Correctional Boot Camps: A Tough Intermediate Sanction - Chapter 10. First-Year Evaluation of the California Youth Authority Boot Camp MENU TITLE: Evaluation: CA Youth Authority Boot Camp Series: NIJ Report Published: February 1996 18 pages 39,656 bytes First-Year Evaluation of the California Youth Authority Boot Camp by Jean Bottcher and Teresa Isorena Jean Botcher and Teresa Isorena are with the California Youth Authority's Research Division. They have been working on the LEAD evaluation since the program's inception in 1992 and expect to complete their final evaluation report at the end of 1996. This chapter assesses the California Youth Authority (CYA) boot camp's first year of operation. The program, called LEAD, was established in 1992 with a legislative mandate to reduce recidivism and provide a cost-effective sentencing option. The camp serves CYA's nonviolent and least serious offenders and was designed with an experimental evaluation to be completed in 1996. This chapter documents the program's accomplishments, including the implementation of an officer training model that has helped develop confidence, teamwork, and discipline at the camp; a generally safer, healthier institutional environment, but with some emotional and abuse problems; and short-term reductions in length of confinement, although insufficient as yet to offset increased short-term program costs. The following issues relevant to policymakers are discussed: (1) the forces that work against cost-saving goals in a prison system, (2) the problems of designing a creative, dynamic program in a prison setting, and (3) the need for a solid treatment and training core with aftercare support. Also documented are inherent problems that will make LEAD difficult to sustain over the long term. The initiative for developing a California Youth Authority (CYA) boot camp came from Joe Sandoval, Secretary of the Youth and Adult Correctional Agency (YACA), the State agency that oversees California's prison system. A former head of California's State police, Sandoval was appointed Secretary of the Correctional Agency by Governor Pete Wilson. The Wilson administration had good reasons for seeking out programs that would lower the State's incarceration costs, which increased dramatically during the 1980's. CYA's population of first commitments from juvenile court increased from 2,190 in 1980 to 2,433 in 1990, and wards' average length of incarceration increased from 14 months in 1982 to 25 months in 1992. CYA's estimated per capita incarceration costs had also risen rapidly, from $19,953 in 1980 to $30,783 in 1990. In response to Sandoval's request, the CYA established a departmental committee in May 1991 to develop a boot camp program.1 State Senator Robert Presley sponsored a bill mandating that CYA develop and evaluate a boot camp program. The bill was signed by Governor Wilson in February 1992. The Presley Bill specified program and evaluation parameters, eligibility criteria, and policies regarding participating wards for the program, which was named LEAD, an acronym for the qualities the program was designed to promote: leadership, esteem, ability, and discipline. The legislation set two major program goals for LEAD: (1) to provide the State a cost-effective sentencing option, and (2) to reduce recidivism. It also specified that the program be a "treatment continuum" beginning with a 4-month, highly structured institutional component with extensive treatment and military training and concluding with a 6-month intensive parole component with a wide range of services and a relapse-management strategy. Judges were permitted to recommend but not select wards for LEAD, and CYA was expected to make every effort to retain participants. In exceptional situations wards could be retained in LEAD's institutional phase for an additional 30 days, but they were not to be penalized solely on the basis of not completing the program. The law also specifically stipulated that staff not use corporal punishment or "degrading, humiliating, or inhumane" training methods. Preparations Planning for LEAD began in earnest in summer 1992. Weekly planning sessions were held at the Preston School of Industry in Ione, California, the institution selected to house the program's "boot camp" phase. A collaboration was established during this period with the California National Guard, which was seeking peacetime service opportunities. Captain Robert Bradley of the Guard took an enthusiastic interest in LEAD, eventually providing the program training materials and staff training. He also persuaded the CYA to base the program's instruction and drill procedures on the U.S. Army officer training model rather than a boot camp model in response to the potential problems of inmate abuse. Military training for institutional staff began in midsummer 1992 at Mather Air Force Base in Sacramento, California, and in September the first platoon of 15 wards entered the program at Preston. The CYA's parole branch then conducted a series of planning sessions that culminated in a manual of guidelines and a 2-day training session for selected field parole agents who were to work with LEAD parolees. The CYA implemented the parole program when the first platoon of LEAD graduates was paroled in January 1993. The Superintendent of Preston and CYA parole officials encouraged research efforts and provided unusual access to the LEAD program, as did all other CYA staff at LEAD and other locations. Evaluation