The Impact of Vocational Education and Training Programs on Recidivism: A Systematic Review of Current Experimental Evidence

Abstract

Although the association between unemployment and offending is well established, relatively little is known about the impact of vocational education and training programs on re-offending, with much of the previous work in this area failing to control for, or correct, selection bias. This article reports the findings of a systematic review, which considers the findings of only those studies that have used experimental or quasi-experimental designs to evaluate vocational training and employment program outcomes for adult offenders. The analysis identifies key features, based on these studies, of those programs associated with the best outcomes and recommends selection criteria for those who are most likely to benefit from prison vocational education and training.

Keywords

employment vocational training prisoners recidivism

Introduction

It is well known that prisoners, ex-prisoners, and ex-offenders often experience considerable difficulty in obtaining and maintaining paid employment. Metcalf, Anderson, and Rolfe (2001), for example, estimated that offenders in the United Kingdom are rejected for about half of all job vacancies, with the refusal rate rising to 90% for those with serious convictions. Another U.K. study reported that nearly two thirds of all prisoners are assessed as having problems in the area of employment, education, and training (Harper & Chitty, 2005). In the United States, Nally, Lockwood, Ho, and Knutson (2014) concluded that over one third of offenders never find employment following release from prison. This may, in part, be a result of both specific job-related issues (such as inadequate skills or limited previous experience) and broader factors such as the impact of disability, substance use dependency, and limited literacy and numeracy (Graffam, Shinkfield, & Hardcastle, 2008), although social and state contexts are also likely to be important in understanding differences between jurisdictions. The United States, for example, incarcerates a significantly greater proportion of its citizens than any other Organisation for Economic Co-Operation and Development (OECD) country (almost 700 per 100,000), with the United Kingdom and similar countries having much lower rates of imprisonment (around 150 per 100,000; International Centre for Prison Studies [ICPS], 2015) and a more “residualized” prisoner population in terms of educational attainment, lack of employment history, levels of disability, and so on. Notwithstanding this, it is clear that levels of education, training, and employment for ex-prisoners everywhere are much lower than for the general population.

Prison administrations seek to address these problems by offering a range of different pre- and post-release services that aim to improve the employability of ex-offenders and ex-prisoners. These include work readiness training, vocational education and training, and job placement to improve skill sets and address poor work histories. It is often assumed that offenders who find employment will be less likely to re-offend than those who do not (Cherney & Fitzgerald, 2014; Krienert & Fleisher, 2004; Phipps, Korinek, Aos, & Lieb, 1999). Often implicit in this thinking is the idea that engagement with employment will strengthen social bonds and ties to social institutions that act as key protective factors against crime (see Sampson & Laub, 1977). In the words of Visher, Winterfield, and Coggeshall (2005),

. . . a good job not only provides the means for basic survival, but also is a key element in rebuilding self-esteem, attachment to conventional lifestyle, and a sense of belonging in the community. Work organizes daily behavior and patterns of interaction, and becomes an important source of informal social control for ex-offenders. (pp. 295-296)

Given that prison training programs have the potential to be offered to large numbers of prisoners (Christofferson, 2014), they have become an attractive option for correctional administrators who are seeking to reduce re-offending and re-incarceration rates. There is, however, only limited evidence to support claims that participation reduces the risk of re-offending, with one of the most commonly cited studies in this area published nearly 20 years ago by Gendreau, Goggin, and Gray (1998). The only previously published systematic review of outcome research in this area, by Visher et al. (2005), identified only eight studies that examined the outcomes of community-delivered vocational training and employment programs with incarcerated populations, and even these were described as “quite disparate in terms of primary intervention and target population” (p. 310). They were nonetheless able to conclude that interventions had “no [statistically] significant effect on the likelihood of arrest among ex-offenders” (Visher et al., 2005, p. 309). This somewhat disappointing conclusion may, of course, have reflected the inability of ex-prisoners to secure meaningful paid employment, rather than point to problems in the quality of training provided, but highlights the need for more rigorous evaluations of these interventions. The aim of this study, conducted some 10 years after the publication of the Visher et al. review, is to establish just how much progress has been made in documenting the impact of post-release vocational training and employment programming on subsequent offending.

Method

A systematic review methodology was chosen to summarize the best available research on program outcomes as it uses transparent procedures to find, evaluate, and synthesize the results of relevant work and minimizes researcher bias in an area of practice in which many have a personal investment or interest. Accordingly, this review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Moher, Liberati, Tetzlaff, Altman, & the PRISMA Group, 2009). Eligible studies were identified through several search strategies, as shown in Figure 1, designed to identify abstracts published between January 2000 and March 2015. The following databases were searched: PsycINFO, CINCH, Academic Search Complete, Criminal Justice Abstracts, Education Source, SocINDEX, Web of Science, Sociological Abstracts, Medline (OVID), Social Work Abstracts (EBSCO), Legal Source, Social Science Citation Index, and Business Source Complete. The following keyword combinations were used: Offender* or prisoner* or imprisonment or parolee* or inmate or ex-prisoner* or jail AND Employ* or vocation* or training or education* or technical.

Figure 1.

Flowchart of the study selection process.

Reference sections from previous reviews were then hand-searched, and a web search was conducted using the same keyword combinations to identify unpublished work or work in progress. Studies were identified as eligible for inclusion as long as they contained text describing the trial in detail. A number of corresponding authors were also contacted to source relevant articles not yet published.

A total of 778 abstracts were identified using this methodology. A review of the titles and abstracts of these articles identified 68 articles that potentially reported the outcomes of employment and/or vocational training programs for prisoners and/or ex-offenders. The full text of each of these was then assessed and read in full by two authors of this article. Studies were retained if the following criteria were met: (a) participants were prisoners or ex-offenders who had been arrested, convicted, or incarcerated in connection with a criminal charge before becoming a study participant; (b) participants were adults; (c) one or more groups of participants had participated in a community-based or prison-based program that included a job placement, job training, or a vocational education/training component (although other components may have also been included); (d) program outcomes were compared with those of one or more control groups that received either “treatment as usual” or no “treatment”; (e) both the intervention and comparison groups consisted of those who volunteered to receive the program; and (f) the study utilized random assignment. The decision to retain only those studies that used experimental or quasi-experimental designs was made in an attempt to minimize selection bias (a common methodological challenge in evaluation research) by ensuring that the baseline characteristics of those assigned to the intervention group reflected, on average, those of a comparison or control group. We acknowledge that this method does not allow inclusion of many studies that provide insight into the mechanisms of change in vocational training and employment programs but is important in answering a question of causality—in this case, the evidence that exists to show that program participation leads to reduced re-offending.

