The Effects of Prison-Based Educational Programming on Recidivism and Employment

Abstract

This study evaluated the effectiveness of prison-based educational programming by examining the effects of obtaining secondary and post-secondary degrees on recidivism and post-release employment outcomes among offenders released from Minnesota prisons between 2007 and 2008. Obtaining a secondary degree in prison significantly increased the odds of securing post-release employment but did not have a significant effect on recidivism or other employment measures such as hourly wage, total hours worked, or total wages earned. Earning a post-secondary degree in prison, however, was associated with greater number of hours worked, higher overall wages, and less recidivism.

Full text

The Prison Journal
2014, Vol. 94(4) 454 –478
© 2014 SAGE Publications
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0032885514548009
tpj.sagepub.com
Article
The Effects of
Prison-Based
Educational
Programming on
Recidivism and
Employment
Grant Duwe
1
and Valerie Clark
1
Abstract
This study evaluated the effectiveness of prison-based educational
programming by examining the effects of obtaining secondary and post-
secondary degrees on recidivism and post-release employment outcomes
among offenders released from Minnesota prisons between 2007 and 2008.
Obtaining a secondary degree in prison significantly increased the odds of
securing post-release employment but did not have a significant effect on
recidivism or other employment measures such as hourly wage, total hours
worked, or total wages earned. Earning a post-secondary degree in prison,
however, was associated with greater number of hours worked, higher
overall wages, and less recidivism.
Keywords
education, prison, employment, recidivism
1
Minnesota Department of Corrections, St. Paul, USA
Corresponding Author:
Grant Duwe, Research Director, Minnesota Department of Corrections, 1450 Energy Park
Drive, Suite 200, St. Paul, MN 55108-5219, USA.
Email: grant.duwe@state.mn.us
548009TPJXXX10.1177/0032885514548009The Prison JournalDuwe and Clark
research-article2014
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 455
Most offenders in Minnesota state prisons will eventually re-enter society,
but more than one third of those offenders will be convicted of a new felony
offense within 3 years of release (Minnesota Department of Corrections
[MnDOC], 2013). Because incarceration disproportionately affects the young
and the under-educated, many released offenders lack the education and basic
job skills it takes to reintegrate back into society (Tolbert, 2012; Western,
Kling, & Weiman, 2001). Formerly incarcerated men are employed an aver-
age of 9 fewer weeks per year than men who have never been incarcerated
(Western & Pettit, 2010). They also earn 11% less per hour and about 40%
less per year. Besides slashing potential earnings, a history of incarceration
stifles upward economic mobility (Western & Pettit, 2010). Thus, prison-
based education and career training may be a key component of successful
prisoner reentry.
Educational programming is currently available in all Minnesota state cor-
rectional facilities, and more than 9,000 inmates were enrolled in educational
programming between July 2011 and June 2012 (MnDOC, 2013). The prom-
inence of education in prisons is likely due to the well-documented relation-
ship between low educational achievement and antisocial behaviors. Several
studies have linked poor academic performance among adolescents to juve-
nile delinquency and future offending, although the direction of the causal
relationship remains unclear (e.g., Farrington, 2005; Hagan & McCarthy,
1997; Huizinga, Loeber, Thornberry, & Cothern, 2000; Maguin & Loeber,
1996; Moffitt, 1993). A large proportion of adult offenders lack a GED or
high school (HS) diploma (Harlow, 2003).
Although corrections administrators usually value educational program-
ming (Adams et al., 1994), these programs require funding from prison bud-
gets that have not kept pace with growing prison populations and operations
costs. Despite the fourfold increase in corrections spending between 1987
and 2007 (Pew Center on the States, 2008), corrections departments are being
forced to eliminate non-essential services, such as educational programming
(Lillis, 1994). Moreover, legislatures are reluctant to allocate education funds
to this unpopular demographic, as evidenced by the elimination of Pell grants
for prisoners (Batiuk, Lahm, McKeever, Wilcox, & Wilcox, 2005; Tewksbury
& Taylor, 1996).
Policy makers and the general public may view prison educational pro-
gramming as a waste of tax dollars to an undeserving population, but these
programs may offer public safety benefits and future savings in corrections
spending. If participation in prison education programs reduces recidivism,
the public is safer and future inmate populations could be reduced. Moreover,
by increasing employment opportunities for offenders, states can increase tax
revenues.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

456 The Prison Journal 94(4)
Given the current high rates of unemployment in the United States, educa-
tional achievement and career training for offenders may be more important
than ever. Unemployment rates appear to directly correspond with levels of
education. According to the Bureau of Labor Statistics (2012), the unemploy-
ment rate is highest for individuals who have less than an HS diploma (12.5%
as of April 2012), and lowest for individuals who hold a bachelor’s degree or
higher (4% as of April 2012). With or without educational attainment, the
employment prospects of offenders are already weak. A felony record dimin-
ishes the likelihood of future employment (Berstein & Houston, 2000), and
many offenders have unstable work histories (Visher et al., 2004).
The Present Study
This research examines correctional education programming by analyzing
the effects of earning secondary and post-secondary degrees in prison on
recidivism and post-release employment. Offenders in this study were
released from Minnesota prisons between 2007 and 2008. Propensity score
matching (PSM) was used to reduce observable selection bias. Recidivism
and post-release employment data were collected through the end of 2010.
The MnDOC mandates educational programming for all offenders who do
not have at least a GED or HS diploma. Completion of a GED/HS diploma is
required for employment within MnDOC facilities. MnDOC’s overall educa-
tional goal is to not only ensure that all offenders have at least a GED/HS
diploma on release, but to also prepare inmates for enrollment in post-sec-
ondary education. On intake into prison, MnDOC staff verifies whether new
inmates hold a GED/HS diploma by contacting the diploma-granting institu-
tion. New inmates also take the Test of Adult Basic Education (TABE). Based
on GED/HS diploma status and TABE scores, inmates are directed into sec-
ondary or post-secondary education programs.
In the ensuing section, we review existing research on prison-based edu-
cational programming. Next, we discuss the data and methods used in this
study. We then conclude by discussing the implications of the results for cor-
rectional policy and practice.
Prior Research on Prison Educational
Programming
Although existing research has evaluated the effects of prison education pro-
grams on post-release outcomes, the results of these studies have been mixed,
and many suffer from methodological shortcomings (Batiuk et al., 2005;
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 457
Cho & Tyler, 2010). For example, in their evaluation of the effects of cor-
rectional education programs on recidivism in three states (including
Minnesota), Steurer, Smith, and Tracy (2003) found that offenders who par-
ticipated in educational programs while imprisoned had lower rates of recidi-
vism. However, the authors of this study failed to differentiate between types
of educational programs (i.e., secondary and post-secondary), use multivari-
ate analyses to control for other relevant factors, or construct a suitable com-
parison group. Similarly, although Lockwood, Nally, Ho, and Knutson (2012)
reported in their recent evaluation that prison education and post-release
employment reduced recidivism, their non-experimental design lacked a
comparison group. Furthermore, the study did not include important controls
such as prior criminal history or participation in other prison programming.
In one of the earliest and most notable reviews of research on correctional
programming, Martinson (1974) and Lipton, Martinson, and Wilks (1975)
found that prison education programs could be effective but were of limited
value. While many observers assumed that offenders were incapable of
achieving academic success (Adams et al., 1994), Martinson’s and Lipton
et al.’s reviews of prison education program evaluations revealed that offend-
ers were willing to participate in these programs. So long as the teachers and
offenders were invested, prison education programs could improve academic
performance among prisoners but were not found to have a significant effect
on recidivism.
Although the authors (Lipton et al., 1975; Martinson, 1974) claimed they
reviewed only the most rigorous program evaluations (e.g., they included
treatment and control groups, had clearly defined and measurable outcomes),
other researchers have questioned whether the studies they included could be
expected to have reliable results due to implementation issues (Gottfredson,
1979; Palmer, 1978; Van Voorhis & Brown, 1996; Wholey, 1979). More
recent meta-analyses of prison education research have produced promising
results, although the effect sizes are usually modest. Adams et al.’s (1994)
review of more than 90 studies on prison education programs revealed that
prison education reduces the likelihood of recidivism, especially for offend-
ers with the largest education deficits. Aos, Miller, and Drake (2006) found
that basic adult education programs in prison reduced recidivism by more
than 5%, and prison-based vocational programs reduced recidivism by more
than 12% (based on the results of 3 studies). Wilson, Gallagher, and
MacKenzie’s (2000) meta-analysis of 33 evaluations of prison-based educa-
tion programs showed modest increases in post-release employment and
reductions in recidivism for participants.
The above meta-analyses produced positive findings, but not all of the
reviewed studies looked at pre-college, vocational, and college-level
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