Design and Methods There are two legislatively mandated components to the program's evaluation design. This chapter presents findings from the first component, which was an implementation and process evaluation to determine whether the boot camp was set up as mandated and to describe how the program is functioning in practice. The second component is an experimental impact evaluation, which will measure the success of the program in achieving its goals.2 Wards were randomly selected and placed either in LEAD or in a control group, with the control wards sent to other CYA programs using standard procedures. The research team attended planning meetings at Preston and with parole staff during summer and fall 1992, and subsequently visited the operating boot camp on numerous occasions to observe the living unit, participate in staff meetings, attend graduation ceremonies, observe program activities, interview staff, and interview wards, dropouts, and control wards. Researchers also contacted parole offices to interview wards and talk to staff, and they visited both reception centers to discuss screening procedures with casework staff. Selected wards and control wards were interviewed periodically toward the end of their incarceration periods. Researchers used an open-ended interview guide that included questions about what they liked and disliked about their programs and about feelings and attitudes that LEAD is attempting to cultivate. Wards also completed surveys at various times during their incarceration on their perceptions of the program and its impact on them, as well as a prison environment inventory designed by Kevin Wright (1985) but modified for a juvenile population.3 Evaluation surveys were administered to 122 LEAD participants and 72 control wards.4 An analysis of the characteristics of the LEAD and control wards used in the evaluation indicated that the two groups were quite comparable. Immediately following LEAD's first year of operation, interviews were conducted with 9 parole and 21 institutional staff, including 13 of the 14 counselors originally with the program. Questions focused on descriptive program information and on subjective assessments. Program Description Findings The findings on LEAD's first year of operation address the screening and selection of wards, the institutional phase, and the intensive parole phase. Screening and selecting LEAD wards. LEAD's legislative mandate included the following eligibility criteria: an age of at least 16 (subsequently amended to 14), a juvenile court commitment, no previous LEAD involvement, no commitments for serious violence, a history or risk of substance abuse (which has not been used to exclude wards from the program), medical clearance for strenuous physical activity, informed consent, and parole board approval. In addition, the CYA and the parole board established policies for other eligibility criteria that excluded wards with a primary need for the department's special mental health programs, wards who were illegal aliens, and wards who engaged in serious incidents of violence in the 6 months prior to entry that involved, or likely involved, substantial injury. Referrals were considered from two sources: juvenile court first commitments and parole violators. Fifty-nine percent of first commitment referrals were automatically rejected on the basis of the violent offense criterion and the much less restrictive age criterion in the first year of screening. Of the wards who survived screening, about 85 percent were approved by the parole board.5 For juvenile court first commitments, only 12 percent were accepted after both CYA screening and parole board review. For parole violators, only 5 percent of referrals were ruled eligible. The screening and selection process was occasionally contentious during the first year. Staff at screening centers had difficulty identifying a sufficient number of eligible wards and were discouraged when wards were rejected by the parole board. Some program staff complained about "inappropriate" referrals of wards they believed should have been excluded due to their belligerence, lack of motivation, or need for more intensive psychological counseling. During the first year 51 wards were referred to the program through judicial recommendations, 15 of whom did not meet the established age or offense criterion. Of the remaining wards, 19 were found eligible and 17 were not. Overall during first-year screening, eligible wards were, on average, 17.5 years old and 40 percent Hispanic, 29 percent African American, 25 percent white, and 6 percent other. Of their commitment offenses, 69 percent were against property, 10 percent against persons, 7 percent drug-related, and 14 percent other. Treatment and training in LEAD's institutional phase. LEAD's living unit or lodge, Hawthorne, occupies a central location at Preston, one of the oldest CYA institutions, located in the rolling foothills of the Gold Rush town of Ione, California. Three classrooms, a gym, and a common dining room are located a short march away. A path runs beside much of the institution's perimeter security fence and, were it not for the hills, wards using it would be able to see the lodge from almost any vantage point along its course. No other living unit at the institution could be more of a "fish bowl." The boot camp lodge, like most Preston buildings, is made of brick and stone and was built in the 1950's. Its setting is attractive and