005
0.000
0.005
0.005
0.010
0.000
0.008
0.000
0.009
0.000
0.000
0.018
0.000
0.000
0.000
0.000
0.000
0.000
0.010
0.000
0.000
0.000
0.000
0.011
0.000
0.014
0.000
0.000
0.000
0.000

0.000
0.004
0.009
0.005
0.009
0.005
0.010
0.010
0.005
0.005
0.000
0.005
0.011
0.000
0.011
0.006
0.006
0.006
0.006
0.006
0.000
0.006
0.006
0.000
0.006
0.006
0.014
0.000
0.011
0.000
0.013
0.000
0.000
0.027
0.000
0.000
0.000
0.000
0.000
0.000
0.016
0.000
0.000
0.000
0.000
0.018
0.000
0.023
0.000
0.000
0.000
0.000

0.000
0.009
0.015
0.017
0.021
0.023
0.026
0.028
0.029
0.030
0.030
0.031
0.033
0.033
0.035
0.036
0.037
0.037
0.038
0.039
0.039
0.040
0.040
0.040
0.041
0.041
0.043
0.043
0.045
0.045
0.047
0.047
0.047
0.051
0.051
0.051
0.051
0.051
0.051
0.051
0.054
0.054
0.054
0.054
0.054
0.056
0.056
0.060
0.060
0.060
0.060
0.060

0.000
0.004
0.006
0.004
0.006
0.004
0.006
0.006
0.005
0.005
0.000
0.005
0.006
0.000
0.006
0.005
0.005
0.005
0.005
0.005
0.000
0.005
0.005
0.000
0.005
0.005
0.007
0.000
0.008
0.000
0.009
0.000
0.000
0.013
0.000
0.000
0.000
0.000
0.000
0.000
0.010
0.000
0.000
0.000
0.000
0.011
0.000
0.014
0.000
0.000
0.000
0.000

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-8. Life Table for UPSTATE Total Arrests (Felonies and Misdemeanors).


0.000
0.004
0.006
0.005
0.007
0.005
0.007
0.007
0.005
0.005
0.000
0.005
0.008
0.000
0.008
0.006
0.006
0.006
0.006
0.006
0.000
0.006
0.006
0.000
0.006
0.006
0.010
0.000
0.011
0.000
0.013
0.000
0.000
0.019
0.000
0.000
0.000
0.000
0.000
0.000
0.016
0.000
0.000
0.000
0.000
0.018
0.000
0.023
0.000
0.000
0.000
0.000

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
102

344
343
340
339
337
333
327
323
317
315
311
304
301
297
295
292
288
281
277
273
271
265
258
250
248
241
238
233
224
217
213
200
191
173
154
141
134
128
111
89
70
62
59
58
53
41
25
25
25
25
17
1

0
0
0
1
1
1
1
3
2
1
6
2
2
1
2
4
4
2
2
0
4
4
3
0
4
1
3
7
6
4
8
9
16
19
10
7
3
15
22
18
8
3
1
3
11
14
0
0
0
7
15
1

344.0
343.0
340.0
338.5
336.5
332.5
326.5
321.5
316.0
314.5
308.0
303.0
300.0
296.5
294.0
290.0
286.0
280.0
276.0
273.0
269.0
263.0
256.5
250.0
246.0
240.5
236.5
229.5
221.0
215.0
209.0
195.5
183.0
163.5
149.0
137.5
132.5
120.5
100.0
80.0
66.0
60.5
58.5
56.5
47.5
34.0
25.0
25.0
25.0
21.5
9.5
0.5

1
3
1
1
3
5
3
3
0
3
1
1
2
1
1
0
3
2
2
2
2
3
5
2
3
2
2
2
1
0
5
0
2
0
3
0
3
2
0
1
0
0
0
2
1
2
0
0
0
1
1
0

0.003
0.009
0.003
0.003
0.009
0.015
0.009
0.009
0.000
0.010
0.003
0.003
0.007
0.003
0.003
0.000
0.011
0.007
0.007
0.007
0.007
0.011
0.020
0.008
0.012
0.008
0.009
0.009
0.005
0.000
0.024
0.000
0.011
0.000
0.020
0.000
0.023
0.017
0.000
0.013
0.000
0.000
0.000
0.035
0.021
0.059
0.000
0.000
0.000
0.047
0.105
0.000

0.997
0.991
0.997
0.997
0.991
0.985
0.991
0.991
1.000
0.991
0.997
0.997
0.993
0.997
0.997
1.000
0.990
0.993
0.993
0.993
0.993
0.989
0.981
0.992
0.988
0.992
0.992
0.991
0.996
1.000
0.976
1.000
0.989
1.000
0.980
1.000
0.977
0.983
1.000
0.988
1.000
1.000
1.000
0.965
0.979
0.941
1.000
1.000
1.000
0.954
0.895
1.000

0.997
0.988
0.986
0.983
0.974
0.959
0.950
0.942
0.942
0.933
0.930
0.926
0.920
0.917
0.914
0.914
0.904
0.898
0.891
0.885
0.878
0.868
0.851
0.845
0.834
0.827
0.820
0.813
0.810
0.810
0.790
0.790
0.782
0.782
0.766
0.766
0.748
0.736
0.736
0.727
0.727
0.727
0.727
0.701
0.686
0.646
0.646
0.646
0.646
0.616
0.551
0.551

0.002
0.004
0.002
0.002
0.004
0.007
0.004
0.004
0.000
0.005
0.002
0.002
0.003
0.002
0.002
0.000
0.005
0.003
0.003
0.003
0.003
0.005
0.009
0.003
0.005
0.004
0.004
0.004
0.002
0.000
0.010
0.000
0.004
0.000
0.008
0.000
0.009
0.006
0.000
0.005
0.000
0.000
0.000
0.013
0.007
0.020
0.000
0.000
0.000
0.015
0.032
0.000

0.002
0.004
0.002
0.002
0.005
0.008
0.005
0.005
0.000
0.005
0.002
0.002
0.003
0.002
0.002
0.000
0.005
0.004
0.004
0.004
0.004
0.006
0.010
0.004
0.006
0.004
0.004
0.004
0.002
0.000
0.012
0.000
0.006
0.000
0.010
0.000
0.012
0.008
0.000
0.006
0.000
0.000
0.000
0.018
0.011
0.030
0.000
0.000
0.000
0.024
0.056
0.000

0.003
0.006
0.007
0.007
0.009
0.011
0.012
0.013
0.013
0.014
0.014
0.014
0.015
0.015
0.015
0.015
0.016
0.017
0.017
0.018
0.018
0.019
0.020
0.020
0.021
0.021
0.022
0.022
0.022
0.022
0.023
0.023
0.024
0.024
0.025
0.025
0.026
0.027
0.027
0.029
0.029
0.029
0.029
0.033
0.035
0.043
0.043
0.043
0.043
0.051
0.076
0.076

0.002
0.003
0.002
0.002
0.003
0.003
0.003
0.003
0.000
0.003
0.002
0.002
0.002
0.002
0.002
0.000
0.003
0.002
0.002
0.002
0.002
0.003
0.004
0.002
0.003
0.002
0.003
0.003
0.002
0.000
0.004
0.000
0.003
0.000
0.005
0.000
0.005
0.004
0.000
0.005
0.000
0.000
0.000
0.009
0.007
0.014
0.000
0.000
0.000
0.015
0.031
0.000

Vera Institute of Justice

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-9. Life Table for Greenlight Felony Arrests. 


0.002
0.003
0.002
0.002
0.003
0.003
0.003
0.003
0.000
0.003
0.002
0.002
0.002
0.002
0.002
0.000
0.003
0.003
0.003
0.003
0.003
0.003
0.004
0.003
0.004
0.003
0.003
0.003
0.002
0.000
0.005
0.000
0.004
0.000
0.006
0.000
0.007
0.006
0.000
0.006
0.000
0.000
0.000
0.013
0.011
0.021
0.000
0.000
0.000
0.024
0.056
0.000

97

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
102
104
106

98

278
276
273
272
272
269
267
264
261
258
256
254
251
245
243
240
237
237
235
232
232
228
226
224
219
217
213
207
203
195
179
166
147
128
120
114
107
94
82
70
51
45
43
41
39
37
28
16
16
15
15
10
2
2

0
0
0
0
2
2
1
1
2
1
1
2
4
1
1
2
0
1
2
0
0
2
1
2
2
3
4
2
5
15
13
18
17
7
6
6
13
11
11
18
5
1
1
0
2
8
12
0
1
0
5
8
0
2

Vera Institute of Justice

278.0
276.0
273.0
272.0
271.0
268.0
266.5
263.5
260.0
257.5
255.5
253.0
249.0
244.5
242.5
239.0
237.0
236.5
234.0
232.0
232.0
227.0
225.5
223.0
218.0
215.5
211.0
206.0
200.5
187.5
172.5
157.0
138.5
124.5
117.0
111.0
100.5
88.5
76.5
61.0
48.5
44.5
42.5
41.0
38.0
33.0
22.0
16.0
15.5
15.0
12.5
6.0
2.0
1.0

2
3
1
0
1
0
2
2
1
1
1
1
2
1
2
1
0
1
1
0
4
0
1
3
0
1
2
2
3
1
0
1
2
1
0
1
0
1
1
1
1
1
1
2
0
1
0
0
0
0
0
0
0
0

0.007
0.011
0.004
0.000
0.004
0.000
0.008
0.008
0.004
0.004
0.004
0.004
0.008
0.004
0.008
0.004
0.000
0.004
0.004
0.000
0.017
0.000
0.004
0.014
0.000
0.005
0.010
0.010
0.015
0.005
0.000
0.006
0.014
0.008
0.000
0.009
0.000
0.011
0.013
0.016
0.021
0.023
0.024
0.049
0.000
0.030
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.993
0.989
0.996
1.000
0.996
1.000
0.993
0.992
0.996
0.996
0.996
0.996
0.992
0.996
0.992
0.996
1.000
0.996
0.996
1.000
0.983
1.000
0.996
0.987
1.000
0.995
0.991
0.990
0.985
0.995
1.000
0.994
0.986
0.992
1.000
0.991
1.000
0.989
0.987
0.984
0.979
0.978
0.977
0.951
1.000
0.970
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000

0.993
0.982
0.978
0.978
0.975
0.975
0.968
0.960
0.957
0.953
0.949
0.945
0.938
0.934
0.926
0.922
0.922
0.918
0.914
0.914
0.899
0.899
0.895
0.883
0.883
0.879
0.870
0.862
0.849
0.844
0.844
0.839
0.827
0.820
0.820
0.813
0.813
0.804
0.793
0.780
0.764
0.747
0.729
0.694
0.694
0.673
0.673
0.673
0.673
0.673
0.673
0.673
0.673
0.673

0.004
0.005
0.002
0.000
0.002
0.000
0.004
0.004
0.002
0.002
0.002
0.002
0.004
0.002
0.004
0.002
0.000
0.002
0.002
0.000
0.008
0.000
0.002
0.006
0.000
0.002
0.004
0.004
0.006
0.002
0.000
0.003
0.006
0.003
0.000
0.004
0.000
0.005
0.005
0.007
0.008
0.009
0.009
0.018
0.000
0.011
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.004
0.006
0.002
0.000
0.002
0.000
0.004
0.004
0.002
0.002
0.002
0.002
0.004
0.002
0.004
0.002
0.000
0.002
0.002
0.000
0.009
0.000
0.002
0.007
0.000
0.002
0.005
0.005
0.008
0.003
0.000
0.003
0.007
0.004
0.000
0.005
0.000
0.006
0.007
0.008
0.010
0.011
0.012
0.025
0.000
0.015
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.005
0.008
0.009
0.009
0.009
0.009
0.011
0.012
0.012
0.013
0.013
0.014
0.015
0.015
0.016
0.016
0.016
0.017
0.017
0.017
0.019
0.019
0.019
0.020
0.020
0.020
0.021
0.022
0.022
0.023
0.023
0.023
0.024
0.025
0.025
0.026
0.026
0.027
0.029
0.031
0.034
0.038
0.041
0.046
0.046
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049
0.049

0.003
0.003
0.002
0.000
0.002
0.000
0.003
0.003
0.002
0.002
0.002
0.002
0.003
0.002
0.003
0.002
0.000
0.002
0.002
0.000
0.004
0.000
0.002
0.004
0.000
0.002
0.003
0.003
0.004
0.002
0.000
0.003
0.004
0.003
0.000
0.004
0.000
0.005
0.005
0.007
0.008
0.009
0.009
0.012
0.000
0.010
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-10. Life Table for TSP Felony Arrests.


