The Use of a Multi-State Broad-Band Risk Assessment in Predicting Sexual Recidivism

• In Event: Risks, Needs, and Responsivity in Juvenile Court Outcomes

Abstract

Objective
As risk assessments have become more prominent in the justice system, and specialized tools have been developed sexual offending recidivism. However, these specialized tools present substantial prediction issues and demonstrate a low level of predictive validity when in non-development samples.
Data/Methods
We address potential improvements in these models using a large, US representative, 10-state sample of assessed youth (N = 494,050). We present novel assessment development procedures using a multi-level model predicting youth sexual offending.
Results
Findings provide performance of developed assessment models for youth with, and those without, a prior sex offense. Further, we assess the potential role of jurisdiction in prediction.
Implications
The current study fills a much-needed gap, identifying improved predictive accuracy using a more robust and powerful sample. Further, we provide an understanding of the similarities and differences between youth assessed with, and without, a prior sex offense history, giving rise to discussion of specialization and progression of offending for justice involved youth.

Authors

• Addison R. Kobie, University of Nebraska at Omaha

• Alex Kigerl, University of Nebraska at Omaha

• Zachary Hamilton, University of Nebraska at Omaha