nicely landscaped, but fixtures are old and furnishings spartan. The lodge is always kept clean and neat, and its two offices are usually busy. Observers found that wards were usually completing assigned paperwork, shining shoes, ironing clothes, or listening to or waiting for orders. Wards rarely watched television, which was available only for educational purposes. LEAD's institutional program is run primarily by 12 "TAC" officers, who are classified as youth counselors and occupy the primary staff positions in all CYA living units. TAC stands for teach, advise, and counsel, key elements of the officer role as defined in the TAC Standard Operating Procedures. The procedures, which are in large part taken directly from Army manuals, stipulate that the most important element of the LEAD program environment is the TAC officer. This LEAD focus on staff leadership is due primarily to the selection of a military officer training model as opposed to a basic boot camp model. It is the TAC officer's role to teach leadership and its attendant good qualities. The TAC officers' role in LEAD is an element of the program that explicitly and theoretically relates to reducing recidivism, which can happen in two mutually reinforcing ways. One way is through the "referent power" of the TAC's, the possibility that wards will identify with TAC's and emulate their good qualities. The other way is by internalizing the program's discipline as self-discipline. According to the operating procedures, this process is made possible when wards experience successes in training and when they identify with their officers. Toby and Pearson made similar observations in their study of juvenile boot camps.6 TAC officers report directly to a sergeant, who reports to a captain. LEAD also has a full-time institutional parole agent, a psychologist, three teachers, standard group supervisors who stand watch at night, two part-time liaison agents from the CYA parole branch, and an office assistant. The program administrator at Preston in charge of special units, who reports to the facility's superintendent, is responsible for the administration of LEAD. The CYA budgeted and enhanced yearly cost of the LEAD program was $532,500 in fiscal year 1992-93. Enhanced costs are those over and above the standard living unit costs of a 50- to 60-bed unit. A review of LEAD's budget indicated that staffing the additional positions mentioned above has been the program's biggest expense. LEAD's overall budget, including the institutional phase, was $1,086,300. The other enhanced costs included additional parole agents, an additional parole board member, and research staffing. LEAD's treatment and training program is organized in monthly phases, visually reinforced by color-coded T-shirts and caps. At monthly ceremonies, wards are promoted in color and receive merit awards. The education program has a practical life skills focus (e.g., on health and employment issues the first month) and includes individualized services such as GED (general equivalency diploma) testing and special education screening and tutoring. TAC officers provide supervisory assistance in classrooms. During the program's fourth and final month, wards perform community service activities, prepare for their parole board appearance, and complete their life plans (narrative accounts of their past, present, and expected future lives that include reappraising values and developing plans to handle such issues as drug use, sexual relationships, finances, and work). Daily activities begin with reveille at 6 a.m. and conclude with showers and cleanup at 10 or 11 p.m., depending on the platoon. Other activities include 2-hour evening substance abuse training sessions that use a 12-step model, daily physical training, drill and ceremony exercises, and lodge cleanup duties. Of particular note are tutoring arrangements with community volunteers and the program's voluntary 6- to 8-session bereavement therapy group, which was developed and conducted by Dr. Gary Mackelburg, the first LEAD psychologist, for wards who had experienced the loss of close family or friends. During the first full year of program operation, 180 wards entered LEAD in monthly 15-member platoons, and 107 wards were paroled. Because of the program's graduated intake process, parolees represented 71 percent of the wards LEAD could have paroled during the year. Forty-five wards were dropped from the program, of whom 42 came from the 10 platoons that could have been paroled. Based on these figures, the program's dropout rate was 28 percent for the first year. (One ward was demoted for 30 days at the end of the first year but was eventually paroled.) Parole program. Guidelines for LEAD's parole program are based on a philosophy of providing the highest quality of services available to meet the needs of parolees. The underlying goal is to use every legitimate strategy in the book to help LEAD parolees avoid revocation and future criminal involvement. The LEAD parole program involves both an institutional phase and a field parole phase. Two "liaison agents" from the parole program work part-time at the institutional site, acting as a crucial communications link among wards, institution personnel, and field parole staff. These agents are to ensure the promised "continuum of treatment" from institution to field parole; they also supervise LEAD parolees on intensive parole. Specified service levels for intensive parole in urban or suburban