Results

A total of 12 studies met the inclusion criteria; these are summarized in Table 1. These studies reported data from seven independent evaluations. The first of these was undertaken in 1975 (Uggen, 2000), with the most recent initiated in 2009 (Cook, Kang, Braga, Ludwig, & O’Brien, 2014). Seven of the studies were reported in academic journals and five were unpublished government agency reports available from study or government websites. All of the studies involved programs conducted in the United States. Each of the identified studies is outlined below, reported in chronological order by date of program commencement. We report the key characteristics of each study separately, before bringing together common and disparate elements in the discussion.

Table 1.

Summary of Independent Studies on Prisoner/Offender Employment Programs and Vocational Training, 2000 to 2014.

Study	Country	Participants	Type of interventions	Outcome measures	Key findings
National Supported Work Demonstration Project (Uggen, 2000)	The United States	3,000 people with an official arrest history assigned to treatment or control group	National Supported Work Demonstration Project—Participants offered minimum-wage jobs (mainly in construction and service industries) in crews of 8 to 10 workers led by a counsellor/supervisor	Employment, crime, arrests	Those aged 27 and older were less likely to report crime and arrest when provided with marginal employment opportunities than when such opportunities are not provided. Among young participants, those in their teens and early 20s, the experimental job treatment had little effect on crime
ESEO (Bierens & Carvalho, 2011)	The United States	2,045 participants who had been incarcerated at an adult federal, state, or local correctional facility for at least 3 months and had been released within 6 months of program participation	Participants assigned randomly to control or treatment group. ESEO program—Primarily consists of job-related counselling crisis intervention, counselling, and, whenever necessary, re-employment assistance, during a period of 6 months after the first placement	Recidivism	ESEO has an effect on recidivism, but this effect depends on age and location
Employment intervention for drug court clients (Leukefeld, Webster, Staton-Tindall, & Duvall, 2007; Webster, Staton-Tindall, Dickson, Wilson, & Leukefeld, 2014)	The United States	500 drug-involved offenders (65% male, M = 30 years); non-violent charges, a self-admitted drug problem	The employment intervention consisted of a series of individual and group sessions with a study-hired employment specialist Individual sessions incorporated behavioral contracting, motivational interviewing, and strengths-based case management with a focus on employment barriers and resolving issues that impede employment success Group sessions also used motivational interviewing as well as structured stories and thought-mapping to address employment needs	Addiction Severity Index; the key 12-month employment outcomes measures included typical work status since the baseline interview, days paid for legal employment in the past 30 days and since the baseline interview (past year), and income from legal employment in the past 30 days and since the baseline interview (past year)	When employment, legal work, illegal work, and employment problems were examined for 1 year and 30 days at follow-up, there were statistically significant effects for jobs in the past year, days worked at a legitimate job in the past year as well as 30 days, and income from a legitimate job in the past year (Leukefeld et al., 2007) The intervention was successful in increasing the number of days worked in the past year but no other employment outcomes. However, analyses focusing on work trajectories showed that those who were on a negative pre-baseline work trajectory were most helped by the intervention. Specifically, participants with a negative pre-baseline work trajectory, who were randomly assigned to the tailored employment intervention, were more likely to be in the workforce, earn more income from legal employment, and more likely be on a positive work trajectory at 12-month follow-up than those with a negative pre-baseline work trajectory and assigned to the control group (Webster et al., 2014)
The CEO (Bloom, Redcross, Zweig, & Azurdia, 2007; Redcross, Bloom, Azurdia, Zweig, & Pindus, 2009; Redcross, Millenky, Rudd, & Levshin, 2012; Zweig, Yahner, & Redcross, 2011)	The United States	977 former prisoners: 568 in the program group and 409 in the control group	Individuals who were assigned to the CEO program group were eligible for all of CEO’s services, including the pre-employment class, the transitional job, job coaching, job development, parenting classes, and post-placement services	Employment, recidivism, cost–benefit analysis; recidivism rates among participants at low, medium, and high risk of re-offending	The 3-year evaluation found that CEO substantially increased employment early in the follow-up period but that the effects faded over time. The initial increase in employment was due to the temporary jobs provided by the program. After the first year, employment and earnings were similar for both the program group and the control group. CEO significantly reduced recidivism, with the most promising impacts occurring among a subgroup of former prisoners who enrolled shortly after release from prison (the group that the program was designed to serve). Among the subgroup that enrolled within 3 months after release, program group members were less likely than their control group counterparts to be arrested, convicted of a new crime, and re-incarcerated. The program’s impacts on these outcomes represent reductions in recidivism of 16% to 22%. In general, CEO’s impacts were stronger for those who were more disadvantaged or at higher risk of recidivism when they enrolled in the study. CEO’s benefits to society outweighed its costs under a wide range of assumptions. Financial benefits exceeded costs for taxpayers, victims, and participants. The majority of CEO’s benefits came in the form of reduced criminal justice system expenditures (Redcross et al., 2012). For former prisoners in the high-risk subgroup, CEO reduced the probability of re-arrest, the probability of re-conviction, and the number of re-arrests in Year 2 after random assignment. Also, for the high-risk subgroup, the impact of CEO in reducing the number of re-arrests in Years 1 and 2 combined was marginally significant. There were no statistically significant program impacts on recidivism in Year 1 for those in the high-risk subgroup. The impact of CEO on high-risk former prisoners did not emerge until the second year following random assignment to the program. There were few program impacts on recidivism—of any type—for former prisoners in the low- and medium-risk subgroups (Zweig et al., 2011)
TJRD (Jacobs, 2012; Redcross et al., 2010)	The United States	Men 18 years and older, released from state prison within 90 days prior to enrolment in the study; 912 participants randomly assigned to the TJRD transitional jobs program, and 901 were referred to the job search program in their city and were offered basic search and placement assistance but were not offered a transitional job	TJRD transitional jobs program provided participants with temporary, minimum-wage jobs that offered 30 to 40 hr of paid work each week. Assistance was provided to look for unsubsidized jobs following transitional jobs program	Employment, earnings, and recidivism	The transitional jobs programs substantially increased employment early in the study period by providing jobs to many who would not otherwise have worked. However, the gains faded as men left the transitional jobs, and the programs did not increase regular (unsubsidized) employment either during or after the program period The transitional jobs programs did not affect key measures of recidivism over the 2-year follow-up period
Employment-focused re-entry program (Farabee, Zhang, & Wright, 2014)	The United States	217 ex-prisoners (179 male, 38 female)	The re-entry program is a 4-week course followed by employment placement assistance and computer labs (accessible during business hours) that all program participants can use for resume writing and job-seeking activities	Employment, residential stability, criminal activity (No. of times arrested in last 12 months)	No statistically significant differences were found for any of the between-group comparisons on any of the major intervention outcomes
Milwaukee Safe Streets Prisoner Release Initiative (Cook, Kang, Braga, Ludwig, & O’Brien, 2014)	The United States	217 ex-offenders with a history of violence or gang involvement	The Milwaukee Safe Streets Prisoner Release Initiaive (PRI) provided wraparound services to address the various challenges facing prisoners. These services included 6 months of programming prior to release and a variety of supports for post-release employment	Effects on employment and earnings, adequacy of earnings, recidivism	Increased employment rates and earnings during the period when ex-offenders are eligible for subsidized jobs were observed, and these gains persisted throughout the year The intervention reduces the likelihood of re-arrest. The likelihood that the treatment group is re-imprisoned during the first year after release is lower than for controls (22% vs. 26%), but the difference is not statistically significantly different from zero

