A r c h i v e d  I n f o r m a t i o n

Recidivism Study Summaries

The following abstracts are the result of the literature review conducted as a precursor to the OCE/CEA Recidivism Study.


OCE does not have copies of these studies. Please see the citation to determine the source of each study.

1. Anderson, Dennis B., Sara L. Anderson, and Randall E. Schumacker. "Correctional Education A Way to Stay Out: Recommendations for Illinois and a Report of the Anderson Study." Illinois Council on Vocational Education. 1988.

Summary: The chief purpose of this study was "to determine how many releasees obtained employment, especially in areas in which they received vocational training while incarcerated." 760 releasees were studied for 12 months. There were four research groups: vocational training, vocational and academic training, no vocational training, and only academic work.

Vocational and vocational and academic groups had higher employment and fewer arrests than the other groups. Those who received no vocational training had the highest criminal rate; the academic only group had the lowest employment rate, but the more education the students had received the higher employment and lower crime rate they had.

Study Strengths: One of the best studies to examine the issue of employment, especially its connection to vocational training. The recommendations derived from the study are very sound and would be of interest to other states.

Study Weaknesses: The main objective of the study was to determine how many releasees were finding work in the areas for which they were trained, but the parole officers did not or could not obtain this information from the releasees.

Comment: A worthwhile study, but it shows the difficulties inherent in using parole officers to track inmates.

2. Beck, Allen J., and Bernard E.Shipley. "Recidivism of Prisoners Released in 1983." Bureau of Justice Statistics, U.S. Department of Justice, April, 1989. Pp. 1-13.

Summary: 16,000 prisoners from 11 states were included in this important study on recidivism. The study reports that an "estimated 62.5% were rearrested for a felony or serious misdemeanor within 3 years." This study examines recidivism rates from a number of angles. Education was one characteristic examined. The data on education supports the contention that education can have an ameliorative effect on the recidivism rate; offenders with an education of 8th grade or less were rearrested at a rate of 61.9% (19.3% of the total population), while those who were high school graduates (25.8% of the total population) had a rearrest rate of 57.4%. Individuals with some college, while making up a small percentage of the total population (6.8%), had an even lower rearrest rate of 51.9%.

Study Strengths: This highly sophisticated study provided two key points of methodology for the evaluation instrument: the decision to use three measures of recidivism (rearrest, reconviction, and reincarceration) and the decision to use State Identification Bureaus and the FBI to track released offenders. The study provides an important, relevant overview of who recidivates and when they do so.

Study Weaknesses: Except for a quick glance at the education levels of released prisoners, there is no discussion here on education, or on any other programming. It is not clear how the information on educational levels of offenders was determined.

3. Blackburn, F.S. "The Relationship Between Recidivism and Participation in a Community College Program for Incarcerated Offenders." Journal of Correctional Education. Volume 32, Issue 3. pp. 23-25.

Study Strengths: One of first studies to attempt to establish a control group; studied large number (243) over a long period (releasees from 1970-78). Drew comparisons between subjects using six variables - race, age at release, date of release, LESS (law encounter severity scale), EDS (environmental deprivation scale) and MBR (maladaptive behavior record).

Study Weaknesses: Unable to do random assignment or random sampling; did not analyze length of college experience (minimum 12 hours completed only criteria).

Comments:Might use the six variables in developing model if the LESS, EDS and MBR are still available or are effective measures; did not interview students or do case studies for insights.

4. Downs, Elizabeth A., Monaco, Kathryn Rabold, and Sheila Ortego Shreiber. "Evaluating the Effects of Vocational Education on Inmates: A Research Model and Preliminary Results." Yearbook of Correctional Education. Correctional Education Association. 1989. pp. 249-262.