0.003
0.003
0.002
0.000
0.002
0.000
0.003
0.003
0.002
0.002
0.002
0.002
0.003
0.002
0.003
0.002
0.000
0.002
0.002
0.000
0.004
0.000
0.002
0.004
0.000
0.002
0.003
0.003
0.004
0.003
0.000
0.003
0.005
0.004
0.000
0.005
0.000
0.006
0.007
0.008
0.010
0.011
0.012
0.018
0.000
0.015
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
102

113
113
113
112
111
110
107
106
106
106
104
101
100
100
100
99
98
97
94
93
93
93
92
91
90
89
88
78
61
50
48
48
46
44
43
41
41
39
38
38
36
32
31
31
31
29
29
25
23
23
19
7

0
0
0
0
0
2
0
0
0
1
2
0
0
0
1
0
0
3
1
0
0
0
0
1
0
1
10
17
11
2
0
1
2
1
2
0
2
1
0
2
2
1
0
0
2
0
4
1
0
4
11
7

113.0
113.0
113.0
112.0
111.0
109.0
107.0
106.0
106.0
105.5
103.0
101.0
100.0
100.0
99.5
99.0
98.0
95.5
93.5
93.0
93.0
93.0
92.0
90.5
90.0
88.5
83.0
69.5
55.5
49.0
48.0
47.5
45.0
43.5
42.0
41.0
40.0
38.5
38.0
37.0
35.0
31.5
31.0
31.0
30.0
29.0
27.0
24.5
23.0
21.0
13.5
3.5

0
0
1
1
1
1
1
0
0
1
1
1
0
0
0
1
1
0
0
0
0
1
1
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
1
0
0
1
0

0.000
0.000
0.009
0.009
0.009
0.009
0.009
0.000
0.000
0.010
0.010
0.010
0.000
0.000
0.000
0.010
0.010
0.000
0.000
0.000
0.000
0.011
0.011
0.000
0.011
0.000
0.000
0.000
0.000
0.000
0.000
0.021
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.057
0.000
0.000
0.000
0.000
0.000
0.000
0.041
0.000
0.000
0.074
0.000

1.000
1.000
0.991
0.991
0.991
0.991
0.991
1.000
1.000
0.991
0.990
0.990
1.000
1.000
1.000
0.990
0.990
1.000
1.000
1.000
1.000
0.989
0.989
1.000
0.989
1.000
1.000
1.000
1.000
1.000
1.000
0.979
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.943
1.000
1.000
1.000
1.000
1.000
1.000
0.959
1.000
1.000
0.926
1.000

1.000
1.000
0.991
0.982
0.974
0.965
0.956
0.956
0.956
0.946
0.937
0.928
0.928
0.928
0.928
0.919
0.909
0.909
0.909
0.909
0.909
0.900
0.890
0.890
0.880
0.880
0.880
0.880
0.880
0.880
0.880
0.861
0.861
0.861
0.861
0.861
0.861
0.861
0.861
0.861
0.812
0.812
0.812
0.812
0.812
0.812
0.812
0.779
0.779
0.779
0.721
0.721

0.000
0.000
0.004
0.004
0.004
0.005
0.005
0.000
0.000
0.005
0.005
0.005
0.000
0.000
0.000
0.005
0.005
0.000
0.000
0.000
0.000
0.005
0.005
0.000
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.025
0.000
0.000
0.000
0.000
0.000
0.000
0.017
0.000
0.000
0.029
0.000

0.000
0.000
0.004
0.005
0.005
0.005
0.005
0.000
0.000
0.005
0.005
0.005
0.000
0.000
0.000
0.005
0.005
0.000
0.000
0.000
0.000
0.005
0.006
0.000
0.006
0.000
0.000
0.000
0.000
0.000
0.000
0.011
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.029
0.000
0.000
0.000
0.000
0.000
0.000
0.021
0.000
0.000
0.039
0.000

0.000
0.000
0.009
0.012
0.015
0.017
0.020
0.020
0.020
0.021
0.023
0.025
0.025
0.025
0.025
0.026
0.027
0.027
0.027
0.027
0.027
0.029
0.030
0.030
0.031
0.031
0.031
0.031
0.031
0.031
0.031
0.036
0.036
0.036
0.036
0.036
0.036
0.036
0.036
0.036
0.048
0.048
0.048
0.048
0.048
0.048
0.048
0.056
0.056
0.056
0.076
0.076

0.000
0.000
0.004
0.004
0.004
0.004
0.005
0.000
0.000
0.005
0.005
0.005
0.000
0.000
0.000
0.005
0.005
0.000
0.000
0.000
0.000
0.005
0.005
0.000
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.017
0.000
0.000
0.000
0.000
0.000
0.000
0.016
0.000
0.000
0.028
0.000

Vera Institute of Justice

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-11. Life Table for UPSTATE Felony Arrests. 


0.000
0.000
0.004
0.005
0.005
0.005
0.005
0.000
0.000
0.005
0.005
0.005
0.000
0.000
0.000
0.005
0.005
0.000
0.000
0.000
0.000
0.005
0.006
0.000
0.006
0.000
0.000
0.000
0.000
0.000
0.000
0.011
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.021
0.000
0.000
0.000
0.000
0.000
0.000
0.021
0.000
0.000
0.038
0.000

99

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100

344
344
344
343
341
338
335
333
327
315
308
295
289
280
275
272
266
260
257
243
241
235
228
223
223
218
215
183
177
170
165
156
142
126
112
95
89
86
74
60
48
42
39
39
37
29
16
15
14
14
9

0
0
1
1
2
1
1
1
6
6
7
4
5
2
0
2
0
1
9
2
1
3
2
0
0
1
27
6
6
3
9
13
15
12
16
5
2
9
13
10
6
1
0
1
7
13
0
1
0
5
9

100 Vera Institute of Justice

344.0
344.0
343.5
342.5
340.0
337.5
334.5
332.5
324.0
312.0
304.5
293.0
286.5
279.0
275.0
271.0
266.0
259.5
252.5
242.0
240.5
233.5
227.0
223.0
223.0
217.5
201.5
180.0
174.0
168.5
160.5
149.5
134.5
120.0
104.0
92.5
88.0
81.5
67.5
55.0
45.0
41.5
39.0
38.5
33.5
22.5
16.0
14.5
14.0
11.5
4.5

0
0
0
1
1
2
1
5
6
1
6
2
4
3
3
4
6
2
5
0
5
4
3
0
5
2
5
0
1
2
0
1
1
2
1
1
1
3
1
2
0
2
0
1
1
0
1
0
0
0
0

0.000
0.000
0.000
0.003
0.003
0.006
0.003
0.015
0.019
0.003
0.020
0.007
0.014
0.011
0.011
0.015
0.023
0.008
0.020
0.000
0.021
0.017
0.013
0.000
0.022
0.009
0.025
0.000
0.006
0.012
0.000
0.007
0.007
0.017
0.010
0.011
0.011
0.037
0.015
0.036
0.000
0.048
0.000
0.026
0.030
0.000
0.063
0.000
0.000
0.000
0.000

1.000
1.000
1.000
0.997
0.997
0.994
0.997
0.985
0.982
0.997
0.980
0.993
0.986
0.989
0.989
0.985
0.977
0.992
0.980
1.000
0.979
0.983
0.987
1.000
0.978
0.991
0.975
1.000
0.994
0.988
1.000
0.993
0.993
0.983
0.990
0.989
0.989
0.963
0.985
0.964
1.000
0.952
1.000
0.974
0.970
1.000
0.938
1.000
1.000
1.000
1.000

1.000
1.000
1.000
0.997
0.994
0.988
0.985
0.971
0.953
0.950
0.931
0.924
0.912
0.902
0.892
0.879
0.859
0.852
0.835
0.835
0.818
0.804
0.793
0.793
0.776
0.768
0.749
0.749
0.745
0.736
0.736
0.731
0.726
0.714
0.707
0.699
0.691
0.666
0.656
0.632
0.632
0.602
0.602
0.586
0.569
0.569
0.533
0.533
0.533
0.533
0.533

0.000
0.000
0.000
0.002
0.002
0.003
0.002
0.007
0.009
0.002
0.009
0.003
0.007
0.005
0.005
0.007
0.010
0.003
0.008
0.000
0.009
0.007
0.005
0.000
0.009
0.004
0.010
0.000
0.002
0.004
0.000
0.003
0.003
0.006
0.003
0.004
0.004
0.013
0.005
0.012
0.000
0.015
0.000
0.008
0.009
0.000
0.018
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.002
0.002
0.003
0.002
0.008
0.009
0.002
0.010
0.003
0.007
0.005
0.006
0.007
0.011
0.004
0.010
0.000
0.011
0.009
0.007
0.000
0.011
0.005
0.013
0.000
0.003
0.006
0.000
0.003
0.004
0.008
0.005
0.005
0.006
0.019
0.008
0.019
0.000
0.025
0.000
0.013
0.015
0.000
0.032
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.003
0.004
0.006
0.007
0.009
0.012
0.012
0.014
0.015
0.016
0.017
0.017
0.018
0.020
0.020
0.021
0.021
0.022
0.023
0.023
0.023
0.024
0.024
0.025
0.025
0.025
0.026
0.026
0.026
0.026
0.027
0.028
0.029
0.029
0.032
0.033
0.036
0.036
0.040
0.040
0.042
0.044
0.044
0.054
0.054
0.054
0.054
0.054

0.000
0.000
0.000
0.002
0.002
0.002
0.002
0.003
0.004
0.002
0.004
0.002
0.003
0.003
0.003
0.003
0.004
0.002
0.004
0.000
0.004
0.004
0.003
0.000
0.004
0.003
0.004
0.000
0.002
0.003
0.000
0.003
0.003
0.004
0.003
0.004
0.004
0.007
0.005
0.008
0.000
0.011
0.000
0.008
0.009
0.000
0.017
0.000
0.000
0.000
0.000

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-12. Life Table for Greenlight Parole Revocations (Reimprisonment). 


0.000
0.000
0.000
0.002
0.002
0.002
0.002
0.003
0.004
0.002
0.004
0.002
0.004
0.003
0.003
0.004
0.005
0.003
0.005
0.000
0.005
0.004
0.004
0.000
0.005
0.003
0.006
0.000
0.003
0.004
0.000
0.003
0.004
0.006
0.005
0.005
0.006
0.011
0.008
0.013
0.000
0.018
0.000
0.013
0.015
0.000
0.032
0.000
0.000
0.000
0.000

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
102
104
106

278
278
278
277
277
275
270
269
266
260
252
247
240
235
231
227
224
221
220
213
210
208
201
198
195
192
190
166
165
160
144
131
113
97
91
81
78
71
59
50
32
29
27
25
25
23
17
11
11
10
10
7
1
1

0
0
1
0
0
2
0
0
4
6
2
5
1
1
3
0
1
0
4
2
1
3
2
0
0
0
21
0
4
15
10
15
15
5
10
3
7
11
9
18
3
1
0
0
1
6
6
0
1
0
3
6
0
1

278.0
278.0
277.5
277.0
277.0
274.0
270.0
269.0
264.0
257.0
251.0
244.5
239.5
234.5
229.5
227.0
223.5
221.0
218.0
212.0
209.5
206.5
200.0
198.0
195.0
192.0
179.5
166.0
163.0
152.5
139.0
123.5
105.5
94.5
86.0
79.5
74.5
65.5
54.5
41.0
30.5
28.5
27.0
25.0
24.5
20.0
14.0
11.0
10.5
10.0
8.5
4.0
1.0
0.5

0
0
0
0
2
3
1
3
2
2
3
2
4
3
1
3
2
1
3
1
1
4
1
3
3
2
3
1
1
1
3
3
1
1
0
0
0
1
0
0
0
1
2
0
1
0
0
0
0
0
0
0
0
0

0.000
0.000
0.000
0.000
0.007
0.011
0.004
0.011
0.008
0.008
0.012
0.008
0.017
0.013
0.004
0.013
0.009
0.005
0.014
0.005
0.005
0.019
0.005
0.015
0.015
0.010
0.017
0.006
0.006
0.007
0.022
0.024
0.010
0.011
0.000
0.000
0.000
0.015
0.000
0.000
0.000
0.035
0.074
0.000
0.041
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

1.000
1.000
1.000
1.000
0.993
0.989
0.996
0.989
0.992
0.992
0.988
0.992
0.983
0.987
0.996
0.987
0.991
0.996
0.986
0.995
0.995
0.981
0.995
0.985
0.985
0.990
0.983
0.994
0.994
0.993
0.978
0.976
0.991
0.989
1.000
1.000
1.000
0.985
1.000
1.000
1.000
0.965
0.926
1.000
0.959
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000

1.000
1.000
1.000
1.000
0.993
0.982
0.978
0.967
0.960
0.953
0.941
0.934
0.918
0.906
0.902
0.890
0.882
0.878
0.866
0.862
0.858
0.841
0.837
0.825
0.812
0.803
0.790
0.785
0.780
0.775
0.759
0.740
0.733
0.725
0.725
0.725
0.725
0.714
0.714
0.714
0.714
0.689
0.638
0.638
0.612
0.612
0.612
0.612
0.612
0.612
0.612
0.612
0.612
0.612

0.000
0.000
0.000
0.000
0.004
0.005
0.002
0.006
0.004
0.004
0.006
0.004
0.008
0.006
0.002
0.006
0.004
0.002
0.006
0.002
0.002
0.008
0.002
0.006
0.006
0.004
0.007
0.002
0.002
0.003
0.008
0.009
0.004
0.004
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.013
0.026
0.000
0.013
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.004
0.006
0.002
0.006
0.004
0.004
0.006
0.004
0.008
0.006
0.002
0.007
0.005
0.002
0.007
0.002
0.002
0.010
0.003
0.008
0.008
0.005
0.008
0.003
0.003
0.003
0.011
0.012
0.005
0.005
0.000
0.000
0.000
0.008
0.000
0.000
0.000
0.018
0.039
0.000
0.021
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.005
0.008
0.009
0.011
0.012
0.013
0.014
0.015
0.017
0.018
0.018
0.019
0.020
0.020
0.021
0.021
0.022
0.023
0.023
0.024
0.025
0.025
0.026
0.026
0.026
0.027
0.028
0.029
0.030
0.030
0.030
0.030
0.030
0.032
0.032
0.032
0.032
0.039
0.050
0.050
0.055
0.055
0.055
0.055
0.055
0.055
0.055
0.055
0.055
0.055

0.000
0.000
0.000
0.000
0.003
0.003
0.002
0.003
0.003
0.003
0.003
0.003
0.004
0.003
0.002
0.003
0.003
0.002
0.004
0.002
0.002
0.004
0.002
0.004
0.004
0.003
0.004
0.002
0.002
0.003
0.005
0.005
0.004
0.004
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.012
0.017
0.000
0.013
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

Vera Institute of Justice

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-13. Life Table for TSP Parole Revocations (Reimprisonment).