458 The Prison Journal 94(4)
educational programs separately. Also, not all of the studies looked at both
recidivism and post-release employment outcomes. Individual studies of pre-
college prison education have produced mixed results.
In a recent study, Cho and Tyler (2010) examined recidivism and post-
release employment outcomes among more than 13,137 released offenders
who participated in Adult Basic Education (ABE) programs while in Florida
prisons. ABE participation did not have a significant effect on recidivism, but
it did significantly improve post-release employment outcomes. The authors
found differences between offenders who chose to stay in ABE classes and
those who voluntarily dropped out. Offenders who chose to complete ABE
classes tended to come from disadvantaged backgrounds. Compared with all
ABE participants, offenders who completed ABE classes did not fare much
better in the job market. However, when comparing ABE completers to ABE
participants who were involuntarily removed, the completers were able to
work longer hours for higher wages after release from prison.
Cho and Tyler (2010) also found that post-release earnings were espe-
cially improved when ABE classes were completed without interruption and
when the offenders pursued GED diplomas. The authors reported that the
average ABE participant earned nearly US$600 more per year than ABE
dropouts in the second year after release.
Anderson (1995) found that GED diploma programs reduced the likeli-
hood of recidivism in a 2-year follow-up of more than 18,000 offenders
released from Ohio prisons in 1992. Participation in ABE programs, how-
ever, did not significantly affect the likelihood of recidivism. Anderson also
found that the negative effects of GED diploma programs on recidivism were
greater for certain groups. Males, younger offenders, African Americans,
offenders with no prior history of incarceration, and offenders who commit-
ted less serious offenses benefitted the most from involvement in GED pro-
grams. ABE programs benefited females more than males, older offenders
more than younger offenders, prisoners with a limited history of incarcera-
tion more than those with length incarceration histories, and inmates serving
longer sentences more than those serving shorter sentences.
Using the same sample of 18,000 offenders released from Ohio prisons,
Anderson (1995) found that college-level academic programs and vocational
training significantly reduced recidivism, especially for certain groups.
Females, younger offenders, persons incarcerated for drug or non-violent
offenses, and offenders with no prior history of incarceration benefited the
most from college-level academic training.
In a more recent analysis on the effects of college-level prison educational
programming, Batiuk and colleagues (2005) found that college-level aca-
demic programming significantly reduced the likelihood of recidivism.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 459
Not only did college-level programming have the strongest effect on recidi-
vism, but it was also the only type of educational programming that had a
significant effect in an analysis that also included HS and GED programming
and vocational training.
One common criticism of educational program evaluations is that research-
ers fail to explain the connection between prison education and recidivism
(Batiuk, Moke, & Rountree, 1997). One theory is that prison educational
achievement increases the likelihood of employment, which in turn decreases
the likelihood of recidivism. Post-release employment keeps offenders occu-
pied and provides them with a disincentive to engage in offending. Batiuk
et al. (1997) provided support for this explanation. The authors found that
post-release employment mediated the relationship between college-level
educational programming in prison and a reduction in recidivisms. College-
level educational programming in prison increased the likelihood of post-
release employment, which in turn decreased the likelihood of recidivism.
Data and Method
This study uses a retrospective quasi-experimental design to determine
whether the completion of prison-based educational programming has had an
impact on recidivism and post-release employment. The effectiveness of edu-
cational programming was evaluated by comparing recidivism and employ-
ment outcomes between offenders who earned secondary (GED or HS
diploma) or post-secondary degrees (e.g., associates degrees and diplomas/
certificates from career/technical programs) in prison and matched compari-
son groups of prisoners who did not earn educational degrees while
incarcerated.
The population for this study contained 9,394 individual offenders released
from Minnesota prisons between January 2007 and December 2008. This
2-year period was selected because individual-level employment data on
Minnesota prisoners did not first become available until 2007. In addition, to
allow a sufficient follow-up period for the recidivism and employment analy-
ses, this study includes offenders released through 2008.
Of the 9,394 offenders, 38% (3,582) entered prison without a secondary
degree (i.e., GED or HS diploma). Of these offenders, 1,212 (33%) earned a
secondary degree in prison. To estimate the effects of earning a secondary
degree on recidivism and employment, we used PSM to individually match
offenders who earned a GED or HS diploma in prison with a comparison
group of offenders released from prison without a secondary degree.
We examined the impact of earning a post-secondary degree on recidivism
and employment by using PSM to individually match offenders who obtained
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