Note. ESEO = Employment Services for Ex-Offenders; CEO = Center for Employment Opportunities; TJRD = Transitional Jobs Reentry Demonstration.

National Supported Work Demonstration Project

The National Supported Work Demonstration Project (Uggen, 2000) was a large-scale experimental employment program implemented in nine U.S. cities over 40 years ago, from March 1975 to July 1977. The program was designed to target offenders, drug users, and youth dropouts who had a history of both chronic and recent unemployment. To be eligible, offenders were required to have been incarcerated in the 6 months prior to enrolment. Those with an official arrest history were referred by criminal justice, social service, and job-training agencies, and randomly assigned to either an intervention or control group. Those who received the intervention were offered minimum-wage jobs (mainly in the construction and service industries) in crews of 8 to 10 workers led by a counsellor/supervisor. Members of both groups reported work, crime, and arrest information at 9-month intervals for up to 3 years.

Uggen (2000), in a re-analysis of the original data, examined recidivism (defined as first self-reported arrest after program entry) among 3,105 participants (split into two groups: those aged 26 years and under, and those aged 27 years and over) who had an official arrest history. Although a previous analysis concluded that the program had no short- or long-term effect on criminal activity (Piliavin & Gartner, 1981), this later analysis revealed the impact of the program to be age-related. The program had little effect on crime for the young group but reduced re-arrests by 8% for those aged 27 and older.

Employment Services for Ex-Offenders (ESEO) Program

The ESEO (Bierens & Carvalho, 2011) was a National Institute of Justice–sponsored experiment to evaluate the impact of re-employment programs for recently released prisoners, which ran from 1980 to 1985. Three established programs (the Comprehensive Offender Resource System in Boston, the Safer Foundation in Chicago, and Project JOVE in San Diego) involving a total of 2,045 prisoners participated. To be eligible, prisoners had to (a) voluntarily accept program services; (b) have been incarcerated in an adult federal, state, or local correctional facility for at least 3 months and be released within 6 months of program participation; and (c) exhibit a pattern of income-producing offenses. Participants were assigned randomly to either the intervention or the control group. Both groups received normal services (orientation, screening, evaluation, support services, job development seminar, and job search coaching), with the intervention group also being assigned to a follow-up specialist who provided support during the job search and for 180 days following job placement. This service mainly involved crisis intervention, counselling, and, whenever necessary, re-employment assistance.

Although previous analyses of ESEO program data using ordinary least squares (OLS) regressions had reported negligible program effects (Milkman, 2001), Bierens and Carvalho (2011) used a bivariate mixed proportional hazard model which showed that reductions in recidivism were contingent on age and location—Recidivism reduced for ex-offenders who were over the age of 27 in San Diego and Chicago, and over the age of 36 in Boston, but increased the risk of recidivism for the other ex-offenders in the program group. It was therefore concluded that these programs can reduce recidivism, but only when the age of ex-offenders is taken into account.

Employment Intervention for Drug Court Clients

A tailored employment intervention for 500 drug-involved offenders recruited between March 2000 and December 2002 from two Kentucky drug courts was the subject of this evaluation (Leukefeld, Webster, Staton-Tindall, & Duvall, 2007; Webster, Staton-Tindall, Dickson, Wilson, & Leukefeld, 2014). Study eligibility was based on the standard criteria for entry into the drug court program (non-violent charges, a self-admitted drug problem, consent to a urine test, and an Addiction Severity Index classification as a drug abuser). Following the baseline interview, each participant was randomly allocated to one of two groups: the employment intervention group or the “drug court as usual” comparison group. The employment intervention consisted of a series of individual and group sessions delivered by an employment specialist who had experience in both employment and substance abuse counselling which aimed to develop job readiness and life skills. The intervention was delivered in three phases: obtaining employment, maintaining employment, and upgrading employment (see Leukefeld, 2003, for a full description of the intervention).

In the evaluation reported by Leukefeld et al. (2007), the percentage of upgrading sessions attended was calculated for each participant. Participants were then split into two groups: Those below the median were allocated to the “low-upgrading” group (n = 118) and those above the median were allocated to the “high-upgrading” group (n = 120). Along with the control group (n = 239), data from these groups were used in the analysis.