0.000
0.000
0.000
0.000
0.003
0.003
0.002
0.003
0.003
0.003
0.004
0.003
0.004
0.004
0.002
0.004
0.003
0.002
0.004
0.002
0.002
0.005
0.003
0.004
0.005
0.004
0.005
0.003
0.003
0.003
0.006
0.007
0.005
0.005
0.000
0.000
0.000
0.008
0.000
0.000
0.000
0.018
0.027
0.000
0.021
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

101

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
102

113
113
113
112
112
112
110
110
108
107
105
102
100
97
97
95
95
95
92
91
90
90
89
87
86
86
84
70
53
41
37
37
36
34
31
29
29
28
27
27
24
22
21
21
21
20
20
18
18
18
14
7

0
0
1
0
0
0
0
2
1
1
0
1
2
0
1
0
0
0
0
1
0
1
2
0
0
2
14
17
12
4
0
0
2
1
2
0
1
1
0
3
2
0
0
0
0
0
2
0
0
4
7
7

102 Vera Institute of Justice

113.0
113.0
112.5
112.0
112.0
112.0
110.0
109.0
107.5
106.5
105.0
101.5
99.0
97.0
96.5
95.0
95.0
95.0
92.0
90.5
90.0
89.5
88.0
87.0
86.0
85.0
77.0
61.5
47.0
39.0
37.0
37.0
35.0
33.5
30.0
29.0
28.5
27.5
27.0
25.5
23.0
22.0
21.0
21.0
21.0
20.0
19.0
18.0
18.0
16.0
10.5
3.5

0
0
0
0
0
2
0
0
0
1
3
1
1
0
1
0
0
3
1
0
0
0
0
1
0
0
0
0
0
0
0
1
0
2
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0

0.000
0.000
0.000
0.000
0.000
0.018
0.000
0.000
0.000
0.009
0.029
0.010
0.010
0.000
0.010
0.000
0.000
0.032
0.011
0.000
0.000
0.000
0.000
0.012
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.027
0.000
0.060
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.046
0.000
0.000
0.048
0.000
0.000
0.000
0.000
0.000
0.000
0.000

1.000
1.000
1.000
1.000
1.000
0.982
1.000
1.000
1.000
0.991
0.971
0.990
0.990
1.000
0.990
1.000
1.000
0.968
0.989
1.000
1.000
1.000
1.000
0.989
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.973
1.000
0.940
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.955
1.000
1.000
0.952
1.000
1.000
1.000
1.000
1.000
1.000
1.000

1.000
1.000
1.000
1.000
1.000
0.982
0.982
0.982
0.982
0.973
0.945
0.936
0.926
0.926
0.917
0.917
0.917
0.888
0.878
0.878
0.878
0.878
0.878
0.868
0.868
0.868
0.868
0.868
0.868
0.868
0.868
0.845
0.845
0.794
0.794
0.794
0.794
0.794
0.794
0.794
0.794
0.758
0.758
0.758
0.722
0.722
0.722
0.722
0.722
0.722
0.722
0.722

0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.000
0.000
0.005
0.014
0.005
0.005
0.000
0.005
0.000
0.000
0.015
0.005
0.000
0.000
0.000
0.000
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.012
0.000
0.025
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.018
0.000
0.000
0.018
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.000
0.000
0.005
0.015
0.005
0.005
0.000
0.005
0.000
0.000
0.016
0.006
0.000
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.014
0.000
0.031
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.023
0.000
0.000
0.024
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.000
0.013
0.013
0.013
0.013
0.015
0.022
0.024
0.025
0.025
0.027
0.027
0.027
0.031
0.032
0.032
0.032
0.032
0.032
0.033
0.033
0.033
0.033
0.033
0.033
0.033
0.033
0.040
0.040
0.051
0.051
0.051
0.051
0.051
0.051
0.051
0.051
0.060
0.060
0.060
0.067
0.067
0.067
0.067
0.067
0.067
0.067
0.067

0.000
0.000
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.005
0.008
0.005
0.005
0.000
0.005
0.000
0.000
0.008
0.005
0.000
0.000
0.000
0.000
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.012
0.000
0.017
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.018
0.000
0.000
0.018
0.000
0.000
0.000
0.000
0.000
0.000
0.000

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-14. Life Table for UPSTATE Parole Revocations (Reimprisonment). 


0.000
0.000
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.005
0.008
0.005
0.005
0.000
0.005
0.000
0.000
0.009
0.006
0.000
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.014
0.000
0.022
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.023
0.000
0.000
0.024
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100

344
344
344
343
341
338
335
333
327
315
308
295
289
280
275
272
266
260
257
243
241
235
228
223
223
218
215
183
177
170
165
156
142
126
112
95
89
86
74
60
48
42
39
39
37
29
16
15
14
14
9

0
0
1
2
2
2
2
6
12
7
10
5
8
4
2
4
3
1
13
2
2
7
4
0
3
2
29
6
6
5
9
14
15
13
17
6
3
12
14
11
6
2
0
2
8
13
1
1
0
5
9

344.0
344.0
343.5
342.0
340.0
337.0
334.0
330.0
321.0
311.5
303.0
292.5
285.0
278.0
274.0
270.0
264.5
259.5
250.5
242.0
240.0
231.5
226.0
223.0
221.5
217.0
200.5
180.0
174.0
167.5
160.5
149.0
134.5
119.5
103.5
92.0
87.5
80.0
67.0
54.5
45.0
41.0
39.0
38.0
33.0
22.5
15.5
14.5
14.0
11.5
4.5

0
0
0
0
1
1
0
0
0
0
3
1
1
1
1
2
3
2
1
0
4
0
1
0
2
1
3
0
1
0
0
0
1
1
0
0
0
0
0
1
0
1
0
0
0
0
0
0
0
0
0

0.000
0.000
0.000
0.000
0.003
0.003
0.000
0.000
0.000
0.000
0.010
0.003
0.004
0.004
0.004
0.007
0.011
0.008
0.004
0.000
0.017
0.000
0.004
0.000
0.009
0.005
0.015
0.000
0.006
0.000
0.000
0.000
0.007
0.008
0.000
0.000
0.000
0.000
0.000
0.018
0.000
0.024
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

1.000
1.000
1.000
1.000
0.997
0.997
1.000
1.000
1.000
1.000
0.990
0.997
0.997
0.996
0.996
0.993
0.989
0.992
0.996
1.000
0.983
1.000
0.996
1.000
0.991
0.995
0.985
1.000
0.994
1.000
1.000
1.000
0.993
0.992
1.000
1.000
1.000
1.000
1.000
0.982
1.000
0.976
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000

1.000
1.000
1.000
1.000
0.997
0.994
0.994
0.994
0.994
0.994
0.984
0.981
0.978
0.974
0.970
0.963
0.952
0.945
0.941
0.941
0.926
0.926
0.921
0.921
0.913
0.909
0.895
0.895
0.890
0.890
0.890
0.890
0.884
0.876
0.876
0.876
0.876
0.876
0.876
0.860
0.860
0.839
0.839
0.839
0.839
0.839
0.839
0.839
0.839
0.839
0.839

0.000
0.000
0.000
0.000
0.002
0.002
0.000
0.000
0.000
0.000
0.005
0.002
0.002
0.002
0.002
0.004
0.006
0.004
0.002
0.000
0.008
0.000
0.002
0.000
0.004
0.002
0.007
0.000
0.003
0.000
0.000
0.000
0.003
0.004
0.000
0.000
0.000
0.000
0.000
0.008
0.000
0.011
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.002
0.002
0.000
0.000
0.000
0.000
0.005
0.002
0.002
0.002
0.002
0.004
0.006
0.004
0.002
0.000
0.008
0.000
0.002
0.000
0.005
0.002
0.008
0.000
0.003
0.000
0.000
0.000
0.004
0.004
0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.012
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.003
0.004
0.004
0.004
0.004
0.004
0.007
0.008
0.008
0.009
0.010
0.011
0.013
0.013
0.014
0.014
0.016
0.016
0.016
0.016
0.017
0.018
0.019
0.019
0.020
0.020
0.020
0.020
0.020
0.022
0.022
0.022
0.022
0.022
0.022
0.027
0.027
0.033
0.033
0.033
0.033
0.033
0.033
0.033
0.033
0.033
0.033

0.000
0.000
0.000
0.000
0.002
0.002
0.000
0.000
0.000
0.000
0.003
0.002
0.002
0.002
0.002
0.003
0.003
0.003
0.002
0.000
0.004
0.000
0.002
0.000
0.003
0.002
0.004
0.000
0.003
0.000
0.000
0.000
0.003
0.004
0.000
0.000
0.000
0.000
0.000
0.008
0.000
0.010
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

Vera Institute of Justice

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-15. Life Table for Greenlight Parole Revocations (Technical Revocations only). 


0.000
0.000
0.000
0.000
0.002
0.002
0.000
0.000
0.000
0.000
0.003
0.002
0.002
0.002
0.002
0.003
0.003
0.003
0.002
0.000
0.004
0.000
0.002
0.000
0.003
0.002
0.004
0.000
0.003
0.000
0.000
0.000
0.004
0.004
0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.012
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

103

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
102
104
106

278
278
278
277
277
275
270
269
266
260
252
247
240
235
231
227
224
221
220
213
210
208
201
198
195
192
190
166
165
160
144
131
113
97
91
81
78
71
59
50
32
29
27
25
25
23
17
11
11
10
10
7
1
1

0
0
1
0
2
5
0
3
4
8
5
6
1
3
3
1
3
1
6
3
2
5
2
1
2
0
22
0
4
15
12
17
15
6
10
3
7
12
9
18
3
2
1
0
2
6
6
0
1
0
3
6
0
1

104 Vera Institute of Justice

278.0
278.0
277.5
277.0
276.0
272.5
270.0
267.5
264.0
256.0
249.5
244.0
239.5
233.5
229.5
226.5
222.5
220.5
217.0
211.5
209.0
205.5
200.0
197.5
194.0
192.0
179.0
166.0
163.0
152.5
138.0
122.5
105.5
94.0
86.0
79.5
74.5
65.0
54.5
41.0
30.5
28.0
26.5
25.0
24.0
20.0
14.0
11.0
10.5
10.0
8.5
4.0
1.0
0.5

0
0
0
0
0
0
1
0
2
0
0
1
4
1
1
2
0
0
1
0
0
2
1
2
1
2
2
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0

0.000
0.000
0.000
0.000
0.000
0.000
0.004
0.000
0.008
0.000
0.000
0.004
0.017
0.004
0.004
0.009
0.000
0.000
0.005
0.000
0.000
0.010
0.005
0.010
0.005
0.010
0.011
0.006
0.006
0.007
0.007
0.008
0.010
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.038
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

1.000
1.000
1.000
1.000
1.000
1.000
0.996
1.000
0.992
1.000
1.000
0.996
0.983
0.996
0.996
0.991
1.000
1.000
0.995
1.000
1.000
0.990
0.995
0.990
0.995
0.990
0.989
0.994
0.994
0.993
0.993
0.992
0.991
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.962
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000

1.000
1.000
1.000
1.000
1.000
1.000
0.996
0.996
0.989
0.989
0.989
0.985
0.968
0.964
0.960
0.951
0.951
0.951
0.947
0.947
0.947
0.938
0.933
0.924
0.919
0.909
0.899
0.894
0.888
0.883
0.876
0.869
0.861
0.861
0.861
0.861
0.861
0.861
0.861
0.861
0.861
0.861
0.828
0.828
0.828
0.828
0.828
0.828
0.828
0.828
0.828
0.828
0.828
0.828

0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.000
0.004
0.000
0.000
0.002
0.008
0.002
0.002
0.004
0.000
0.000
0.002
0.000
0.000
0.005
0.002
0.005
0.002
0.005
0.005
0.003
0.003
0.003
0.003
0.004
0.004
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.016
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.000
0.004
0.000
0.000
0.002
0.008
0.002
0.002
0.004
0.000
0.000
0.002
0.000
0.000
0.005
0.003
0.005
0.003
0.005
0.006
0.003
0.003
0.003
0.004
0.004
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.019
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.000
0.000
0.004
0.004
0.007
0.007
0.007
0.008
0.011
0.012
0.012
0.014
0.014
0.014
0.014
0.014
0.014
0.016
0.016
0.017
0.018
0.019
0.020
0.021
0.021
0.022
0.023
0.024
0.025
0.025
0.025
0.025
0.025
0.025
0.025
0.025
0.025
0.025
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040
0.040

0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.000
0.003
0.000
0.000
0.002
0.004
0.002
0.002
0.003
0.000
0.000
0.002
0.000
0.000
0.003
0.002
0.003
0.002
0.003
0.004
0.003
0.003
0.003
0.003
0.004
0.004
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.016
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-16. Life Table for TSP Parole Revocations (Technical Revocations only).


