FORUM on Corrections Research

The Youth Management Assessment: Assessment of young offenders at risk of serious reoffending

Society has recently shifted its concern about traditional "juvenile delinquency" to a more specific focus on the serious violence committed by a small number of youthful offenders. The media⁽²⁾ has fuelled this shift to the point that concern about violent youth crime is now shared by correctional professionals, politicians, policy makers and the general public. Whether or not increases in the rate of violent youth crime during the past decade⁽³⁾ reflect changes in youth behaviour or in law enforcement practice, it commands our attention - both in terms of correctional policy and practice.

This article examines a recent attempt to address this problem - the development and implementation of an assessment instrument designed to identify young offenders at risk of serious reoffending - by illustrating the instrument's theoretical foundations and structure, and by demonstrating its utility. Background Remarkably the prediction of violent youthful behaviour has not received widespread attention. However, Ontario corrections front-line staff (particularly those in open-custody facilities) needed a mechanism to assess the potential of a youthful offender to cause serious harm to another offender, to a staff member or to himself or herself.

Such an instrument had to be able to capture the wide range of violent antisocial behaviour committed by adolescents. It had to provide structure, a common methodology and guidelines for the user, while permitting staff discretion in determining the threat imposed by each young offender.

The assessment procedure also had to recognize that an offender's status may improve or deteriorate over time. Therefore, the instrument had to be capable of adaptation and had to include dynamic risk factors that would reflect any change in offender risk level.

Finally, the prediction of serious violent behaviour is difficult because the behaviour occurs infrequently Although this "base rate problem" might suggest a clinical assessment approach to the prediction of these behaviours, empirical evidence indicates that a statistical approach is usually superior.⁽⁴⁾ Clinical assessment should not, however, be completely ignored.⁽⁵⁾ The new instrument, therefore, combined the two approaches. The Youth Management Assessment The Youth Management Assessment was designed to perform a number of functions:

• to provide a standardized overall perspective of youthful offenders that accurately indicates their risk of causing serious personal injury;
• to reduce the risk of potential harm to staff and other offenders, and to increase the safety of all residences and institutions that accommodate young offenders;
• to provide a standardized language for communicating about a young offender with other staff, settings or agencies;
• to assist staff in preparing young offender predisposition reports and recommendations and, in so doing, to improve the overall appropriateness of young offender sentences;
• to provide security, management and treatment direction for young offenders at risk of serious self-injury; and
• to monitor the progress of young offenders during the course of their sentences.

The instrument assesses 12 individual risk factors⁽⁶⁾ grouped into four general categories: offences (3 factors), problematic behaviour (6 factors), personal characteristics (2 factors) and administrative concerns (1 factor).⁽⁷⁾

The assessment is completed in a four-step process that results in a risk assessment and strategy for the young offender. First, the offender's personal characteristics and background history are evaluated and potential risk factors flagged. Next, the risk factors are verified. The assessor then makes a professional decision as to whether to declare the client "at risk" - in accordance with the assessment's specific guidelines or by exercising an assessor override. Finally, the assessor chooses specific offender-management strategies to ensure the progress of the offender and the safety of others.

The initial identification of possible risk factors is fact-based and provides a statistical component (total number of factors) to the assessment. The verification of the risk factors, however, consists of a more clinical inves-tigation of particular circumstances, patterns and other issues that may aggravate or mitigate the factor(s).

Based on this more intensive clinical investigation, which follows a specific protocol for each of the 12 factors, the assessor determines whether the offender is currently "at risk." Guidelines (based on preliminary research) require an offender to be declared "at risk" if the assessor can verify that more than two of the basic factors are suggestive of risk. An offender is to be designated "not at risk" if fewer than two of the factors can be verified. The Youth Management Assessment and probation Following preliminary pilot work to refine the instrument, the Youth Management Assessment was administered by probation and parole officers to 202 young offenders in conjunction with the preparation of predisposition reports (which can be ordered by youth court judges before sentencing).