Summary:This study is included here because it is frequently quoted in the literature, but it is important to note that it was a pilot study. Its conclusions, which are negative, are based on a population of which over 60% had received a very small degree (5-9 credit hours) of exposure to educational programming. The vocational training program itself had only been in existence a year and may have needed further development before it was fair to test it. So the results are not what is important, but rather the study design. The comparison groups were carefully matched and then separated into high and low risk categories. A database check was then run on their parole status and conclusions of success or failure drawn from that.

Study Strengths: The control for several variables chosen for their relation to parole success and the careful matching of the comparison groups. Compares students with offenders who wanted education and never received it, thus countering the problem of self-selection.

Study Weaknesses: The small number studied (66) and the lack of exposure to educational programming means that their conclusions are not scientifically based.

Comments:Good discussion of 1970s research.

5. Elliott, Michael J. "Department of Corrections Program Performance Indicators October, 1988 to September, 1994: Alternative Solutions to Metropolitan Chicago's Problems of Crime, Employment, Welfare & Education." Roosevelt University, 1994.

Summary:Report identifies performance indicators, particularly those involving costs, which are impacted by reduced recidivism of those released from Illinois prison after attending Roosevelt University program - wages, federal and state income taxes paid, social security paid, fringe benefits, consumer spending and sales taxes; numbers are fairly large (138 releasees) and long term (covers 1989 to present). Recidivism rate is 4.5%.

Study Weaknesses: Students are "selected" and then compared to Illinois general recidivism rate; program is academic with no transition or vocational program indicated (in a way it is a strength because it indicates that even an academic program has positive effect by itself); not much on how the cost figures were arrived at; does not speak to other impacts such as less crime, better family life, or a drug free life.

Comments:This is a good study for use in identifying cost savings to public in several areas which few studies look at in any depth.

6. Eisenberg, Michael. "Five Year Outcome Study: Factors Associated with Recidivism." Texas Department of Criminal Justice. December 1991.

Summary:A long term study (5 years) with large random sample of Texas inmates (1,539); studied many variables - socio-demographic, criminal history, risk factors, offense, post-release and educational data; determined recidivism as "returned to DOC" during study period. Attainment of education grade compared to recidivism.

Study Strengths: Recognizes that degree of educational attainment may influence recidivism rate. A long term, large sample study that analyzes many variables.

Study Weaknesses: No comparison of vocational attainment, postsecondary attainment or career and job history with recidivism; no attempt to isolate "self selection" or create a control group per se.

Comments:Study useful because it establishes strong link with GED or high school attainment and reduced recidivism for all age groups, the largest difference with the younger offenders, but a significant difference with all age groups; difference holds true for minority offenders as well.

7. Flanagan, Timothy J. "Prison Education Research Project Final Report." Sam Houston State University Criminal Justice Center. September 1994.

Summary: This study reviews many other studies from various states. It uses large numbers of Texas releasees between 1991-1992, studies history of Texas Department of Criminal Justice and the problems with putting inmates in programs given overcrowding and early release. In its summary of other recidivism studies it rates them by quality of design - the use of control groups, control matching, random assignment, statistical controls and tests of significance. The study concludes that Adult Basic Education is very effective in reducing recidivism. This study and the other studies reviewed demonstrate that though studies are often weak in design, they show consistent and positive results for society. It identifies characteristics making education programs successful, including that programs are extensive, separate from rest of prison atmosphere, involve follow-up after release and teach skills relevant to market. It also points out predictors of recidivism - type of crime, presence of prior convictions, employment stability, age at first arrest, living/family arrangements, current income, drug history and alcohol history.

Study Strengths: Reviews many studies: uses large numbers of releasees between 1991-1992; studies history of TDCJ and problems with putting inmates in programs given overcrowding and early release; rates previous studies by quality of design (do they use control group); control matching, random assignment, statistical controls and tests of significance; concludes that ABE is very effective in reducing recidivism; shows studies, though often weak in design, demonstrating that education achieves consistent positive results for society; points out moral development education does not have good research base showing impact; identifies characteristics which allow education programs to be successful - extensive, separate from rest of prison, follow-up after release, attract appropriate clients, teach skills relevant to market; points out predictors of recidivism - type of crime, presence of prior convictions, employment stability, age at first arrest, living/family arrangements, current income, drug and alcohol history.

