Threat Assessment Tools for the Individual Assessment of Terrorism: A Content Evaluation of the MLG, VERA-2, and HCR-20 V3 Alana N. Cook 1,2,3 , Stephen D. Hart 1,3 , D. Elaine Pressman 4 , Steven Strang 5 , & Yan L. Lim 1 1 Simon Fraser University, 2 BCMHSUS Forensic Psychiatric Service Commission, 3 ProActive ReSolutions, 4 Carlton University, 5 RCMP Contact: alanac@sfu.ca BACKGROUND AND AIM • Terrorism represents a significant concern for public safety internationally with ongoing terrorism violence on almost a weekly basis (United Nations 2015). Threat assessment professionals can contribute to the response and prevention to this violence by assessing and managing violence risk of evaluees who poses a threat to commit a terrorist act. While there have been significant advancements in the development, evaluation, and implementation of tools for violence risk, these efforts have largely ignored for terrorist violence until recently. • Two promising tools were developed in recent years to guide threat assessment professionals in the assessment and management of terrorist risk. The first tool is the Violent Extremism Risk Assessment or VERA guidelines (Pressman, 2009) and second edition of these guidelines, VERA-2 (Pressman & Flockton, 2010). The VERA and VERA-2 were designed specifically for assessment and management of risk for terrorist violence. The second tool is the Multi-Level Guidelines or MLG (Cook, Hart, & Kropp 2014, 2015). The MLG was designed for assessment and management of group-based violence generally, including terrorism. • The VERA-2 and MLG are both structured professional judgment (SPJ) tools based on evidence (research) and consensus (expert opinion) and grounded in best-practice models of threat assessment. Research to date suggests that both the VERA-2 and MLG have demonstrated utility, or usefulness, for their purposes, reliability of the application of the tools by academics and practitioners, and content validity (Beardsley & Beech 2013; Pressman & Flockton 2012; Cook, 2014). • AIM: The current study examined the content overlap of the two recently developed tools (VERA-2 and MLG) as well as a tool designed to assess individual risk for general violence, the third version of the HCR-20 V3 (Douglas et al. 2013). METHOD • A content review was examined by mapping the risk factors of each tool onto the other tools by two independent raters. The content analysis was completed by examining the risk item title and definitions. After independent ratings for content analysis consensus ratings were made. MLG. The MLG (Multi-level Guidelines; Cook et al 2013, 2014) • The MLG is a SPJ tool designed to assess group-based risk for violence. The MLG is administered in seven systematic and standardized steps that guide evaluators through gathering information about a case (Step 1), rating the presence and relevance of 20 risk factors across four domains (Individual, Individual-Group, Group, Group Societal; Steps 2 and 3), developing a case formulation (Step 4) and scenario planning (Step 5) to inform future risk management plans (Step 6). In the final administration step evaluators make summary risk ratings about violence risk (rated as low, moderate, or high; Step 7). VERA-2. The VERA-2 (Violence Extremism Risk Assessment Verision 2; Pressman & Flockton 2010). • The VERA-2 is a SPJ tool designed to assess risk for terrorists and violent political extremists, though may be applied to violent extremist using violence for other ideological reasons (e.g., religious). The VERA-2 is comprised of 31 indicators across four sectors (Beliefs and Attitudes, Context and Intent, History and Capability, Commitment and Motivation) and a fifth factor to consider protective indicators. Evaluators consider each of the 31 indicators on the basis of multiple sources of information. Indicators are rated as low, moderate or high (with protective factors inversely rated). A final risk judgement is made after consideration of all risk and protective indicators. (METHOD CON’T) HCR-20 V3 . The HCR-20 V3 (Historical, Clinical, Risk Management-20, version 3; Douglas et al. 2013) • The HCR-20 is a SPJ tool designed to assess general risk for violence. The HCR-20 V3 is the most widely used risk assessment tool and has a large body of research to support the reliability and validity of the tool. The HCR-20 V3 is administered in seven systematic and standardized steps identical to that of the MLG. The HCR-20 V3 was not developed to assess for terrorism violence. Conceptually, authors have argued and demonstrated that the HCR-20 does not contain the necessary and sufficient risk factors to assess and manage terrorist violence (Gudjonsson, 2009; Pressman, 2009; Pressman & Flockton, 2012). RESULTS • The results were mapped onto four potential domains to consider in the assessment and management if risk for any form of group-based violence, including terrorism based on a comprehensive review of the literature (Cook, 2014). For a detailed content overlap map of the MLG and VERA-2 see Figure 1. • The Individual domain includes factors relevant only to the individual; • The Individual-Group domain considers factors relevant to the individual’s identity, attitudes, and role related to the group; • The