Correctional Boot Camps: A Tough Intermediate Sanction - Chapter 16. Boot Camps and Prison Crowding MENU TITLE: Boot Camps and Prison Crowding Series: NIJ Report Published: February 1996 13 pages 21,188 bytes Boot Camps and Prison Crowding by Dale G. Parent Dale Parent is an Associate at Abt Associates Inc. in Cambridge, Massachusetts. He has been studying boot camps since 1989. Prison boot camps reduce crowding only under a limited set of difficult-to-achieve conditions. Population simulations suggest that to reduce prison crowding boot camps must recruit offenders who otherwise would be imprisoned, offer big reductions in prison terms for completing a boot camp, minimize washout and return-to-prison rates, and operate on a large scale. Few boot camps meet these conditions. Many limit eligibility to nonviolent first offenders, select offenders who otherwise would receive probation, and intensively supervise graduates, thus increasing return-to-prison rates for technical violations. In most jurisdictions, boot camps appear more likely to increase correctional populations and costs rather than reduce them. These findings have important implications for how those programs should be designed and operated. If boot camps are intended to reduce prison populations, they should have a large capacity; they should select participants from among offenders already committed for relatively long prison terms; and they should implement policies to minimize in-program and postrelease failures. In jurisdictions that run boot camps, most correctional officials report that their boot camps are intended to reduce prison populations, prison crowding, and correctional costs.1 This chapter explores conditions that must be met to achieve these goals and assesses the extent to which boot camps are likely to achieve them. First, boot camps' impacts on population levels and crowding are examined, and then cost issues are discussed. The chapter concludes by describing key implications and design features of boot camps that will increase the chances of affecting prison population levels. States' prison populations are determined by factors that are mostly outside the control of officials who run prisons, including sentencing laws passed by legislatures, enforcement priorities set by police, charging and bargaining practices instituted by prosecutors, sentencing patterns followed by judges, supervision and revocation patterns of probation agencies, and release and revocation practices of parole boards. Prison officials cannot exercise discretion legally vested in these other officials. In the short term, correctional officials have one discretionary decision that can be used to affect prison population levels. Within limits set by law, they can shorten the duration of confinement for selected offenders who have been committed to their custody. To some extent they can adjust durations of confinement by granting or withholding various credits for aspects of program participation and behavior. Correctional officials can release some prisoners early on furloughs or work release, but other forms of release, such as parole, must be approved by other agencies. For some inmates, release before serving a minimum period of confinement is prohibited by law. Thus, if boot camps are to reduce prison populations or crowding, inmates admitted to boot camps must serve substantially shorter prison terms than they would have served otherwise. In the long term, correctional officials might be able to reduce prison admissions if boot camp graduates returned to prison at lower rates than comparable parolees, because the graduates either had fewer new convictions or lower revocation rates for rule violations. Impact on Prison Bedspace In 1991, a simulation model was developed to identify conditions that must be met for prison boot camps to reduce the number of prison beds needed. The model can be used when planning a boot camp to understand how it should be designed to reduce populations. It can also be used once a boot camp is in operation to see whether the program is having the expected impact on prison bedspace.2 Researchers used the model to assess the population impact of Louisiana's boot camp3 and to assess the population impacts of boot camps in jurisdictions participating in a multisite evaluation.4 In the analysis of the model reported, boot camps' impacts on prison population levels were assessed in relative terms. For example, as the probability of imprisonment varies, so does the number of prison beds lost or, conversely, saved. The objectives of the simulation were to find out how different boot camp design options affected the need for prison space and to determine how many prison beds a State would need to provide after implementing a boot camp that had particular design options, compared to the number it would have had to provide if its past population trends continued. For this objective, other factors influencing prison populations were assumed to be constant, which would allow the identification of the independent effect of a particular configuration of boot camp policies on prison population levels. The objective of the simulation was not to predict the total number of prisoners a jurisdiction would have in the future. Such a model would be much more complex and require development of a substantial data base on overall sentencing practices and offender characteristics. These analyses suggest that