Reentry and supervision: one is impossible without the other.
Supervision, like reentry, is focused on helping the offender to become a productive citizen; supervision usually consists of the period after release through the period of correctional control. Attention to the issues related to initial stabilization in the community generally can benefit from the same set of practices that have been defined in proactive supervision, or a supervision model that is focused on facilitating offender change. This model of supervision, described in Tools of the Trade: A Guide to Incorporating Science Into Practice, (1) essentially provides a framework for supervision that reinforces the same goals ascribed to reentry: offender change and reduction in recidivism.
Reframing Supervision To Be an Intervention
The article Supervision--Exploring the Dimensions of Effectiveness (2) discusses how face-to-face contacts can become an intervention through the use of a supervision process with defined steps and goals. Such a supervision process would consist of moving the offender through different stages of supervision: engagement, early change and sustained change. Engagement includes the process of assessment, development of a case plan that targets the offender's needs based on assessments, and clarification of expectations (e.g., program requirements, sanctions, incentives). Early change refers to the use of formal controls (e.g., supervision contacts, treatment services) that target the offender's needs and the use of informal controls to begin the process of building pro-social networks in the community. The last phase is sustaining the progress through the development of the pro-social networks and involvement in step-down services. Underlying this is an environment that is supportive of offender change and uses communication strategies that support the change process. The importance of this model is that the goal is to alter supervision to be consistent with where offenders are in their change process.
Results From Proactive Community Supervision in Maryland
The model illustrated in Figure 1 was implemented in Maryland to alter the context of supervision. The Proactive Community Supervision (PCS) project received specialized funding to reduce the caseload sizes for intensive supervision from 100 to 55 in the four pilot areas. The evaluation study used an individual match design that compared the outcomes of 548 offenders--274 randomly selected offenders supervised in PCS offices and 274 matched offenders in offices that use the traditional supervision model.
[FIGURE 1 OMITTED]
The purpose of the study was to determine whether the implementation of the PCS core concepts had an impact on offender outcomes, particularly on rates of rearrest and warrants for violation of probation/parole. Selection criteria for the study included:
* Started supervision during calendar year 2004;
* Served a minimum of six months on supervision (to ensure exposure to the core of the PCS process);
* Rated high risk for recidivism using the Maryland Division of Parole and Probation's risk screener; (3) and
* On active parole, probation or mandatory release.
The number of cases selected from each office was proportionate to the total number of PCS cases entering supervision at each pilot site. Each individual offender in a PCS office was matched to an individual offender in a non-PCS office on the following criteria: gender, race, age, type of supervision and category of instant offense.
The PCS model adopts the tenets of science-based research to reframe supervision services for offenders. The emphasis of the model is on the nature and intent of the contacts between the offenders and their probation/parole agents. The contacts are central for the agent to facilitate change in three ways: to engage the offender in a change process that focuses on obtaining prosocial skills; to begin the offender change process by using supervision tools and treatment interventions to address criminogenic traits; and to assist the offender in desistance and sustaining change through positive involvement with community support networks such as family/associates, mentors and civic associations. Simply put, the agent's role has been broadened from merely monitoring the offender to facilitating the offender's involvement in pro-social activities that focus on building skills to be productive in society. The behavioral manager role encompasses both the law enforcement and social work skills that are needed in protecting the public. (4)
The PCS model reflects the underlying behavioral management theoretical model and incorporates its main ingredients in the following components of the process:
* Use of the Level of Service Inventory-Revisited (LSI-R) instrument to assess the criminogenic risk and need factors affecting the offender's involvement in criminal behavior;
* Development of a case plan that is responsive to the criminogenic traits identified in the assessment process, which includes the LSI-R and other objective information, and that addresses goals that are specific to the offender's criminogenic traits. The focus is on identifying a purpose for the supervision plan that is specific to the goals and emphasizes desistence;
* Referral to the appropriate array of treatment, educational, vocational and other services to assist the offender in developing new skills to be a productive citizen;
* Use of the supervision process to help the offender learn about the triggers (e.g., people, places or situations) that affect involvement in criminal behavior;
* Use of incentives and sanctions to shape offender behaviors;
* Timely communication with the offender to review progress on the case plan and achievement of supervision goals; and
* Emphasis on desistence from criminal lifestyle and conduct.
Table 1 compares the PCS and non-PCS groups at the time of intake according to selected characteristics. The matching criteria were intended to define key behavioral characteristics linked to recidivism. There was also equivalency between PCS and non-PCS individuals on variables that were not used in the individualized match. (5) For employment status, marital status, education level and prior arrest history, there are no statistically significant differences between the PCS and non-PCS groups.