Group domain includes factors related to group process and structure; • The Group-Societal level includes factors that are external or peripheral to the group that impact risk. (RESULTS CON’T) Individual-Group Domain: • The MLG and VERA-2 overlap in this domain was clustered in three key overlapping content areas: extremist orientation, commitment to group/cause, and negative attitudes toward other groups. The MLG and VERA-2 did not have any overlap with the HCR-20 V3 in the Individual-Group domain. There were factors on the VERA-2 in the Individual-Group domain that did not overlap in content with the MLG: CI2 (Identification of target for attack) and CI3 (Personal contact with violent extremist). This content is incorporated in the MLG not as risk factors but as part of scenario planning (i.e., Step 4). Similarly, the type of contact the person has with violence extremist is part of describing the individuals role or affiliation with the extremist group is incorporated in Step 1 of the MLG. Group Domain: • The tools did not overlap in content related to the Group domain. The MLG was the only tool with content related to the Group domain. Group-Societal Domain: • The tools did not overlap in content related to the Group-Societal domain. The MLG was the only tool with content related to the Group-Societal domain. Other Considerations on Content Overlap: • Importantly, both the MLG and VERA-2 allow for the inclusion of additional risk factors present and relevant to the risk posed by the evaluee in a given case to allow for additional, case specific factors above and beyond the basic or core risk factors considered in every assessment. • The VERA-2 also has protective factors in addition to the risk factors considered in the assessment of violence risk. The MLG does not explicitly include protective factors. Some content of protective factors of the VERA-2 are captured in the MLG as the absence of the respective risk factors. Also, protective factors are considered in the MLG as part of the development of risk management plans in Step 6. SUMMARY AND CONCLUSIONS 1. In sum, all tools cover Individual domain factors. The VERA-2 overlaps in content focused on the Individual domain, but provides unique at the Individual and Individual Group domain and the MLG at the Individual, Individual-Group, Group, and Group- Societal domain compared to the HCR-20 V3 . 2. Comparing the VERA-2 and MLG, there appeared to be eight overlapping content areas in the Individual and Individual-Group domains, indicating that risk assessment tools in this area should, at a minimum, address these key content areas. 3. There are risk factors on both tools that do not overlap. • The MLG additionally included factors related to mental health and substance use problems, as well as factors that reflect Group and Group-Societal domains. • The VERA-2 additionally included items such as identification of a target and personal contact with violent extremist, as well as six protective factors. Although the MLG does not include these factors items separately, the administration of the tool does include potential targets and the nature of contact with the group. The MLG is also applied so that the risk factors are rated dynamically such that any changes in attitudes and willingness would still be captured as changes over time in the presence and relevance of related risk factors. 4. There is a clear and urgent need for ongoing evaluations to continue to develop and evidence-base for assessors and policy makers on these (or other) tools to inform the development of best-practice tools in the area of threat assessment of terrorist violence. One critical area for research is to evaluate these and other tools in operation contexts. Researchers, threat assessment professionals, policy makers, and organization will need to work together toward this aim. Individual Domain: • The MLG and VERA-2 overlapped in content related to Individual domain factors. These clustered in five key overlapping content areas: individual history of violence, adverse childhood experiences, problems fitting in with society, capacity for violent actions, and antisocial orientation. • There were factors on the MLG in the Individual domain that did not overlap in content with the VERA-2, I5 (Mental Disorder) and I6 (Substance Use). • The MLG and VERA-2 also overlap in content of the HCR-20 V3 . • The content overlap with the HCR-20 V3 was exclusive to the Individual domain and also almost exclusively includes the Historical Items on the HCR-20 V3 . The MLG covers all Historical factors with the exception of H10 (Treatment or Supervision Response); and the VERA-2 covers all Historical factors with the exception of H3 (Other Antisocial Behavior), H5 (Substance Use), and H6 (Major Mental Disorder). • Neither the VERA-2 nor MLG has content that substantially overlaps with the Clinical and Risk Management factors of the HCR-20 V3 . The MLG has some overlap with C2 (Violent Ideation or Intent), C3 (Symptoms of Major Mental Disorder), and R3 (Personal Support) form the HCR-20 V3 , but the VERA-2 did not overlap with any Clinical or Risk Management factors. This project was funded by a research grant from the Canadian Network for Terrorism, Security, & Society.