0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.000
0.003
0.000
0.000
0.002
0.004
0.002
0.002
0.003
0.000
0.000
0.002
0.000
0.000
0.004
0.003
0.004
0.003
0.004
0.004
0.003
0.003
0.003
0.004
0.004
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.019
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
54
56
58
60
62
64
66
68
70
72
74
76
78
80
82
84
86
88
90
92
94
96
98
100
102

113
113
113
112
112
112
110
110
108
107
105
102
100
97
97
95
95
95
92
91
90
90
89
87
86
86
84
70
53
41
37
37
36
34
31
29
29
28
27
27
24
22
21
21
21
20
20
18
18
18
14
7

0
0
1
0
0
0
0
2
1
2
1
2
3
0
2
0
0
0
1
1
0
1
2
0
0
2
14
17
12
4
0
1
2
3
2
0
1
1
0
3
2
0
0
0
0
0
2
0
0
4
7
7

113.0
113.0
112.5
112.0
112.0
112.0
110.0
109.0
107.5
106.0
104.5
101.0
98.5
97.0
96.0
95.0
95.0
95.0
91.5
90.5
90.0
89.5
88.0
87.0
86.0
85.0
77.0
61.5
47.0
39.0
37.0
36.5
35.0
32.5
30.0
29.0
28.5
27.5
27.0
25.5
23.0
22.0
21.0
21.0
21.0
20.0
19.0
18.0
18.0
16.0
10.5
3.5

0
0
0
0
0
2
0
0
0
0
2
0
0
0
0
0
0
3
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0

0.000
0.000
0.000
0.000
0.000
0.018
0.000
0.000
0.000
0.000
0.019
0.000
0.000
0.000
0.000
0.000
0.000
0.032
0.000
0.000
0.000
0.000
0.000
0.012
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.046
0.000
0.000
0.048
0.000
0.000
0.000
0.000
0.000
0.000
0.000

1.000
1.000
1.000
1.000
1.000
0.982
1.000
1.000
1.000
1.000
0.981
1.000
1.000
1.000
1.000
1.000
1.000
0.968
1.000
1.000
1.000
1.000
1.000
0.989
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
1.000
0.955
1.000
1.000
0.952
1.000
1.000
1.000
1.000
1.000
1.000
1.000

1.000
1.000
1.000
1.000
1.000
0.982
0.982
0.982
0.982
0.982
0.963
0.963
0.963
0.963
0.963
0.963
0.963
0.933
0.933
0.933
0.933
0.933
0.933
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.922
0.880
0.880
0.880
0.838
0.838
0.838
0.838
0.838
0.838
0.838
0.838

0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.000
0.000
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.000
0.015
0.000
0.000
0.000
0.000
0.000
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.021
0.000
0.000
0.021
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.000
0.000
0.000
0.010
0.000
0.000
0.000
0.000
0.000
0.000
0.016
0.000
0.000
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.023
0.000
0.000
0.024
0.000
0.000
0.000
0.000
0.000
0.000
0.000

0.000
0.000
0.000
0.000
0.000
0.013
0.013
0.013
0.013
0.013
0.018
0.018
0.018
0.018
0.018
0.018
0.018
0.025
0.025
0.025
0.025
0.025
0.025
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.027
0.048
0.048
0.048
0.061
0.061
0.061
0.061
0.061
0.061
0.061
0.061

0.000
0.000
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.000
0.007
0.000
0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.005
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.021
0.000
0.000
0.021
0.000
0.000
0.000
0.000
0.000
0.000
0.000

Vera Institute of Justice

SE Hazard
Rate

SE Prob
Density

SE Cum
Survival

Hazard Rate

Probability
Density

Cumulative
Survival

Proportion
Surviving

Proportion
Terminating

# Terminal
Events

# Exposed
to Risk

# W/Drawn
during
Interval

# Entering
Interval

Interval Start

Table C-17. Life Table for UPSTATE Parole Revocations (Technical Revocations only). 


0.000
0.000
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.000
0.007
0.000
0.000
0.000
0.000
0.000
0.000
0.009
0.000
0.000
0.000
0.000
0.000
0.006
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.023
0.000
0.000
0.024
0.000
0.000
0.000
0.000
0.000
0.000
0.000

105

48

56

64
Weeks at Risk

72

80

88

96

104

0
40

0.2
32

100

0.3

24

200

0.4

16

300

0.5

8

400

0.6

0

500

0.7

U P I nterva l N

TSP I nterva l N

700

600

G L I nterva l N

TSP Surv iva l

U P Surv iva l

G L Surv iva l

800

0.8

0.9

1.0

Figure C-1. Cumulative survival (life table) for total arrests, and sample Ns by risk interval (N entering
interval), for total followup period.

106 Vera Institute of Justice

P rop ortion S u rv iv in g

Cumulative Survival Figures

N E n terin g In terv al

Cumulative Hazard

0.00

0.05

0.10

0.15

0.20

0.25

0

4

8

12

UP Cum Haz

TSP Cum Haz

GL Cum Haz

16

20

28

Weeks at Risk

24

32

36

44

48

52

0.1386

0.156

0.2087

Vera Institute of Justice

40

Figure C-2. Cumulative hazard rates (life table) for total arrests, one-year post release.


107

24

32

40

48

56

64

72

80

88

96

104

Weeks at Risk

0

16

0.3

8

100

0.4

0

200

0.5

400

300

U P I nterva l N

TSP I nterva l N

0.6

0.7

G L I nterva l N

U P Surv iva l

500

0.8
TSP Surv iva l

600

0.9

G L Surv iva l

700

1.0

Figure C-3. Cumulative survival (life table) for felony arrests and sample Ns by risk interval (N
entering interval) for total followup period.

108 Vera Institute of Justice

P rop o rtion Surviving

Cumulative Hazard

0

0.05

0.1

0.15

0.2

0.25

0

4

8

12

UP Cum Haz

TSP Cum Haz

GL Cum Haz

16

20

28

Weeks at Risk

24

32

36

44

48

Vera Institute of Justice

40

Figure C4. Cumulative hazard rates for felony arrests, one year post-release.


52

0.0641

0.0693

0.099

109

32

40

48
56
Weeks at Risk

64

72

80

88

96

104

0

0.3
24

100

0.4

16

200

0.5

8

300

0.6

500

600

400

0

UP Interval N

GL Interval N

UP Survival
TSP Interval N

TSP Survival

GL Survival

700

0.7

0.8

0.9

1.0

110 Vera Institute of Justice

Proportion Surviving

Figure C-5. Cumulative survival (life table estimates) for parole revocations (reimprisonment)
and sample Ns by risk interval (N entering interval) for full followup period.

N Entering Interval

Cumulative Hazard

0

0.05

0.1

0.15

0.2

0.25

0

4

8

12

UP Cum Haz

TSP Cum Haz

GL Cum Haz

16

20

28

Weeks at Risk

24

32

36

44

48

Vera Institute of Justice

40

Figure C6. Cumulative hazard rates (life table estimates) for parole revocations
(reimprisonment) for one year after release.

52

0.0708

0.1179

0.1443

111

8

16

24

32

40

48

56

64

Weeks at Risk

72

80

88

96

104

0

0.4

0

100

0.5

UP Interval N

TSP Interval N

300

200

GL Interval N

UP Survival

0.6

0.7

400

0.8
TSP Survival

500

0.9

GL Survival

600

1

112 Vera Institute of Justice

Proportion Surviving

Figure C-7. Cumulative survival (life-table estimates) for revocations independent of arrest and
sample Ns by risk interval (N entering interval) for full followup period.

N Entering Interval

Tables: Logistic Regression Models

Table C-15. Logistic Regression of Demographic Predictors and
Criminal History on Total Arrests.
Variables
GL/TSP
GL/Upstate
Exp (B)
Exp (B)
b
b
-.05
.95***
- .05
.95***
Age at release
-.02
.98
- .01
.99
Education
Ethnicity
--------NH White/Other
.30**
-1.21
.88
-.13
Black
.37*
- .99
.90
-.11
Hispanic
.08
1.08***
.10
1.10***
Prior arrests
Primary offense
--------Robbery
.82
- .20
.79
-.23
Violent
1.49
.40
.64
1.89***
Drugs
1.24
.22
.34
1.40
Property
.88
- .13
.49**
-.71
Other
Substance abuse
------None
--1.61
.47
.71
2.03**
Alcohol
1.22
.20
1.19
.17
Drugs
1.67*
.51
1.30
.27
Alcohol and drugs
-.01
.99
- .08
.92***
Age at first arrest
-.37
.69**
.40
.67
Study Group
*p<.10; **p<.05; ***p<.01; Entries are unstandardized coefficients.

Vera Institute of Justice

113

Table C-16. Logistic Regression of Demographic Predictors and
Criminal History on Felony Arrests.
Variables
GL/TSP
GL/Upstate
B
Exp (B)
b
Exp (B)
Age at release
Education
Ethnicity
NH White/Other
Black
Hispanic
Prior arrests
Primary offense
Robbery
Violent
Drugs
Property
Other
Substance abuse
None
Alcohol
Drugs
Alcohol and drugs
Age at first arrest
Study Group

-.04
.05

.96**
1.05

-.02
.01

.98
1.01

---.17
.02
.04

--.85
1.02
1.04***

---1.10
-.99
.04

--.33**
.37*
1.04***

---.15
.49
.64
-.35

--.86
1.64*
1.89**
.70

---.06
.33
.64
.12

--.95
1.39
1.90*
1.13

--.24
.38
.06
-.02
-.27

--1.28
1.46
1.06
.98
.76

--.23
.61
.35
-.06
-.53

--1.26
1.84**
1.42
.94*
.59*

*p<.10; **p<.05; ***p<.01; b’s are unstandardized coefficients. 


114 Vera Institute of Justice

Table C-17: Logistic Regression of Demographic Predictors and
Criminal History on Parole Revocations.
Variables
GL/TSP
GL/Upstate
(N=586)
(N=428)
B
Exp (B)
b
Exp (B)
-.02
.98
-.04
.96**
Age at release
.02
1.02
-.03
.97
Education
Ethnicity
--------NH White/Other
.94
-.07
.91
-.10
Black
.72
-.33
.75
-.29
Hispanic
.05
1.05***
.07
1.07***
Prior arrests
Primary offense
--------Robbery
1.04
.04
.84
-.18
Violent
.76
-.28
.96
-.04
Drugs
1.32
.27
1.28
.24
Property
1.15
.14
1.14
.13
Other
Substance abuse
------None
--.88
-.13
1.13
.12
Alcohol
1.15
.14
1.34
.30
Drugs
1.19
.17
1.23
.20
Alcohol and drugs
-.02
.98
-.01
.99
Age at first arrest
-.20
.82
-.68
.51**
Study Group
-2 Log-Likelihood
Model Chi-Square
*p<.10; **p<.05; ***p<.01; b’s are unstandardized coefficients.

Vera Institute of Justice

115

Appendix D. Instruments and Forms
The Time 1 interview conducted in the first one to two months after release focused on the issues and
experiences during the first few months in the community, a period that has been described as part of the
“moment of transition.”159 The interview instrument asked questions related to program satisfaction about
the in-prison component of the programming, post-release relationships with family, use of community
services, participant perceptions of parole officers and parole supervision, and other social-psychological
based measures such as depression, anxiety, and locus of control. The survey instrument is derived from
an early version of an interview instrument designed by Urban Institute researchers for use in their 11state study of reentry processes and outcomes. Although comparability across states was an issue because
we perceive real value in being able to make these types of comparisons, we made extensive
modifications to the instrument to address the underlying issues in our study. We significantly shortened
the original instrument and added key measures from other sources which we discuss here.
Survey Instrument Scales: Design and Reliability

This section describes several of the scales used in the post-release survey instruments. For most of the
additions to the instrument, we relied almost exclusively on previously developed scales and measures
(we note that many of the questions and measures in the Urban Institute instrument were already
components in scales that we added—we discuss those as they relate to the specific scales included). The
discussion in this section, especially in regard to scale measures, reliability, and validity is based on prior
work—not on our results. Scale assessments for the current study are shown in the report text. It is
important to note that in many cases, we often reduced existing scales by several items. The primary
reason for this is that in trial administrations, our original survey instrument took almost two hours to
administer. This was much too long, and dropping items from multiple scales helped significantly reduce
the administration time. It was also our experience that we quickly tested the patience of our respondents
when too many similarly-oriented questions were included—thus we made the decision to reduce the
number of items where it seemed reasonable to do so under the assumption that it would be better to
reduce scale reliability and validity somewhat, but still obtain a completed interview. Too often,
respondents cut an interview short when their patience grew thin with what they see as redundant
questions. In all cases, we show the entire scale and designate which items were drawn from the existing
scale for use in our study. In most cases, we did not have inter-item coefficients to make distinctions
regarding which items were most important to the scales. Thus, in making decisions about which items to
delete from each scale, we relied on comparisons to other scales with similar items to see which were
most frequently used, and in some cases, relied specifically on the judgments of the principal investigator
and research staff.

159

See especially, Nelson, Deess and Allen, First Month Out, 1999; Travis, Jeremy, Amy L. Solomon and Michelle Waul.
From Prison to Home: The Dimensions and Consequences of Prisoner Reentry. Washington, DC: Urban Institute, 2001.
116 Vera Institute of Justice

RAND: The Medical Outcomes Study (MOS) Social Support Survey

This social support survey was developed for patients in the Medical Outcomes Study conducted by
RAND, a two-year study of patients with chronic conditions.160 The items in the MOS survey are short,
simple, and easy to understand. Multi-trait scaling analyses from prior research supported the
dimensionality of four functional support scales (emotional/informational, tangible, affectionate, and
positive social interaction) and the construction of an overall functional social support index. These
support measures are distinct from structural measures of social support and from related health measures.
They are reliable (all alphas > 0.91), and are fairly stable over time. Selected construct validity hypotheses
were supported. Table D-1 illustrates the questions that we used to assess support.
Table D-1: RAND MOS Social Support Survey: Emotional / Informational Support Scale. 

Emotional/Informational Support
* 	 You have someone you can count on to listen to you when you need to talk
You have someone to give you information to help you understand a situation
You have someone to give you good advice about a crisis
* 	 You have someone to confide in or talk to about yourself or your problems
* 	 You have someone whose advice you really want
* 	 You have someone to share your most private worries and fears with
* 	 You have someone to turn to for suggestions about how to deal with a personal problem
* 	 You have someone who understands your problems
Tangible support
* 	 You have someone who would provide help or advice on finding a place to live.
* 	 You have someone who would provide support for dealing with a substance abuse problem.
* 	 You have someone who would provide you with some financial support.

