FORUM on Corrections Research

Situating risk assessment in the reintegration potential framework

by Larry Motiuk and Ralph Serin¹
Research Branch, Correctional Service of Canada

Over recent decades, criminal justice researchers have highlighted conceptual and methodological advances in risk assessment technology, distinguishing between “statistical” and “clinical” prediction, and between “static” and “dynamic” factors. Throughout the 1970s and 1980s, because of the relatively low predictive accuracy of the available risk assessment instruments, researchers focused on false positives — that is, incorrect placement or release of offenders who subsequently succeeded. By the late 1980s and early 1990s, improvements in the accuracy of various risk scales had shifted concerns to false negatives — that is, prematurely released offenders who subsequently failed. This latter research was fueled by public concerns for community safety and new legislation developed to allay those concerns.

Today, because society has a low tolerance for false negatives, decision makers are overly concerned about failures, and this excessive concern may be impeding reintegration efforts. Correctional decision errors can be minimized by improving systematic risk/needs assessment strategies and standards of practice. Such a strategy should help us to identify and release offenders with good potential for successful reintegration. Offenders with high reintegration potential might be defined as low-risk offenders or moderate-risk offenders who are manageable in the community with prescriptive intervention and appropriate supervision.

This article identifies several issues relating to reintegration potential that apply to all correctional jurisdictions to help decision makers determine where to make judicious changes to increase reintegration. This approach should also ensure that offenders are released safely and in a timely manner, consistent with each jurisdiction’s mission statement. This article argues that risk assessment can be bound into reintegration efforts in a way that minimizes decision errors. It also recognizes that jurisdictions may vary in their tolerance of false negatives, particularly with specific types of offenders or specific types of failure.

Reintegration

A recent report by the Canadian Centre for Justice Statistics presents a one-day snapshot of provincial and federal offenders.² Although the federal jurisdiction differs from provincial and territorial jurisdictions in the risk/needs profiles of their inmates and the proportion of violent offenders in their correctional population, there is sufficient range to suggest developing differentiated strategies according to risk/need factors. Reintegration encompasses a broad range of decisions intended to: place offenders in the least restrictive setting possible, grant temporary absence or conditional release, and invoke suspension or revocation of conditional release when necessary. Each of these correctional practices, however, is also an index of reintegration success, and significantly affect an offender’s movement through his or her sentence. For instance, Luciani, Motiuk and Nafekh³ reviewed case management decisions regarding placement in minimum security using different cutoffs for the Custody Rating Scale (CRS), an initial security classification tool used by the Correctional Service of Canada. Modifying CRS cutoffs reduced the number of days served, with minimal impact on the number of escapes. Specifically, offenders placed directly in minimum security spent fewer days incarcerated than offenders placed directly in medium security. Offenders who went straight to minimum security also benefited from a higher rate of parole grants.

“Decarceration” has been defined as the selection of offenders suitable for early release.⁴ Identification of suitable offenders raises questions about selection criteria, risk scales, and using treatment and program information in correctional and parole decision making. Although this emphasis on early release is central to reintegration, other factors may also influence the likelihood of decisions to grant discretionary release. For instance, Motiuk and Belcourt⁵ found that an offender who had received a temporary absence was highly likely to be subsequently granted parole. This suggests gains in one area of reintegration may yield increases in another.

Risk and security classification instruments

Most jurisdictions have specific strategies for assessing offender risk. Although an overview of these approaches is beyond the scope of this paper, such work is in progress.⁶ Evidence indicates that the various risk scales used in Canadian corrections are highly intercorrelated — that is, the choice of a specific risk assessment instrument is mainly an operational issue, since none has been proved to be markedly superior. In fact, a thorough review of the offender assessment literature strongly encourages assessment strategies that use several instruments.⁷

Other issues to consider from a systems perspective are content and process: Do the instruments reflect sufficient content to meet the guidelines for risk assessment?⁸ In the case of objective classification instruments, it has been proposed that they be used as anchors, and that the use of case-specific information be increased. For process, it is important that staff who prepare reports for decision makers clearly articulate how they have integrated risk factors and estimates of recidivism into their assessments. Staff in several jurisdictions have received comprehensive risk assessment training to ensure that they understand that risk assessment scales are valuable but no substitute for sound correctional decision making.

Offender risk/needs assessments are done to inform staff about an offender’s requirements and criminogenic needs. This allows decisions to be made regarding specific treatment targets and the appropriate intensity and mode of intervention for an offender. Recently, a systems approach to community-based offender risk management led to the development and implementation of risk/needs instruments such as the Level of Service Inventory — Ontario Revised (LSI-OR) in Ontario, and the Community Risk/Needs Management Scale (CRNMS) in federal corrections. These dynamic risk assessment tools produce indications of reassessment needs as well as case needs. The LSI-OR and the CRNMS have been demonstrated to be correlated to recidivism. Such an assessment strategy can be incorporated into guidelines for preventive supervision that would yield gains in reintegration — that is, monitoring, supervision and intervention could be gradually increased to coincide with an offender’s time of increased risk. Even modest reductions in suspensions and revocations would increase the number of offenders safely serving their sentences in the community.