Table 1. Group Characteristics for PCS and Non-PCS at Time of Intake Population Group Characteristics PCS Non-PCS In Each Group Gender -- Male 82.8% 82.8% 274 Race -- Black 85.0% 85.0% 274 Employment Status -- Unemployed 61.1% 69.5% 167 Age -- Over 30 53.3% 54.0% 274 Marital Status -- Single 82.5% 78.6% 154 Education Categ ories -- At Least High 62.1% 60.6% 132 School Diploma Mean Number of Arrests Prior to 7.41 6.83 259 Supervision Intake Supervision -- Probation 88.0% 87.6% 274 Supervision -- Parole/Mandatory Release 12.0% 12.4% 274
The first question was to determine the degree to which PCS was implemented. The researchers focused on whether the following actions were taken: completion of the LSI-R, identification of the offender's typology and triggers of criminal behavior, development of a case plan, and implementation of the case plan.
The first core step is the completion of a LSI-R-based assessment. The supervision plan and progress monitoring cannot occur without substantive information from the LSI-R or another intensive assessment tool; otherwise, the probation/parole agent is using a "generic" model of supervision instead of focusing the supervision on addressing the criminogenic traits of the individual offender. In this cohort, 70 percent of the offenders had an LSI-R. The average LSI-R score was 15.6, with a standard deviation of 7.9. This average score corresponds to a rating of medium risk in the community, according to the national standards for the LSI-R. (6) About 30 percent of the offenders scored under 10, which indicates low risk; 29 percent scored 11 to 17, indicating medium risk; 28 percent scored 18 to 25, which is moderate to high risk; and 13 percent scored 26 or more, reflecting high risk. (7)
Table 2 presents a comparison of the implementation of the PCS model across the four sites in terms of the responsibilities assigned and completed. The unidtified group was considered to approximate traditional supervision. The PCS message of accomplishing supervision goals appears to be translating into more definitive actions taken by both agents and offenders. Agents and offenders develop a case plan that includes a list of responsibilities. The number of responsibilities the agents and offenders specify and then take action on signify the degree to which the PCS model impacts the productivity of supervision. Offenders generally take action on 90 percent to 200 percent more responsibilities in the case plan than traditional supervision offenders.
Table 2. Development and Implementation of Responsibilities in Case Plans Typology Drug Typology Disassociated Involved DV/MH Violent Unidentified (n=71) (n=36) Sex * (n=30) (n=121) Mean Number of 3.4 6.4 4.3 2.9 0.5 Agent Responsibilities in Case Plan Mean Number of 3.8 5.0 5.0 4.0 2.0 Offender Responsibilities in Case Plan Mean Number of 2.3 4.2 4.2 2.5 0.3 Responsibilities On Which Agent Took Action Mean Number of 1.7 2.2 2.2 2.0 0.9 Responsibilies On Which Offender Took Action * Combines domestic violence (DV), mental health (MH) and sex offender (Sex) typologies
The Impact of PCS: Comparing Outcomes
The PCS and non-PCS groups were compared according to three outcomes: positive drug tests, new arrests and requests for warrants (proxy for technical violation or non-criminal behavior acts that are violations of orders of release). The analysis plan uses two main strategies to determine whether the PCS process has differential outcomes from the non-PCS process. Bivariate and logistic regression models were used to determine whether differences between the PCS and non-PCS groups were statistically significant. The logistic regression models were used to test for statistical significance controlling for variables that were not used in the case selection process that might be relevant to outcomes, such as the number of days under supervision and the offender's criminal history (number of prior arrests).
Drug Test Results. Court-ordered drug-testing conditions were in place for 48.9 percent of the PCS offenders and 62.8 percent of the non-PCS offenders. Statistical tests indicated that there were no differences between the two groups in terms of failure to appear for testing or positive urines.
Table 3 reports the findings from a logistic regression that determined participation in PCS had no effect on the likelihood of having positive urine, controlling for the influence of time on supervision and prior arrests. The results of these analyses suggest that PCS did not have an effect on the drug consumption patterns of those in the sample with a court-ordered mandate for urinalysis. (8)
Rearrest Rates. The PCS group was statistically less likely to be arrested than the non-PCS group; 40.9 percent of the non-PCS group was arrested compared with 32.1 percent of the PCS group. Individuals in the PCS group had a mean number of arrests that was 30.7 percent lower than the non-PCS group.
Two logistic regression models were developed to assess the impact of PCS on rearrests. In these models, researchers controlled for length of time on supervision and prior arrests, since these were not part of the original selection criteria. The results in Table 3 illustrate that offenders participating in PCS were significantly less likely to be rearrested during the study period. The odds ratio of 0.62 means PCS offenders have 38.3 percent less chance of being rear-rested for new criminal behavior than the non-PCS group.
Table 3. Results for the PCS vs. Non-PCS Cases on Key Outcomes: Drug Testing Results, Rearrests and Warrants for Violation of Probation/Parole PCS Non-PCS Odds Ratio Drug Testing Results Positive Urine 22.8% 25.4% 1.12 Failure to Appear for Urinalysis 50.0% 53.6% Rearrest for New Crime New Arrest * 32.1% 40.9% 0.62 Mean Number of New Arrests * 0.45 0.65 Warrants for Violation Of Probation/Parole Warrant Filed * 20.1% 29.2% 0.60 * p < 0.05 Note: For drug test results, the sample size was 94 PCS offenders and 106 non-PCS offenders. For failure to appear, the sample size is 99 PCS and 121 non-PCS. For all other measures, the sample size was 274 PCS and 274 non-PCS.
