FORUM on Corrections Research

Prison industry programs and inmate institutional behaviour

Much of the research assessing the impact of prison work programs focuses on the post-release effects of having employed inmates in prison industrial programs during their incarceration. The results have generally indicated that prison-based programs are, at best, marginally effective in bringing about long-term offender reform.⁽²⁾

The impact of prison industry employment on offenders while they are incarcerated is often overlooked. However, if inmate employment is to reduce recidivism, it must necessarily instill self-discipline, positive work habits and a positive self-image. Such changes should show up in inmate behaviour patterns.

It may, therefore, be useful to examine the effect of prison industry participation on inmate behaviour before assessing the relationship between prison industry and recidivism. Examining the nature and efficacy of short-term behaviour changes may provide clues about how correctional work programs can produce post-release benefits.

This article summarizes a recent research study that explored whether prison industry participation can improve inmate institutional behaviour. The study proposed that such participation may produce better adjusted inmates. If prison industry programs can provide incentives to improve inmate behaviour, then we have, at the very least, taken a small step toward mitigating the overwhelmingly negative influences of prison environments.
What do prison industry programs offer? Prison industry programs offer a number of rewards to inmates. Work programs often pay higher wages than other prison programs and offer opportunities to earn additional money through production incentives and bonus programs.

Prison industry programs also allow inmates to obtain valuable work experience. At least one third of U.S. inmates were unemployed before their incarceration. Any work experience obtained in prison may provide some hope of finding a job after release. Industry programs can also provide environmental advantages. The industry shops may help relieve boredom for some inmates and ease stress for others.⁽³⁾ This setting may even provide a sanctuary for inmates preoccupied with safety concerns.⁽⁴⁾

Prison industry participation can produce less tangible rewards. Prison industry programs are a productive enterprise. This not only helps pass an inmate's time constructively, it also provides some meaningful activity within a relatively stagnant existence.

Further, industry participation involves interaction with civilian supervisors. Anecdotal evidence suggests that inmates are likely to view these individuals differently because they are not part of the prison power structure and don't discipline inmates. This interaction can, therefore, give inmates support and positive reinforcement, as well as an opportunity to interact in an environment similar to real-world work situations.

Involvement in meaningful activity and positive interaction with civilian supervisors should also tend to increase an inmate's self-esteem.

The combined effects of the tangible and intangible rewards provided by correctional industry programs may help to reduce prison misconduct and improve inmate institutional adjustment.⁽⁵⁾

Further, most prison industry programs in New York State have waiting lists of inmates eager to work. It is, therefore, reasonable to expect that industry participation may steer inmates away from misconduct to avoid jeopardizing their jobs. Finally, the time spent in the industry shops may simply replace idleness that would otherwise result in mischief.
Data description The data for this research were drawn from a larger study examining the effectiveness of prison industry in New York State.⁽⁶⁾ Inmates randomly selected for the study had been employed in the industry program for at least six consecutive months during 1981-1982. Participants were identified from prison industry payroll records in five New York State maximum security prisons.

A control group of inmates who had not participated in correctional industry was also drawn from the same five facilities during the same period.⁽⁷⁾

The research examined inmate behaviour before and after industry participation.⁽⁸⁾ The final study sample was made up of 1,077 inmates-511 industry participants and 566 control group members.
Sample relationships The industry participants and control group members were compared in several areas before examining the effect of prison industry employment on inmate misconduct.

There were no significant differences between the two groups as to race, marital status, area of residence, employment, drug and alcohol use, offence type, prior arrests and time served.

However, industry participants did have lower educational levels, were serving longer sentences and were about one year older (on average) than the control group members at the time of incarceration for their current offence.

Further, industry-employed inmates averaged just 2.2 institutional infractions per year, while the control group members averaged 3.3. This difference was statistically significant.
Selection concerns To determine whether prison industry participation had a positive effect on inmate behaviour, it was necessary to compare behaviour before (time1) and after (time2) program participation.

However, industry participants committed significantly fewer institutional infractions than control group members during both their entire term of incarceration and at time1 (2.5 infractions per year versus 3.1 for the control group). These two differences point to the possibility of sample selection bias.

The potential for sample selection bias arises whenever subjects are assigned non-randomly to treatment and control groups.

In this case, although both groups were selected randomly from the general prison population, the groups differed as to institutional infraction rates and other factors (such as age) prior to group assignment. Factors related to these differences may have influenced the original selection of inmates for the correctional industry program.

Further, other unmeasured variables may influence both whether inmates are placed in the industry program and their behaviour patterns (such as emotional maturity or work motivation).

Selection bias hinders efforts to determine the effects of the correctional industry program because it is possible that the inmates in the program were selected for their better overall adjustment and would have improved their behaviour faster than those not selected, regardless of program participation.

In short, selection bias makes it difficult to distinguish program effects from pre-existing selection effects.

Selection bias may also result from sources apart from behaviour differences. For example, individual differences (such as motivation) and institutional differences (such as inmate job availability) may affect the selection process.

However, the effect of these variables is unknown because these factors were not measured. This should be considered in evaluating the results of this study.

