Response to 'The Use of Meta-Analysis to Compare and Select Offender Risk Instruments'

Journal article

A complex meta-analysis often addresses more than one principal question, which may be published in two or three separate papers. The Singh 2011 review compares risk assessment instruments and asks the question whether there are differences in their predictive accuracy using multiple performance statistics (Singh, Grann, & Fazel, 2011). It relies on how the research literature has used these instruments, which likely reflects the intended population and practical considerations. However, there were two other key questions that we tried to address when we reviewed risk assessment tools, which required separate publications. In a 2012 publication, using mostly the same data but with updated information, we stratified the performance of these instruments according to whether they focused on violent, sexual, and general criminal offending separately (Fazel, Singh, Doll, & Grann, 2012). So a key criticism of the current paper – that stratification by type of offender was not done - has been addressed to a large extent in the 2012 review. Furthermore, the current paper conflates the aims of the 2011 and 2012 reviews. The 2011 review was not primarily intended to examine the overall accuracy of risk assessment tools. Rather that was the main question of the 2012 review according to whether the particular instrument was focused on the assessment of violent, sexual, and criminal recidivism.

Authors: Fazel, S

• Publication status: Published
• Peer review status: Reviewed (other)
• Publisher: Taylor and Francis
• Journal: International Journal of Forensic Mental Health
• Volume: 16
• Issue: 1
• Pages: 16-17
• Publication date: 2017-01-01
• Acceptance date: 2016-09-20
• DOI: 10.1080/14999013.2016.1261965
• EISSN: 1932-9903
• ISSN: 1499-9013
• Keywords: Article in Press