Statistically significant associations were found between the level of intervention received and a number of variables, including full-time and part-time employment patterns in the previous year, days worked at a legitimate job in the previous year, and income from a legitimate job source in the past year. For example, participants in the high-upgrading group reported more full-time employment (83%) than those in either the low-upgrading group (54%) or the control group (59%). They also reported working more days in the past year (M = 275.8, SD = 109.9) and in the previous 30 days (M = 23.4, SD = 9.9) than those in both the low-upgrading group (M = 185.7, SD = 109.5; M = 13.4, SD = 13.7) and the control group (M = 215.4, SD = 126.1; M = 19.4, SD = 12.4). In addition, those who received more of the employment intervention reported less alcohol, marijuana, cocaine, and multiple substance use than those in other groups, as well as less criminal activity at follow-up. Specifically, participants in the high-upgrading group reported less: property damage, handling of stolen goods, stealing an item worth more than US$50, selling marijuana, or having friends who sold drugs than those in either the low-upgrading or the control group. This suggests that substance-involved ex-offenders who remain engaged in an employment intervention have more positive outcomes across a range of areas than those who drop out early or do not receive a specialist service.

Data from this study were further analyzed by Webster et al. (2014) using intent-to-treat analyses to examine 12-month differences in employment outcomes between the intervention group and the control group. Overall, these analyses indicated that the intervention was successful in increasing the number of days worked in the past year (210.1 vs. 199.9 days) but not other employment outcomes. However, further analyses revealed that those with a negative pre-baseline work pattern, who were assigned to the intervention, were more likely to be in the workforce, earn more income from legal employment, and more likely to be on a positive work trajectory at 12-month follow-up than those with a similar history assigned to the control group. Therefore, this intervention appears to be particularly efficacious among those lacking recent employment success.

The Center for Employment Opportunities (CEO)

The CEO (Bloom, Redcross, Zweig, & Azurdia, 2007; Redcross, Bloom, Azurdia, Zweig, & Pindus, 2009; Redcross, Millenky, Rudd, & Levshin, 2012; Zweig, Yahner, & Redcross, 2011) in New York City, a comprehensive employment program for former prisoners, is one of four sites in the Enhanced Services for the Hard-to-Employ Demonstration and Evaluation project, sponsored by the U.S. Department of Health and Human Services. Between January 2004 and October 2005, 977 ex-prisoners who were referred by their parole officer and met the study eligibility criteria were randomly assigned to one of two groups: the program group (n = 568), who were provided with access to CEO’s jobs and other services, and a control group (n = 409), who were offered basic job search assistance at CEO along with other services in the community.

The CEO program began with a pre-employment class lasting 5 days after which participants were immediately placed in a transitional job in one of CEO’s work crews. Participants worked 7 hours a day, 4 days per week, and were paid each day for the work performed that day. Participants met with their job coaches (case managers) and job developers once a week and were given assistance in finding a permanent position when they were deemed “job-ready.” They were also able to participate in other services, such as a fatherhood program and parenting classes.

At 12-month follow-up, the evaluation data indicated that members of the program and control groups were equally likely to be employed, with similar earnings. Among those who came to CEO within 3 months of release, program group members were less likely to have had parole revoked, to have been convicted of a felony, and to have been re-incarcerated (Bloom et al., 2007). At 2-year follow-up, the program group was less likely than the control group to have been convicted of a crime, to have been admitted to prison for a new conviction, or to have been incarcerated for any reason (Redcross et al., 2009). A 3-year follow-up, however, found that although CEO participation substantially increased employment early in the follow-up period, the effects faded over time as program group members attempted to transition to unsubsidized jobs. It appeared that the initial increase in employment was due to the temporary jobs provided by the program. The effect on re-offending was nonetheless positive, albeit modest. Recidivism rates were reported to be 70% over the 3-year follow-up period for the control group, compared with 65% for the program group. The most promising effects were observed for a subgroup of former prisoners who enrolled shortly after release from prison. A cost–benefit analysis also concluded that CEO’s financial benefits exceeded costs for taxpayers, victims, and participants (Redcross et al., 2012).

Zweig et al. (2011), using the same data, sought to extend these analyses by using a regression-based approach to identify subgroups of former prisoners for whom CEO had the greatest impact. Their goal was to differentiate former prisoners into low-, medium-, and high-risk subgroups depending on their risk of recidivism as assessed prior to study participation. For former prisoners in the high-risk subgroup, CEO reduced the probability of re-arrest in the 2 years following random assignment, the probability of re-conviction, and the number of re-arrests. However, there were no statistically significant program impacts on recidivism for those in the high-risk subgroup after the first year. Recidivism outcomes were better for those in the high-risk subgroup, and few changes in recidivism—of any type—were reported for those in the low-risk and medium-risk subgroups. This study also reported that other factors such as pre-program employment, housing, or children did not contribute to recidivism outcomes.

Transitional Jobs Reentry Demonstration (TJRD)

The TJRD (Jacobs, 2012; Redcross et al., 2010) project set out to establish whether transitional jobs programs are more effective than simpler, cheaper programs that provide basic job search and referral services and not subsidized employment. During 2007 to 2008, 1,813 men from four cities in the Midwestern United States who were aged 18 or older, and who had been released from state prison within the previous 90 days, were enrolled in the study. All types of criminal histories were accepted, and eligible participants had to be willing to work full-time and could not have worked in a transitional job in the past year. There was some variation between the sites as to how participants were recruited and when random assignment occurred. However, each of the sites largely utilized random assignment methods, and those randomly assigned to the intervention group were offered a temporary, minimum-wage job that provided 30 to 40 hr of paid work each week as well as other support services, such as pre-employment classes, job coaching, job search assistance, job placement, and post-placement services. Participants at two sites were also offered retention bonus payments (up to US$1,500) for getting and holding unsubsidized jobs.

Follow-up studies at both 1 year (Redcross et al., 2010) and 2 years (Jacobs, 2012) showed that the programs substantially increased employment early in the study by providing jobs to many who would not otherwise have worked. However, the gains faded as men left transitional work, and the programs did not increase regular (unsubsidized) employment either during or after the program period, although the highest placement rates in unsubsidized jobs were in the sites that offered retention bonuses. At the end of the second year, only about one fifth of each group were employed in the formal labor market with no statistically significant differences between the two groups.