The observable bias in this study results from the industry participants having significantly lower institutional infraction rates than the control group members-both over their entire sentences and at time1. The sample was, therefore, divided into two groups with high infraction rates to attempt to compensate for this bias.
Inmate behaviour Official records of inmate involvement in updating prison rules were used to measure inmate adjustment. Annual institutional infraction rates were calculated for:

• the entire period of incarceration;
• time1 (the period before beginning correctional work for industry participants and before the designated division point for control group members); and
• time2 (the period after industry participation began for industry participants and after the designated division point for control group members).

First, institutional infraction rates for the entire sample were examined. The sample was then divided into two groups of high-rate infractors.

The first group was made up of the 50% of sample offenders with the highest infraction rates-those with annual infraction rates of two or more at time1. The second group was made up of the 20% of sample offenders with the highest infraction rates-those with annual infraction rates of five or more at time1.

The initial comparison of rules violations showed that industry participants had lower infraction rates than control group members both before (time1) and after (time2) starting their correctional jobs.

Industry participants also committed significantly fewer institutional infractions than control group members within the first group of high-rate infractors (top 50%) at both time1 and time2.

However, given that program participants committed fewer infractions before beginning their correctional job (time1), we cannot infer how industry participation affected inmate behaviour at time2.

The sample ultimately had to be reduced to the top 20% of infractors to arrive at a point where the infraction rates of industry participants and control group members were not significantly different at time1.

But, industry participants still incurred significantly fewer infractions (3.3) than control group members (5.0) at time2. This suggests that prison industry participation may positively affect inmate behaviour- at least for inmates with high levels of institutional infractions.

Several multivariate models also estimated time2 behaviour for the entire sample and the two groups of high-rate infractors to provide further insight into the effect of industry participation, while controlling for other influential factors.⁽⁹⁾

The most stringent test examined the effect of industry participation at time2 while controlling for institutional infraction rates at time1, as well as other influential factors.

The results showed that industry participation was not significantly related to lower institutional infraction rates for the total sample at time2. Industry participation was, however, significantly related to lower infraction rates at time2 for both high infraction groups.

Although some selection bias may still exist, this research shows that correctional industry participation has a significant positive impact on the behaviour of inmates with relatively high rates of institutional rule infractions during the early portion of their sentence.

Therefore, prison industry participation was an important factor in modifying the behaviour of inmates prone to relatively high rates of prison rules violation.
Something to build on Although it is difficult to clearly identify treatment effects associated with prison industry participation, this research shows that the inmates involved in prison industry were better behaved to begin with and displayed better behaviour over time.

Therefore, although involvement in the program may or may not have directly brought about improved behaviour, prison industry appears to have provided a venue for well-adjusted inmates to do their time and avoid potential problems.

However, industry participation was associated with significant behaviour improvement for inmates with high rates of institutional infractions. These findings support the argument that a reward or incentive mechanism may be associated with prison industry participation.

In short, prison industry programs may provide a structure that promotes and reinforces positive inmate behaviour. Exploring the dynamics of these types of short-term behaviour changes may provide a foundation for more realistic and successful long-term offender outcomes.



⁽¹⁾135 Western Avenue, Albany, New York 12222.

⁽²⁾A. Basinger, Are Prison Work Programs Working: The Impact of Prison Industry Participation on Recidivism Rates in Ohio, Master's Thesis, School of Public Administration, Ohio State University, 1985. See also C. Johnson, The Effects of Prison Labor Programs on Post-Release Employment and Recidivism, Doctoral Dissertation, Florida State University, 1984. And see Utah Governor's Task Force on Correctional Industries, The Challenge of Correctional Industries (Salt Lake City: Utah Department of Corrections, 1984.) And see K. Maguire, T. J. Flanagan and T. P. Thornberry, "Prison Labor and Recidivism," Journal of Quantitative Criminology, 44 (1988):3-18. For the single exception to the generally negative findings, see W. G. Saylor and G. G. Gaes, "PREP Study Links UNICOR Work Experience With Successful Post-Release Outcome," Federal Bureau of Prisons Research Forum, 1 (1992):1-8.

⁽³⁾H. Toch, K. Adams and J. D. Grant, Coping: Maladaptation in Prisons (New Brunswick: Transaction Publishers, 1989).

⁽⁴⁾Toch, Adams and Grant, Coping: Maladaptation in Prisons.

⁽⁵⁾Evidence exists that "meaningful participation in conventional activities can turn attitudes in a prosocial direction." See Toch, Adams and Grant, Coping: Maladaptation in Prisons.

⁽⁶⁾T. J. Flanagan, T. P. Thornberry, K. Maguire and E. McGarrell, The Effect of Prison Industry Employment On Offender Behaviour: Report Of The Prison Industry Research Project (Albany: Hindelang Criminal Justice Research Center, 1988).

⁽⁷⁾The six-month exposure period was selected after consultation with industry personnel about appropriate minimum exposure to the program. The 1981-1982 period was chosen so that most of the sample would have been released by the end of data collection in 1986. Therefore, institutional adjustment records cover the entire period of incarceration for most inmates.

⁽⁸⁾The start date of inmate industry assignments was a naturally occurring intervention point for industry participants. A two-stage multivariate matching procedure was used to match each industry participant to a member of the control group to determine a point at which to divide each control group member's prison term into "time1" and "time2."

⁽⁹⁾The multivariate models were estimated using Ordinary Least Squares regression. The following variables were included in all models: age, race, education, marital status, pre-prison occupation, offence, sentence length, number of prior arrests, pre-prison drug and alcohol use, and area of residence prior to arrest.