460 The Prison Journal 94(4)
a post-secondary degree in prison with a comparison group of prisoners with
a secondary degree who did not earn a post-secondary degree while incarcer-
ated. Among the 9,394 offenders released from prison during the 2007-2008
period, 62% (5,812) had a secondary degree at the time of their most recent
admission to prison. Of the 5,812 offenders, 545 earned a post-secondary
degree in prison. In addition, there were 148 offenders who obtained both a
secondary degree and a post-secondary degree in prison prior to their release
to the community. The PSM analyses that examine the effects of earning a
post-secondary degree thus include the 693 offenders who earned this type of
degree in prison with 5,267 offenders admitted to prison with a secondary
degree but were released without obtaining a post-secondary degree.
Dependent Variables
As discussed above, two main outcome measures—recidivism and post-
release employment—were used to assess the effectiveness of educational
programming. The following section discusses how each outcome measure
was operationalized.
Recidivism. In this study, recidivism was defined as a (a) rearrest, (b) recon-
viction, (c) reincarceration for a new sentence, or (d) supervision revocation
for a technical violation. It is important to emphasize that the first three recid-
ivism variables strictly measure new criminal offenses. In contrast, technical
violation revocations (the fourth measure) represent a broader measure of
rule-breaking behavior. Offenders can have their supervision revoked for vio-
lating the conditions of their supervised release. Because these violations can
include activity that may not be criminal in nature (e.g., use of alcohol, failing
a community-based treatment program, failure to maintain agent contact,
failure to follow curfew, etc.), technical violation revocations do not neces-
sarily measure reoffending.
Recidivism data were collected on offenders through December 31, 2010.
Considering that offenders in this study were released between 2007 and
2008, the follow-up time for the offenders examined in this study ranged
from 24 to 36 months. Data on arrests and convictions were obtained elec-
tronically from the Minnesota Bureau of Criminal Apprehension.
Reincarceration and revocation data were derived from the Correctional
Operations Management System (COMS) database maintained by the
MnDOC. The main limitation with using these data is that they measure only
arrests, convictions, or incarcerations that took place in Minnesota. As a
result, the findings presented later likely underestimate the true recidivism
rates for the offenders examined here.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 461
Post-release employment. Data on post-release employment were obtained
from the Minnesota Department of Employee and Economic Development
(DEED). The main caveat with using these data is that it does not capture any
labor (or compensation for that labor) not reported to DEED, which can occur
in situations where employees are paid “under the table” for their labor. Still,
the DEED data provide important information not only on whether offenders
obtained employment, but also on how much they worked and the extent to
which they were compensated. Because the employment data are compiled
on a quarterly basis, information was not available on the specific date(s)
when offenders entered and/or exited a job. As a result, the post-release
employment measures included (a) any employment (dichotomized as “1”
for employment and “0” for no employment), (b) total number of hours
worked, (c) total wages earned, and (d) hourly wage.
Educational Programming Variables
The main objective of this evaluation is to determine whether prison-based
educational programming has had an impact on recidivism and post-release
employment. For the secondary degree variable, offenders who earned a
GED or HS degree in prison were assigned a value of “1,” whereas those in
the comparison group received a value of “0.” For the post-secondary degree
variable, offenders who earned this type of degree were given a value of “1,”
whereas those in the comparison group were assigned a value of “0.”
Independent Variables
The independent, or control, variables included in the statistical models were
those that were not only available in the COMS database but also might theo-
retically have an impact on recidivism and post-release employment. A
description of the covariates used in the statistical models can be found in
Table 1.
PSM
PSM is a method that estimates the conditional probability of selection to a
particular treatment or group given a vector of observed covariates
(Rosenbaum & Rubin, 1985). The predicted probability of selection is typi-
cally generated by estimating a logistic regression model in which selection
(0 = no selection; 1 = selection) is the dependent variable whereas the predic-
tor variables consist of those that theoretically have an impact on the selec-
tion process. Once estimated, the propensity scores are then used to match
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

462 The Prison Journal 94(4)
Table 1. Logistic Regression Models for Educational Degree Selection.
Predictors Predictor description
Secondary
Post-
secondary
Coefficient Coefficient
Male Male = 1; female = 0 0.264 −0.818**
Minority Minority = 1; White = 0 −0.968** −0.068
Age at release
(years)
Offender age in years at time of
release from prison
−0.059** −0.038**
Prior supervision
failures
Number of prior revocations while
under correctional supervision
−0.105* −0.058
Prior convictions Number of prior felony convictions,
excluding index conviction(s)
0.033* −0.001
Metro commit Twin Cities metropolitan area = 1;
Greater Minnesota = 0
−0.191* −0.003
Offense type Person offense serves as the
reference
Property Property offense = 1; non-property
offense = 0
−0.525* 0.126
Drugs Drug offense = 1; non-drug offense
= 0
−0.551* −0.171
Criminal
sexual
conduct
Sex offense = 1; non-sex offense = 0 −0.165 −0.199
Felony DWI Felony DWI offense = 1; non-Felony
DWI offense = 0
−0.505* −0.099
Other Other offense = 1; non-other offense
= 0
−0.264 −0.133
Admission type New commitment serves as the
reference
Probation
violator
Probation violator = 1; new
commitment or release violator = 0
0.242* −0.242*
Release
violator
Release violator = 1; new
commitment or probation violator
= 0
−0.999** −0.887**
Length of stay
(months)
Number of months between prison
admission and release dates
0.052** 0.058**
Discipline Number of discipline convictions
received during imprisonment prior
to release
−0.029 −0.062**
CD treatment Entered chemical dependency
treatment during current prison
sentence
−0.087 −0.136
(continued)
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 463
individuals who received educational degrees with those who did not. Thus,
an advantage with using PSM is that it can simultaneously “balance” multiple
covariates on the basis of a single composite score.
PSM reduces selection bias by creating a counterfactual estimate of what
would have happened to offenders had they not earned a secondary or post-
secondary degree. PSM has several limitations, however, that are worth noting.
First, and foremost, because propensity scores are based on observed covariates,
PSM is not robust against “hidden bias” from unmeasured variables that are
associated with both the assignment to treatment and the outcome variable.
Second, there must be substantial overlap among propensity scores between the
two groups for PSM to be effective (Shadish, Cook, & Campbell, 2002); other-
wise, the matching process will yield incomplete or inexact matches. Finally, as
Rubin (1997) pointed out, PSM tends to work best with large samples.
Although somewhat limited by the data available, an attempt was made to
address potential concerns over unobserved bias by including as many theo-
retically relevant covariates (22) as possible in the propensity score model.
Predictors Predictor description
Secondary
Post-
secondary
Coefficient Coefficient
Sex offender
treatment
Entered sex offender treatment
during current prison sentence
0.105 −0.482
Supervision type Supervised release serves as the
reference
ISR ISR = 1; non-ISR = 0 −0.289* −0.146
Work release Work Release = 1; non-Work
Release = 0
0.160 −0.247
CIP CIP = 1; non-CIP = 0 1.899** −0.360
Discharge Discharge = 1; released to
correctional supervision = 0
−1.166** −0.234
Release year Year in which first released from
prison for instant offense
−0.216* −0.132
Constant 435.589 263.463
n 3,582 5,960
Log-likelihood 3,515.174 3,640.498
Nagelkerke R
2
.359 .200
Note. CD = chemical dependency; ISR = intensive supervised release; CIP = Challenge
Incarceration Program.
*p < .05. **p < .01.
Table 1. (continued)
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

