The predictive validity of risk assessment tools used in Australia for female offenders: A systematic review

• Published in: Aggression and violent behavior 2020-07, Vol.53, p.101428-15, Article 101428
• Main Authors: Gower, Menna, Spiranovic, Caroline, Morgan, Frank, Saunders, Julie
• Format: Article
• Language: English
• Subjects: Antisocial personality disorder; Base rates; Female offenders; Females; Gender differences; Indigenous peoples; Offenders; Predictive validity; Recidivism; Representativeness; Risk assessment; Risk management; Systematic review; Validity; Women
• ISSN: 1359-1789; 1873-6335
• DOI: 10.1016/j.avb.2020.101428

Assessing an offender's level of risk is a vital step in the determination of treatment services; however, risk assessment tools have been consistently developed using a male cohort. Geraghty and Woodham (2015) conducted a review to determine the validity of risk assessment tools for females and found the Level of Service Inventory (LSI) to be the most valid. The LSI has been further developed and additional risk assessment tools have been included in the suite of assessments across various international jurisdictions; therefore, an updated review is necessary to encompass subsequent research. The current review is an international investigation into the relevance of risk assessment tools for females; however, Australia has not carried out extensive research in this area, particularly Western Australia which has an overrepresentation of Aboriginal offenders. As such, the focus was on studies that included risk assessment tools used in Australia including the Level of Service/Risk, Need, Responsivity (LS/RNR), the Violence Risk Scale (VRS), the Psychopathy Checklist (PCL-R) and the Historic, Clinical and Risk 20 (HCR-20). Five databases were searched and included OVID Embase, OVID Medline, OVID PsycInfo, OVID full text and PsycARTICLES and elicited 610 studies with a further three sourced from gray literature. The articles were assessed based on pre-determined criteria which resulted in 18 studies reviewed in detail. It was identified that all four risk assessment tools were valid for use with female offenders; however, this was not consistent with the findings from a single Australian study, likely due to the combination of a small sample and low base rates of reoffending. Further research is needed on gender comparisons to consider potential gender differences in factors and items that form the risk assessment tools so that treatment and risk management pans may be optimised for female offenders.