the following conditions must be met if boot camps are to reduce the need for prison beds: o The probability of imprisonment for those admitted to boot camps must be very high. o Boot camp graduates must get a substantial and real reduction in their term of imprisonment. o In-camp failure rates must be low. o Boot camps must operate on a large scale. High probability of imprisonment. This condition refers to the odds that offenders admitted to the boot camp would have been imprisoned if the boot camp did not exist. Given typical in-program and postrelease failure rates, the model indicates that 80 percent of those admitted to a boot camp would have to receive a prison term just to begin reducing required prison beds. Larger reductions require even higher probabilities of imprisonment. In many States, boot camps are restricted to nonviolent first offenders. This restriction was lifted from earlier shock probation laws, under which young, inexperienced offenders (who presumably were most impressionable and most likely to be deterred) were sent to a regular prison for 90 days and then returned to community supervision. Evaluations failed to show any deterrent effect in shock probation.5 Variants of these deterrence-based eligibility criteria appear in most boot camp programs. If boot camp participants are selected at time of sentencing, deterrence-based eligibility criteria make it difficult to achieve a high probability of imprisonment; no State imprisons 80 percent of its nonviolent, first-time felons. Reduction in term of imprisonment. If boot camp graduates serve 6 months rather than 24 months, each boot camp graduate will save the system 18 person-months of confinement. A small reduction--6 months versus 8 months--will not reduce bedspace needs appreciably. Duration of imprisonment varied substantially among different State programs in the multisite study. Particularly important is the comparison between the time that shock incarceration graduates spent in prison and the time they would have spent (e.g., the time the shock-eligibles spend in prison). Deterrence-based eligibility criteria also make it difficult to achieve substantial reductions in durations of confinement. In most States, the few nonviolent first offenders who are imprisoned are likely to be released after having served very short prison terms. Minimization of in-program failure rates. In typical programs, 30 to 40 percent of those admitted fail to complete the program and serve regular prison terms, quickly eroding savings in prison bedspace. If these first three conditions are not met, the simulation shows that boot camps will increase prison populations. Boot camps can cause substantial increases (relative to their capacity) depending on how they are configured. If their participants have a low probability of imprisonment--for example, if only 10 percent of them would have been in prison if the boot camp did not exist--a typical 200-bed boot camp will increase the prison population by more than 500 inmates, due mainly to the 30 to 40 percent who will fail during the program and who will serve prison terms as a consequence. If the boot camp lasts just 90 days, then four times a year it can cycle in a new group of participants, of whom 30 to 40 percent will become regular prison inmates. In this example, 90 percent of these in-program failures would not have been in prison at all if the boot camp did not exist. Operation on a large scale. If jurisdictions achieve the three conditions described above, they still must operate large-scale boot camps if they want to reduce prison bedspace requirements. The exact size of boot camps needed to achieve a desired reduction varies with the probability of imprisonment, the amount of reduction in time served for completing the boot camp, and the failure rates. But the effect of a small (e.g., 50-bed) boot camp in a moderate-sized prison system will not be discernible. If jurisdictions cannot run boot camps that conform to the three conditions described above, running small boot camps at least will minimize the increase in prison bedspace that will result. Long-Term Impact on Prison Populations In the long-term, boot camps might reduce prison populations if graduates return to prison at lower rates than comparable offenders who did not participate in boot camps before their release from prison. At present, the evidence of boot camps' impact on return-to-prison rates is inconclusive. Most studies show that boot camp graduates do not return to prison at significantly different rates than members of comparison groups.6 To date, however, none of the studies has randomly assigned inmates to experimental and control groups, so outcomes may have been affected by inmate self-selection. (For example, some eligible inmates volunteer for boot camps and others do not; some enter and complete boot camps while others drop out after a few weeks.) Also, none of the studies has tested the impact of different models of aftercare on outcomes. Boot camp graduates frequently are intensively supervised, even though they usually are not the high-risk or high-need offenders for whom intensive supervision was originally intended. Intensive surveillance may detect more rule violations that lead to revocation, thus itself inflating return-to-prison rates.7 Limits on Population Reduction When design features and practices in existing boot camps are compared with the findings of the simulation