An Examination of Risk, Need, and Protective Factors Among Saskatchewan Young Offenders: Implications for Risk Management and Community Reintegration
Lovatt, Kristine May
MetadataShow full item record
The present study examines the validity, reliability, and psychometric properties of three established forensic measures for youth, the Psychopathy Checklist: Youth Version (PCL: YV; Forth, Kosson, & Hare, 2003), the Structured Assessment of Violence Risk in Youth (SAVRY; Borum, Bartel, & Forth, 2002), and the Level of Service Inventory: Saskatchewan version (LSI-SK; Luong & Wormith, 2011). This study also examines the validity, reliability, and psychometric properties of two newer tools, the Violence Risk Scale – Youth Version (VRS-YV; Lewis, Stockdale, Gordon, & Wong, 2014) and the Structured Assessment of Protective Factors for Violence Risk: Youth Version (SAPROF-YV; de Vries Robbé, Geers, Stapel, Hilterman, & de Vogel, 2015). The VRS-YV is a violence risk assessment tool designed to bridge the gap between assessment and treatment by incorporating a modified version of Prochaska, DiClemente, and Norcross’ (1992) transtheoretical model of change into the tool, allowing for assessment of treatment-based change. The SAPROF-YV was designed to provide an empirical measure of protective factors in youth, as the role of protective factors in risk assessment has garnered increasing attention in recent years (de Vries Robbé & Willis, 2017). Finally, the present study examined the predictive validity of the above tools with diverse populations; specifically, Indigenous youth and females. The total sample consisted of 451 youth who had received assessment and/or treatment services from a community mental health facility in Saskatoon, Saskatchewan, however, not all files had enough information to score the tools. The PCL: YV, SAVRY, VRS-YV, and SAPROF-YV were each rated from file information on 257 young offenders (197 males, 60 females; 174 Indigenous, 57 non-Indigenous) who had enough information in file to rate the tools. The LSI-SK is rated by youth workers on all young offenders in Saskatchewan, so this information was gathered directly from files when available. The forensic measures and their component/factors demonstrated good internal consistency (α = 0.84 – 0.99) and inter-rater reliability (ICC = 0.73 – 0.97). There was also good convergent validity amongst the five tools, with correlations in the medium to large range but, there was generally no incremental validity amongst various tool pairings. Male and female scores were generally similar across all five measures and component/factor scores. However, Indigenous youth tended to score significantly higher than non-Indigenous youth, except on measures of protective factors where non-Indigenous youth scored higher than Indigenous youth. Recidivism information was available for almost the entire sample (n = 444) over a mean follow-up time of approximately nine years. Predictive accuracy evidence was obtained for all five tools with respect to youth and adult violent, non-violent, and any recidivism. Area under the curve (AUC) values varied based on the tool and the type of recidivism, but generally fell in the small (AUC = 0.57 – 0.63) to moderate (AUC = 0.64 – 0.70) range for the total sample, with predictive accuracy for adult recidivism typically better than for youth recidivism. Survival analyses further supported the ability of the VRS-YV, SAVRY, and PCL: YV to differentiate between low, medium, and high-risk offenders (depending on recidivism type). Predictive accuracy for diverse sub-groups varied; they tended to be similar in magnitude for males and females, although values were often not significant for females. However, predictive accuracy was better for violent recidivism for Indigenous youth and non-violent recidivism for non-Indigenous youth. Sufficient treatment information was available for a small portion of youth (n = 89), which was used to rate the VRS-YV post-treatment. There was a small but significant degree of change and change results tended to be in the expected direction, but was only significantly associated with decreased rates of non-violent recidivism. The present study provides further evidence for the use of established forensic measures to assess risk in youth. It also provides further evidence that the newer VRS-YV is a valid tool to assess risk and predict recidivism, performing comparably to the other tools included in this study, with the added unique ability to assess treatment-based change. This study also speaks to the role that an empirical measure of protective factors, the SAPROF-YV, may play in the assessment of young offender risk. Finally, this study demonstrated that these tools can have predictive validity for future recidivism among diverse groups such as Indigenous youth and female youth. Limitations and future directions are discussed.
DegreeDoctor of Philosophy (Ph.D.)
SupervisorOlver, Mark; Stockdale, Keira
CommitteeGagnon, Michelle; Marche, Tammy; Luther, Glen; Chirkov, Valery
Copyright DateNovember 2020