464 The Prison Journal 94(4)
In addition, this study later demonstrates there was substantial overlap in
propensity scores between the treated and untreated offenders. Furthermore,
the sample size limitation was addressed by assembling a large number of
cases on which to conduct the propensity score analyses.
Matching for secondary degree. Propensity scores were calculated for the
1,212 offenders who earned a secondary degree in prison and the 2,370 pris-
oners in the comparison group pool by estimating a logistic regression model
in which the dependent variable was obtaining a secondary degree. The pre-
dictors were the 22 control variables used in the statistical analyses (see Table 1).
Even though the difference in mean propensity score between both groups
was statistically significant at the .01 level (see Table 2), there was substantial
overlap in propensity scores. Indeed, the vast majority of offenders in both
groups (87% for secondary degree and 98% for those without a secondary
degree) had propensity scores less than 0.80.
After obtaining propensity scores for the 3,582 offenders, a greedy match-
ing procedure was used to match the offenders who earned a secondary
degree in prison with those who did not. Using a relatively narrow caliper of
0.10, matches were found for 910 (75%) of the 1,212 offenders who earned a
secondary degree in prison. Table 2 presents the covariate and propensity
score means for both groups prior to matching (“total”) and after matching
(“matched”). In addition to tests of statistical significance (“t-test, p-value”),
Table 2 provides a measure (“Bias”) developed by Rosenbaum and Rubin
(1985) that quantifies the amount of bias between the treatment and compari-
son samples (i.e., standardized mean difference between samples), where
X
t
and S
t
2
represent the sample mean and variance for the treated offenders and
X
c
and S
c
2
represent the sample mean and variance for the untreated offend-
ers. If the value of this statistic exceeds 20, the covariate is considered to be
unbalanced (Rosenbaum & Rubin, 1985).
Bias
c
=
()
+
()
100
2
22
XX
SS
t
tc
-
,
As shown in Table 2, the matching procedure reduced the bias in pro-
pensity scores between both groups by 96%. Whereas the p-value was .00
in the unmatched sample, it was .29 in the matched sample. In the
unmatched sample, there were nine covariates that were significantly
imbalanced (i.e., the bias values exceeded 20). But in the matched sam-
ple, covariate balance was achieved given that no covariates had bias
values greater than 20.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 465
Table 2. Propensity Score Matching and Covariate Balance for Secondary Degree.
Variable Sample
Secondary
mean
Comparison
mean Bias (%)
Bias
reduction
t-test
p-value
Propensity score Total 51.84% 24.63% 102.84 .00
Matched 43.86% 42.92% 4.05 −96.07% .29
Male Total 94.06% 91.39% 8.64 .01
Matched 93.08% 93.52% 1.42 −83.52% .71
Minority Total 49.34% 67.81% 30.83 .00
Matched 57.03% 59.45% 4.00 −87.01% .30
Age at release (years) Total 30.85 34.44 30.91 .00
Matched 31.58 31.23 3.15 −89.83% .41
Prior supervision failures Total 0.68 1.44 51.34 .00
Matched 0.81 0.83 1.40 −97.28% .72
Prior convictions Total 4.18 4.62 10.55 .00
Matched 4.21 4.29 2.05 −80.59% .59
Metro Total 39.77% 53.00% 21.91 .00
Matched 43.74% 46.48% 4.50 −79.46% .24
Property offenders Total 17.00% 22.74% 11.99 .00
Matched 19.23% 19.67% 0.91 −92.42% .81
Drug offenders Total 28.47% 26.33% 3.90 .17
Matched 24.29% 24.29% 0.00 −100% 1.00
Sex offenders Total 10.81% 11.10% 0.76 .79
Matched 11.43% 9.89% 4.04 431.03% .29
DWI offenders Total 6.44% 4.30% 7.54 .01
Matched 6.48% 6.48% 0.00 −100% 1.00
Other offenders Total 13.20% 13.08% 0.29 .92
Matched 13.74% 14.40% 1.55 436.89% .69
Probation violators Total 30.36% 25.86% 8.11 .00
Matched 34.07% 35.16% 1.87 −76.91% .62
Release violators Total 4.37% 31.43% 69.87 .00
Matched 5.71% 5.60% 0.39 −99.45% .92
Length of stay (months) Total 23.19 13.10 55.48 .00
Matched 20.37 19.71 3.71 −93.31% .34
Institutional discipline Total 2.71 1.98 18.32 .00
Matched 2.73 2.61 2.99 −83.67% .44
CD Treatment Total 28.55% 13.63% 29.09 .00
Matched 20.55% 21.54% 1.99 −93.16% .61
Sex offender treatment Total 2.97% 2.74% 1.12 .70
Matched 2.97% 2.31% 3.28 193.96% .38
Intensive supervised release Total 23.76% 22.07% 3.27 .25
Matched 25.93% 25.82% 0.20 −93.73% .96
Work release Total 10.73% 8.40% 6.35 .02
Matched 10.44% 11.76% 3.45 −45.62% .37
CIP Total 9.49% 1.56% 25.93 .00
Matched 4.29% 4.07% 0.89 −96.55% .82
Discharge Total 2.72% 21.86% 57.06 .00
Matched 3.63% 3.85% 0.95 −98.33% .81
Release year Total 2,007.44 2,007.43 2.37 .41
Matched 2,007.44 2,007.45 1.08 −54.26% .73
Note. Total Secondary Degree n = 1,212. Total Comparison n = 2,370. Matched Secondary Degree n = 910.
Matched Comparison n = 910. CD = chemical dependency; CIP = Challenge Incarceration Program.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