As expected, there was a consistently lower rate of verified risk factors than possible risk factors - some of the possible factors are usually eliminated during the verification inspection (see Table 1). Not surprisingly, considering the evidence as to peer influence on antisocial youths,⁽⁸⁾ third-party influence was the most common risk factor (on both the possible and verified scales), followed by current violent offence.

Table 1

Percentage of Young Offenders with Possible and Verified Risk Factors (202 Young Offenders)
Risk factor	Risk factor possible	Risk factor verified
Offences
Previous convictions for a violent offence	15%	11%
Current violent offences	22%	18%
Outstanding charges	12%	7%
Problematic behaviour
Other physical or sexual assault(s)	16%	12%
Assault(s) on authority figures	14%	10%
Weapons incidents	11%	6%
Fire-setting incidents	6%	4%
Escapes, attempted escapes and serious unlawlly-at-large incidents	8%	7%
Suicide attempts or self-injury	10%	6%
Personal characteristics
Mental disorder	9%	4%
Third party influence	31%	20%
Administrative concerns
Absence or lack of information	13%	10%

Further analyses revealed a strong relationship between risk at all three stages of assessment and the offenders' ultimate sentence (see Figure 1). However, although judges were not given assessment results, it should be noted that probation officers who complete the assessments also complete predisposition reports.



Figure 1

It is, therefore, very possible that the results of the Youth Management Assessment affected the content of the predisposition reports provided to the courts.

Regardless of the possible influence of the assessment on the judicial decisions, it is clear that the sentencing practices of Ontario youth court judges are consistent with the assessment of offender risk through the Youth Management Assessment.

Further examination also revealed that the young offenders eventually placed in secure custody are at greater risk than other young offenders because they are more likely to be verified as "at risk" for current violent offence, other assaultive incidents, fire setting and unlawfully at large on the Youth Management Assessment (see Table 2). Offenders on probation were, on the other hand, at lowest risk because they were least likely to have been involved in assaultive incidents or to have a history of being unlawfully at large. Open custody The Youth Management Assessment was also administered to a group of 157 young offenders in seven open-custody settings. As the previous analysis suggests, a comparison with the pre-sentence sample of young offenders (discussed earlier) indicated that young offenders in open custody present a greater risk than the general young offender population.

Four risk factors (third-party influence, missing information, other assaultive incidents and current offence) were verified for at least 20% of the open-custody offenders, and another six factors were verified for an additional 10% of the sample. Further, young offenders in open custody had significantly more potential and verified risk factors, and were more likely to be declared "at risk" (53%) than the young offenders at the pre-sentence stage.

These findings led to the implementation of the assessment across Ontario. During the first year of operation, 2,056 Youth Management Assessments (including retests) were administered to 1,469 young offenders.

Analyses of these data indicate that female young offenders had more risk factors present at both the possible (3.83) and verified (3.22) stages than their male counter-parts, who had averages of 3.25 and 2.40, respectively.

These differences occurred primarily because female offenders were verified more often than male offenders on the following risk factors: unlawfully at large (36% versus 17%), suicide/self-injury (20% versus 9.37%), third-party influence (42% versus 29%) and absence or lack of information (63% versus 49%). Consequently, more female offenders (69.74%) were found to be "at risk" on the Youth Management Assessment than male offenders (57.57%). All these differences were statistically significant.

Table 2

Subsequent Sentences Young Offenders with Verified Risk Factors (199 Young Offenders)
Verified risk factor	Probation	Open Custody	Secure Custody
Offences
Previous convictions for a violent offence	9%	15%	13%
Current violent offences	14%	15%	32%
Outstanding charges	6%	10%	11%
Problematic behaviour
Other physical or sexual assault(s)	8%	15%	22%
Assault(s) on authority figures	7%	15%	15%
Weapons incidents	7%	3%	11%
Fire-setting incidents	2%	0%	13%
Escapes, attempted escapes and serious unlawlly-at-large incidents	3%	10%	17%
Suicide attempts or self-injury	8%	8%	4%
Personal characteristics
Mental disorder	4%	3%	9%
Third party influence	17%	21%	28%
Administrative concerns
Absence or lack of information	7%	15%	15%

Secure custody The validity of the Youth Management Assessment was also examined in a co-educational, secure-custody setting. During a one-year period, social work staff administered the assessment to 142 young offenders within three days of their admission to the institution.