For the purposes of our instrument, we have sampled questions solely from the Emotional /
Informational Support Scale (alpha = 0.96). Furthermore, we specify the support system in question as
family (i.e., our question states “You have someone in your family you can count on to listen to you when
you need to talk”). We do this primarily because of the Greenlight project’s emphasis on the family
reintegration segment of the program. In addition to the six items adapted from the emotionalinformational support scales, we also included three items from a post-release interview instrument
developed by the Urban Institute that gauges “tangible” support specifically for a criminal justice
population. A higher score for an individual scale indicates more support. Items sampled by our Time 1
Interview are marked with asterisks. The RAND support scale uses a five-point scale (Values are 1

160

For more information, see http://www.rand.org/health/surveys/mos.descrip.html.
Vera Institute of Justice

117

through 5; None of the time, A little of the time, Some of the time, Most of the time, All of the time)
designed to assess how often a certain kind of support is available to the respondent.
TCU – CEST Survey (Client Evaluation of Self and Treatment)

This instrument was developed by the Institute of Behavioral Research at Texas Christian University and
was designed for use with participants in drug treatment programs.161 The survey provides client selfratings on motivation, psychological and social functioning—all representing indicators that have been
shown to predict outcomes during and following treatment—as well as including questions about the
treatment process and social support. The Greenlight Project Time 1 Interview, in its use of some
measures from the CEST survey, takes the emphasis off treatment. We sampled questions from the
following scales and subscales shown in Tables D-2 and D-3. The complete TCU scales from which we
drew relevant questions are also shown below. Again, those items used in the Time 1 Interview are noted
with asterisks.
The reliability coefficients shown in Table D-2 for each of the TCU scales are based on prior work.162
There are substantive differences across subpopulations in the inter-item coefficients such that again, we
relied on a combination of the strength of the coefficients and the judgments of the research team to make
decisions about which items to retain and which to discard.
Table D-2. TCU Treatment Research Scales
Scale
Subscale

Alpha Reliability

Social Functioning:

Social Conformity
Risk-Taking
Hostility

.63 - .65
.70 - .76
.78 - .83

Motivation:

Desire for Help

.72 - .82

Psychological Functioning:

Depression
Anxiety
Decision Making
Self-Efficacy

.77 - .78
.77 - .82
.74 - .81
.66

161

Knight, K., Holcom, M., & Simpson, D. D. TCU Psychosocial Functioning and Motivation Scales: Manual on
Psychometric Properties. Fort Worth: Texas Christian University, Institute of Behavioral Research, February, 1994.
Downloaded from http://www.ibr.tcu.edu/pubs/datacoll/cjforms.html#CorrRTForms on Thursday, May 08, 2003.
162
Knight, Holcom and Simpson, TCU Psychosocial, 1994:26-37.
118 Vera Institute of Justice

Table D-3. Social Functioning Scales
Social Conformity Scale
* 	 You feel people are important to you.
* 	 You feel honesty is required in every situation.
You have trouble following rules and laws. (R)
You depend on “things” more than people. (R)
* 	 You keep the same friends for a long time.
* 	 You work hard to keep a job.
* 	 Your religious beliefs are very important in your life.
Taking care of your family is very important.
Hostility
* 	 You feel mistreated by other people.
You like others to feel afraid of you.
You have urges to fight or hurt others.
* 	 You have a hot temper.
* 	 Your temper gets you into fights or other trouble.
* 	 You get mad at other people easily.
* 	 You have carried weapons, like knives or guns.
You feel a lot of anger inside you.
Risk-Taking
* 	 You like to take chances.
You like the “fast” life.
You like friends who are wild.
* 	 You like to do things that are strange or exciting.
* 	 You avoid anything dangerous. (R)
* 	 You only do things that feel safe. (R)
* 	 You are very careful and cautious. (R)

Two of the items taken from the TCU scales have been adapted to the specific needs of the Greenlight
Time 1 interview. The third question in the “Desire for Help” scale (Table D-4) now reads “You have
given up friends and hangouts that get you in trouble.” The fifth question in the same scale reads “You
are tired of the problems caused by the crimes you committed.” These are relatively simple changes that
we assume will not greatly alter the reliability of the scale.

Vera Institute of Justice

119

Table D-4. Motivation—TCU Desire for Help Scale
You need help in dealing with your drug use. 

It is urgent that you find help immediately for your drug use. 

* 	 You will give up your friends and hangouts to solve your drug problems.
* 	 Your life has gone out of control.
* 	 You are tired of the problems caused by drugs.
* 	 You want to get you life straightened out.

Table D-5. TCU Psychological Functioning Scales
Depression Scale
You feel interested in life. (R)
* 	 You feel sad or depressed.
You feel extra tired or run down.
You worry or brood a lot.
You have thoughts of committing suicide.
* 	 You feel lonely.
Anxiety Scale
* 	 You have trouble sleeping.
* 	 You have trouble concentrating or remembering things.
You feel afraid of certain things, like elevators, crowds, or going out alone.
* 	 You feel anxious of nervous.
* 	 You have trouble sitting still for long.
You feel tense or keyed-up.
You feel tightness or tension in your muscles.
Decision Making Scale
* 	 You consider how your actions will affect others.
You plan ahead.
You think about probable results of your actions.

120	 Vera Institute of Justice

* 	 You think about what causes your current problems.
You think of several different ways to solve a problem.
* 	 You have trouble making decisions. (R)
* 	 You make good decisions.
* 	 You make decisions without thinking about consequences. (R)
You analyze problems by looking at all the choices.
Self-Efficacy Scale
* 	 You have little control over the things that happen to you. (R)
* 	 What happens to you in the future mostly depends on you.
* 	 There is little you can do to change many of the important things in your life. (R)
* 	 There is really no way you can solve some of the problems you have. (R)
You can do just about anything you really set your mind to do.
* 	 Sometimes you feel that you are being pushed around in life. (R)
* 	 You often feel helpless in dealing with the problems of life. (R)
Furthermore, two questions were added to those selected from the TCU scales. “It is important to pay
off your debts” relates to issues of social conformity. This item has not been used in pre-existing scales
and a discussion of its predictive value is included in the body of the report. The second statement,
“Everything you do is an effort,” though not included in the TCU depression scale, has been used in preexisitng studies of depression, such as the CES-D (Center for Epidemiological Studies Depression)
survey.
TCU’s CEST scales are scored on a five-point Likert scale (From strongly agree to strongly disagree).
Scores for each subscale are obtained by summing responses to the corresponding set of items (after
reversing scores on reflected items – items marked with an “R” above – by subtracting the item response
from “6”), dividing the sum by the number of items in the subscale and multiplying by 10.
DAST-10 (DRUG ABUSE SCREENING TEST)

The DAST-10 gathers information regarding drug use/abuse, or problems related to drug use, not
including alcohol. It does not provide information on the types of drugs used, or on the frequency or
duration of drug use. The interview form includes separate questions to obtain this information. In
DAST, there is a question regarding multiple drug use, and some of the problems that can result from
drug use/abuse are surveyed: marital-family relationship problems and physical problems (medical
symptoms and conditions).
Vera Institute of Justice

121

The DAST-10 asks a series of ten "yes or no" questions. It is a unidimensional scale and yields one
score ranging from 0 to10, which is computed by summing all items that are endorsed in the direction of
increased drug problems. Though in the original, all items are keyed for a “yes” response, the Greenlight
interview rephrases an item so that it is keyed for a "No" response: "Have you been able to stop using
drugs when you want to?" Higher scores relate to a higher degree of problems related to drug abuse, with
6 or above indicating a substantial level of problems.
With the 20-item DAST, an internal consistency coefficient of .92 was obtained for a sample of 256
drug or alcohol-abuse clients. Adequate validity has been demonstrated by the fact that the DAST attained
85 percent overall accuracy in classifying clients according to DSM-III diagnosis, and was also
demonstrated by significant correlations of the DAST scores with frequency of various types of drugs
used during the preceding 12 months. Finally, although the DAST-10 usually refers to drug use within the
past 12 months, because our study sample was interviewed immediately after their release from prison,
we use the instrument with respect to the respondent’s entire history of drug use.

122 Vera Institute of Justice

TIME 1 INSTRUMENT 


TRANSITIONAL SERVICES
PROGRAMMING IN NEW YORK
STATE:
EVALUATING
PROJECT GREENLIGHT

st
1 Post-Release Interview

(Newest Version as of 1/17/03)

Vera Institute of Justice

123

PARTICIPANT TRACKING FORM
Project Greenlight
Respondent’s Full Name____________________________________________________________
Date of Birth (mm/dd/yyyy)________________________________________________________
Street Address_____________________________________________________________________
City____________________________________State_____________Zip______________________
Interview Date (mm/dd/yyyy)________________________________________________________
Vera ID

F F F

We may be interested in speaking with you again in the future to see how you are doing. Any
conversation we have in the future would be entirely up to you, and we would reimburse you just like
we pay you for this interview. You could also refuse to talk with us if you decide you don’t want to be
interviewed at that time. The contact people will only be asked how to contact you; we will not reveal
any information about why we want to speak with you or your participation in this research.

1. Are you willing to provide contact information?

(0) NO

(1) YES

Please provide us with the name, address and telephone number of one person who will know how to
locate you in the future.
Name____________________________________________________________________
Street_______________________________________________________________________
______________________________________________________________________
City_______________________________________State________________Zip___________
Telephone________________________________________

Interviewer:

The consent form and tracking form must be completed before beginning 

survey.

F

Check this box if consent form has been completed.


F

Check this box if tracking form has been completed.
Record the start time:
_________________ am F / pm F

124 Vera Institute of Justice

F F F

VERA ID

Before we begin, I want to thank you again for agreeing to speak with me today. The primary reason for this
interview is to hear about how your life has been since you were released from prison last month. I’ll be asking
you questions about different experiences you’ve had since your release, your opinions about various things,
and how easy or hard it has been for you to find a place to live, to get a job, and so forth. Throughout the
interview, I’ll be taking notes about what you tell me on this form. If any question I ask you is too personal or
makes you feel uncomfortable, please let me know and I will skip it and go on to the next one. I also want to
remind you that you can refuse to answer a question at any time and you can stop participating at any time.
Finally, I want to emphasize to you that everything you tell me today will be kept in the strictest confidence. I
will be asking you about past crimes and drug use and your responses to these questions will be kept
completely confidential. The only exception to this is if you tell me that you intend to hurt yourself or to commit
a specific crime in the future, in which case I may have to report it. Also, if you tell me about a child being
abused, I am obligated by law to report that abuse to the State. Otherwise, everything you tell me will be kept
completely confidential. Do you have any questions before we get started?
Okay, I’d like to start by asking you a few questions about the day you were released from prison.

If R is in prison, questions refer to last release from Queensboro Correctional or Upstate
facility.

1.

What day were you released? 

________ / _______ / _______________ (mm/dd/yyyy) 


2.

Check
only
ONE

box

Read All
Responses

4.
Check
only
ONE

box
Read All
Responses

98 – Don’t Know

Was anyone at the prison to meet you when you were released?

†
†
3.

‫׃‬

0 - No
1 - Yes

Where did you go immediately after release?

†
†
†
†
†
†
†
†

1 - To parole office
2 - To your own home (you own or rent) 

3 - To a family member's home 

4 - To girlfriend/ partner’s home
5 - To a friend’s home
6 - To new or potential employer
7 – To the Department of Motor Vehicles\Medicaid Office
8 - Other Î [specify: ___________________________________________ ]

Where did you sleep your first night out of prison? 


†
†
†
†
†
†

1 - In your own house or apartment (you rent or own) 

2 - At a family member's house or apartment 

3 - At a friend's house or apartment 

4- A girlfriend/ partner’s home
5 - At a shelter
6 - At a hotel / motel / rooming house

Vera Institute of Justice 125

†
†
†
5.	

7 - On the street
8 - Other Î [specify: ___________________________________________ ]
98 - Don't know

How long did you spend there? 


________ Days


† 1 - Still living there
† 98 - Don't know
What kinds of identification did you have on the day you were
released from prison?

No

Yes

6.

Did Respondent have any identification? 	

†

†

7.

Driver’s License 	

†

†

8.

State ID card (non-driver’s license) 	

†

†

9.

Department of Corrections ID card	

†

†

10.

Birth Certificate 	

†

†

11.

Social Security Card 	

†

†

12.

Medicaid Card (or letter) 	

†

†

13.

OtherÄ[Specify:__________________________________]	

†

†

What kinds of identification do you have now? 	

No

Yes

14.

Does Respondent have any identification?

†

†

15.

Driver’s License

†

†

16.

State ID card (non-driver’s license)

†

†

17.

Department of Corrections ID card

†

†

18.

Birth Certificate

†

†

19.

Social Security Card

†

†

20.

Medicaid Card (or letter)

†

†

21.

OtherÄ[Specify:__________________________________]

†	

†

22.

How long after your release did you get your Medicaid card?

1 – I got it before I was released
2 – It was at home on the day I was released
3 – Within 2 weeks
4 – 3 to 4 weeks
5 – More than a month
6 – I never recieved it
23.

How much money did you have with you when you were released?

126 Vera Institute of Justice




[Include only the money you had with you—include cash, checks, money orders, etc.]
$___________ [ IF ZERO, SKIP TO Q 28 ]

98 = Don't know

Where did you get the money you had when you were
released?

No

Yes

24.

Family member(s) sent it

†

†

25.

Friend(s) sent it

†

†

26.

From your prison work or program participation

†

†

27.