locations are two contacts per week for the first 60 days and one per week for the next 4 months, two random drug tests per month, and other services as needed. Rural locations call for somewhat reduced numbers of contacts. LEAD parole agents reported providing parolees the following interventions and services beyond routine contacts: substance abuse counseling (46 percent of LEAD parolees), employment referrals (42 percent), temporary detention combined with counseling or "restructuring" (36 percent), alternative placements (22 percent), a 60-day residential drug treatment program (13 percent), required community volunteer work (10 percent), electronic monitoring (4 percent), and a full-time day reporting program (2 percent). Program Assessment Findings The report's discussion of program assessment is framed by two questions: Are the program's legislatively mandated goals guiding its operation? How is the program working? Findings discussed below focus on the activities undertaken to fulfill the goals and on perceptions of the staff and wards interviewed. Final assessment of whether LEAD is achieving its mandated goals to reduce costs and recidivism must await completion of the impact evaluation. Are the Program's Legislatively Mandated Goals Guiding Its Operation? The LEAD program is a rich assortment of treatment and training activities that, compared to other CYA living units, are set in a relatively safe, positive, and disciplined environment. The program design is not notably theoretical. Implicitly, it is based on the following assumptions: o By employing diverse activities along with individualized treatment, the program will reach more wards. o Living in a structured environment will rub off as self-discipline. o A variety of constructive skills, positive attitudes, and knowledge will "produce" less criminal activity. The TAC procedures previously discussed, which suggest how wards can internalize new values and behaviors through relationships with TAC officers, do, however, add a theoretical dimension to LEAD programming. Reducing recidivism. LEAD's first mandated goal was to reduce recidivism. The evaluation found that the program, as specified in the enabling legislation, had been implemented. It also addressed whether the program was designed and implemented with the explicit goal of reducing recidivism. Virtually all staff expressed the goal of reducing recidivism by means of some specific treatment goals. Although institutional staff most commonly said that the goal of the program was to save money, they mentioned only treatment and training as methods to achieve that goal. There were some variations in focus among the boot camp staff, the parole staff, and the wards. Among the boot camp staff, the most commonly mentioned methods of achieving LEAD goals were military training and structure (mentioned by 62 percent); educational training (43 percent); the 12-step drug abuse program (43 percent); motivational techniques and high staff expectations (38 percent); physical training (33 percent); good staff and role modeling (27 percent); and discipline and structure (27 percent). Most parole staff cited reducing recidivism as a major goal and thought that it could be achieved through intensive supervision and more parole services. A third of the parole staff also said that forming closer relationships with parolees would help reduce recidivism. Wards indicated in interviews that they generally liked the program and found it beneficial. Those who completed the institutional boot camp phase typically thought of it as an accomplishment in their lives. One ward pointed specifically to a "sense of accomplishment" and added that "most Youth Authority kids haven't done anything difficult, anything to have pride about. It is instilled here. This is the best thing that's happened to a lot of people." Analyses of selected items from the ward survey and prison environment inventory indicate that, compared to control wards, LEAD wards thought staff were more supportive and offered a more challenging program. They also felt they had developed more self-discipline and lived in a safer environment. Cost savings. LEAD's second mandated goal was cost savings. So far, data collected in this study have indicated that the program can "produce" differences in custody over the short term, but these differences are not yet sufficient to cover the additional costs of running an intensive program like LEAD. Cost savings can be achieved in two ways: by reducing lengths of confinement and by lowering rates of recidivism. Some experimental study data on confinement are available on this issue, but they are based on very short-term analyses. Data from the first seven platoons and their controls showed that 81 percent of LEAD wards, compared to 50 percent of control wards, were paroled before the end of the first year. On average, the LEAD wards were in custody 8.8 months, compared to 10.7 months for the control wards. Time in custody was measured from a ward's admission to a reception center through September 30, 1993, and included any custody on parole for any reason including residential drug treatment.7 Cost savings were then estimated by multiplying the average difference in custody for the two groups (10.7 minus 8.8 = 1.9) by the number of wards in less custody (1.9 x 103 = 195.7 months) to get months saved; and then multiplying that number by the average costs per month of custody (195.7 x $2,303 = $450,697) to