The transitional jobs programs also did not systematically affect key measures of recidivism over the 2-year follow-up period: arrest, conviction, admission to prison, or average number of days spent in prison in Year 2 (Jacobs, 2012). About one third of participants in both the program and control groups were re-arrested, and about one sixth were convicted of a crime during this time; there was no statistically significant difference between groups in either outcome.

Employment-Focused Re-Entry Program

In this study, conducted in Southern California from November 2008 to November 2010, the impact of an employment-focused re-entry program on employment, housing, and recidivism outcomes was assessed (Farabee, Zhang, & Wright, 2014). The program had been in existence since the early 1990s and utilized the STRIVE employment model (see http://striveinternational.org/what-we-do/), which is offered by more than 25 community-based organizations in the United States. Participants attend job readiness training classes and complete an employment readiness 3- to 4-week course that is followed by employment placement assistance and access to computer labs that can be used for job-seeking activities. All ex-offenders were eligible to apply for the program, as long as they had no convictions for sexual offenses or arson, although a negative drug test and a valid form of identification were required. Those who met the program admission criteria and had been released from jail or prison during the preceding 180 days completed a baseline interview and were randomized to the program (n = 115) or control (n = 102) group. Control participants were provided a list of community resources and given a voucher for a free meal.

Outcomes were based on self-report interviews conducted 1-year post-randomization and arrest records reflecting a 2-year follow-up period. The 1-year outcomes were similar for both groups. For example, the percentage of program participants who were employed full-time was 29.8% compared with 27.1% of the control group. No statistically significant differences in participant re-arrest rates were reported, with 45.2% of program participants re-arrested compared with 49.4% of the control group. Both groups also showed similar patterns in their rate of re-incarceration (47.1% for the program group and 48.2% for the control group). In addition, both groups had similar housing arrangements during the 12-month period. No effects were observed for other outcomes, such as substance use, education, and overall health.

Milwaukee Safe Streets Prisoner Release Initiative

A key feature of the Milwaukee Safe Streets Prisoner Release Initiative was to “reach in” and to provide inmates with assessment and 6 months of needed services before their release from prison, along with employer subsidies for the first 6 months following release (Cook et al., 2014). The intervention was multifaceted and included vocational skills assessment, access to soft skills training and vocational training, as well as participation in the Breaking Barriers cognitive-reality program (a 12- to 16-week program designed to change behavior, thinking, and attitudes known to contribute to criminality, and to address dynamic risk factors for criminal behavior). Program participants were also given access to reach-in services, alcohol and drug treatment, and remedial education, and for some minimum security offenders, work release (subsidized employment) was a possibility toward the end of their sentence.

Inmates were eligible to participate if they were aged 35 or under, male, scheduled to be released to the Milwaukee initiative with at least 6 months of community supervision, and had a history of violence or gang involvement. Between January and August 2009, 236 offenders were randomized to either the program group (n = 106) or a control group (n = 130), where they were placed on a waitlist for the same services (although a small percentage did complete in-prison treatment programs and work release).

The impact of this program was assessed by the Wisconsin Department of Corrections. Intention-to-treat analysis was used to examine program outcomes over the year following prison release. The results showed that the program group worked more than the control group in this time and received higher median earnings. These positive effects on employment were sustained after the first 6 months, as the program group was more likely to be employed in the third and fourth quarters of the year following release. By the end of the first year, overall arrest rates and re-imprisonment rates for the program group were lower than those of control participants (by 9 and 4 percentage points, respectively, equal to 13%-15% of the respective control means). However, the difference between the average number of arrests between the treatment and control groups during the first year released from prison was not statistically significant.

Discussion

The aim of this systematic review was, using strict inclusion criteria, to establish the current status of evidence relevant to understanding the impact of vocational education and training and/or post-release employment programs delivered to prisoners, ex-prisoners, and/or ex-offenders. Given that only seven independent employment programs were identified as meeting the study inclusion criteria, Visher et al.’s (2005) conclusion that rigorous evaluations of employment interventions for former prisoners are sorely needed is still most apposite. Thus, this study serves as a reminder that despite the continuing appeal and substantial investment of public funds into these programs, the evidence base to support program delivery has not improved substantially over the past decade. There are, in short, still insufficient data from which to draw any firm conclusions about effectiveness. Of the studies identified in the searches, the first three predate Visher et al.’s study, with one (Uggen, 2000) reporting an analysis of data collected from a trial conducted over 40 years ago. Furthermore, the diversity among those programs that have been subject to methodologically rigorous evaluation makes it difficult to clearly identify those aspects that are most effective, and, of course, all of the programs evaluated were delivered in one country (the United States). For these reasons, it is considered inappropriate to aggregate the findings of these different studies to produce an overall effect size, although this would be possible if studies that fell outside of our eligibility criteria (such as those reported pre-2000 by Wilson, Gallagher, & Mackenzie, 2000) were included.

In particular, the requirement that program outcomes had to be compared with those of one or more control groups which received either “treatment as usual” or “no treatment” limited the number of studies that could be included. The purpose of this was to help ensure that preexisting differences between the treatment and control groups were accounted for such that clear statements about causality could be made. Rather than including a larger number of less rigorous studies, we were purposely selective to avoid a misimpression that volume of research can be loosely equated with degree of clarity regarding “what works.” Nonetheless, other evaluation methodologies have been used by other researchers and produced findings that are of interest. Maré and Hyslop (2011, cited in Christofferson, 2014), for example, used logistic regression to show that pre-release program participation led to reductions in re-offending in cohorts of prisoners in New Zealand released between 2007 and 2008. This method allowed these researchers to control for differences in risk of re-offending, although it neglects other factors that have the potential to systematically influence both employment and re-offending outcomes.