Other Î [specify: ___________________________________________ ]

†

†

Now I’d like to ask you some questions about different programs that prisons sometimes offer, like GED
classes and job training programs. For each one, I’m going to ask you if you participated in a particular
program and for how long. Do not include any programs you participated in during your last three
months in prison (for example, a DOCS pre-release program or Project Greenlight).
Column B
[If yes] How long did you participate in
this program…
(in months)?
[0=Did not participate]

Did you participate in this program?

28.

GED / adult basic education

98 – Don’t know

†
29.

English as a second language (ESL)

†
30.

Life skills (finances, budgeting, time mgmt, banking, etc.)

31.

Substance abuse treatment

Counseling, including mental health

Anger management / violence prevention

HIV/AIDS education

37.

Other Î [specify:__________________________________________ ]

______ ______ ______

98 – Don’t know

†
36.

______ ______ ______

98 – Don’t know

†
35.

______ ______ ______

98 – Don’t know

†
34.

______ ______ ______

Trade or job training (e.g. Building maintenance, floor-covering, etc) 98 – Don’t know

†
33.

______ ______ ______

98 – Don’t know

†
32.

______ ______ ______

98 – Don’t know

†
Employment readiness (e.g., how to interview)

______ ______ ______

98 – Don’t know

†

______ ______ ______

98 – Don’t know
______ ______ ______

______ ______ ______

The next few questions are about programs in prison that were specifically designed to help prisoners prepare
for release and that you may have participated in during your last three months of incarceration. The
following questions refer only to your last 3 months in prison.

Vera Institute of Justice 127

If R is in prison, questions refer to last release from Queensboro Correctional or Upstate
facility.

38.	

In the last three months before you were released from prison, did you participate in any activities or programs
that were designed to help prepare you for release or help you after you were released?
[Interviewer: Project Greenlight or DOCS Pre-release: activities to prepare for release and help after being released
from prison—drug treatment, help with obtaining identification, community referrals, etc]

No [skip to Q62] 	
†	
39.	

Yes
†

Was participation in pre-release programs voluntary or required?

†
†
†

0 - Voluntary
1 - Required
98 - Don't know

[Interviewer: If Q38=No, ask only columns A and D for Q40-51]
I’d like to know a little bit more about the kinds of information you were provided in the pre-release activities or programs you may have
participated in during the last 3 months before you were released (for example, after you were transferred to Queensboro). I’m going to
read you a list of things that you might have done since you were released, like find a place to live or find a job, and I’d like you to tell me if
you were provided with information to help you with those things. I’d also like to know how helpful you found that information and if a
community program or provider has helped you with any of those things since release.
Column A
Column B
Column C
Column D
Has a community
How many days in the Did your pre-release How helpful was the preprogram or provider
past month have you preparation cover [topic]? release program in
helped you with…?
spent in your
helping you…?
[Skip if B=No, DK]
community…?

40.

41.

Cognitive behavior/ Life skills
(e.g. problem solving, anger
management, budgeting)

†
†
†

Finding a place to live
97 – R had one lined up

_______ _______

†	
42.

Finding a job
97 – R had one lined up

†	
43.

_______ _______

Continuing your education
_______ _______

44.

Getting custody of your children 	
97 – R has no children (<18)	

†	
45.

_______ _______

Getting drug or alcohol treatment 	
from a treatment provider
_______ _______

46.

Attending support groups like NA, 	
AA, or Ex-Offender Support

128 Vera Institute of Justice

†
†
†
†
†
†
†
†
†
†
†
†
†
†
†	
†

No
Yes
Don’t know
No
Yes
Don’t know
No
Yes
Don’t know
No
Yes
Don’t know
No
Yes
Don’t know
No
Yes	
Don’t know
No

Very
Moderately
Slightly
Not at all
Very
Moderately
Slightly
Not at all
Very
Moderately
Slightly
Not at all
Very
Moderately
Slightly
Not at all
Very
Moderately
Slightly
Not at all
Very
Moderately
Slightly
Not at all
Very

†
†
†
†
†
†
†
†
†
†
†
†
†
†
†
†

No
Yes
Don’t know
No
Yes
Don’t know
No
Yes
Don’t know
No
Yes
Don’t know
No
Yes
Don’t know
No

Networks 	

47.

Accessing health care, not
including drug or alcohol
treatment

48.

†
†	

_______ _______

†
†
†

_______ _______

Getting counseling (including 	
family, mental health)

†
†
†

_______ _______

49.

Getting financial help or 	
assistance (Social Service,
Welfare, Food stamps) 	

50.

†
†
†	

_______ _______

Obtaining identification

†
†
†

_______ _______

51.

†
†
†

Other(s) Î specify:
___________________________

52.

_______ _______

Yes
Don’t know
No
Yes
Don’t know
No
Yes	
Don’t know	

No
Yes
Don’t know
No
Yes
Don’t know
No
Yes
Don’t know

Moderately
Slightly
Not at all
Very
Moderately
Slightly
Not at all
Very
Moderately
Slightly
Not at all

†
†
†
†
†

Yes
Don’t know
No
Yes
Don’t know

† No
† Yes
Don’t know

Very
Moderately
Slightly
Not at all
Very
Moderately
Slightly
Not at all
Very
Moderately
Slightly
Not at all

†
†
†

†
†
†

No
Yes
Don’t know

No
Yes
Don’t know

Did the pre-release program refer you to any community-based programs after release?

0 No [skip to Q62]	

1 Yes

†	

†

IF R WAS REFERRED TO COMMUNITY PROGRAMS, ASK:
Column A
Column B
Did the pre-release program Specify the name of
refer you to any of the following
the program
community programs?

53.

Substance abuse treatment

1

0 – No

† † † † †

†
54.	

†

†
1 - Yes

†

Job training program
0 – No

†
56.

1 - Yes

Job interview (e.g. UPS)
0 – No

55.	

Column C
What type of
referral were
you given?
[see codes]

1 - Yes

†

Employment service
0 – No

1 - Yes

1

2

2

3 4

3 4

5

5

† † † † †
1

2

3 4

5

† † † † †
1

2

3 4

5

Column D 	

Column E

Did you
meet with them?

Why did you not
meet with them?
[see codes]

0-No 1-Yes

† †

1

2

3

4 5 6

7 8 98

† † † † † † † † †

0-No 1-Yes

† †

1

2

3

4 5 6

7 8 98

† † † † † † † † †

0-No 1-Yes

† †

1

2

3

4 5 6

7 8 98

† † † † † † † † †

0-No 1-Yes

1

2

3

4 5 6

† †
Vera Institute of Justice 129

7 8 98

†

†

† † † † †

Housing (half-way house,
57.	shelter,
etc.)
0 – No

1 - Yes

†

†

†

59. Anger management
0 – No

†

1 - Yes

1 - Yes

61. Other

0-No 1-Yes

† †

2

3	4

5

2

3	4

1

2

3	4

1 - Yes

2

3	4

† †

5

† †

5

5

† †

3.
4.

You were only given the name of organization
You were given the organization name, address, and/or phone
number
You were given a specific name of someone at the organization
Someone scheduled an appointment for you

5.

Other Î [specify: __________________________________

3

4 5 6

2

3

4 5 6

1

2

3

4 5 6

1

2

3

4 5 6

1

2

3

4 5 6

1.
2.
3.
4.
5.
6.

You lost the information that was given to you
You contacted another program that offered similar
assistance
You didn't think you needed assistance from this
organization
You were too busy
The organization was too far away
You tried to contact them

facility.
Where are you currently living?
Homeless / on the street
Your own house or apartment [that you rent or own by yourself or with others]
A family member’s house or apartment
A friend’s house or apartment
A public shelter (e.g. Bellevue)
Transitional housing (e.g Ready, Willing and Able, The Castle, Salvation Army)
Other Î [specify: ___________________________________________ ]

How long have you lived there? [How long have you been homeless]
Less than a week
1 to 2 weeks
3 to 4 weeks
More than a month 


98 -

Don’t know 


130 Vera Institute of Justice

7 8 98

For Column E
Codes for reasons for not meeting with them
SELECT ALL THAT APPLY.

If R is in prison, questions refer to last release from Queensboro Correctional or Upstate

63.	

7 8 98

† † † † † † † † †

your current living situation — that is, the place where you spend most of your nights now.

0123456-

7 8 98

† † † † † † † † †

Okay, we’re going to switch gears a bit here and talk about your life now. The next few questions are about

Choose
ONE where
R
spends
most time

7 8 98

† † † † † † † † †

7. Other Î [specify: _____________________ __
98. Don't know

62.

7 8 98

† † † † † † † † †

0-No 1-Yes

For Column C
Codes for type of referral
CHOOSE ONE.
1.
2.

1

0-No 1-Yes

† †

2

† † † † † † † † †

0-No 1-Yes

† † † † †

†

1

0-No 1-Yes

† † † † †

1

0 – No

5

† † † † †

†

†

3 4

† † † † †

60. Family support or counseling
†

1

1

†

0 – No

2

† † † † †

Educational service (continuing
58.your
education)
0 – No
1 - Yes

†

1

† † † † † † † † †

64.

Is this where you were living when you were first released from prison?

7899 -

65.

No
Yes [Skip to Q64]
Don't know

Since your release from prison, how many different places have you lived?

1
†

2
†

3
†

4
†

5
†

6
†

The next few questions are about the people that you live with now.

66.

Who do you currently live with?

Nobody Î [SKIP TO Q67]
Spouse / Partner
Girlfriend / fiance'
Parent / Step-parent

Æ
READ
ALL
ALOUD
Check
all that
apply

Brother / Sister
Other family
Friend
Other
______________________________
__

No

Yes

Don’t know

Refused

67.

Do any of the people you currently live with use illegal
drugs?

†

†

†

†

68.

Do any of the people you currently live with get drunk often?

†

†

†

†

Now I’d like to ask you some questions about your relationships with your friends and family.
Married or
Common law
Marriage

Living Together

Divorced,
Separated,
Widowed

Never Married

†

†

†

†

69.

What is your current marital status?

70.


How many close family relationships do you have now? By close, I mean people who look out for you, who would
do you a favor even when you don't ask them to, and who will listen to you and offer you advice when you need it?

0
†
71.
Æ
READ
ALL
ALOUD
Check
all that
apply

1
†

2
†

3
†

4
†

5
†

5+
†

Of your close friends and family members, who have you had contact with in the last 30 days?

0 - Nobody
1 - Spouse
2 - Girlfriend / fiance'
3 - Parents
4 – Brother / Sister]

5- R’s Children
6 – Other extended family (e.g. grandmother, aunt, 

uncle)

7 – Friend(s)

8 - Someone R met in prison
†
 9 - Other Î [specify: 

______________________________________

I’m going to read you some statements that may describe how you feel about your family now. After I read
each one, please tell me if you strongly agree, agree, disagree, or strongly disagree with the statement.
Strongly
agree

Agree

Not Sure

Disagree

Strongly
disagree

Vera Institute of Justice 131

Strongly
agree

Agree

Not Sure

Disagree

Strongly
disagree

72.

You have someone in your family you can count on to listen to
you when you need to talk.

†

†

†

†

†

73.

You have someone in your family to talk to about yourself or
your problems.

†

†

†

†

†

74.

You have someone in your family whose advice you really
want.

†

†

†

†

†

75.

You have someone in your family to share your most private
worries and fears with.

†

†

†

†

†

76.

You have someone in your family to turn to for suggestions
about how to deal with a personal problem.

†

†

†

†

†

77.

You have someone in your family who understands your
problems.

†

†

†

†

†

78.

You have someone in your family who would provide help or
advice on finding a place to live.

†

†

†

†

†

79.

You have someone in your family who would provide support
for dealing with a substance abuse problem.

†

†

†

†

†

80.

You have someone in your family who would provide you with
some financial support.

†

†

†

†

†

For the next part of the interview, I’m going to ask you some questions about friends you have now.

81.

How many close friends, who are not family members, do you have? Again, by close, I mean people who will look out for you,
who would do you a favor even when you don’t ask them to, and who will listen to you and offer you advice when you need it.

Indicate how many:

[If “0” Skip to Q91]

0
†

1
†

2
†

3
†

4
†

5
†

For the next set of questions, tell me whether this applies to all, most, half, some, or none of your close
friends.
None

Some

About
Half

Most

All

Don't
Know

82.

How many have ever been in prison? 


†

†

†

†

†

†

83.

How many are currently employed? 


†

†

†

†

†

†

84.

How many do you think have ever committed a theft or an

assault?

†

†

†

†

†

†

85.

How many can you hang out with and know that you won't get in

trouble? 


†

†

†

†

†

†

86.

How many do you think have ever used illegal drugs?


†

†

†

†

†

†

87.

How many are not involved in any criminal activity?

†

†

†

†

†

†

132 Vera Institute of Justice

5+
†

88.

None

Some

About
Half

Most

All

Don't
Know

†

†

†

†

†

†

Of the close friends you had before prison, how many do you
hang out with now?

Please respond to the following statements by stating whether you strongly agree, agree, are not sure,
disagree, or strongly disagree. [Use response card]
Strongly
agree

Agree

Not
Sure

Disagree

Strongly
disagree

89.

Your friends have been supportive after your release
from prison.

†

†

†

†

†

90.

Your friends sometimes convince you to do things
you know you shouldn't be doing.

†

†

†

†

†

The next few questions are about the neighborhood where you live now. Please respond to the following
statements about how you feel about the neighborhood you live in now.
Strongly
agree

Agree

Not
Sure

Disagree

Strongly
disagree

91.

Your neighborhood is a safe place to live.

†

†

†

†

†

92.

It is hard to stay out of trouble in your neighborhood.

†

†

†

†

†

93.

Drug selling is a major problem in your neighborhood.

