Violence Risk Assessment with Youth Who Have Sexually Offended: A Psychometric Examination of the Violence Risk Scale: Youth Sexual Offender Version (VRS:YSO)
Date
2013-04-25
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
ORCID
Type
Degree Level
Doctoral
Abstract
The present research was archival in nature. It was divided into three phases involving a comprehensive psychometric examination of a newly developed risk assessment measure designed to assess risk for sexual violence, guide treatment planning, assess readiness to change, and evaluate whether positive changes in risk are linked to reductions in recidivism risk among sexually offending youth–the Violence Risk Scale: Youth Sexual Offender version (VRS:YSO). Phase I focused on an examination of the basic psychometric properties of a young offender version of the Violence Risk Scale–Sexual Offender Version (VRS–SO; Wong et al., 2003), subsequently referred to as the VRS:YSO. The tool revision was based on a thorough review of the literature on male youth and adults who have sexually offended and selected readings on youth violence, as well as a review of a measure designed to assess risk for violence among violent youth (i.e., the VRS:YV). The developers of the VRS–SO completed the structural revisions to the tool. The resulting risk assessment measure included six static and 17 dynamic risk variables that are empirically and/or theoretically associated with an increased risk for sexual violence among youth. As part of its initial psychometric examination, the VRS:YSO was rated retrospectively from the comprehensive files of 99 male sexually offending youth who had received outpatient sex offender services (i.e., assessment and/or treatment) from the Saskatoon Health Region, Child and Youth Services (CYS)–Young Offender Program (YOP) from 1995 to 2008. Overall, the VRS:YSO showed good-to-excellent interrater reliability, sound item properties (i.e., internal consistency and item-total correlations), and a factor structure that is consistent with research on sexually offending youth and adults, as well as other measures designed to assess risk for sexual violence in youth.
Phase II focused on the validation of the VRS:YSO through examining the concurrent, postdictive, predictive, and incremental validity of the measure. Moreover, the psychometric properties of existing specialized risk assessment measures (i.e., J-SOAP-II, ERASOR, and J-SORRAT-II) were examined to inform the limited, albeit growing, literature on risk assessment with sexually offending youth. All measures were rated from the same comprehensive youth files as in Phase I. Youth were followed-up for an average of 11.83 years (SD = 3.42, range = 3.89-17.41) starting from their first contact with the community post-adjudication (i.e., release from custody or commencement of a community sentence). Overall reconviction rates were 8% for sexual, 24% for any violent (sexual and nonsexual), and 37% for general (any) offending. In sum, there was good preliminary evidence for the concurrent and, to some extent, postdictive validity of the VRS:YSO. Furthermore, there was good evidence for the predictive validity of the measure, particularly with respect to any violence (sexual and nonsexual) and general (any) recidivism. There was also encouraging evidence, albeit offset seemingly by a small post-treatment N and thus restricted power, on the potential value of the therapeutic change score in the prediction of recidivism risk. These findings supported the value of the VRS:YSO as a new specialized risk assessment measure for sexually offending youth, particularly as it uniquely includes a systematic rubric for assessing change. Moreover, there was good evidence for the predictive validity of the J-SORRAT-II and certain components of the J-SOAP-II for sexual recidivism. There was also good evidence for the predictive validity of the J-SOAP-II and the ERASOR for violent (sexual and nonsexual) and general (any) recidivism.
Lastly, Phase III examined the role of psychopathy-related personality features (as measured by the PCL:YV) in the criminal and treatment outcomes of youth who have sexually offended. The PCL:YV was rated from the same comprehensive youth files (in conjunction with the other measures referenced above). Overall, the PCL:YV significantly predicted violent (sexual and nonsexual) and general (any) reoffending, particularly among offenders with any peer/adult victims, but not sexual reoffending. Increasing psychopathy-related personality features were significantly associated with non-completion of treatment. Clinical implications of research findings, along with limitations and future research directions for each of the three phases of this research were discussed.
Description
Keywords
risk assessment, sex offenders, sexual violence, youth violence, sexually offending youth, juvenile sex offenders
Citation
Degree
Doctor of Philosophy (Ph.D.)
Department
Psychology
Program
Psychology