Our reading of the studies identified in this study allows us to make some tentative observations about the outcomes of more contemporary programs that have been subject to evaluation. First, there remains a lack of empirical support linking program participation with subsequent employment success. This seems important given that the underlying logic of these programs is that it is engagement with employment that reduces the risk of re-offending. Both the CEO and TJRD programs offered temporary, wage-paying jobs and job placement services to former prisoners shortly after their release from prison, and although both programs led to initial improvement in employment rates, they were generally unsuccessful in increasing unsubsidized employment. The absence of private-sector job opportunities may have contributed to this, although it may also be that the duration of the transitional job phase was too short to allow ex-offenders to develop the skills needed to make them employable. However, contributing causes are not discernible. Financial incentives in the form of retention bonuses did, however, appear to help (Jacobs, 2012; Redcross et al., 2010; Redcross et al., 2012). Of course, such observations remain speculative due to the paucity of rigorously derived data.

The studies reviewed here further suggest that the impact of vocational training and employment program participation is likely to be enhanced for those who undertake programs soon after release from prison. This suggestion is clearly supported by the evaluation of the CEO program, which showed that those prisoners who entered the CEO program within 3 months of their release benefitted the most and were less likely to recidivate (be re-arrested, convicted of crime, or incarcerated; Bloom et al., 2007; Redcross et al., 2012). A survey of CEO participants showed that they felt more connected to staff than did participants in other programs (Redcross et al., 2009; Redcross et al., 2012), and this may have improved their chances of success. The small work crews offered as part of this program may also have allowed ex-offenders to interact within a positive and supportive environment, and to receive peer support, and increased their sense of connection. Of course, work environments are also social environments. Degree of social integration may well play a critical part in an ex-offender’s employment success.

There is some evidence to indicate that programs are most effective for high-risk ex-offenders (see also Christofferson, 2014). The impact of the CEO program, on both employment and recidivism outcomes, was strongest for those ex-offenders who were assessed at high risk (including those with lengthier criminal histories, without high school credentials). High-risk ex-offenders who took part in the CEO program were less likely to be re-arrested, had fewer re-arrests, and were less likely to be reconvicted of crimes than high-risk ex-offenders who did not take part in the program (Bloom et al., 2007; Redcross et al., 2009; Redcross et al., 2012; Zweig et al., 2011). Moreover, those in the low-risk category who took part in the CEO program demonstrated similar outcomes to the control group.

Related to this are the findings that older offenders (27 years and above) benefit more from employment interventions. The National Supported Work Demonstration Project, a program in which ex-offenders were provided with minimum-wage jobs, was found to reduce recidivism among those aged 27 years of age and older but had little effect on the re-arrest rates of those who were younger (Uggen, 2000). The ESEO program evaluation reported a similar pattern of results, with Bierens and Carvalho (2011) reporting reductions in recidivism for ex-offenders over the age of 27 in two of the sites (San Diego and Chicago) and over the age of 36 in the other (Boston). Moreover, recidivism for younger participants increased following program participation. Of course, simply being in the program may provide a possible explanation for these findings, but generally age and offending behavior are negatively correlated, and there is some evidence from other studies that younger and older offenders are less successful post-release than mid-aged offenders (Bouffard, MacKenzie, & Hickman, 2000; Bushway & Reuter, 2002).

The geographical difference in outcomes for the ESEO program is worth commenting upon as it may reflect variations in the nature or quality of the job search assistance provided and/or variations in jobs availability across differing local economies. The evaluation authors speculated, for example, that the focus of the location, which produced the poorer outcomes, was on obtaining lesser quality jobs, and this might have influenced the results. It may also be a result of the quality of implementation; Lipsey and Wilson’s (1998) meta-analysis of programs for young offenders clearly shows that good delivery is more important than a good “model” that is poorly delivered across a range of different offender programs (see also Mercy & Hammond, 2001). Local economic and employment environment and job availability may also explain geographical differences in outcomes.

There is evidence that particular employment interventions can benefit ex-offenders with substance use problems. A tailored employment intervention was shown to have a positive effect on the full-time and part-time employment patterns of non-violent offenders with substance use addictions attending a drug court program. The authors also reported less substance use and less criminal activity at follow-up (Leukefeld et al., 2007). Another key finding from this program was that the employment intervention was particularly effective for those participants who lacked recent employment success. Following participation in the program, those individuals with a negative pre-baseline work trajectory were more likely to be in the workforce and earn more income than similar individuals who did not receive the program (Webster et al., 2014).

When offered in isolation from other services, there is reason to suspect that employment readiness programs will be ineffective. Farabee et al.’s (2014) evaluation of an employment-focused re-entry program reported similar 12-month outcomes for both the treatment group and the control group. However, more holistic, comprehensive programs appear to hold more promise (see also Bushway & Reuter, 2002). There is evidence that prison “reach-in” services that help prepare inmates for productive lives after release, together with multifaceted wraparound services, increase the likelihood of employment upon release and decrease the likelihood of arrest and re-imprisonment in the year following release (Cook et al., 2014). It is worth noting here that a key feature of this intervention was a 12- to 16-week cognitive skills program designed to change behavior, attitudes, and dynamic risk factors associated with criminal behavior, as well as alcohol and drug treatment, remedial education, and soft skills training. Disaggregating relative effects of each component is a common problem for attempts to interpret program effects.

There are, of course, a number of limitations to this systemic review that warrant discussion. One of the most obvious of these, perhaps, relates to the systematic review methodology and the possibility that the way in which the searches were constructed (choice of key terms and databases) overlooked potentially useful studies. Certainly, the findings of this review have relatively little to say about the mechanisms by which change occurs or in relation to key questions regarding whether it is employment itself or the skills that are acquired during training that assist the desistance process. It is likely to also be a case that finding employment may prove insufficient to affect behavior change if other risk factors are not addressed, and the field has largely neglected the role that individual differences (such as age) have to play in determining the outcomes of different forms of intervention. Very little is known about factors mediating the beneficial effects of programs, or the interplay between different causal factors, including moderator and mediation effects (Kraemer, Stice, Kazdin, Offord, & Kupfer, 2001). These studies also say relatively little about whether it is vocational training and/or employment support programs provided either in-prison or post-prison (or a combination of these) that have the most potential to influence subsequent offending. Nonetheless, the use of random allocation in program evaluation is widely considered to produce the quality of evidence that is needed to inform the development of evidence-based practice in crime prevention (see Sherman, Farrington, Welsh, & McKenzie, 2002). In this particular context, this helps to overcome the rather significant issues that are associated with selection bias, whereby the characteristics of those prisoners and ex-offenders that influence their selection into programs also influence their subsequent offending.

