Journal Issue: Juvenile Justice Volume 18 Number 2 Fall 2008
Current Practices for Assessing Risk and Amenability in Juvenile Justice
In several areas of juvenile justice processing, more structured methods for screening and assessment are beginning to be put into practice. Most commonly, locales have devised risk assessment instruments that tally items and calculate an overall risk score regarding the appropriateness of institutional placement or detention, with some locales even requiring such an instrument.26 Items and scores are derived for a specific locale, based on a combination of local data about re-arrest or re-institutionalization and local values regarding the acceptable level of community risk from juvenile crime.27 The logic behind these measures is simple: the more risk factors endorsed, the more likely the juvenile is to re-offend, and therefore the more justification there is to detain or place the juvenile in an institution. Overall, actuarial instruments of this sort have been shown to be moderately predictive of re-arrests.28
Developing and testing these sorts of tools for specific points in juvenile justice processing, however, can be costly.29 Therefore, many locales simply adopt tools used by a comparable city or state. Although such an approach almost inevitably increases consistency within the court, it does not achieve the full payoff of implementing structured judgment approaches. A locale's failure to develop local standards undercuts the considerable potential gain to be had from using its own data and a consensus process to develop instruments tailored to its juvenile justice system.30
An actuarial decision-making system can often be developed as part of a broad-based reform of a community's juvenile justice system.31 Having stakeholders identify appropriate risk indicators and choose thresholds for particular actions, such as detention, can increase collaboration and produce a shared sense of mission. Carefully introducing actuarial risk assessments can also effectively address the continuing issue of disproportionate minority confinement in a community.32 Regardless of the strength of the underlying causes of minority overrepresentation (a subject covered in the article by Alex Piquero in this volume), a community can implement detention standards fairly and objectively using a structured instrument, thereby affecting the rates of minority adolescents locked up in the process.
One comprehensive approach to developing assessment instruments, the combined risk-and-need approach, goes beyond calculating a single score of how likely a juvenile might be to re-offend to include an assessment of protective factors or treatment needs.33 Rather than treating risk as a stable characteristic of the adolescent, the risk-and-need approach assumes that risk might be lowered by particular interventions or by careful monitoring in the community. An adolescent with a drug or alcohol problem may be at higher risk for re-offending but may also be a good candidate for positive community adjustment if that problem can be addressed effectively. The risk-and-need assessment strategy goes beyond simply sorting adolescents into lower or higher risk groups by providing information about how to select interventions to reduce risk. It allows the evaluator to use results from the assessment to identify treatment interventions specific to an individual's needs rather than just to offer a binary decision about the need for incarceration.
Although structured decision making and risk-and-need assessment strategies are not yet used widely in the juvenile justice system, they are beginning to be put into practice. Mental health clinicians, for example, are increasingly likely to integrate structured clinical judgment into their practice when doing assessments for courts.34 Used properly, these approaches can increase the scientific soundness and practical utility of their assessments.35 Numerous interview and rating systems, such as the Early Assessment Risk List (EARLS) or the Structured Assessment of Violence Risk in Youth (SAVRY), now provide systematic methods for assessing the future risk of violence with acceptable predictive accuracy.36 Several brief, self-report measures related to risk of future violence, such as the Antisocial Process Screening Device, are also predictive of antisocial behavior and likelihood of successful involvement in treatment.37 Although there is no self-report measure of treatment amenability, there are measures of motivation to change, a key component of treatment amenability. Such measures as the University of Rhode Island Change Assessment (URICA) and Treatment Motivation Questionnaire (TMQ) have variable predictive ability, and few studies have examined motivation to change in adolescents.38
We know of one interview-based rating system for assessing treatment amenability: the Risk, Sophistication, and Treatment Instrument (RST-I).39 Preliminary data show that it is predictive of important juvenile justice and clinical outcomes, such as treatment compliance.40 It is, however, relatively new, and more research is needed on its predictive validity, especially in comparison with other risk-and-need measures.
Research on the utility and validity of risk-and-need instruments in court systems in general remains rather limited. Only a few structured risk-and-need instruments have undergone careful recalibrations and repeated validity testing.41 The benefit of introducing these systems at this point may simply be to increase the uniformity of decision making at certain stages in juvenile justice processing, but that is still a plus: more consistent application of decision-making rules has been shown to increase overall accuracy even if the model used is less than optimal.42 Much of the value of using these risk-and-need systems may lie in the fact that they are used at all, rather than in the exact specificities or sensitivities of their algorithms.43
One important limit of some of these highly structured instruments that must be addressed before they can be adopted widely is that they may not be equally valid across different racial and ethnic groups. Although there is some evidence that comprehensive risk-and-need assessments may predict outcomes equally well across gender and ethnicity, screening instruments may not do as well. For example, brief risk measures predict recidivism better for whites and males than for blacks and females.44 Clinical assessments too appear to predict recidivism better for white samples than for ethnically diverse samples.45 Two papers in this volume discuss these matters in more detail. Elizabeth Cauffman addresses the assessment of female offenders, and Alex Piquero addresses the assessment of youths from diverse ethnic backgrounds.