466 The Prison Journal 94(4)
Matching for post-secondary degree. Propensity scores were calculated for the
693 offenders who earned a post-secondary degree in prison and the 5,267
prisoners in the comparison group pool by estimating a logistic regression
model in which the dependent variable was obtaining a post-secondary
degree (see Table 1). Similar to the analyses for secondary degree, there was
substantial overlap in propensity scores (i.e., 96% of those in the post-sec-
ondary group had scores lower than 0.60 compared with 99% in the compari-
son group pool). After calculating propensity scores for the 5,960 offenders,
the greedy matching procedure was used, once again, to match the offenders
who earned a post-secondary degree in prison with those who did not. Using
the same caliper of 0.10, matches were found for all 693 offenders who
earned a post-secondary degree in prison. Table 3 presents the covariate and
propensity score means for both groups prior to matching (“total”) and after
matching (“matched”).
As shown in Table 3, the matching procedure reduced the bias in propen-
sity scores between post-secondary and comparison group offenders by 98%.
In the unmatched sample, there were five covariates that were significantly
imbalanced. In the matched sample, however, none of the covariates had bias
values greater than 20.
Analytical Procedures
In analyzing recidivism, survival analysis models are preferable in that they
utilize time-dependent data, which are important in determining not only
whether offenders recidivate but also when they recidivate. As a result, this
study uses a Cox regression model, which uses both “time” and “status” vari-
ables in estimating the impact of the independent variables on recidivism. For
the analyses presented here, the “time” variable measures the amount of time
from the date of release until the date of first rearrest, reconviction, reincar-
ceration, technical violation revocation, or December 31, 2010, for those who
did not recidivate. The “status” variable, meanwhile, measures whether an
offender recidivated (rearrest, reconviction, reincarceration for a new crime,
and technical violation revocation) during the period in which she or he was
at risk to recidivate. In the analyses presented below, Cox regression models
were estimated for each of the four recidivism measures.
As noted above, the DEED data are compiled on a quarterly basis and,
thus, do not provide information on the specific date(s) when offenders
entered and/or exited employment. Because employment start date infor-
mation would be needed to use Cox regression, multiple logistic regres-
sion was used to assess the impact of educational programming on
obtaining employment. Considering that logistic regression assumes the
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 467
Table 3. Propensity Score Matching and Covariate Balance for PSD.
Variable Sample PSD mean
Comparison
mean Bias (%)
Bias
reduction
t-test
p-value
Propensity score Total 22.56% 10.19% 70.29 .00
Matched 22.56% 22.38% 0.96 −98.63% .83
Male Total 88.46% 91.51% 8.12 .01
Matched 88.46% 89.18% 1.86 −77.15% .67
Minority Total 36.94% 39.19% 3.79 .25
Matched 36.94% 37.23% 0.49 −87.07% .91
Age at release (years) Total 33.92 35.46 13.32 .00
Matched 33.92 33.83 0.81 −93.92% .85
Prior supervision failures Total 0.90 1.51 35.55 .00
Matched 0.90 0.85 3.41 −90.40% .43
Prior convictions Total 5.30 5.72 7.77 .02
Matched 5.30 5.49 3.59 −53.85% .41
Metro Total 40.69% 44.33% 6.02 .07
Matched 40.69% 40.12% 0.95 −84.28% .83
Property offenders Total 21.36% 21.40% 0.08 .98
Matched 21.36% 24.24% 5.65 6,988.37% .20
Drug offenders Total 27.99% 26.24% 3.20 .33
Matched 27.99% 23.09% 9.08 183.82% .04
Sex offenders Total 10.10% 12.61% 6.57 .06
Matched 10.10% 10.25% 0.41 −93.83% .93
DWI offenders Total 8.23% 8.77% 1.59 .63
Matched 8.23% 9.96% 4.97 213.68% .26
Other offenders Total 12.55% 13.31% 1.86 .58
Matched 12.55% 13.13% 1.42 −23.57% .75
Probation violators Total 20.20% 25.61% 10.68 .00
Matched 20.20% 19.77% 0.88 −91.79% .84
Release violators Total 7.79% 29.26% 51.25 .00
Matched 7.79% 6.20% 5.00 −90.24% .25
Length of stay (months) Total 28.81 14.76 72.21 .00
Matched 28.81 28.21 2.81 −96.10% .53
Institutional discipline Total 259.88% 168.63% 25.12 .00
Matched 259.88% 256.13% 0.97 −96.12% .83
CD treatment Total 31.46% 22.95% 15.42 .00
Matched 31.46% 32.47% 1.77 −88.53% .69
Sex offender treatment Total 3.75% 3.97% 0.94 .78
Matched 3.75% 4.04% 1.23 31.22% .78
Intensive supervised release Total 21.36% 20.52% 1.68 .61
Matched 21.36% 22.37% 2.00 18.93% .65
Work release Total 14.43% 11.58% 6.81 .03
Matched 14.43% 14.29% 0.33 −95.23% .94
CIP Total 4.91% 5.77% 3.16 .35
Matched 4.91% 3.75% 4.55 43.92% .29
Discharge Total 4.33% 13.61% 29.35 .00
Matched 4.33% 3.61% 2.98 −89.86% .49
Release year Total 2,007.46 2,007.42 6.78 .04
Matched 2,007.46 2,007.48 3.07 −54.66% .49
Note. Total PSD n = 693. Total Comparison n = 5,267. Matched PSD n = 693. Matched Comparison n =
693. PSD = post-secondary degree; CD = chemical dependency; CIP = Challenge Incarceration Program.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

468 The Prison Journal 94(4)
lengths of follow-up periods do not vary among offenders, the follow-up
period was capped at 24 months, or eight quarters, for all offenders (i.e.,
for the most recently released offenders, eight was the maximum number
of quarters for which DEED data were available). Because the remaining
employment variables (total numbers of hours worked, total wages earned,
and hourly wage) were ratio-level measures, ordinary least squares (OLS)
regression was used to estimate the impact of educational programming on
these three outcomes.
Results
In Table 4, we present recidivism and post-release employment results for
offenders who earned secondary and post-secondary degrees in prison as well
as for those in the comparison groups. Offenders who obtained a secondary
degree had the same rearrest rate as prisoners in the comparison group,
although they had slightly lower rates of reconviction and reincarceration for
a felony offense. Secondary degree offenders had a higher technical violation
revocation rate, however, than those in the comparison group. Offenders who
earned a post-secondary degree in prison had lower rates of recidivism than
their comparison group counterparts for all four measures.
Post-release employment data show that 60% of offenders who earned
secondary degrees in prison found employment within the first 2 years com-
pared with 50% in the comparison group. The employment rate for offenders
Table 4. Recidivism and Employment by Educational Degree.
Outcomes
Secondary
degree
Secondary
comparison
Post-secondary
degree
Post-secondary
comparison
Recidivism
Rearrest 58.5% 58.5% 54.1% 59.3%
Reconviction 41.3% 43.1% 37.8% 43.4%
Reincarceration 17.3% 21.0% 14.6% 18.6%
Revocation 41.5% 37.8% 34.3% 38.4%
Employment
Employment 59.5% 49.8% 71.0% 68.3%
Total Hours 885 767 1,255 1,057
Total Wages US$10,533 US$9,082 US $16,380 US $13,432
Hourly Wage
a
$11.91 $15.49 $12.05 $12.09
n 910 910 693 693
a
Hourly wage calculated only for offenders who obtained post-release employment
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 469
who earned post-secondary degrees (71%) was slightly higher than that of the
prisoners in the comparison group (68%). Offenders in both educational
degree groups fared better than their comparison group counterparts regard-
ing total hours worked and total wages earned. Among offenders who
obtained employment, those who obtained a secondary degree had a lower
hourly wage than their counterparts in the comparison group. There was no
difference in hourly wage, however, for offenders who earned a post-second-
ary degree in prison and those in the comparison group.
These findings suggest that obtaining educational degrees in prison may
have an impact on the outcomes measured, particularly post-release employ-
ment. It is possible, however, that the observed recidivism and employment
differences are due to other factors such as time at risk, prior criminal history,
discipline history, or post-release supervision. To statistically control for the
impact of these other factors on reoffending, Cox regression models were
estimated for each of the four recidivism measures. In addition, logistic and
OLS regression models were estimated to assess the impact on post-release
employment.
The Impact of Education on Recidivism
The results in Table 5 indicate that, controlling for the effects of the other
independent variables in the statistical model, obtaining a secondary degree
in prison did not have a significant effect on any of the four recidivism mea-
sures. It is worth noting, however, that although the effect for new offense
reincarceration was not statistically significant at the .05 level, it did approach
statistical significance (p = .06). The results also showed that the hazard ratio
was significantly greater for all four recidivism measures for males, minority
offenders, younger offenders, offenders with more prior supervision failures
and convictions, offenders with a metro-area county of commitment, offend-
ers with shorter lengths of stay in prison, and those who incurred institutional
discipline convictions.
The results in Table 5 also suggest that, net of the effects of the other pre-
dictors in the model, earning a post-secondary degree in prison significantly
decreased the risk of reoffending, lowering the hazard by 14% for rearrest,
16% for reconviction, and 24% for new offense reincarceration. Obtaining a
post-secondary degree did not have a significant impact on technical violation
revocations, although this finding approached statistical significance (p = .13).
Many of the significant predictors in the secondary degree analyses pre-
sented in Table 5 were also significant for the post-secondary degree analy-
ses. Metro-area county of commit, however, was not a significant predictor in
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