Study Weaknesses: Population studied not representative of entire TDCJ population; use of grade level only as achievement measure does not take into account milestones such as degrees and certificates which other studies show as linked to reduced recidivism; no attempt to establish a control group per se to eliminate effects of self selection.

Comments: Very useful study for many of the reasons cited under strengths.

8. Gainous, Fred J., "Alabama: Correctional Education Research." Department of Postsecondary Education, 1992. n.p.

Summary: Long term study of graduates released to community from J.F. Ingram and six other Alabama colleges providing correctional education to prisoners; compares them to overall state average for recidivism; Ingram offers job placement and post-release counseling (not defined); decrease of 2% in recidivism would result in major increase in tax revenue (not explained, but an interesting concept to pursue); study ties into State's employment and job market data.

"The average recidivism percentage of all correctional institutions combined is approximately five percent (5%) for those inmates completing courses during the study period (87-91). This figure is significantly below the DOC recidivism figure for the entire prison population that is generally considered to average around thirty-five percent (35%). "

Study Weaknesses: No mention of how students are selected for program; self-selection not controlled; no comparable group selected from general inmate population to compare to graduates of correctional education; kinds of courses offered are not distinguished in study (GED, AA or AB degrees).

Comments: Study background information would be helpful; such large difference in recidivm between graduates of education programs and other Inmates raises lots of questions. Most other studies do not show such a large difference. Since the population characteristics are not discussed, this may be the major problem (i.e. that students have been "selected" for success). On the other hand the programs at J.F. Ingrain are quite good (opinion of reviewer after two visits to Alabama). Also, not much else to do except work or go to school.

9. Harer, Miles D. "Recidivism Among Federal Prison Releasees in 1987: A Preliminary Report." Federal Bureau of Prisons, Washington D. C., March 10, 1994.

Summary:This is probably one of the best recidivism studies to date measuring most any kind of variable possible. It also links risk assessments (US. Parole Commission's Salient Factors and U. S. Sentencing Commission's Criminal History Score) with recidivism showing a strong positive relationship between education participation and completion to a reduction in recidivism. There is also a strong relationship with job training and placement and reduced recidivism. Inmates were tracked through Interstate Identification Index (III) and FBI's National Criminal Information Center. There is data to show that any strong suggestion to inmates to participate in school, drug programs, or work may actually reduce recidivism, rebutting idea that volunteer participation is the only way one can benefit from programs.

This study found that the more education the inmate had upon entry into the system, the less likely the inmate was to recidivate. The highest recidivism rate was 54.6% for individual with some high school and the lowest rate was 5.4% for college graduates. Recidivism rates also decreased according to how much education a student received during incarceration. Inmates who did not take education classes recidivated at 44.1%, while individuals who completed at least one course every 6 months of their incarceration recidivated at a rate of 35.5%.

There are future research recommendations concerning the effects of education on "normalization" (program participation counters the forces of imprisonment and inculcates law-abiding norms) as well as selective knowledge attainment and effects on recidivism and other post release successes. Further study is recommended linking "risk" and "need" and methods used to meet demands of each.

There are good recommendations regarding, the kind of data to collect about an inmate's drug history, pre-prison schooling, employment, marital status, visit history, risk factors, program participation, prison experiences and post-release results. These will be helpful to other agencies in constructing an overall data collection system.

Study Strengths: Extremely large sample (1,205); long term study (3 years), numerous variables studied (pre-release, prison and post-release as well as criminal history, age, race, substance-abuse history, and risk profiles); review of many kinds of programs - furlough, education, and prison industry; strong statistical base.

Study Weaknesses: Control group is basically comparison among inmates with similar characteristics save one (i.e. education or training).