get estimated savings. The actual savings would probably not be as large as this figure indicates, however, because differences of this type would occur "at the margin," where estimated costs are typically lower than at the base. Furthermore, calculating actual cost savings would require including additional costs of the program for both its parole and institutional phases. Over the course of the evaluation, many factors were identified that work against the goal of cost savings. One factor is a belief in the efficacy of institutional treatment for a very needy ward population. In staff interviews, for example, 8 of 21 institutional staff spontaneously mentioned LEAD's short length as a problem. Even parole staff commented on the program's short institutional program, with one agent remarking that LEAD "needs to be longer to sink in." A second factor is that wards are often transferred out of the program when they are disruptive. Of the 180 wards in the first 12 platoons, at least 30 were dropped from the program for disruption, assaultiveness, or lack of interest and effort. Such transfers, sometimes called "bus therapy," are essential to the program's overall health and safety, but it is difficult to know where to draw the line. A third factor might be termed the availability of special psychological treatment programs, to which five LEAD wards were transferred over the course of the first year. In contrast, not one ward from the control group was transferred to such programs, which suggests that LEAD's strong emphasis on treatment has unintentionally produced additional institutional treatment programming. Another factor is the parole board, which retains the power to deny parole following program completion. Despite a general understanding that the parole board would refer LEAD graduates to parole, one graduate was denied parole based on the need for further psychological treatment, and another graduate was denied because of poor program performance. Program efficacy was also hampered by parole revocations after offenders were released. Most of the revoked parolees were given one or more chances before revocation, but some simply exhausted the patience and ingenuity of parole agents. Wards themselves sometimes appeared to be willing partners in the extension of incarceration. They occasionally told interviewers that the program might be better if it were longer, and two demoted wards said the extra time was beneficial to them. One ward who resigned from LEAD was apparently partly motivated by his need for a place to stay longer. Finally, at the department level, lengths of incarceration in the CYA were beginning to decline as LEAD was being implemented, and if these trends continue, they will dilute any cost savings projected for LEAD. Electronic monitoring, for example, is now used as a method of early release to parole. Nine control wards in the study gained early release to parole by submitting to electronic monitoring. In contrast, no LEAD or former LEAD wards were released early with electronic monitoring and only three LEAD wards were placed on electronic monitoring to prolong their parole. In short, maintaining savings in the LEAD program will be an uphill struggle. At the levels of institutional and parole programming, the cost savings goal often seems to conflict with treatment-oriented goals. How Is the Program Working? What are the prospects or promising characteristics of LEAD, and what problems or limitations emerged during its first year of operation? Positive results. In response to an open-ended question about the positive characteristics of LEAD, most staff (76 percent) mentioned military structure. Wards also considered LEAD's military milieu and leadership training to be important features. One of the program's most successful training techniques is the rotation of platoon leadership among all the wards. Wards expressed a clear awareness of how this technique promotes responsibility, self-confidence, and teamwork. Many staff mentioned enriched staffing patterns as a positive characteristic. On average, LEAD parolees had about two more parole agent contacts per month than did control wards and about one more drug test per month. Moreover, wards more often thought that staff cared about them and wanted them to succeed. More than half the staff considered the educational program to be a benefit, and wards spontaneously mentioned their classes in interviews, noting, for example, the practicality of learning about parenting and budgeting and the confidence gained by giving speeches in front of class. Another area identified by staff and wards was the program's focus on physical training. An analysis of ward interview data found that wards liked physical training and drill and ceremony exercises more than anything else in the program. Wards said that they had never felt better or been in better physical shape and that physical conditioning improved their performance in other program activities. After making many observations and reviewing survey data and interviews, researchers concluded that LEAD living units were safer and had much less gang activity than the living units of the general ward population. On every survey measure of violence or gang activity, for example, LEAD ward responses indicated less violence and gang-related hostility than control ward responses. There were no significant differences, however, between the two groups in responses to questions regarding the