All Resources

​​Illustrates lessons learned from surveying all 50 states and the District of Columbia on how risk and need assessments are used by criminal justice agencies.

Describes the implications of a study that shows how machine learning algorithms could lead to better criminal justice outcomes for both stakeholders and justice-involved populations.

Describes the development of the MnSCORE as an effort to reduce first-time sexual offending and assesses the tool’s interrater reliability and predictive validity.

Examines the predictive accuracy of the Static-2002 on people convicted of sex offenses in four different countries and supports the use of Static-2002 in applied assessments.

Measures the validity of the LSI-RSV for use with people in Australia with a mental illness and suggests that the tool is a good predictor of recidivism among people with mental disorders but not people with a dual diagnosis.

​​Compares the interrater reliability and predictive validity of the SVR-20 and Static-99 in assessing risk and predicting recidivism for sexual violence.

Summarizes the development and validation of the VPRAI and provides instructions for how to use the tool.

​​Conducts a reliability and validation study of the LSI-R to assess the tool’s usefulness as a component of Pennsylvania’s parole decision making guidelines.

​​This checklist lays out action items for tool selection, validation, assessment process, and structured decision making stages.

​​Identifies a set of predictors that can be used by an agency to identify individuals who will be challenging to supervise, with the goal of developing an instrument for informing immediate, risk-anticipated security and treatment assignments for people under community supervision.