Grant A Blake1, James R P Ogloff2, Won Sun Chen3. 1. Centre for Forensic Behavioural Science, Swinburne University of Technology and Forensicare, Melbourne, Australia. Electronic address: gablake@swin.edu.au. 2. Centre for Forensic Behavioural Science, Swinburne University of Technology and Forensicare, Melbourne, Australia. 3. Department of Statistics, Data Science and Epidemiology, Swinburne University, Australia.
Abstract
BACKGROUND: Since the 1990s, standardized competency to stand trial assessment measures have been developed to improve the reliability and validity of expert assessment and opinion in legal settings. This study aimed to meta-analyze the psychometric properties of the four 'second generation' assessment tools: the MacCAT-CA, MacCAT-FP, ECST-R and CAST*MR. METHOD: A systematic search of published and unpublished studies was undertaken on PsycNet, PubMed, Scopus, Web of Science and Science Direct up until February 2018. Test manuals were sourced and attempts at finding grey literature included contacting study authors and test developers. RESULTS: Eleven studies were included in the final analysis. The MacCAT-CA, ECST-R and CAST*MR demonstrated acceptable subscale internal consistencies and interrater reliability. The MacCAT-CA and CAST*MR discriminated between competent and incompetent defendants with large effects. The MacCAT-FP had insufficient reliability data and poor discrimination. There was insufficient data to meta-analyze the factor structure, sensitivity, specificity, or effect sizes between types of clinical presentation on any tool. CONCLUSIONS: Very few studies have investigated the psychometric properties of the selected tools. Many studies were excluded for not providing original data or simply re-reporting the psychometric properties stated in test manuals. Whilst the MacCAT-CA and ECST-R have promising psychometric properties, ongoing construct validation is required.
BACKGROUND: Since the 1990s, standardized competency to stand trial assessment measures have been developed to improve the reliability and validity of expert assessment and opinion in legal settings. This study aimed to meta-analyze the psychometric properties of the four 'second generation' assessment tools: the MacCAT-CA, MacCAT-FP, ECST-R and CAST*MR. METHOD: A systematic search of published and unpublished studies was undertaken on PsycNet, PubMed, Scopus, Web of Science and Science Direct up until February 2018. Test manuals were sourced and attempts at finding grey literature included contacting study authors and test developers. RESULTS: Eleven studies were included in the final analysis. The MacCAT-CA, ECST-R and CAST*MR demonstrated acceptable subscale internal consistencies and interrater reliability. The MacCAT-CA and CAST*MR discriminated between competent and incompetent defendants with large effects. The MacCAT-FP had insufficient reliability data and poor discrimination. There was insufficient data to meta-analyze the factor structure, sensitivity, specificity, or effect sizes between types of clinical presentation on any tool. CONCLUSIONS: Very few studies have investigated the psychometric properties of the selected tools. Many studies were excluded for not providing original data or simply re-reporting the psychometric properties stated in test manuals. Whilst the MacCAT-CA and ECST-R have promising psychometric properties, ongoing construct validation is required.