The assessment results were then compared with a standard intake form (the Initial Placement Report) -a 29-item checklist that is completed by admissions staff as part of the intake process in all of Ontario's provincial correctional facilities.

Staff members who completed the Youth Management Assessment were not aware of the intake form results when they conducted their assessments. Yet, the total number of items endorsed on the intake reports correlated (r=.38, p<.001) with the number of verified risk factors on the assessment.

The Youth Management Assessment was also examined with respect to subsequent problem behaviour (during the course of a one-year follow-up period). Evidence of problem behaviour in the institution was obtained by examining an offender database (the Offender Management System), which tracks all offender incident reports. These incidents are categorized as either verbal (such as a threat) or physical (such as a fight) and are broken down in terms of the adversary/victim (other offenders, staff or self).

Although the assessment did not predict verbal or physical incidents involving other offenders, it was predictive of verbal (r=.33, p<.001) and physical (r=.37, p<.001) incidents against staff. It was also correlated with verbal threats of self-harm (r=.32, p<.001) and incidents of self-harm (.15, p<.08).

When gender was taken into consideration, a number of risk-by-gender interactions were revealed (see Figure 2). For example, "at risk" male offenders committed significantly more offender assaults than did any other group, while "at risk" female offenders were more verbally abusive to staff than any other group. Further, although they were few in number ⁽¹⁹⁾, female young offenders in secure custody were rated at significantly greater risk than their male counterparts.



Figure 2

A successful combination Preliminary studies suggest that the Youth Management Assessment provides a reasonable means of evaluating young offender risk of committing serious personal injury offences while in the community or in custody.

This use of a routine screening protocol followed by the application of diverse clinical material may, in the end, prove to be superior to both the mechanistic, standardized approach and the unstandardized clinical approach. This should not come as a surprise, since the objective is the prediction of rare events that have varied and interactive origins.

Finally, it is important to realize that the assessment is a dynamic and, hopefully, reactive tool. It was designed explicitly to be readministered at periodic intervals and includes a final section that assists staff in developing an offender-management plan that addresses key problem areas. This section was specifically designed to weave risk assessment into the offender supervision and service process. Together, they comprise the two most important ingredients in the case management of young offenders.



⁽¹⁾Ministry of the Solicitor General and Correctional Services, Box 4100, 200 First Avenue West,
North Bay, Ontario P1B 9M3.
⁽²⁾N. R. Gibbs, "Murder in Miniature," Time, 144, 12 (1994): 14-19.
⁽³⁾R. R. Corrado and A. Markwart, "The Need to Reform the YOA in Response to Violent Young Offenders: Confusion, Reality or Myth?" Canadian Journal of Criminology, 36, 3 (1994): 343-378.
⁽⁴⁾I. Monahan and H. Steadman, Violence and Mental Disorder: Developments in Risk Assessment (Chicago: University of Chicago Press, 1994).
⁽⁵⁾J. S. Wormith and C. S. Goldstine, "The Clinical and Statistical Prediction of Recidivism," Criminal Justice and Behaviour, 11 (1984): 3-34.
⁽⁶⁾The risk factors were taken primarily from a risk indicators study, undertaken by the Ministry of Community and Social Services. See A. W. Leshied, D. A. Andrews and R. D. Hoge, Youth at Risk: A Review of Ontario Young Offenders, Programs and Literature that Support Effective Intervention (Toronto: Ministry of Community and Social Services, 1992). Social science literature on violent and suicidal youth was also reviewed.
⁽⁷⁾A copy of the Youth Management Assessment can be obtained from the author.
⁽⁸⁾M. Warr, "Age, Peers and Delinquency," Criminology, 31 (1993): 17-40.