10. Holloway, Jerry and Paul Moke. "Post Secondary Correctional Education: An Evaluation of Parolee Performance." Wilmington College, Wilmington. Ohio. May 14, 1986.

Summary: Studied graduates of AA degree program who were paroled during 1982-83. This group of 95 students was compared to a randomly selected group of high school graduates (including GED) received inside or outside of prison and a third group of randomly selected inmates (106 in all) who had no GED or high school education and were released during the same time period. Findings included that more education led to likelihood of employment after release, lower recidivism for higher levels of education (college lower than high school and both lower than non high school graduates); higher risk inmates do better with education programs on recidivism and employment measures.

Study Strengths: Some correction for self-selection; used high risk members of the prison population who also completed college to see if there was a difference between them and lower risk college completers, also uses a group that wanted to complete college and was not able to do more than 2 quarters of work.

Study Weaknesses: Study used methods of correcting for self-selection, but question of self-selection still remains; only covers about 12 months after release

Comments:This is a good study because it addresses the issues of variables that are important, self-selection, and motivation.

11. Jeffords, Charles R. and Scott McNitt. "The Relationship Between GED Attainment and Recidivism: An Evaluation Summary." Texas Youth Commission. September, 1993.

Summary:This study examined 1,717 Texas youths who were released between July 1, 1990 and June 30, 1992. 475 had received their GED while incarcerated and 1, 242 had not. This last group served as the control group. Statistical controls were employed and the youths were tracked for one year. The authors concluded that "the recidivism rates of youth attaining a GED were significantly lower than those who did not, even after controlling for various demographic, behavioral, and criminal history characteristics." The treatment group had a 41.3 % rearrest rate and a 10.1% reincarceration rate as compared to a 53.5% rearrest rate and 19.1% reincarceration rate for the control group.

Study Strengths: The study controlled for a large number of variables in an attempt to isolate the attainment of the GED as the chief difference between the treatment and control group.

Study Weaknesses: There was no correction for self-selection or random assignment. More information on the programming involved, and on background of the youths would have been helpful.

Comments: A good study on an important topic, which has not received much study.

12. Jenkins, H. David., Jennifer Pendry, and Stephen J. Steurer. "A Post Release Follow-up of Correctional Education Program Completers Released in 1990-1991." Maryland State Department of Education. 1993.

Summary: This is a long-term (18 months) study with various subgroups (Adult Basic Education, GED, vocational education and post secondary students). There are comparison groups for control with sufficient numbers in 3 of the 4 group cells (post secondary has less than 10 in the cell). There is a good telephone data collection instrument with parole officers giving the major information on inmate post release success factors. The study concludes that there is a positive and significant benefit of education for students at all levels (ABE, GED, vocational education) when compared to similar inmates who have not had education while incarcerated. The post secondary group contained no recidivists, but the number of inmates was only 9. The benefits include reduced recidivism and increased employability and higher wages.

Study Strengths: Various subgroups (ABE, GED, VocEd, Postsecondary students); good data collection instrument with parole officer input; No control group per se; comparison was made to recidivism rate among other randomly selected inmates who were released during the same time period and who did not complete or did not participate in education programs.

Study Weaknesses: Postsecondary group small; no control for self-selection of students; no comparison to other variables such as sentence length or type of crime; no data bank matching by social security number.

Comments: Study with direct interview of parole officers to answer questions regarding recidivism and employment records of releasees; compares education records with postrelease records; also does good review of other research in area.

13. Little. Gregory L., et. al., "Treating Drug Offenders with Moral Reconation Therapy: A Three Year Recidivism Report." Psychological Reports. Correctional Counseling, Memphis, TN. 1991, 69, 1151-1154.