model, it appears that most adult boot camps are now designed and operated so that they are more likely to increase rather than reduce prison populations. In many jurisdictions, enabling legislation restricts participation to nonviolent first offenders and allows judges to select participants at the time of sentencing, practices almost certain to yield a pool of offenders who probably would not have been imprisoned if the boot camp were not available. Many existing boot camps also have moderately high in-program and postrelease failure rates. Even if boot camps have been carefully designed to reduce prison bedspace requirements, the broad factors that drive prison population can easily overwhelm any population reductions achieved by boot camps. When the New York State Department of Correctional Services (NYSDOCS) designed boot camps in 1987, officials made all the appropriate choices to reduce populations. NYSDOCS, not judges, identified candidates for boot camps from among regular incoming prison inmates, so their probability of imprisonment was 100 percent. The legislature approved boot camps for offenders convicted of moderately serious crimes, who had somewhat longer prior records. Hence, most New York inmates who complete a boot camp shave 18 to 24 months from their prison terms. Finally, officials ran a large boot camp operation--providing 1,750 beds in 5 separate boot camps. NYSDOCS staff credibly estimate that by 1992 their boot camps caused the prison population to be 1,540 less than it would have been otherwise. Yet, between 1988 and 1992, New York's prison population grew from approximately 41,000 to nearly 58,000--an increase more than 11 times greater than the relative reduction from boot camps.8 At best, New York's boot camps slightly lessened the extent of prison population increase. If officials seriously expect to reduce or eliminate prison crowding, then boot camps should be only one small piece of a much larger and more fundamental strategy to structure discretionary sentencing choices and to provide an appropriate range of sanctioning options. If prison populations are at or near equilibrium, properly designed boot camps by themselves may be able to effect a slight reduction. However, if the prison population is rising, boot camps are not likely to stop or reverse that trend. Impact on Correctional Costs Boot camps that have sparse programming during the institutional phase cost about the same to operate as regular prisons on a per-inmate, per-day basis because the staffing levels and facility costs are very similar. Boot camps with richer programming (e.g., intensive drug treatment, individual counseling, extensive education, or other programs) cost more to operate.9 Thus, to reduce total operating costs, boot camps must substantially shorten participants' duration of confinement.10 Measuring operating cost savings is difficult. Officials often incorrectly estimate operating cost savings by multiplying the estimated reduction in total person-days of confinement by the average daily cost of running regular prisons. They calculate average daily costs by dividing the total cost of running a prison for a year by the number of person-days of confinement it provides in the year. This approach leads to substantial overestimation of savings. In the short term, population reductions are likely to save only the costs related to reductions in goods and services "consumed" by offenders, such as food, clothing, and health care. In most prisons, these short-term costs amount to only a few dollars a day. Staffing patterns (and costs) do not change when a facility's population varies by a few inmates. To save long-term costs, populations must be reduced enough that a cellblock or even an entire prison can be closed, or that construction of a new facility can be averted. Averted costs are especially difficult to demonstrate because it is not known whether a legislature would have authorized additional prison construction if the boot camp had not been operating. Finally, many practitioners now advocate an expanded aftercare program for boot camp graduates that provides a longer period of supervision, greater continuity in treatment, and more rigorous provision of support services. It is difficult to evaluate the cost of such an aftercare program, but it will cost more than the type of aftercare now offered to most boot camp graduates. Implications For officials considering establishing boot camps to reduce prison populations or costs, these findings have sobering implications. They suggest that officials should first consider whether they should rely on boot camps for this purpose or invest their energy and resources in broader and more promising strategies. For those who decide to proceed with boot camp development, this section describes the author's assessment of how boot camps should be designed and operated. It is based on a review of the research, talking with correctional officials, and visiting programs. If officials want boot camps to reduce prison populations, crowding, or correctional costs in the short term, then they should select boot camp participants from among offenders already committed for imprisonment. This is the most effective way to recruit offenders who have a very high probability of imprisonment. This method, of course, is not foolproof. Judges might alter their sentencing practices and imprison selected offenders so that they may be