†

†

†

†

†

94.

You think your neighborhood is a good place for you to live.

†

†

†

†

†

95.

You think your neighborhood is a good place for you to find a job.

†

†

†

†

†

96.

Living in this neighborhood makes it hard for you to stay out of
prison.

†

†

†

†

†

97.

You care about what your neighbors think of your actions.

†

†

†

†

†

98.

If there is a problem in your neighborhood, people who live there
can get it solved.

†

†

†

†

†

99.

You expect to live in this neighborhood for a long time.

†

†

†

†

†

The next few questions are about activities and organizations that you might participate in. For each activity or
organization, please tell me how often you attend or participate and how important the organization or activity
is to you.

Do you participate in or attend….?

100.
101.

Church, synagogue, mosque, or other
religious institution

Column A
Column B
[If yes] How many days
per month do you attend ? [If yes] How important or valuable is this organization to you?
[0 = Does not participate]

Very

Moderately

Slightly

Not At All

†

†

†

†

†

†

†

†

†

†

†

†

_______ _______

Recreational club (e.g. YMCA, gym)

_______ _______
102.

Sports team or activities (e.g.
basketball, coaching)

_______ _______
Vera Institute of Justice 133

Do you participate in or attend….?

103.

Column A
Column B
[If yes] How many days
per month do you attend ? [If yes] How important or valuable is this organization to you?
[0 = Does not participate]

Ex-offender group

†

†

†

†

†

†

†

†

†

†

†

†

_______ _______
104.

Self-help groups (for example, AA or NA)

_______ _______
105.

Other community groups or activities
[Specify:________________________]

_______ _______

We’d like to know more about what kinds of financial obligations people face when they leave prison. The next
few questions are about your current financial obligations.

Do you owe money each month for any of the
following….?

106.

Housing or Rent

107.

Basic living expenses (food, utilities)

108.

Supervision fee

How much money
do you owe each
month for …?
[0 = None]

Lump Sum
Owed or
accumulated
debts

R Does Not
Pay

$
† 98 – Don’t know
$
† 98 – Don’t know
$
† 95 – Waived
† 98 – Don’t know

109.

Alimony or Child support

110.

Other debt Î [specify:]

$
† 98 – Don’t know
$

________________________________________]
† 98 – Don’t know

If R is in prison, questions refer to last release from Queensboro Correctional or Upstate
facility.
The next questions are about looking for work since your release from prison.

111.

112.

Have you spent any time looking for a job since you were released from prison?
(If “Yes”, skip to Q113)

0- No

1- Yes

†

†

Why haven't you looked for work since you were released?

1 - I had a job/contact lined up when I was released
2 - Health problems / disabled
3 - I do not want to work
4 – I am in school
5 – I am in vocational or other job training
6 - Other Î [specify: ___________________________________________ ]
134 Vera Institute of Justice

98 - Don't know 

113.

About how many hours per week have you spent looking for work since you were released?
_______________ hours per week

114.

How many employers have you contacted since your release?
_______________ employers

What have you done to find a job since you were released?

No

Yes

115.

Talked to friends or relatives

†

†

116.

Talked to your parole officer

†

†

117.

Used newspaper ads

†

†

118.

Used an employment agency

†

†

119.

Answer help-wanted signs

†

†

120.

Went to a former employer

†

†

121.

Sent a resume or called an employer

†

†

122.

Used a community service agency

†

†

†

†

123.
Other Î [specify: ______________________________________________________]

124.

What is your current employment status?

9 - Employed full-time (35+ hours/week)
10 -Employed part-time
11 -Vocational or other job training
12 -Student
13 -Retired/disability
14 -Unemployed and looking for work ¼ [Skip to Q137]
15 -Unemployed and not looking for work¼[Skip to Q137]
The next questions are about your recent work situation.

125.
126.

How many weeks have you worked since your
release from prison?
How many jobs are you working at, including
self-employment?

0

1

2

3

4

5

5+

†

†

†

†

†

†

†

†

†

†

†

†

†

†

Please respond to the following questions about the work you are doing now or a job you’ve had since you were
released.
Strongly
Strongly
† 97 – R has not been employed since release
agree
Agree
Not Sure
Disagree
disagree
[Skip to Q134]

127.

You like the work you are doing.

†

†

†

†

†

Vera Institute of Justice 135

†

97 – R has not been employed since release
[Skip to Q134]

Strongly
agree

Agree

Not Sure

Disagree

Strongly
disagree

128.

You are not happy with the amount you are paid for the
work you do.

†

†

†

†

†

129.

You don't get along with the people you work with.

†

†

†

†

†

130.

You'd be happy if you were at this job one year from
now.

†

†

†

†

†

131.

You think this job will give you better opportunities in
the future.

†

†

†

†

†

132.

You get along with the people you work for.

†

†

†

†

†

133.

The people you work for don't treat you fairly.

†

†

†

†

†

Now I’d like to ask you several questions about your work history.

134.

Have you ever been fired from a job,
including self-employment?

0

1

2

3

4

5

5+

†

†

†

†

†

†

†

[If yes, how many times have you ever been 

fired from a job?]


135.
Æ
READ
ALL
ALOUD

136.
Æ
READ
ALL
ALOUD

137.

What is the longest you have worked at one job?

12345-

Less than 6 months
6 months to 1 year
1+ to 2 years
2+ to 5 years
5+ years

What was your usual employment pattern in the three years before you were sent to prison?

1 – Full-time (35+ hours/week or regularly employed)
2 – Part-time
3 – Student
4 – Military Service
5 – Retired/disability
6 – Unemployed
7 – In prison orjail (or other institution)
Have you ever been unemployed for a year or more?

0 – No 

1 – Yes 

Now I’m going to ask you a little about your financial situation since you were released. 

Can you tell me if any of the following have been a source of income in the past month, and how much you 

received from each in the last month? 

Have you received any money in the last month from….?

136 Vera Institute of Justice

[If yes] How much did you
receive in the last month
[0=None]

[If yes] How much did you
receive in the last month
[0=None]

Have you received any money in the last month from….?

138.

Legal employment

$

139.

Spouse, family or friends

$

140.

Food stamps

$

141.

Social Security Benefits (any kind)

$

142.

Medicaid / Medicare

$

143.

Public assistance (including Public housing aid or vouchers)

$

144.

Under the table employment

$

Illegal activities
95 –Refused

$

145.

146.

$
Other Î [specify:___________________________,
e.g. Worker’s compensation, Veteran’s or Military benefits ]

Have you had a significant period of time in which you have…? Please remember that all information is
confidential, and that you may refuse to answer any question.
Past 30 days

147.
148.
149.
150.
151.
152.
153.

154.
Æ
READ
ALL
ALOUD
Check
all that
apply

Experienced serious depression, sadness,
hopelessness, loss of interest.
Experienced serious anxiety / felt uptight,
unreasonably worried, unable to feel relaxed.
Experienced trouble understanding, concentrating, or
remembering.
Experienced trouble controlling violent behavior
including episodes of rage or violence.
Experienced thoughts of suicide or attempted suicide.

Spent time in a hospital or other institution for mental
or emotional problems?
Have you ever been prescribed medications for
mental or emotional problems?

Lifetime

Refused

0-No

1-Yes

0-No

1-Yes

95

†

†

†

†

†

0-No

1-Yes

0-No

1-Yes

95

†

†

†

†

†

0-No

1-Yes

0-No

1-Yes

95

†

†

†

†

†

0-No

1-Yes

0-No

1-Yes

95

†

†

†

†

†

0-No

1-Yes

0-No

1-Yes

95

†

†

†

†

†

0-No

1-Yes

0-No

1-Yes

95

†

†

†

†

†

0-No

1-Yes

0-No

1-Yes

95

†

†

†

†

†

What kind of health coverage or insurance do you have?

No insurance or health coverage
Medicaid
VA
Private insurance or Blue Cross
Other Î [specify: ___________________________________________ ]
Vera Institute of Justice 137

Don't know
The next set of questions asks about your experiences with parole or mandatory supervision.

Very
likely

Likely

Not Sure

Unlikely

Very
Unlikely

155.	

How likely do you think it is that you will violate at
least one of the conditions of your supervision? 

Do you think it is...


†

†

†

†

†

156.	

How likely do you think it is that you will get
caught if you violate the conditions of your
supervision? Do you think it is...

†

†

†

†

†

Please respond to the following statements about your parole status. [Use the response card]
Strongly
agree

Agree

Not Sure

Disagree

Strongly
disagree

157.

Your parole officer has been helpful with your transition back to
the community.

†

†

†

†

†

158.

Being under supervision will help you stay out of prison.

†

†

†

†

†

159.

Being under supervision will help you stay drug free.

†

†

†

†

†

160.

Being under supervision will help you stay crime free.

†

†

†

†

†

161.

Your parole officer seems trustworthy.

†

†

†

†

†

162.

Your parole officer gives you correct information.

†

†

†

†

†

163.

Your parole officer acts too busy to help you.

†

†

†

†

†

164.

Your parole officer treats you with respect.

†

†

†

†

†

165.

Your parole officer acts professionally.

†

†

†

†

†

166.

Your parole officer doesn’t listen to you.

†

†

†

†

†

For each of the following conditions of supervision, please indicate whether it is one of your conditions and, if it
is, whether you have violated it regardless of being caught since your release from prison. Again, please
remember that anything you tell me about your past behavior is confidential and that you may refuse to answer
any question.
General conditions of parole and mandatory supervision

167.

Report as directed and follow parole officer instructions

168.

Notify parole officer of changes in residence or employment

169.

Notify parole officer if arrested

138 Vera Institute of Justice

Is this a
Condition of supervision?

Have you violated
this condition since
release?

0-No

1-Yes

0-No

1-Yes

†

†

†

†

0-No

1-Yes

0-No

1-Yes

†

†

†

†

0-No

1-Yes

0-No

1-Yes

†

†

†

†

General conditions of parole and mandatory supervision

170.

Will not associate with anyone who has a criminal record

171.

Obey all laws

172.

Not own or possess firearms or other weapons

173.

Not possess or use drugs

174.

Curfew

175.

Attend treatment (drug, alcohol, or mental health)

176.

Participate in a domestic violence or anger management program

177.

Seek and maintain employment

Are there other conditions of supervision that we have not talked
about Î [specify: _________________ _]

179.
Æ
READ
ALL
ALOUD

180.

Have you violated
this condition since
release?

0-No

1-Yes

0-No

1-Yes

†

†

†

†

0-No

1-Yes

0-No

1-Yes

†

†

†

†

0-No

1-Yes

0-No

1-Yes

†

†

†

†

0-No

1-Yes

0-No

1-Yes

†

†

†

†

Condition of supervision?

Special conditions

178.

Is this a
Condition of supervision?

Violated
since release?

0-No

1-Yes

0-No

1-Yes

†

†

†

†

0-No

1-Yes

0-No

1-Yes

†

†

†

†

0-No

1-Yes

0-No

1-Yes

†

†

†

†

0-No

1-Yes

0-No

1-Yes

†

†

†

†

0-No

1-Yes

0-No

1-Yes

†

†

†

†

How often did you see your parole officer in person in your first month after release?

1 - once
2 - twice
3 - 3 times
4 - 4 times
5 - 5 times
16 -Not at all

Î [SKIP TO Q198 ]

On average, how long do you meet with your parole officer?

1234-

Less than 5 minutes
5 to 30 minutes
31 minutes to 1 hour
More than 1 hour
One
Time

Two
Times

Three
Times

Four
Times

Five
Times

Not at
all

181.

How often have you spoken with your parole officer on the
phone in the last month?

†

†

†

†

†

†

182.

How often has your parole officer visited you at the place
you live or work in the last month?

†

†

†

†

†

†

Vera Institute of Justice 139

The next set of questions asks about how easy or hard it has been for you to do various things since you’ve
been released, like re-establishing contact with old friends. For each one, please tell me if it has been very
easy, pretty easy, pretty hard, or very hard. If the question does not apply to you, tell me it does not apply and
why.
[Use response card]
Very
easy

Pretty
easy

Not
Sure

Pretty
hard

Very
hard

†

†

†

†

†

183.

How easy or hard has it been to avoid a parole violation?

184.

How easy or hard has it been to stay out of prison?

†

†

†

†

†

185.

How easy or hard has it been to not commit crimes?

†

†

†

†

†

186.

How easy or hard has it been to make enough money to support
yourself?

†

†

†

†

†

187.

How easy or hard has it been to find a place to live?

†

†

†

†

†

188.

How easy or hard has it been to find a job?

†

†

†

†

†

189.

How easy or hard has it been to keep a job?

†

†

†

†

†

†

†

†

†

†

†

†

†

†

†

†

†

†

†

†

97 - R is not on parole

97 - R has committed crimes

97 - R already had a place to live lined up.
97- R already had a job lined up.
97 - R is unemployed.

190.

How easy or hard has it been to renew relationships with your
family?

96 - R has not tried to renew relationships
191.

How easy or hard has it been to renew relationships with your
children?

96 - R has not tried 

97 – R has no children 

192.

How easy or hard has it been to not use drugs?

97 - R does not use drugs

If R is in prison, questions refer to last release from Queensboro Correctional or Upstate
facility.
Now I’d like to read to you some statements about your opinions or views on various things. After I read each
one, please let me know whether you strongly agree, agree, are not sure, disagree, or strongly disagree with
each statement. [Interviewer use response card]
Strongly
agree

Agree

Not Sure

Disagree

Strongly
disagree

193.

You have little control over the things that happen to you. 


†

†

†

†

†

194.

You have trouble making decisions 


†

†

†

†

†

195.

You feel people are important to you. 


†

†

†

†

†

196.

You feel anxious or nervous.


†

†

†

†

†

197.