470 The Prison Journal 94(4)
Table 5. Impact of Secondary and Post-Secondary Degrees on Time to First
Recidivism Event.
Rearrest Reconviction Reincarceration Revocation
SD PSD SD PSD SD PSD SD PSD
HR HR HR HR HR HR HR HR
Secondary
degree
0.994 0.982 0.817 1.107
Post-secondary
degree
0.860* 0.844* 0.759* 0.870
Male 1.701** 1.469** 1.943** 1.996** 2.512** 3.644** 1.996** 1.510*
Minority 1.386** 1.348** 1.264** 1.272* 1.466** 1.638** 1.412** 1.323**
Age at release
(years)
0.970** 0.967** 0.970** 0.968** 0.966** 0.969** 0.982** 0.975**
Prior supervision
failures
1.094** 1.189** 1.092** 1.175** 1.094** 1.263** 1.053** 1.020
Prior convictions 1.158** 1.056** 1.149** 1.052** 1.200** 1.073** 1.233** 1.267**
Metro commit 1.379** 1.130 1.224** 0.891 1.255* 0.772 1.299** 1.074
Offense type
Property 1.167 1.085 1.101 1.030 1.117 0.997 0.917 1.021
Drugs 1.029 1.152 0.890 1.021 0.803 0.755 0.908 0.963
Criminal
sexual
conduct
0.444** 0.556** 0.347** 0.515** 0.454** 0.282** 1.346* 1.926**
Felony DWI 0.997 0.818 0.996 0.845 1.350 0.936 0.869 1.406
Other 1.123 0.898 1.069 0.907 1.029 0.569* 0.985 1.273
Admission type
Probation
violator
0.934 0.852 0.952 0.936 0.677** 0.602** 1.080 0.967
Release
violator
1.081 1.118 1.155 0.907 1.041 0.675 0.795 1.084
Length of stay 0.974** 0.984** 0.974** 0.982** 0.977** 0.975** 0.986** 0.992
Institutional
discipline
1.076** 1.075** 1.062** 1.060** 1.071** 1.059* 1.094** 1.100**
CD Treatment 1.124 1.301 1.141 1.410 1.029 1.356 1.283 1.268
Sex offender
treatment
0.571 0.860 0.331 0.914 0.516 1.515 0.921 0.912
Supervision type
ISR 0.926 0.715** 1.040 0.764* 0.966 0.760 2.019** 1.340*
Work release 0.894 0.706** 0.964 0.781 0.779 0.690 2.128** 1.339*
CIP 0.385** 0.444** 0.484* 0.381** 0.562 0.220* 1.176 1.141
Discharge 1.396 1.015 1.416 1.665* 0.720 0.773
Release year 1.019 1.047 0.847* 0.896 0.658** 0.602** 0.941 1.053
n 1,820 1,386 1,820 1,386 1,820 1,386 1,752 1,331
Note. SD = secondary degree; PSD = post-secondary degree; HR = hazard ratio; CD = chemical
dependency; ISR = intensive supervised release; CIP = Challenge Incarceration Program.
*p < .05. **p < .01.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 471
any of the models. In addition, intensive supervised release (ISR), work
release, and participation in the Challenge Incarceration Program (CIP) low-
ered the risk of recidivism.
The Impact of Education on Post-Release Employment
The results from the logistic regression models, which are shown in Table 6,
reveal that obtaining a secondary degree in prison significantly increased the
chances of securing employment within the first 2 years after release from
prison by 59%. The odds of finding a job were significantly greater for
younger offenders, probation violators, chemical dependency (CD) treatment
participants, offenders released to ISR, offenders placed on work release, CIP
participants, and those with an earlier release year. The odds were signifi-
cantly less, however, for those with a metro-area county of commit.
The results also show that earning a post-secondary degree in prison did
not significantly increase the odds of finding post-release employment.
Similar to the results for obtaining a secondary degree, the odds of finding
employment were greater for younger offenders, CD treatment participants,
offenders placed on work release and those with an earlier release year. The
chances of securing post-release employment were significantly less for male
offenders, drug offenders, and those with institutional discipline
convictions.
As shown in Table 7, obtaining a secondary degree did not have a signifi-
cant effect on total hours worked, total wages earned, or hourly wage. The
results further show that post-release employment was negatively associated
with male offenders (hourly wage), minority offenders (total hours and total
wages), prior convictions (total wages), metro commit (total hours), institu-
tional discipline (total hours), and release year (total hours and total wages).
Employment was positively associated, however, with felony driving while
intoxicated (DWI) offenders (total wages), offenders placed on ISR (total
hours and total wages), offenders placed on work release (all three measures),
and CIP participants (total hours and wages).
In Table 7, the findings suggest that although earning a post-secondary
degree did not have a significant impact on hourly wage, it significantly
increased total hours worked and wages earned. Compared with those in the
comparison group, offenders who obtained a post-secondary degree worked
176 more hours in the follow-up period, net of the effects of the control vari-
ables in the model. Moreover, controlling for the other covariates, these
offenders earned US$2,649 more in wages during the follow-up period than
comparison group offenders.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