One final risk assessment issue is the need to address decision errors. Currently, many decision makers use risk assessment tools to estimate the offender’s risk of reoffending. These scales can also be used to inform staff about the probability of types of failure, and the decision errors most associated with the various cutoff scores on the instrument. This method permits staff to consider the likelihood and costs of recidivism. For example, some events, especially sexual reoffences, are relatively rare, but have high costs when they occur. Including decision errors in the risk assessment tool allows consideration of multiple cutoffs to minimize both false positives and false negatives. It is crucial that base rates be known for various types of failure, various types of release (e.g., discretionary, expiration of sentence) and various settings, regions and security levels.

Meeting the full potential for reintegration of offenders

Several indicators show that emphasizing false negatives may be impeding reintegration efforts. By changing the way the CRS is applied, transfers to minimum security and parole grants could be increased, without an increase in escapes. This would minimize false positives, which lead to unnecessary higher-security incarceration, without increasing false negatives, which lead to escapes. Another indicator relates to the results of examination of temporary absence program participation and release of offenders. The number of offenders granted temporary absences has diminished markedly over the past several years, and it is not clear whether this reduction is related to policy changes or concern about failure. When one considers that successful temporary absences predict successful release on parole, then reductions in temporary absence grants tend to reduce parole grants. This situation suggests considerable opportunity for reintegration gains.

What intervention could do

It is generally accepted that effective correctional intervention is important to the reduction of offender risk. Consistent with this view, a concerted effort is required to ensure that correctional programs and intervention are linked to reintegration efforts. If this is to occur, however, core programs must be continuously evaluated and the changes indicated by the evaluations must be integrated. Furthermore, the accreditation of correctional programs must ensure that programs meet standards for content and delivery. Finally, a mechanism is required to incorporate treatment information into decisions regarding reintegration potential.⁹

Application to various jurisdictions

Although the application of approaches noted in this article may vary from jurisdiction to jurisdiction, several themes merit attention. Objective security classification is desirable for good correctional management and to demonstrate that decisions are informed and rationalized. Systematic and objective classification may also reduce decision errors regarding the security level at which offenders are incarcerated. Not only is over-classification more expensive, but it also limits offenders’ release opportunities.

In correctional systems, temporary absence programs are inextricably linked to subsequent discretionary release. Careful attention to temporary absence programs should yield gains in other reintegration areas. Similarly, overly stringent parole-suspension practices will concentrate the offender population in prison. Even a modest reduction in suspensions and probation breaches through better community management might yield substantial gains in the population of offenders living safely in the community.

Correctional programs remain an important risk-reduction strategy, but speculation continues on the best way to integrate treatment into reintegration decisions. We do know, however, that community-based programs tend to be more efficacious, which suggests that they should be used more.

Offence-based guidelines for security classification or discretionary release are unlikely to minimize decision errors as effectively as statistically anchored and case-differential risk/needs assessment strategies do. All jurisdictions must make release decisions before expiration of sentence, preferably through operational applications of research findings (e.g., LSI-OR, Statistical Information on Recidivism Scale). Further, standards of practice are available for guidelines on completing risk assessments,¹⁰ and should be incorporated into correctional practices.

Summary

Incorporating systematic risk/needs assessments and principles into a reintegration potential framework seems both legitimate and potentially fruitful. The process and content reflected in traditional risk/needs assessments are compatible with the goals of reintegration, but require refocusing. This can be done only if correctional staff and decision makers consider the issue of decision errors more carefully. Current research highlights several areas where gains can be made, and indicates that reintegration efforts in one area may produce gains in several other areas. If such gain is, in fact, exponential, then only modest gains might be preferred at the initial stage. This would allow researchers to evaluate the impact of changes in procedures, guidelines and cutoffs designed to improve reintegration efforts. This policy will also appeal to those who argue that reintegration is not indicated for resistant and high-risk offenders. The release of offenders with high reintegration potential is consistent with recent legislation targeting high-risk offenders.

---

1. 340 Laurier Avenue West, Ottawa, Ontario K1A 0P9.

2. F. J. Porporino, National Overview, A Snapshot Profile of All Inmates On-register in Federal and Provincial/Territorial Adult Correctional Facilities on Saturday, October 5th , 1996 (In press).

3. F. P. Luciani., L. L. Motiuk and M. Nafekh, An Operational Review of the Custody Rating Scale: Reliability, Validity and Practical Utility (Ottawa: Correctional Service of Canada, 1996).

4. Porporino,National Overview.

5. L. L. Motiuk and R. Belcourt, Temporary Absence Program Participation and the Release of Federal Offenders (Ottawa: Correctional Service of Canada, 1996).

6. S. L. Brown and R. L. Serin, A Consumer’s Guide to Risk Assessment: A Systematic Review (Manuscript in press, 1998).

7. T. A. Leis, L. L. Motiuk and J. R. P. Ogloff, Forensic Psychology: Policy and Practice in Corrections (Ottawa: Correctional Service of Canada, 1995).

8. Leis, Motiuk and Ogloff, Forensic Psychology.

9. R. C. Serin and S. Kennedy, Treatment Readiness and Responsivity: Contributing to Effective Correctional Programming (Ottawa: Correctional Service of Canada, 1997).

10. Leis, Motiuk and Ogloff, Forensic Psychology.