Predictive validity of the Young Offender Level of Service Inventory for criminal recidivism of male and female young offenders

Risk assessment for juvenile justice: a meta-analysis

Risk assessment instruments are increasingly employed by juvenile justice settings to estimate the likelihood of recidivism among delinquent juveniles. In concert with their increased use, validation studies documenting their predictive validity have increased in number. The purpose of this study was to assess the average predictive validity of juvenile justice risk assessment instruments and to identify risk assessment characteristics that are associated with higher predictive validity. A search of the published and grey literature yielded 28 studies that estimated the predictive validity of 28 risk assessment instruments. Findings of the meta-analysis were consistent with effect sizes obtained in larger meta-analyses of criminal justice risk assessment instruments and showed that brief risk assessment instruments had smaller effect sizes than other types of instruments. However, this finding is tentative owing to limitations of the literature.

Level of Service Inventory-Revised profiles of violent and nonviolent prisoners

Level of Service Inventory-Revised (LSI-R) assessments were conducted on 251 male prisoners beginning prison sentences in six English prisons. The prisoners who were classified as violent, on the basis of either current or previous convictions, scored significantly higher than the nonviolent prisoners on the four LSI-R subscales of Criminal History, Companions, Education and Employment, and Alcohol and Drugs. In addition, those prisoners with a record of violent offenses gave higher total LSI-R scores, indicating an overall higher risk of recidivism. Multivariate analysis using the LSI-R subscales to classify the higher risk violent prisoners and minimum-risk nonviolent prisoners basically confirmed the findings of the univariate analysis. The implications of the findings for informing practice with violent offenders are discussed in terms of screening for intensity of service delivery and identification of targets for change. It is suggested that screening for drug and alcohol problems should be a prerequisite to working with violent offenders.