472 The Prison Journal 94(4)
Similar to results presented for secondary degree, post-release employ-
ment was negatively associated with male offenders (total hours), minority
Table 6. Logistic Regression Models for Post-Release Employment.
Predictors
Secondary degree Post-secondary degree
Odds ratio SE Odds ratio SE
Secondary degree 1.587** 0.101
Post-Secondary degree 1.206 0.125
Male 1.351 0.218 0.594* 0.242
Minority 0.804 0.116 0.860 0.138
Age at release (years) 0.979** 0.007 0.964** 0.008
Prior supervision failures 1.018 0.018 0.912 0.059
Prior convictions 0.951 0.059 1.010 0.017
Metro commit 0.750** 0.108 0.916 0.136
Offense type
Property 0.987 0.165 0.847 0.208
Drugs 0.748 0.167 0.574* 0.215
Criminal sexual conduct 0.789 0.203 0.874 0.269
Felony DWI 1.063 0.276 0.807 0.304
Other 0.860 0.168 1.046 0.230
Admission type
Probation violator 1.549** 0.131 1.072 0.173
Release violator 1.053 0.315 0.680 0.332
Length of stay 1.004 0.005 1.008 0.005
Institutional discipline 0.995 0.018 0.932** 0.023
CD Treatment 1.601** 0.155 1.832** 0.177
Sex offender treatment 1.218 0.354 1.686 0.384
Supervision type
ISR 1.581** 0.133 1.357 0.178
Work release 7.670** 0.218 9.081** 0.328
CIP 5.778** 0.344 1.598 0.405
Discharge 0.963 0.343 0.749 0.372
Release year 0.643** 0.103 0.679** 0.126
Constant 206.787 253.930
n 1,820 1,386
Log-likelihood 2,267.433 1,517.245
Nagelkerke R
2
.165 .177
Note. CD = chemical dependency; ISR = intensive supervised release; CIP = Challenge
Incarceration Program.
*p < .05. **p < .01.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 473
Table 7. Impact of Secondary and Post-Secondary Degrees on Post-Release
Employment.
Total hours Total wages Hourly wage
SD PSD SD PSD SD PSD
Predictors B B B B B B
Secondary
degree
116.596 1,362.642 −0.443
Post-secondary
degree
176.387* 2,649.196* 0.247
Male 110.347 −480.371** 3,200.968 −2,705.783 −11.051** 1.032
Minority −177.580* −312.905** −4,349.971** −6,193.005** 0.864 −1.740**
Age at release
(years)
−4.225 −6.552 −39.868 −42.210 −0.184 −0.085*
Prior supervision
failures
−7.013 −5.806 127.626 −130.783 0.424 −0.027
Prior convictions −79.487 −96.701* −1,543.478** −1,497.041** −0.230 −0.405
Metro commit −264.332** −88.764 −2,224.258 −727.374 1.860 −0.084
Offense type
Property 141.604 −270.743* 2,267.045 −2,698.761 −2.653 1.119
Drugs 108.482 −64.111 2,573.237 153.443 −3.505 −0.571
Criminal sexual
conduct
55.481 −200.669 1,170.708 −4,694.389 −0.031 −1.203
Felony DWI 321.133 −157.298 5,121.191* 48.694 −3.602 −0.259
Other −24.898 −9.928 −113.276 1,960.952 −3.761 0.522
Admission Type
Probation
violator
107.196 7.936 1,189.582 804.647 −0.098 1.673*
Release
violator
−85.418 −109.167 −1,080.260 −63.609 −4.891 −0.352
Length of stay 7.235* 12.061** 117.508* 196.504** 0.057 0.062**
Institutional
discipline
−48.081** −79.318** −570.200** −1,144.693** −0.587 −0.307**
CD Treatment 200.024 124.768 2,064.623 1,291.361 3.705 1.463
Sex offender
treatment
−32.012 37.748 −301.818 824.921 16.340 1.715
Supervision type
ISR 283.891** 271.706* 3,326.874* 2,889.794 0.863 0.070
Work release 860.134** 776.669** 8,956.948** 9,600.721** 9.749** 2.627**
CIP 1,464.450** 478.938* 16,029.020** 5,325.057 1.657 3.359*
Discharge 267.795 −98.683 3,598.602 −2,672.773 0.421 −1.991
Release year −483.793** −524.242** −6,050.838** −8,269.201** −0.169 −2.045**
Constant 971,895.767 1,054,137.617 12,150,000.000 16,620,000.000 360.652 4,114.438
n 1,820 1,386 1,820 1,386 1,820 1,386
Adjusted R
2
.120 .145 .099 .136 .008 .043
Note. SD = secondary degree; PSD = post-secondary degree; CD = chemical dependency; ISR = intensive
supervised release; CIP = Challenge Incarceration Program.
*p < .05. **p < .01.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

474 The Prison Journal 94(4)
offenders (all three measures), younger offenders (hourly wage), prior con-
victions (total hours and total wages), property offenders (total hours), insti-
tutional discipline convictions (all three measures), and release year (all three
measures). Conversely, employment was positively related to longer lengths
of stay in prison (all three measures), ISR (total hours), work release (all three
measures), and participation in CIP (total hours and hourly wage).
Conclusion
The results reported here suggest that earning a secondary degree in prison
significantly improves an offender’s chances of securing post-release employ-
ment. Although obtaining a secondary degree may help offenders “get their
foot in the door” with employers, it does not necessarily lead to better pay or
more consistent employment. Rather, offenders who earned secondary
degrees in prison did not work significantly more hours or earn higher overall
wages than those in the comparison group. In contrast, earning a post-second-
ary degree in prison did not significantly improve an offender’s odds of find-
ing post-release employment, nor did it result in a higher hourly wage. Yet,
offenders who earned these degrees in prison worked significantly more
hours following their release to the community, resulting in a significant
increase in total wages during the follow-up period.
There are likely a few reasons why offenders who earned post-secondary
degrees were more successful at maintaining employment following their
release from prison. First, a secondary degree generally focuses on basic skill
development, whereas a post-secondary degree is geared more toward pro-
viding students with the knowledge required to succeed within a particular
field or discipline. Second, there are likely differences in the types of jobs
available to secondary degree graduates versus those with a post-secondary
degree. For example, offenders with post-secondary degrees may be more
likely to find permanent positions that require higher levels of skill and edu-
cation. Offenders with secondary degrees, however, may be more likely to
find short-term, temporary employment.
Ensuring that offenders obtain employment following their release from
prison is important for a number of reasons. Yet, when it comes to reducing
recidivism, maintaining employment is what appears to be critical. Indeed,
existing research suggests individuals are less likely to commit crime when
they work more often (Uggen, 1999). We observed that inmates who earned
post-secondary degrees not only worked more hours, but they also earned
more total wages, which may have reduced their economic need. Maintaining
employment may also expand informal social control by giving individuals a
greater stake in conformity and involvement in conventional activities, which
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 475
inhibit opportunities for criminal behavior. Furthermore, associating with
others who are employed increases the likelihood that offenders will develop
or maintain pro-social values, beliefs, and attitudes.
Although the findings suggest that prison-based educational programming
can produce positive recidivism and employment outcomes, it is worth not-
ing the limitations with this study. First, because post-release programming
data were not available, we were unable to determine whether released
offenders obtained educational degrees in the community during the follow-
up period. Second, we were also unable to control for prior work history due
to the absence of pre-incarceration employment data. Still, it is important to
note that earning a secondary or post-secondary degree did not have a signifi-
cant effect on hourly wage, which weakens the argument that the results
observed here were due to the fact that offenders who earned degrees had
more impressive prior work histories than offenders in the comparison group.
If true, then it would be reasonable to expect that this pre-incarceration dif-
ference, if it exists, would result in a significantly higher hourly wage for
offenders who earned secondary and post-secondary degrees, which was not
the case.
Despite these limitations, the results suggest, on the whole, that more
emphasis should be placed on increasing offender access to post-secondary
educational opportunities. This is not to say, however, that increasing the
availability of post-secondary educational programming should limit offender
access to secondary education. On the contrary, although we found that
obtaining a secondary degree in prison did not significantly reduce recidi-
vism, it did significantly elevate the odds of finding a job. Moreover, earning
a secondary degree is critical insofar as it is a prerequisite to post-secondary
educational enrollment.
Expanding the availability of post-secondary education for prisoners
would be in step with the ever-increasing educational demands from employ-
ers. Compared with the population in general, released prisoners are gener-
ally at a disadvantage due not only to the educational and employment history
deficits they often have, but also to the harmful effects that prior criminal
history has on obtaining employment. Although obtaining a post-secondary
degree will not erase the stigmatizing mark of a criminal record, it can help
make offenders more competitive in the labor market.
To be sure, investing more in prisoner educational programming, espe-
cially access to post-secondary education, may prove to be more costly in the
short term. Over the long run, however, this investment may produce divi-
dends by increasing offender employment and decreasing the extent to which
offenders recidivate. When released prisoners maintain employment, they
contribute to local, state, and federal tax revenues. And, when offenders
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

