| Summary: | Despite growth in research regarding people with an Intellectual Disability who offend, there has yet to have been a tool that accurately predicts reoffending amongst this population across the broad range of offence types over the short term. This thesis addressed this gap by examining the reliability, ecological validity and predictive validity of the Assessment of Risk and Manageability of Individuals with Developmental and Intellectual Limitations who Offend – Generally (ARMIDILO-G). The ARMIDILO-G, along with the Historical Clinical Risk – 20 (HCR-20), Level of Service Inventory – Revised (LSI-R), Current Risk of Violence (CuRV), Group Risk Assessment Model (GRAM) and Psychopathy Checklist Revised (PCL-R) were administered to 139 people with an Intellectual Disability who have a history of offending and were being supported by a NSW community based forensic disability service. Whilst found reliable, the factor structure of the ARMIDILO-G raised question as to the separation of risk and protective items. Predictive validity of instruments was measured prospectively at three and six months based on official criminal charges, convictions and custodial episodes. Little difference was found in the predictive validity of tools when used in an actuarial manner. The GRAM performed best out of all actuarial measures and reliably predicted reoffending across general, theft and violent offences. The ARMIDILO-G using a structured professional judgment (SPJ) approach demonstrated good predictive validity for general, violent, public order and theft offences. Methodological limitations made it unclear whether the ARMIDILO-G’s performance was owing to the qualities of the tool or assessor. No conclusion could be made as to whether SPJ or actuarial approaches, or static or dynamic variables, are more effective in predicting short term risk of reoffending for those with an Intellectual Disability
|
|---|
