Interrater Reliability and Predictive Validity of the Spousal Assault Risk Assessment - Version 3 (SARA-V3)

Area of Science:

• Forensic Psychology
• Criminology
• Risk Assessment

Background:

• Intimate partner violence (IPV) risk assessment is crucial for public safety and offender management.
• The Spousal Assault Risk Assessment Guide-Version 3 (SARA-V3) is a tool used to assess this risk.
• Evaluating the reliability and validity of risk assessment tools is essential for their effective implementation.

Purpose of the Study:

• To examine the interrater reliability of Conclusory Opinions generated by the SARA-V3.
• To assess the predictive validity of SARA-V3 Conclusory Opinions for future IPV.
• To compare the utility of SARA-V3 Conclusory Opinions with previous versions.

Main Methods:

• A pseudo-prospective design was employed.
• SARA-V3 assessments from 125 Australian males with prior IPV were coded.
• Predictive validity was evaluated using subsequent police-recorded IPV charges over a 4-year period.

Main Results:

• Total SARA-V3 scores demonstrated excellent interrater reliability (ICC = .94).
• Conclusory Opinions showed variable interrater reliability (ICC = .54–.92).
• Conclusory Opinions had limited ability to discriminate future IPV (AUC = .51–.61) with consistently low sensitivity.

Conclusions:

• Conclusory Opinions derived from SARA-V3 have limited predictive utility for recidivism in intimate partner violence.
• The low sensitivity suggests many high-risk individuals may be missed, impacting case prioritization.
• Significant limitations include the inability to rate Victim Vulnerability factors due to data constraints.