What happens to you in the future mostly depends on you. 


†

†

†

†

†

140 Vera Institute of Justice

Strongly
agree

Agree

Not Sure

Disagree

Strongly
disagree

198.

There is little you can do to change many of the important things in
your life.

†

†

†

†

†

199.

You make decisions without thinking about the consequences.

†

†

†

†

†

200.

You feel honesty is required in every situation.

†

†

†

†

†

201.

You feel sad or depressed.

†

†

†

†

†

202.

There is really no way you can solve some of the problems you
have.

†

†

†

†

†

203.

You make good decisions.

†

†

†

†

†

204.

It is important to pay off your debts.

†

†

†

†

†

205.

You feel lonely.

†

†

†

†

†

206.

Sometimes you feel like you're being pushed around in life.

†

†

†

†

†

207.

You often feel helpless dealing with the problems of life.

†

†

†

†

†

208.

Your life has gone out of control.

†

†

†

†

†

209.

You keep the same friends for a long time.

†

†

†

†

†

210.

You have trouble concentrating or remembering.

†

†

†

†

†

211.

You are tired of the problems caused by the crimes you committed.

†

†

†

†

†

212.

You have trouble sleeping.

†

†

†

†

†

213.

You have given up friends and hangouts that get you in trouble.

†

†

†

†

†

214.

You think about what causes your current problems.

†

†

†

†

†

215.

You want to get your life straightened out.

†

†

†

†

†

216.

You consider how your actions will affect others.

†

†

†

†

†

217.

Your religious beliefs are very important in your life.

†

†

†

†

†

218.

You work hard to keep a job.

†

†

†

†

†

219.

You have trouble sitting still for long.

†

†

†

†

†

220.

You feel mistreated by other people.

†

†

†

†

†

221.

You have a hot temper.

†

†

†

†

†

222.

Your temper gets you into fights or other trouble.

†

†

†

†

†

Vera Institute of Justice 141

Strongly
agree

Agree

Not Sure

Disagree

Strongly
disagree

223.

You get mad at other people easily.

†

†

†

†

†

224.

You sometimes carry weapons, like knives or guns.

†

†

†

†

†

225.

You like to take chances. 


†

†

†

†

†

226.

You do things that are strange or exciting 


†

†

†

†

†

227.

You avoid things that are dangerous.


†

†

†

†

†

228.

You only do things that feel safe. 


†

†

†

†

†

229.

You are very careful and cautious. 


†

†

†

†

†

Please respond to the following factors that may have been
important in keeping you out of prison since this last release?

Very
important

Important

Not Sure

Unimportant

Very
unimportant

230.

Having a place to live

†

†

†

†

†

231.

Having a job

†

†

†

†

†

232.

Having access to healthcare

†

†

†

†

†

233.

Having enough money to support yourself

†

†

†

†

†

234.

Not using drugs

†

†

†

†

†

235.

Not drinking alcohol

†

†

†

†

†

236.

Getting support from your family

†

†

†

†

†

237.

Getting support from your friends

†

†

†

†

†

238.

Avoiding certain people and situations

†

†

†

†

†

239.

Other Î
___________________________________________ ]

†

†

†

†

†

240.

Do you fear returning to prison?
0- No
†

1- Yes
†

98- Don’t know
†

For the next part of the interview, I’d like to ask you some questions about how often, if at all, you have used
drugs and alcohol since you’ve been released and over the course of your life. Please answer as accurately
and honestly as you can. Remember, all the information you tell me will be kept completely confidential, and
once again, you may refuse to answer any question.

142 Vera Institute of Justice

Respondent has never used any drugs or alcohol
241.

Alcohol (to the point of being drunk)

242.

Marijuana (or hashish)

243.

Heroin

244.

Cocaine (powder, crack, rock)

245.

Hallucinogens(LSD, Acid)

246.

Other drugs [Specify___________________________________]

247.

More than 1 substance in the same day (including alcohol)

How many days in the
past 30 have you used…?
[95 = Refuse]

How many years have
you used…?
[95 = Refuse]

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

_____ _____

The following questions ask about your history of drug and alcohol use (DAST-10).

No

Yes

248.

Have you ever used drugs other than those required for medical reasons?

†

†

249.

Have you abused more than one drug at a time?

†

†

250.

Have you been able to stop using drugs when you want to?

†

†

251.

Have you had blackouts or flashbacks as a result of drugs?

†

†

252.

Have you ever felt bad or guilty about your drug use?

†

†

253.

Has your spouse (or parents) ever complained about your involvement with drugs?

†

†

254.

Have you ever neglected your family because of your use of drugs?

†

†

255.

Have you ever engaged in illegal activities in order to obtain drugs?

†

†

256.

Have you ever experienced withdrawal symptoms as a result of heavy drug intake?

†

†

257.

Have you ever had medical problems as a result of your drug use (e.g., memory loss, hepatitis,
convulsions, bleeding)?

†

†

The final questions are about your juvenile criminal history, education, and background.

258.

Were you ever arrested before you were 16 years old? [If yes, how many times (0=No arrests)]

0

1

2

3

4

5

5+

†

†

†

†

†

†

†

259.

Were you ever convicted before you were 16 years old? [If yes, how many times (0=No convictions)]

0

1

2

3

4

5

5+

†

†

†

†

†

†

†

Vera Institute of Justice 143

260.	

What is the highest grade you completed in school (do not include grades started but not completed; 0=None,
1=1st grade….12=12th grade, 13=1 year college, etc.)?

__________

261.

[Specify highest grade completed in years]

Have you completed your GED? 

0- No

1- Yes 


†	
262.

†

Were you ever suspended or expelled from school ? 

0- No
1- Yes 


†	

†

263. Do you have any children under age 18? [if yes, how many children under age 18?]
0

1

2

3

4

5

5+

†

†

†

†

†

†

†

Thank you
for answering the questions on the survey. Your
participation is very helpful to us.

144 Vera Institute of Justice

†

0F

†

1M

Date:

Start time:

A.M

End time:
Time to complete:
(minutes)

Location:

Vera

Parole

†

†

Community

Jail

†

P.M.

F

F

F

F

Other
†

†

Completion codes:

Others present during interview?

†
†

0 - No

†

0 - No

1 - Yes

†

1 - Yes Î Write Who:

Poor

Acceptable

Good

Excellent

Respondent's attention to you was

†

†

†

†

Respondent's general understanding of the
questions was

†

†

†

†

Respondent's cooperation throughout most of the
interview was

†

†

†

†

Did R appear to be…

No

Somewhat

Very

Suspicious

†

†

†

Uncommunicative

†

†

†

Depressed or withdrawn

†

†

†

Anxious or nervous

†

†

†

Hostile

†

†

†

Tired or in pain

†

†

†

Drunk

†

†

†

On illegal drugs

†

†

†

How honest was R throughout the interview?

Probably not
honest

Somewhat
honest

Mostly honest

Entirely honest

†

†

†

†

Other interviewer comments (e.g., children or visitors interrupted/distracted R; setting was not
conducive to privacy):

Vera Institute of Justice 145

PAROLE OFFICER QUANTITATIVE/QUALITATIVE INSTRUMENT

TRANSITIONAL SERVICES PROGRAMMING IN NEW YORK STATE: 

EVALUATING PROJECT GREENLIGHT 

Parole Officer Interview 

2003
Introduction:
The following is a series of questions that we’d like to ask you about parolees who are,
or have been, under your supervision. There are only 6 questions we’d like to ask about
each study participant. All questions should be answered in reference to the parolees
first 1-2 months in the community under your supervision.
Vera ID:
1. How well did the parolee meet the reporting requirements of parole supervision?
Poorly	
Average
Very Well
1

2

3

4

5

6

7

ż

ż

ż

ż

ż

ż

ż


2. How well did the parolee meet the general requirements of parole supervision?
Poorly	
Average
Very Well
1

2

3

4

5

6

7

ż

ż

ż

ż

ż

ż

ż


3. 	 What best describes the parolee’s current living situation? 

1 - Shelter


ż
ż
ż
ż
ż

2 - Short-term residential or transitional housing
3 - Permanent housing (own housing or living with someone)
4 – Drug facility
5 - Other housing (please specify:_________________________________)

4. 	 Did the parolee live in a shelter since his release from Queensboro? 

No- 0
Yes -1 


ż	

ż


5. 	 Was the parolee employed during the first three months? 

No- 0
Yes -1 

146 Vera Institute of Justice

ż

ż


6. Which supervision conditions did they violate and how often during first 3 months? 

Parole Supervision Condition

0

1

2

3

4

5 or
more

1 –Contact PO at release

ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż

ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż

ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż

ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż

ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż
ż

ż
ż

ż

ż

ż

ż

ż

ż

ż

ż

ż

ż

ż


2 –Make office/written reports
3 –Won’t leave assigned area
4 –Notify PO of changes
5 –Will reply to PO inquiries
6 –Will Notify PO of an arrest
7 –No criminal acquaintances
8 –Will obey the law
9 –No firearms or weapons
10 –Will waive extradition
11 –No drugs / paraphernalia
12 –Any Special Conditions
13 –Will comply w/ conditions

Vera Institute of Justice 147

The following questions are asking for your perceptions and input on certain practices
and procedures associated with the Project Greenlight and DOCS Transitional Services
Program.
1. 	 Project Greenlight program staff develop a detailed release plan for their program
participants.
a. 	 Have you received that release plan in a timely manner?
b. 	 How have you used the release plan?
c. 	 Have you found the release plan useful? Why or why not?
d. 	 Were specific parts of the release plan more useful than others? Which
ones and why?
e. 	 Were specific parts of the release plan more problematic than others?
Which ones and why?
2. 	 Have you had any interaction or other contact with Greenlight parole officers
inside Queensboro—Ms. Benjamin or Ms. Jordan—or the Greenlight corrections
counselors—Ms. Hairston or Mr. Davis—as part of this project?
a. 	 Is that similar to the contact that you would normally have?
b. 	 What was the general purpose of that interaction?
c. 	 How often does (did) it occur?
d. 	 Does (did) it help you in your responsibilities as a field parole officer? Why
or why not?
3. 	 Have you encountered any major obstacles or problems with the way the 

Greenlight program operates?

a. 	 What were they?
b. 	 How could they have been resolved?
4. 	 Did the parolees under your supervision receive their Medicaid cards?
a. 	 Were there delays in receiving it?
b. 	 Were there difficulties posed by not receiving it? If “yes” what were they?
5. 	 Are there other observations that you can make, or issues you can address that
you think are important?

148 Vera Institute of Justice

CONSENT FORMS
Consent to Participate in Research
Evaluation of Project Greenlight
A Transitional Services Demonstration Project

Purpose of the Study
The Vera Institute of Justice is conducting a study aimed at determining the effectiveness of a new
transitional services program located at the Department of Correctional Services’ (DOCS) Queensboro
correctional facility. The research results will help the Vera Institute of Justice and DOCS determine
whether or not the program works to reduce recidivism and help people reintegrate into society. If the
program is successful, it may be expanded to other DOCS facilities.
Description of Research Procedures
Interviews
The study will involve Vera researchers talking to me after my release from DOCS custody. I will
be contacted and interviewed one month after release and then six months after release. I may also
be asked to participate in other interviews, for example if I am rearrested or re-incarcerated, a
Vera researcher will contact me and ask to interview me again. During the interviews, Vera
researchers will ask questions about my background – including age, education, employment,
family information, and drug use. I will also be asked about my experiences in the transitional
services program and about my experiences adjusting to life outside of prison after my release.
Some of the questions will be very personal, such as questions about my mental health or drug use
history. I do not have to answer any questions I do not want to answer. Each interview will last
approximately one and a half hours and researchers may audio-tape the interviews.

Secondary Data
Vera researchers will also obtain information about me from DOCS, the Division of Parole, the New
York State Division of Criminal Justice Services, and/or the New York City Criminal Justice Agency.
This will include, but not be limited to information about my arrest and criminal records, and information
about the services I received while in DOCS custody. If I am in the Vera Queensboro transitional
services program, researchers will also obtain information about my progress in the program from the
program staff.

Study is Voluntary

I understand that I am under no obligation to participate in this study and that participation in this
study is not a requirement of my involvement in any program. If I choose not to participate, I
understand that I will not be penalized in any way and I will not forfeit benefits to which I am
otherwise entitled. I understand that I may choose to stop participating in the study at anytime
and that during the interviews and focus groups I may choose not to answer questions that make
me feel uncomfortable.
Confidentiality
I understand that all the information I provide will remain private. My information will only be made
available to researchers and will only be used for research purposes. When the results of the study are
reported, my name and other identifying information will not be used. Once the research is concluded, all
data will be destroyed.

Vera Institute of Justice 149

Vera researchers will not ask me about plans for future crimes, however, if I share information about a
future crime Vera researchers may have to disclose this information to law enforcement authorities. In
addition, if I reveal that I intend to hurt myself Vera researchers may also reveal this information.

Risks and Benefits
I understand that I may feel uncomfortable answering some of the interview questions. If this happens, I
may stop the interview or refuse to answer the questions
I understand that there is a risk that information collected by researchers could be wrongfully disclosed,
but that Vera researchers will take precautions to prevent this from happening. I also understand that if
anyone other than Vera research staff requests to see information about me, Vera researchers will refuse
this request.
I understand that I will be compensated for my participation in this interview. For each interview I will
receive a fifteen-dollar metrocard.

Contact Information
I understand that if I have questions about the study or about my participation in the study, I may contact
the study’s lead researcher, James Wilson at (212) 376-3085.
Consent is in effect from _______________ (start date) to _______________ (end date).

Signature _________________________________

Date ________________

Witness __________________________________

Date_________________

150 Vera Institute of Justice