Ways Forward?

The relationship between unemployment and offending/re-offending is well established. Questions remain, however, about the extent to which vocational education and employment assistance programs result in reduced re-offending and whether this is determined by the success that ex-prisoners have in finding post-release employment, particularly ongoing employment. The conclusion of this review is that much more work still needs to be done to answer these questions. This is clearly needed given that the provision of prison vocational education and training programs is a significant area of activity for many correctional administrations.

A number of elements of potentially effective practice can, however, be identified from this review, and these can be used to develop meaningful hypotheses that should be tested in future evaluation efforts.

For example, we might expect programs that offer financial incentives to increase the likelihood of participants keeping an unsubsidized job, whereas those that offer low-wage, low-quality jobs may not prove successful in either reducing recidivism or increasing the likelihood of employment. The most promising programs appear to be those that offer a holistic and comprehensive approach. These programs do not focus only on employment but provide assistance/services in other areas that may impact on a prisoner’s reintegration into the community such as drug and alcohol treatment, housing assistance, cognitive skills, and remedial education. Programs that provide “reach-in” services as well as community-based services for the period immediately following release appear to hold the most promise. Protracted periods of holistic program eligibility/support with a gradual decrease may prove more effective than the currently typical short-term, single-focus, clear endpoint approach.

Collectively, these studies also offer support for the suggestion that not all prisoners and ex-offenders need all of these services and that programs are likely to be much more effective if they are properly targeted to the needs of individual participants. There is evidence, for example, that those who are at high risk of recidivating should be offered programs, along with those who lack recent employment success, and are older than 27 years of age. Employment programs for those who fall outside of these criteria may need to be re-conceptualized, but clearly future evaluations should consider issues relating to the most effective combination of services for specific types of ex-offenders.

In many ways, these conclusions reiterate those of Visher et al. (2005) but take on new significance in a context in which correctional budgets are under increasing fiscal strain, and the need to provide evidence-based correctional programs is perhaps more acute than ever (Craig, Dixon, & Gannon, 2014). Our conclusion is clear that the current evidence relating to the impact of vocational education and training programs on re-offending remains weak, and most current evaluation research is methodologically poor. Existing work in this area has largely failed to provide the type of information that policy makers need regarding which programs to fund, scale up, or discontinue. In addition, the lack of rigor that led to the exclusion of a number of outcome studies fuels concerns about what Travers (2005) referred to as “intellectually uninteresting” evaluations that suffer from “managerial bias.” He referred to small-scale and methodologically weak evaluations conducted for local agencies, with no expectation that they will influence subsequent service delivery.

In our view, what is required are flagship evaluations involving the collection and analysis of data over many years, which are built around strong evaluation methodologies, including the development of program logic models, clearly articulated short-term outcome measures, the assessment of implementation issues and program integrity, and the relative contribution of different program components. As noted at the beginning of this article, social, cultural, economic, and legislative contexts are also potentially important determinants of program outcomes that are not considered in the studies reviewed. A good example of this is the availability of paid employment in the local area in which programs are delivered, as well as the impact of collecting outcome data in periods of high unemployment (following the global financial crisis). We recommend analysis of these contextual factors be built into future studies, especially in jurisdictions other than the United States. The application of current program accreditation processes and standards (such as the Principles of Empirically Supported Interventions; Wampold, Lichtenberg, & Waehler, 2002) or the Correctional Program Assessment Inventory (Gendreau & Andrews, 2001) may also help to identify where programs might be strengthened to meet current standards of evidence-based practice.

Footnotes

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by the Australian Research Council (LP140100329: A Future Beyond the Wall: Improving Post-release Employment Outcomes for People Leaving Prison).

References

*Bierens

H. J.

Carvalho

J. R.

(2011). Job search, conditional treatment and recidivism: The Employment Services for Ex-Offenders program reconsidered. B.E. Journal of Economic Analysis & Policy: Advances in Economic Analysis & Policy, 11(1), 1-38. doi:10.2202/1935-1682.2361

*Bloom

Redcross

Zweig

Azurdia

(2007). Transitional jobs for ex-prisoners: Early impacts from a random assignment evaluation of the Center for Employment Opportunities (CEO) prisoner reentry program (An MDRC working paper). Retrieved from http://www.mdrc.org/publication/early-impacts-evaluation-center-employment-opportunities-ceo-prisoner-reentry-program

Bouffard

J. A.

MacKenzie

D. L.

Hickman

L. J.

(2000). Effectiveness of vocational education and employment programs for adult offenders: A methodology based analysis of the literature. Journal of Offender Rehabilitation, 31, 1-41.

Bushway

Reuter

(2002). Labor markets and crime risk factors. In Welsh

Farrington

(Eds.), Preventing crime (pp. 198-240). New York, NY: Springer.

Cherney

Fitzgerald

(2014). Finding and keeping a job: The value and meaning of employment for parolees. International Journal of Offender Criminology and Comparative Therapy, 60, 21-37. doi:10.1177/0306624X14548858

Christofferson

S. B.

(2014). Prison-based employment interventions: Effects on recidivism. Practice: The New Zealand Corrections Journal, 2, 28-31.

*Cook

P. J.

Kang

Braga

A. A.

Ludwig

O’Brien

M. E.

(2014). An experimental evaluation of a comprehensive employment-oriented prisoner re-entry program. Journal of Quantitative Criminology, 31, 355-382. doi:10.1007/s10940-014-9242-5

Craig

L. A.

Dixon

Gannon

T. A.

(Eds.). (2014). What works in offender rehabilitation: An evidence-based approach to assessment and treatment. Chichester, UK: Wiley-Blackwell.

*Farabee

Zhang

S. X.

Wright

(2014). An experimental evaluation of a nationally recognized employment-focused offender re-entry program. Journal of Experimental Criminology, 10, 309-322. doi:10.1007/s11292-014-9201-z

10.

Gendreau

Andrews

(2001). Correctional Program Assessment Inventory (CPAI-2000). Saint John, Canada: University of New Brunswick.

11.