Summary: Large number in each group (70 treatment and 82 controls); groups very similar in average age, race, sentence length, and all treatment and control chose to enter program. Limited number were treated due to funds; long term follow-up (38 months); treatment took place while incarcerated and as a follow-up and those who got less treatment than others were studied showing they did less well after release

Study Weaknesses: Drug offenders only; no other variables studied such as other crime types, educational levels, etc.; more detail needed on treatment itself; do not know how clients were chosen (i.e. 152 applied and how 70 were selected and others not chosen)

Comments: Shows that this type of therapy can be effective with drug offenders and especially if continued after release; powerful study since it may really have a true control group (i.e. all wanted the treatment, but space limited participants); need to know more about the treatment itself; shows need for follow-up in community for program started behind bars.

14. MacKenzie, Doris Layton, James W. Shaw and Voncile B. Gowdy. "An Evaluation of Shock Incarceration in Louisiana." National Institute of Justice, U.S. Department of Justice, June 1993. pp. 1-7.

Summary: A 1991 study on shock incarceration, this study suggested an alternative methodology to random assignment. This methodology is a quasi-experimental design which uses a nonequivalent control group receiving the same pre and posttests as the treatment group. The results were statistically controlled to enhance the validity of the results.

Study Strengths: This methodology provides a sound mechanism for ensuring valid results without relying on random assignment. The repeated measure design with carefully selected comparison groups and statistical controls was a major influence on the design of the model evaluation instrument.

15. Menon, Ramdas, et.al.. "An Evaluation of Project RIO Outcomes: An Evaluative Report." Texas Employment Commission. July, 1992.

Summary:This large study of a sample size of 12,000 examined the rearrest and reincarceration rates of parolees enrolled in Project RIO, a project to increase the employment of released offenders by providing them with special services. The study looked at employment, the relationship between employment and recidivism, and the project's cost/benefit analysis. Even when certain variables were controlled for, Project RIO participants were more likely to be employed and less likely to recidivate than nonparticipants.

Study Strengths: A number of statistical controls were applied and the sample was very large. It contains a cost/benefit analysis.

Study Weaknesses: Data was collected for only four months and conditions affecting the data could change over time. The study needed to repeat the measures at a later date, but did not. The authors suggest that because they did not differentiate between how long their subjects participated in the program, but rather counted everyone as a participant even if they just signed up, that the study should be viewed as showing strong positive results. It might, however, suggest that the self-selection bias of their sample was very high. Anyone who signed up might have been so determined to work that the program may or may not have helped them to succeed. There was no correction for a self-selection bias discussed in the study.

Comments: An important study because its topic, an employment program, is seeing a great deal of interest in correctional education today. They tracked the employment records by SSN through the Texas Employment Commission and although they did not have every participant's social security number, it is an intriguing and potentially useful tracking component.

16. Porporino, Frank J. and David Robinson. "The Correctional Benefits of Education: A Follow-up of Canadian Federal Offenders Participating in ABE." Journal of Correctional Education. Correctional Education Association. Volume 43, Issue 2, June 1992. 92-98.

Summary: Study uses a large sample (1,736 releasees in 1988) with a fairly long follow-up period (at least 6 months and up to several years). Three groups were studied (completers, non-completers before release and those choosing to drop out). The study shows that Adult Basic Education alone can influence recidivism and other outcomes. This study also indicates that impact of education may be larger on higher risk than lower risk inmates. Other impacts include positive effects on the job, on family life and personal control. The authors looked at previous studies and showed that studies with good designs indicate similar positive impacts. The study also sets an interesting goal for correctional education overall: "to engage the offender in developing a new perspective on the possibility and the possibly rewarding outcomes of a non-criminal lifestyle."

Study Weaknesses: Although study tried to account for the variables which favor completers (they were less violent, fewer early crimes, received full paroles), argument is not convincing because there was no random assignment or attempt to compare groups to regular and similar inmates

Comments: Useful for attempt to measure impact of lowest level students, one of few non-American studies.

17. Ramsey, Charlotte. "The Value of Receiving a General Education Development Certificate While Incarcerated in the South Carolina Department of Corrections on the Rate of Recidivism." South Carolina Department of Corrections. 1988. pp.1-31.