selected for boot camps. Correctional officials might counter this potential by operating "low profile" boot camps, in which they actively seek to avoid publicizing their programs, and by forcefully communicating with judges that the Department of Corrections, not judges, control placement decisions for those committed to prisons. Thus, if population impact is the goal, boot camps should be used as a form of early release. Agencies that operate confinement facilities should run boot camps and select participants. If judges select boot camp participants at time of sentencing, experience strongly suggests that a large percentage of those admitted to boot camps probably would have been given probation, making their expected rate of imprisonment very low. If this is the case, judicial selection of boot camp participants may be incompatible with the goal of reducing costs and prison populations.11 In addition, boot camps should target offenders who otherwise would serve moderately long periods of confinement, not those who will be released in a few months anyway. This is the most effective way to get bigger reductions in time served for offenders who complete the boot camp. Because almost all boot camps' length varies between 3 and 6 months, there is little room to increase the reductions by shortening the boot camp's duration. This implies that boot camps should be run only by correctional agencies that confine offenders for substantial periods of time, for example, 2 or more years. Agencies that confine offenders for short terms--like county jails and juvenile training schools--will find it especially difficult to reduce populations with boot camps because they probably will not be able to reduce confinement terms enough to reduce total person-days of confinement significantly. They will also find it harder to get eligible offenders to volunteer if they can offer only a very small reduction in time served. Another policy implication is that policymakers should avoid deterrence-based eligibility criteria that limit boot camps to first-time, nonviolent offenders. Early boot camps inherited deterrence-based criteria from prior programs like Scared Straight!, and shock probation. Deterrence-based eligibility criteria limit both the probability of imprisonment and the amount of potential discount. Instead, boot camps should target offenders with prior convictions and with moderately serious offenses (e.g., situational assaults), as well as probation or parole violators. In order to minimize in-program failure rates, boot camps should not permit participants to withdraw during the first few weeks of the program. The first few weeks in a boot camp are most difficult, physically and mentally, for inmates, leading to high rates of voluntary withdrawal. A policy of prohibiting withdrawals during the first 3 weeks would allow inmates to adjust to the regimen before considering withdrawing. In addition, boot camps should develop strong behavioral management systems aimed at reducing dropouts due to violations of program rules. They could do this by providing a range of in-program rewards and sanctions (the latter not involving removal from the program). Boot camps should continuously monitor and assess participants' progress toward behavioral or performance objectives, so that they can provide immediate, positive reinforcement to participants who behave appropriately or who excel, and immediate correction to participants who do not attain minimum specified levels. This will let officials intervene before small problems escalate into serious issues that might warrant removal from the program. Of course, officials need the option to remove from the boot camp and return to regular prison those who persistently misbehave. Finally, officials must be willing or able to operate boot camps with a large capacity. Otherwise, their relative impact on prison populations (even under the best conditions) is likely to be negligible. Long-Term Impact on Prison Readmissions If officials intend to reduce prison populations by reducing return-to-prison rates for boot camp graduates, then officials should be certain that the means by which they expect boot camps to reduce return-to-prison rates are clearly stated, plausible, and capable of being implemented as expected. Boot camps may affect return-to-prison rates by two means or by both operating together (see chapters 17 and 18). First, the disciplinary regimen may improve offenders' self-esteem or attitudes toward authority, leading them to change their behavior once they are released. Second, treatment programs such as counseling and drug treatment may be effective in the structured setting of boot camps in dealing with inmate deficiencies that form obstacles to improved behavior. Disciplinary regimen. If the disciplinary regimen is expected to alter offenders' future behavior, the agency should describe the link between the regimen and the expected future behavioral changes. For example, an agency may expect that if staff provide good role models for inmates, then offenders' attitudes and behavior may become more prosocial, and upon release, their new crime and violation rates will decline. Careful selection and thorough training may be essential if staff are to play these roles effectively. Treatment. If treatment is expected to lower the return-to-prison rates, the boot camp should have a plausible method of identifying treatment needs and delivering