existence of racial problems or the extent to which wards liked or cared about each other. Problems and limitations. A series of LEAD review meetings at Preston emphasized the need to develop more structured programming on parole. In particular, LEAD parole agents commented that paroled wards need more jobs, job training, assistance with schooling, and housing alternatives. Parolees described similar needs in interviews, saying that finding stable employment had been their most difficult problem. Parole agents placed over 20 percent of the LEAD parolees in alternative housing during the first year, and in some cases these placements appeared to have been critical to their short-term success. Another issue is the need for an adequate number of candidates for LEAD. An analysis of the overall CYA pool of LEAD candidates (based on age and commitment offense) suggests that many potential participants were not properly tapped by LEAD's screening process. The CYA parole branch also studied the issue of screening parole violators and reported that about 40 percent of current parole violators represent an eligible but untapped pool of candidates. A closely related issue is that of appropriate screening. A third of the institutional staff and over half of the parole agents felt that LEAD wards were too criminally sophisticated, immature, or unmotivated. Administrative steps have been taken to resolve the problems, but interviews with institutional and parole staff suggest that this issue has not been adequately addressed at the program level. The development of the program itself, however, has revealed that virtually all wards present difficult problems in terms of treatment or training, and staff will almost certainly have to develop more realistic expectations of ward accomplishment. However, this adjustment should be made without compromising the staff's ability to use high expectations as an important motivational tool. The most common, and most controversial, criticism among institutional staff was the need for shift rotation to alleviate burnout. LEAD's original design built three yearly shift rotations into the program, but when the union representing staff counselors successfully challenged these rotations, no other comparable staffing arrangements were made. Although a few staff members supported the union's position, nearly all agreed that the program was unusually exhausting and that variations in assignment could provide some relief. Very few staff felt that emotional abuse of wards was a problem at LEAD. They acknowledged the confrontational nature of their military training techniques but tended not to find them excessive or abusive. However, ward responses to questions on emotional abuse presented a different picture. Both LEAD and control wards were likely to feel that some staff got carried away with their power over wards, although neither group typically reported being physically threatened by staff. (The LEAD group was, however, more likely to report physical threats than the control group.) About half of each group reported being embarrassed or humiliated in front of other wards; on the other hand, neither group felt that their program was too harsh or punishing. This information, as well as intimations of degrading techniques used by a few staff, suggest that some of LEAD's confrontational training methods should be toned down. A number of staff criticisms directly or indirectly spoke to the need for administrative refinements at the level of institutional program delivery. About 24 percent of the staff directly mentioned administrative problems, and slightly higher proportions mentioned staff communication problems and conflicts regarding disciplinary procedures. Recent review meetings isolated a leadership-related problem as well-- the program's need for a shared vision and shared goals. For example, some staff mentioned conflicts with management over LEAD's level of ward attrition. If everyone shared the cost savings goal, however, there would be more consensus on the need to contain ward attrition. LEAD functions administratively in much the same way that other living units function, with two levels of staff supervision and standard procedures for most aspects of the program (e.g., grievance and disciplinary procedures). At LEAD, however, daily situations must be handled more creatively to motivate wards and keep them to the program's strict time commitments, and to ensure that wards are not lost to other programs. LEAD supervisors must creatively resolve issues that other programs do not have to deal with, such as being fair with wards in disciplinary situations while at the same time maintaining the program's momentum. Put differently, a certain inconsistency is required in LEAD to handle wards individually and constructively, according to each ward's needs and abilities. Staff also expressed the need to have supervisors on hand more of the time. This need stems from another area that reflects the differences between LEAD and other CYA programs. LEAD thrives on relationships rather than on established procedures and requires an administrative structure somewhat different from that used by other programs. This chapter's discussion of cost savings pointed to the many factors that work to lengthen incarceration time in CYA. Given the instability of ward motivations for a program such as LEAD, this issue will likely be a perennial problem. However, the program's 28-percent attrition rate