Summary: This study reviewed the relationship between receiving a GED and recidivism. It used a sample size of 650 broken down in the following comparison groups:
175 youthful offenders who received the GED while incarcerated
175 youthful offenders who did not receive the GED
100 Adult Basic Education students who received the GED while incarcerated
100 Adult Basic Education students who did not receive the GED
100 straight timers who had achieved an 8.0 reading level ... but were not in school at all while incarcerated.

Among the 350 youthful offenders, 22% who received a GED recidivated, while 35% recidivated who had not received a GED. Among the 200 ABE students, 16% who received a GED recidivated, while 33% who did not receive a GED recidivated. Of the 100 individuals who had not received any programming, 36% recidivated. Tracking was done by database and handchecking records.

Study Strengths: This study is useful in that it looks at GED students. Although GED students make up a sizable portion of inmate students, there are only a handful of studies on this population.

Study Weaknesses: Although several control variables were included, data collection difficulties hindered their use in the analysis. Education information on some of the subjects was not available. Apparently intended for departmental use only, many terms such as "youthful offender" and "straight timer" are not defined.

18. Ross. R. R., et. al., "Reasoning and Rehabilitation." International Journal of Offender Therapy and Comparative Criminology." 32, 1988. pp. 29-35.

Summary: Study compares different types of instruction - cognitive, life skills and probation training in terms of impact or recidivism; defines differences in life and cognitive skills; fairly long term, 18 months, follow-up.

Study weaknesses: Very low numbers (62 divided into three groups); same probation officers teaching cognitive are doing regular probation and life skills as well which may create a bias; also calls into question the author's choice of a quality life skills program vs. cognitive program (inmates did not like this life skills program, but did like the cognitive program)

Comments: While the study looks well structured, it has low numbers and the possible bias on part of the authors creates skepticism. Does define cognitive areas in some detail.

19. Saylor, William G. and Gerald G. Gaes. "Post Release Employment Project: Summary of Preliminary Findings." Federal Bureau of Prisons, Office of Research and Evaluation. Washington D. C. June 27, 1991.

Summary:Widely known as the PREP study; very large sample (7,000+), long term (4 years); studied UNICOR workers and vocational students and matched group with inmates with no such experience and with overall BOP population; involved direct interviews with parole agents as well as other data collection methods; shows UNICOR and vocational students have better institutional and half-way house adjustment, are less likely to recidivate, more likely to be employed and earn more money than comparison group; both treatment and comparison groups were more at risk than general population.

Data was collected on more than 7,000 inmates over four years. Inmates with either UNICOR work experience, vocational training or apprenticeship training, were matched with inmates who had no UNICOR experience or training, who were released at the same time, and who had similar variables to inmates in the study group.

The study examined three questions: institutional adjustment, halfway house outcomes and post-release outcomes. The study group inmates had a better institutional adjustment record than the comparison group and the two groups had a similar level of halfway house success. The study group, however, had a lower rate of parole revocation than the comparison group. The authors state, "although the magnitude of difference seems small, the differences are not only statistically significant, they are substantially meaningful. At the 12-month time period, 10.12 percent of comparison offenders had been revoked, while only 6.59 percent of study offenders had been revoked." For both groups the recidivism rate is significantly lower than it had been for previous FBP recidivism studies. The authors suggest that this difference can be attributed to common variables in the two groups and argue that if they had compared the study group to the average recidivism rate they would have exaggerated the rehabilitative effect of UNICOR and the vocational training programs.

Study Weaknesses: Did not study academic students; did not compare areas where inmates were released for employment data differences; data shows significant but small differences in treatment and comparison groups.

Comments: Excellent study without control groups per se; discussion of selection bias and why it is difficult and unethical to employ random assignment to treatment in this setting, instead rely on large numbers to "tease out" differences between groups.