Gendreau

Goggin

Gray

(1998). Case need domain: Employment. Forum on Corrections Research, 10, 16-19.

12.

Graffam

Shinkfield

A. J.

Hardcastle

(2008). The perceived employability of ex-prisoners and offenders. International Journal of Offender Therapy and Comparative Criminology, 52, 673-685.

13.

Harper

Chitty

(2005). The impact of corrections on reoffending: A review of “what works.” London, England: Home Office. Retrieved from http://webarchive.nationalarchives.gov.uk/20110218135832/rds.homeoffice.gov.uk/rds/pdfs04/hors291.pdf

14.

International Centre for Prison Studies. (2015). World Prison Brief. London, England: International Centre for Prison Studies, University of London. Available from http://www.prisonstudies.org

15.

*Jacobs

(2012). Returning to work after prison: Final results from the Transitional Jobs Reentry Demonstration. The Manpower Demonstration Research Corporation (MDRC). Retrieved from http://www.mdrc.org/publication/returning-work-after-prison

16.

Kraemer

H. C.

Stice

Kazdin

Offord

Kupfer

(2001). How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors. American Journal of Psychiatry, 158, 848-856.

17.

Krienert

J. L.

Fleisher

M. S.

(2004). Crime and employment: Critical issues in crime reduction for corrections. Walnut Creek, CA: AltaMira Press.

18.

Leukefeld

C. G.

(2003). An employment intervention for drug-abusing offenders. Federal Probation, 67(2), 27-31.

19.

*Leukefeld

Webster

J. M.

Staton-Tindall

Duvall

(2007). Employment and work among drug court clients: 12-month outcomes. Substance Use & Misuse, 42, 1109-1126. doi:10.1080/10826080701409701

20.

Lipsey

M. W.

Wilson

D. B.

(1998). Effective intervention for serious juvenile offenders: A synthesis of research. In Loeber

Farrington

D. P.

(Eds.), Serious and violent juvenile offenders: Risk factors and successful interventions (pp. 313-345). Thousand Oaks, CA: SAGE.

21.

Mercy

J. A.

Hammond

W. R.

(2001). Learning to do violence prevention well. American Journal of Preventive Medicine, 20(1, Suppl. 1), 1-2.

22.

Metcalf

Anderson

Rolfe

(2001). Barriers to employment for offenders and ex-offenders (Department of Work and Pensions Report No. 155). Leeds, UK: Centre for Disability Studies.

23.

Milkman

(2001). Employment services for ex-offenders, 1981-1984 (Discussion Paper No. 8619). Boston, MA: Inter-university Consortium for Political and Social Research.

24.

Moher

Liberati

Tetzlaff

Altman

D. G.

, & the PRISMA Group. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. British Medical Journal, 339, Article b2535. doi:10.1136/bmj.b2535

25.

Nally

J. N.

Lockwood

Knutson

(2014). Post-release recidivism and employment among different types of released offenders: A 5-year follow-up study in the United States. International Journal of Criminal Justice Sciences, 9, 16-34.

26.

Phipps

Korinek

Aos

Lieb

(1999). Research findings on adult corrections programs: A review. Olympia: Washington State Institute for Public Policy.

27.

Piliavin

Gartner

(1981). The impact of supported work on ex-offenders. New York, NY: The Manpower Demonstration Research Corporation (MDRC).

28.

*Redcross

Bloom

Azurdia

Zweig

Pindus

(2009). Transitional jobs for ex-prisoners: Implementation, two-year impacts, and costs of the Center for Employment Opportunities (CEO) prisoner reentry program. The Manpower Demonstration Research Corporation (MDRC). Retrieved from http://www.mdrc.org/publication/transitional-jobs-ex-prisoners

29.

*Redcross

Bloom

Jacobs

Manno

Muller-Ravett

Seefeldt

. . . Zweig

(2010). Work after prison: One-year findings from the Transitional Jobs Reentry Demonstration. The Manpower Demonstration Research Corporation (MDRC). Retrieved from http://www.mdrc.org/publication/work-after-prison

30.

*Redcross

Millenky

Rudd

Levshin

(2012). More than a job: Final results from the evaluation of the Center for Employment Opportunities (CEO) transitional jobs program (OPRE Report No. 2011-18). Washington, DC: Office of Planning, Research and Evaluation, Administration for Children and Families, U.S. Department of Health and Human Services. Retrieved from http://www.mdrc.org/publication/more-job

31.

Sampson

R. J.

Laub

J. H.

(1977). A life course theory of cumulative disadvantage and the stability of delinquency. In Thornberry

T. P.

(Ed.), Developmental theories of crime and delinquency (pp. 133-161). New Brunswick, NJ: Transaction.

32.

Sherman

Farrington

Welsh

McKenzie

(Eds.). (2002). Evidence-based crime prevention. New York, NY: Routledge.

33.

Travers

(2005). Evaluation research and criminal justice: Beyond a political critique. The Australian & New Zealand Journal of Criminology, 38, 39-58.

34.

*Uggen

(2000). Work as a turning point in the life course of criminals: A duration model of age, employment, and recidivism. American Sociological Review, 67, 529-546. doi:10.2307/2657381

35.

Visher

Winterfield

Coggeshall

(2005). Ex-offender employment programs and recidivism: A meta-analysis. Journal of Experimental Criminology, 1, 295-315. doi:10.1007/s11292-005-8127-x

36.

Wampold

B. E.

Lichtenberg

J. W.

Waehler

C. A.

(2002). Principles of empirically supported interventions in counseling psychology. The Counseling Psychologist, 30, 197-217.

37.

*Webster

J. M.

Staton-Tindall

Dickson

M. F.

Wilson

J. F.

Leukefeld

C. G.

(2014). Twelve-month employment intervention outcomes for drug-involved offenders. The American Journal of Drug and Alcohol Abuse, 40, 200-205. doi:10.3109/00952990.2013.858722

38.

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.

39.

*Zweig

Yahner

Redcross

(2011). For whom does a transitional jobs program work? Examining the recidivism effects of the Center for Employment Opportunities program on former prisoners at high, medium, and low risk of reoffending. Criminology & Public Policy, 10, 945-972. doi:10.1111/j.1745-9133.2011.00767.x