during the first year may not be unduly high as long as it is kept in check. Because most TAC officers do not have prior military experience and because new staff are hired from time to time, periodic military training is essential. A degree of slackness in current military procedures and drill and ceremony exercises was noted by about 20 percent of staff interviewees. Moreover, ward interviewees typically requested even more disciplined military procedures. Although they were critical of demeaning or humiliating tactics, wards were almost uniformly in favor of tight structure and discipline. Finally, about one-third of parole agents interviewed said that their case counts were too high for intensive supervision, and one-fifth of institutional staff complained about staffing shortages due to limited backup during staff absences and replacement. Staff supervision levels may need to be monitored more closely to ensure that enriched staffing patterns--a hallmark of this program--are continued. Conclusions Although LEAD had some notable accomplishments during its first year, the program will be difficult to sustain if some of its inherent problems and limitations are not resolved. This report concludes with five suggestions for program continuation and refinement: o Work toward some variation or modification of TAC officer assignments to address staff burnout, a serious problem that will sap the vitality of LEAD if left unattended. o Establish periodic program review processes and staff training, as well as CYA departmental administrative support for boot camp staff leaders. Coordinate program services among CYA branches. Experiment with program supervisory patterns and staff leadership strategies. o Continue work on a transitional program in the community that, if possible, includes employment. There is simply overwhelming evidence that wards released to the community lack the opportunities and overall skills required for long-term success. o Integrate the program's cost savings goal with its treatment goals. While requiring a great deal of creativity, meeting both these needs may be possible through greater reliance on the treatment continuum (from institution to parole), more extensive use of relapse management strategies, and possibly policy revisions that permit LEAD parolees to return to the program for brief periods of restructuring. To lower attrition, for example, referrals to parole could in some cases be made contingent on more rigorously defined treatment in the community. Linking parole to more treatment in this way might also provide more assistance to wards who need several chances to succeed in community supervision. Staff who know and care about such wards might be well poised to help them restructure their plans. o Cultivate and maintain positive features of the program. Prominent among these features are the program's safe and constructive environment, and the leadership rotation and TAC mentoring strategies provided in the military officer training model. While the military milieu has something to offer by way of disciplinary structure, it is the positive and nurturing relationships of the officer training model that stand to change wards in a positive and lasting way. Notes 1. Later that year, the committee drafted a budget change proposal, which is the procedure for modifying activities funded by the State, but before it was approved, the CYA participated in discussions with the National Council on Crime and Delinquency, State Senator Robert Presley's office, YACA, the State Assembly Ways and Means Committee (chaired by Assemblyman John Vasconcellos), and the State Youthful Offender Parole Board. At issue were parole board policies, particularly increased lengths of juvenile incarceration and the need for new options for youthful offenders. A compromise was eventually worked out. 2. Program effects will be determined in the impact evaluation by measuring differences over a 30-month period on lengths of incarceration (the primary measure of cost savings) and on subsequent arrests (the primary measure of recidivism). Researchers expect LEAD wards to be incarcerated for shorter periods of time, on average, than control wards, and that this difference will not be offset by more arrests of LEAD parolees over equivalent time periods. 3. Kevin Wright, "Developing the Prison Environment Inventory," Journal of Research in Crime and Delinquency, 22 (1985):257-277. 4. Control wards were generally less accessible because they had been sent to CYA institutions and camps around the State. Staff at those institutions were not asked to administer surveys to ensure greater objectivity. Thus, it was not possible to administer as many surveys to control wards. There were no selection criteria for surveying control wards other than their geographical accessibility. 5. Data on parole board approvals were taken from the program's northern California reception center in Sacramento because such information was not available from the southern reception center or for any parole violators. 6. Jackson Toby, and Frank Pearson, "Boot Camp for Juvenile Offenders: Constructive Intervention," unpublished paper, 1992. 7. It should be noted that the first 7 platoons each had 15 wards, for a total of 105 wards. The experimental study groups through the seventh platoon each had only 103 wards, however, due to screening errors that forced the exclusion of 2 wards from each study group (that is, experimentals and controls).