Short and long-term prediction of recidivism using the youth level of service/case management inventory in a sample of serious young offenders.

Abstract

The present investigation examined the predictive accuracy of the Youth Level of Service/Case Management Inventory (YLS/CMI) for youth and adult recidivism in a Canadian sample of 167 youths (93 males, 74 females) charged with serious offenses who received psychological services from a community mental health outpatient clinic. Youths were followed for an average of 7 years in the community, and predictive accuracy was examined for several recidivism outcomes as a function of gender, ethnicity, and developmental age group. YLS/CMI total scores significantly predicted all recidivism categories in the overall sample (area under the curve values ranged from 0.66 to 0.77) although the instrument as a whole, and its criminogenic needs, demonstrated somewhat stronger and more consistent predictive accuracy for youth outcomes. The YLS/CMI also demonstrated significant predictive accuracy within demographic subgroups. The implications of these findings are discussed in terms of the use of risk-need assessment tools in providing clinical assessment, treatment, and case management services to diverse young offender groups.

An Examination of Criminogenic Needs, Mental Health Concerns, and Recidivism in a Sample of Violent Young Offenders: Implications for Risk, Need, and Responsivity

“…From this perspective, the "central eight" variables of criminal activity are posited to be the strongest predictors of criminal activity, with demographic variables, including gender, having only minimal predictive and explanatory value after the central eight variables are considered (Rettinger & Andrews, 2010). Studies on the YLS/CMI, provide support for the validity of the "central eight" risk factors in predicting young offender outcomes (Olver, Stockdale, & Wong, 2012;Olver, Stockdale, & Wormith, 2009;Schwalbe, 2008). Research advances may be made from further examination of risk factors among different population subgroups including young offenders with mental health and substance use diagnoses.…”

“…Additionally there is strong evidence for internal consistency, interrater reliability, and predictive accuracy of the YLS/CMI with both sexes and with ethnic minorities (Olver et al, 2009;Olver et al, 2012). For the present study sample, the YLS/CMI was rated primarily from court and treatment files (n D 126) at the facility by trained research personnel described in Olver et al (2012) or when available, actual test protocols rated in real time in the course of clinical service delivery were extracted from youth files (n D 41). The clinically rated protocols were rated by registered psychologists experienced in risk assessment and trained on the YLS/CMI.…”

“…The present study is an archival investigation that involved extraction and analysis of psychometric data, including demographic and clinical variables (e.g., mental health diagnosis) from the adolescents' court, treatment, and psychometric test files located at the community facility, and is an extension of Olver et al (2012). The amount of information available for each participant varied, with the most complete files containing psychological assessments including measures of cognitive ability, academic achievement, personality, psychopathology, substance abuse, and criminogenic risk.…”

Gender and Risk Assessment in Juvenile Offenders: A Meta-Analysis

“…Baird et al (2013) also found that the tool predicted recidivism better for boys than girls. While Olver, Stockdale and Wong (2012) found that the instrument worked equally well on some outcomes, they found that it predicted male recidivism better for others. While the study completed by Marshall, Egan, English, and Jones (2006) looked at violent offending against correctional staff rather than recidivism, it revealed that the YLS/CMI was better at predicting male violence than female violence.…”

The Use of the SAVRY and YLS/CMI in Adolescent Court Proceedings: A Case Law Review

“…The emphasis of risk assessments in such cases is often a youths' long-term risk; however, thus far relatively little evidence exists regarding the ability of adolescent risk assessment tools to predict long-term risk (Edens & Cahill, 2007;Worling, Bookalam, & Litteljohn, 2012). Although adolescents' risk may be quite dynamic, some studies do suggest that long-term predictions with these tools may be possible in some cases (Schmidt et al, 2011;Olver, Stockdale, & Wong, 2012).…”