Associations between race and violence risk appraisals using structured professional judgment.

Abstract

Racial disparities in criminal justice processes and outcomes have garnered considerable attention. Whether and how risk assessment instruments might contribute to such discrepancies are critical questions. However, tools based on the widely applied structured professional judgment (SPJ) model have received little attention in this regard. This study investigated relationships between race and violence risk appraisals derived from two commonly used SPJ tools, the Historical-Clinical-Risk Management 20 (HCR-20 V3 ) and the Spousal Assault Risk Assessment Guide (SARA-V3), using a database of 337 diverse individuals' information. Results were mixed with respect to racial disparity. Race was not clearly associated with summary ratings on the HCR-20 V3 . However, being Indigenous predicted high-risk ratings on the SARA-V3 after controlling for risk factors and criminal history, but not after controlling for actuarial risk scores. While the results were promising in many respects, they nonetheless suggest that associations between race and SPJ risk appraisals warrant further empirical study. Public Significance StatementThis study demonstrates that racial disparities may be observed among risk appraisals derived from structured professional judgment tools, based on a sample of Canadian Indigenous and non-Indigenous persons. The results were mixed and do not identify causes of the disparities, but they provide an impetus and clear directions for further study to further the aim of racial justice in risk assessment.

Child and adolescent forensic evaluations: Challenges amid scientific and technological advancements.

“…They suggested a practical approach moving forward, taking a close look at the strengths and weaknesses of all the available alternatives (Spivak & Shepherd, 2021). Hogan et al (2021) see great potential in AI-driven algorithms while emphasizing the need to keep ethics at the fore as we build toward a consensus of use and stress the need for appropriate training. Luxton (2014) noted the potential benefits of AI-based clinical decision support systems in psychological practice: (a) assisting practitioners with extracting relevant data, (b) reducing demands on staff time, and (c) reducing human error.…”

“…Spivak and Shepherd (2021) presented a balanced view of the use of AI in forensic risk assessment while acknowledging and appreciating the concerns presented by Hogan et al (2021). More specifically, they argue that many of the concerns with AI mirror the same ones with current methods of risk assessment (Spivak & Shepherd, 2021).…”

“…They suggested a practical approach moving forward, taking a close look at the strengths and weaknesses of all the available alternatives (Spivak & Shepherd, 2021). Hogan et al (2021) see great potential in AI-driven algorithms while emphasizing the need to keep ethics at the fore as we build toward a consensus of use and stress the need for appropriate training.…”

Special issue on cultural responsiveness in threat assessment.

“…They then propose a method for increasing self-awareness of social locations and offer recommendations for improving cultural responsiveness in TAM training and practice. In the special section’s second article, Hogan and Corabian (2022) present the findings of their study investigating possible bias in violence risk predictions. They outline conceptual and empirical foundations for the potential association between race and violence risk findings when structured professional judgment approaches are used.…”

Consistency and Construct Validity of the Five-Level System for Risk Communication Using Static and Dynamic Tools: An Investigation Using the Static-99R and VRS-SO

“…Risk instruments continue to proliferate (Singh et al, 2014), reflecting not only the various functions they are designed to serve (e.g., to inform frontline police operations, judicial sentencing decisions, treatment services) but also the complexity and variability inherent in concepts of recidivism. Elaborating on the latter point, Hogan et al (2021) recently argued that the breadth of tools available to forensic evaluators reflects, at least in part, the fact that recidivism definitions encompass "behaviors ranging widely in form (e.g., sexual versus nonsexual violence), motivation (e.g., instrumental versus reactive aggression), and severity (e.g., serious threats, pushing, homicide)" (pp. 3-4).…”