Warrant Filed for Violation of Parole/Probation. As shown in Table 3, 20.1 percent of the PCS cases and 29.2 percent of the non-PCS cases had a warrant filed, a statistically significant difference (p < 0.05). Supervision agents for PCS cases filed a warrant when the offender had a median of 17 noncompliant behaviors for missing appointments, failing to inform the agent of location, failing to comply with court-ordered conditions of supervision (such as paying fines or attending drug treatment), testing positive for drugs, absconding from supervision and failing to report for drug testing. Overall, the fewer the number of misbehaviors, the less likely the agent was to file a warrant. (9)
Findings from the logistic regression reveal an odds ratio of 0.60, which means that participation in PCS resulted in a 38 percent reduction in the probability of a warrant being filed for technical violations. Further, this finding was statistically significant, controlling for the number of days on supervision and the number of arrests prior to the current sentence.
What Has Been Learned?
Reentry is the process of helping offenders identify criminal drivers that are likely to increase the odds that they will engage in criminal acts after release and to stabilize them after release from incarceration. The proactive supervision model illustrates that sound use of assessment information can result in supervision plans that assist offenders in the change process. The supervision model and reentry are similar in that the goal is to facilitate the offender in becoming a law-abiding citizen. The use of behavioral-management strategies (i.e., valid risk and needs tools, case plans and compliance management strategies) focuses on resocializing the offender to be a productive member of the community. This strategy for supervision is theoretically sound, as it is based on the literature on behavioral change and conditioning as well as organizational change. PCS results illustrated that a goal-focused reentry and supervision plan can reduce the likelihood of arrest and technical violations that result in warrants. Even more important, the results suggest that more productive, goal-centered contact between the offender and agent can have a positive outcome.
PCS encompasses the key components of effective interventions: targeting high-risk behaviors, focusing on key criminogenic traits, managing progress, using place-based strategies and engaging the organization in change. The results illustrate the following:
* An assessment instrument, in this case the LSI-R, can be used to develop case plans that are meaningful to offenders;
* Case plans can target different goals based on criminogenic traits;
* Offenders will take responsibility for conditions in their case plan if they understand the rationale;
* Offender noncompliance can be managed in a way to reduce warrants for technical violations but ensure public safety; and
* Adherence to case plans can reduce rearrest and technical violation rates.
All of this is what the reentry movement is trying to accomplish--identifying a responsive model to engage theoffender in the change process. This model of supervision does that, and it provides a framework that serves to facilitate offender change.
Author's Note: This project is funded by the Maryland Division of Parole and Probation. All opinions in this article are those of the author and do not represent the position of the agency.
(1) Taxman, F.S., E.S. Shepardson and J.M. Byrne. 2004. Tools of the trade: A guide to incorporating science into practice. Washington, D.C.: U.S. Department of Justice, National Institute of Corrections. Available at www.nicic.org/Library/020095.
(2) Taxman, F.S. 2002. Supervision--Exploring the Dimensions of Effectiveness. Federal Probation, 66(2) 14-27.
(3) The intent was to select individuals whom agency agents had classified during intake as requiring intensive supervision. Individuals classified as intensive at their most recent classification were selected. Note that 79 individuals included in the study, who were classified as standard at the beginning of their parole/probation sentence, were switched to intensive supervision at some later, undetermined point (44 PCS and 35 non-PCS).
(4) Taxman, F.S., E.S. Shepardson and J.M. Byrne, 2004.
(5) Since these other characteristic variables were not used for selection into the study, there were missing data for many of the individuals. In addition to excluding PCS cases with missing data from the comparison in Table 3, the researchers excluded PCS cases for which the non-PCS match had missing data.
(6) Andrews, D. A., James L. Bonta and J. Stephen Wormith. 2004. Level of service/case management inventory: An offender assessment system. North Tonawanda, N.Y.: Multi-Health Systems.
(7) Offenders who scored in the highest category of risk on the LSI-R had the highest rate of rearrest (45 percent) during the observation period, while those who scored in the low risk category had the lowest rate of rearrest (19.3 percent). These findings validate the LSI-R instrument.
(8) All of the logistic regression models were run adding in the covariate of the offender having a court-ordered condition for drug testing. The assignment of a court-ordered condition for drug testing did not explain any variance in the arrest and warrant for violation of probation/parole models, and the covariate was not statistically significant.
(9) The percentages for the non-PCS locations ranged from 16 percent to 40 percent.
Faye S. Taxman, Ph.D., is a professor in the Virginia Commonwealth University Wilder School of Government and Public Affairs. She would like to acknowledge contributions by Judith Sachwald, director of the Maryland Division of Probation and Parole, and Ernest Eley.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||CT FEATURE|
|Author:||Taxman, Faye S.|
|Date:||Apr 1, 2007|
|Previous Article:||Lessons learned from SVORI: program director perspectives on implementing reentry programming.|
|Next Article:||Membership news.|