treatment services. At a minimum, this should include the capacity to assess offenders' needs and to develop treatment plans that address each offender's needs, setting measurable objectives and performance schedules. The correctional treatment services specified in those plans should be delivered to offenders during their stay in the boot camp. The agency should provide the resources (staff, materials, space) and allot sufficient time during the daily schedule of activities to deliver the services. It should monitor offenders' progress toward implementing their treatment plans while in the boot camp and modify plans as required. Because all treatment needs probably cannot be met during a short boot camp program, the agency operating the boot camp should provide for or ensure continuity of treatment during aftercare. If different agencies operate the boot camp and aftercare, both should be involved in planning the total program, and appropriate interagency agreements and protocols should be implemented to make sure that treatment is delivered during aftercare as intended. Those who deliver aftercare should be included in the development of treatment plans and in the preparation for offenders' transitions from institution to community. Finally, the agency operating aftercare should monitor progress on the treatment plan during aftercare and modify it as required. Program officials and aftercare providers should work as advocates to promote offenders' access to programs and services offered by other agencies and to monitor offenders' utilization of those services. They also need to recognize and remove barriers to access and utilization of services. Aftercare should provide surveillance only to the degree that it is justified on a case-by-case basis by a valid risk-assessment instrument. If aftercare consists mostly of intensive surveillance, revocation rates are likely to be high (due mainly to minor or "technical" violations of supervision rules, not new crimes), thus undercutting population reduction objectives. Boot camps that conform to the conditions described in this chapter may reduce somewhat the need for prison bedspace. However, as New York's experience demonstrates, even the best-done boot camps cannot stop prison population growth if imprisonment rates or prison terms are growing rapidly. However, boot camps could be one small part of a broad set of policy initiatives to help structure the use and duration of confinement sentences. Notes 1. Doris L. MacKenzie and Claire Souryal, "Boot Camp Survey: Rehabilitation, Recidivism, Population Reduction Outrank Punishment as Main Goals," Corrections Today (October 1992):92. 2. The simulation model runs under Lotus 1-2-3, version 2.1, and is available free from Dale G. Parent, Associate, Abt Associates Inc., 55 Wheeler Street, Cambridge, MA 02138. 3. Doris L. MacKenzie and Dale G. Parent, "Shock Incarceration and Prison Crowding in Louisiana," Journal of Criminal Justice, 19, no. 3 (1991):225- 237. 4. D.L. MacKenzie and A. Piquero, "The Impact of Shock Incarceration Programs on Prison Crowding," Crime and Delinquency, 40, no. 2 (1994): 222-249. 5. In 1979, Dr. James Finckenauer, an Assistant Professor of Criminal Justice at Rutgers University, conducted an experimental analysis that randomly assigned eligible participants to Scared Straight! (a program to deter juveniles from further delinquent behavior by showing them the rigor and brutality of adult prison) and control groups (no Scared Straight!). At a 6-month followup point, those exposed to Scared Straight! performed worse--41.3 percent engaged in new delinquent behavior, compared to only 11.9 percent of the controls. See J. Finckenauer, Scared Straight! and the Panacea Phenomenon, Englewood Cliffs, New Jersey: Prentice-Hall, 1982. 6. Doris Layton MacKenzie and Claire Souryal. Multisite Evaluation of Shock Incarceration: Executive Summary. Final report to the National Institute of Justice, 1994. 7. See MacKenzie and Parent (1991), and J. Petersilia and S. Turner, Evaluating Intensive Supervision Probation/Parole: Results of a Nationwide Experiment, Research in Brief, Washington, D.C.: U.S. Department of Justice, National Institute of Justice, May 1993. 8. Facility Capacity and Inmate Populations, 12/31/85 to 12/31/94, Program Planning, Research, and Evaluation Unit, New York State Department of Corrections Services, Albany, 1995. 9. See Dale G. Parent, Shock Incarceration: An Overview of Existing Programs, Washington, D.C.: U.S. Department of Justice, National Institute of Justice, June 1989. 10. This assumes that boot camp graduates are not more likely to fail when released to the community (and to trigger the social and criminal justice costs associated with such failures) than are offenders who do not graduate from them. In fact, data from States with "backend" selection procedures generally show that boot camp completers fail at rates equal to or lower than noncompleters or nonparticipants. In States with frontend selection, the situation is less clear. Failure rates for boot camp graduates appear to be somewhat higher, but "comparison" groups may not, in fact, be comparable. In addition, there is strong evidence that many boot camp graduates fail not because they are a more risky population, but because they are more intensively supervised, and, hence, there is a greater chance that minor rule violations will be detected and punished. 11. Parent, 1989.