An examination of change in dynamic risk of offending over time among serious juvenile offenders

• There is a call to consider dynamic risk/need indicators in juvenile assessments.
• We examine how much risk/need indicators change over time and their relation to offending.
• Overall risk scores decrease over time, but not all domains of indicators shifted equivalently.
• Risk indicators were related to later offending, with more recent indicators being more salient.
• Repeated assessment of selected risk indicators would be useful for intervention planning.

PurposeWe examine whether and how much risk/need indicators change over time in a sample of serious adolescent offenders and whether changes in risk are related to self-reported and official record reports of offending in the year following assessment.MethodsGrowth curve and multilevel mixed-effects models are used to examine change through age 18 in a sample of 1354 serious adolescent offenders participating in the Pathways to Desistance Study.ResultsThree primary findings emerge: 1) Compared to the baseline assessment, overall risk/need scores decrease over time. 2) Risk/need does not change in a uniform sequence across domains and time; the form and rate of change differ by domain. 3) Risk/need indicators were related to later offending, with more recent indicators being more powerful for predicting rearrest.ConclusionsThe findings provide empirical support for recent efforts to incorporate routine risk/need assessment into juvenile justice practice. Repeated assessments are likely to identify fluctuations in areas of risk/need that can be used to inform case management and intervention efforts, even for serious offenders.