476 The Prison Journal 94(4)
reoffend less often, they victimize fewer people and are less likely to con-
sume costly criminal justice resources, especially prison.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research,
authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publi-
cation of this article.
References
Adams, K., Bennett, K. J., Flanagan, T. J., Marquart, J. W., Cuvelier, S. J., Fritsch,
E., . . .Burton, V. S. (1994). A large-scale multidimensional test of the effect of
prison education programs on offenders’ behavior. The Prison Journal, 74, 433-
449.
Anderson, D. (1995). Evaluation of the impact of correctional education programs on
recidivism. Columbus: Ohio Department of Rehabilitation and Correction, Office
of Management Information Systems, Bureau of Planning and Evaluation.
Aos, S., Miller, M., & Drake, E. (2006). Evidence-based adult corrections programs:
What works and what does not. Olympia: Washington State Institute for Public
Policy.
Batiuk, M. E., Lahm, K. F., McKeever, M., Wilcox, N., & Wilcox, P. (2005).
Disentangling the effects of correctional education: Are current policies mis-
guided? An event history analysis. Criminal Justice, 5, 55-74.
Batiuk, M. E., Moke, P., & Rountree, P. W. (1997). Crime and rehabilitation:
Correctional education as an agent of change—A research note. Justice Quarterly,
14, 167-180.
Berstein, J., & Houston, E. (2000). Crime and work: What we can learn from the low-
wage labor market. Washington, DC: Economic Policy Institute.
Bureau of Labor Statistics. (2012). Economic news release: Table A-4. Employment
status of the civilian population 25 years and over by educational attainment.
Retrieved from http://www.bls.gov/news.release/empsit.t04.htm
Cho, R. M., & Tyler, J. H. (2010). Does prison-based adult basic education improve
postrelease outcomes for male prisoners in Florida? Crime & Delinquency, 59,
975-1005.
Farrington, D. P. (2005). Childhood origins of antisocial behavior. Clinical Psychology
& Psychotherapy, 12, 177-190.
Gottfredson, M. (1979). Treatment destruction techniques. Journal of Research in
Crime & Delinquency, 16, 39-54.
Hagan, J., & McCarthy, B. (1997). Intergeneratioal sanction sequences and trajec-
tories of street-crime amplification. In I. H. Gotlib & B. Wheaton (Eds.), Stress
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

Duwe and Clark 477
and adversity over the life course: Trajectories and turning points (pp. 73-90).
Cambridge, UK: Cambridge University Press.
Harlow, C. W. (2003). Education and correctional populations [Special report].
Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics.
Huizinga, D., Loeber, R., Thornberry, T. P., & Cothern, L. (2000). Co-occurrence
of delinquency and other problem behaviors (Juvenile Justice Bulletin).
Washington, DC: U.S. Department of Justice, Office of Juvenile Justice and
Delinquency Prevention.
Lillis, J. (1994). Prison education programs reduced. Corrections Compendium, 19,
1-4.
Lipton, D. S., Martinson, R., & Wilks, J. (1975). The effectiveness of correctional
treatment: A survey of treatment evaluation studies. Albany: New York Office of
Crime Control Planning.
Lockwood, S., Nally, J., Ho, T., & Knutson, K. (2012). The effect of correc-
tional education on postrelease employment and recidivism: A 5-year fol-
low-up study in the state of Indiana. Crime & Delinquency, 58, 380-396.
doi:10.1177/0011128712441695
Maguin, E., & Loeber, R. (1996). Academic performance and delinquency. In M.
Tonry (Ed.), Crime and justice: A review of research (Vol. 20, pp. 145-264).
Chicago, IL: University of Chicago Press.
Martinson, R. (1974). What works? Questions and answers about prison reform. The
Public Interest, 34, 22-54.
Minnesota Department of Corrections. (2013). Performance report: Fiscal year 2012.
St. Paul: Author. Retrieved from http://www.doc.state.mn.us/PAGES/files/large-
files/Publications/legislative-reports/2012DOCPerformanceReport.pdf
Moffitt, T. E. (1993). Adolescent-limited and life-course-persistent antisocial behav-
ior: A developmental taxonomy. Psychological Review, 100, 674-701.
Palmer, T. (1978). Correctional intervention and research: Current issues and future
prospects. Lexington, MA: Lexington Books.
Pew Center on the States. (2008). One in 100: Behind bars in American 2008.
Washington, DC: The Pew Charitable Trusts.
Rosenbaum, P. R., & Rubin, D. B. (1985). Constructing a control group using mul-
tivariate matched sampling methods that incorporate the propensity score. The
American Statistician, 39, 33-38.
Rubin, D. B. (1997). Estimating causal effects from large data sets using propensity
scores. Annals of Internal Medicine, 127, 757-763.
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-
experimental designs for generalized causal inference. Boston: Houghton
Mifflin.
Steurer, S. J., Smith, L., & Tracy, A. (2003). Three state recidivism study. Lanham,
MD: Correctional Education Association.
Tewksbury, R., & Taylor, T. M. (1996). The consequences of eliminating Pell Grant
eligibility for students in post-secondary correctional education programs.
Federal Probation, 60, 60-63.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from

478 The Prison Journal 94(4)
Tolbert, M. (2012). A reentry education model: Supporting education and career
advancement for low-skill individuals in corrections. Washington, DC: U.S.
Department of Education, Office of Vocational and Adult Education.
Uggen, C. (1999). Ex-offenders and the conformist alternative: A job quality model
of work and crime. Social Problems, 46, 127-151.
Van Voorhis, P., & Brown, K. (1996). Evaluability assessment: A tool for program
development in corrections [Report]. Washington, DC: National Institute of
Corrections.
Visher, C. A., LaVigne, N., & Travis, J. (2004). Returning home: Understanding the
challenges of prisoner reentry. Maryland Pilot Study: Findings from Baltimore.
Washington, DC: Urban Institute, Justice Policy Center.
Western, B., Kling, J. R., & Weiman, D. F. (2001). The labor market consequences of
incarceration. Crime & Delinquency, 47, 410-427.
Western, B., & Pettit, B. (2010). Collateral costs: Incarceration’s effect on economic
mobility. Washington, DC: The Pew Charitable Trusts.
Wholey, J. S. (1979). Evaluation: Promise and performance. Washington, DC: The
Urban Institute.
Wilson, D. B., Gallagher, C. A., & MacKenzie, D. L. (2000). A meta-analysis of
corrections-based education, vocation, and work programs for adult offenders.
Journal of Research in Crime & Delinquency, 37, 347-368.
Author Biographies
Grant Duwe is the director of Research and Evaluation, Minnesota Department of
Corrections. The author of the book, Mass Murder in the United States: A History
(McFarland & Co., Inc.), his recent work has been published in Criminology and
Public Policy, Journal of Criminal Justice, Criminal Justice Policy Review, Criminal
Justice Review, Sexual Abuse: A Journal of Research and Treatment, Criminal Justice
and Behavior, and Crime & Delinquency.
Valerie Clark is a research analyst, Minnesota Department of Corrections. Her
research has been published in Journal of Interpersonal Violence, Duquesne Law
Review, and Criminal Justice Policy Review. She is the author of a forthcoming book
on adolescent intimate partner violence.
at DEPT OF CORRECTIONS on October 22, 2015tpj.sagepub.comDownloaded from