PROTOCOL: Multiagency programmes with police as a partner for reducing radicalisation to violence.

Abstract

Violent radicalisation is a complex problem, complicated by the lack of a clear terrorist profile and variation in the risk factors that predict violent extremism across individuals and groups (Campelo, Oppetit, Neau, Cohen, & Bronsard, 2018; Carlsson et al., 2020; Desmarais, Simons-Rudolph, Brugh, Schilling, & Hoggan, 2017; Wolfowicz, Litmanovitz, Weisburd, & Hasisi, 2019). While models of understanding radicalisation vary (Borum, 2015; Christmann, 2012; Desmarais et al., 2017; Horgan, 2008; Koehler, 2017; Kruglanski, Bélanger, & Gunaratna, 2019; Sarma, 2017), it is broadly defined as the process by which a person adopts extremist views and moves towards committing a violent act (Irwin, 2015; Jensen, Atwell Seate, & James, 2018). Radicalisation has been linked with individual and group engagement in terrorist attacks against innocent civilians (Wilner & Dubouloz, 2010), as well as individuals entering conflict zones to join formal extremist groups to engage in violent combat (Lindekilde, Bertelsen, & Stohl, 2016). As a result, radicalisation has become a key focus for counterterrorism and violence prevention interventions. The complex and varied nature of individuals' progression from radicalisation to violence presents challenges for designing and evaluating appropriate interventions and policy responses (Hafez & Mullins, 2015; Helmus et al., 2017; Horgan & Braddock, 2010; Horgan, 2008; Jensen et al., 2018; Kruglanski et al., 2019). This level of complexity has driven national counterterrorism policy agendas to adopt intersectoral and multiagency responses that aim to address various radicalisation processes and risks (Beutel & Weinberger, 2016). These multiagency responses often involve partnerships and collaborations between various different agencies and entities (Hardy, 2018), such as governmental agencies, private businesses, community organisations and service providers. Multiagency interventions can provide a framework for pooling and sharing resources to address a common problem (Crawford, 1999; Rosenbaum, 2002), such as radicalisation to violence. Yet they can be challenging to implement, and their effectiveness may be influenced by the quality and nature of the collaboration between agencies (see Berry, Briggs, Erol, & van Staden, 2011, for review; Atkinson, 2019; Gittell, 2006; Kelman, Hong, & Turbitt, 2013; McCarthy & O'Neill, 2014; Rosenbaum 2002). Multiagency interventions may be conceptualised on a continuum, with activities ranging from minimal collaboration to a wholistic integration of agencies and organisations (Atkinson, 2019). As a result, the outcomes of multiagency interventions may vary depending on the where the intervention falls—or is perceived to fall—on this collaborative continuum (Atkinson, 2019). Partnerships can enhance formal and informal communication, trust, respect, shared goals and knowledge (Bond & Gittell, 2010). Conversely, partnership-based interventions may highlight a number of shortcomings in service delivery including the disjointed nature of services, the need for significant stakeholder buy-in, the isolation for some of the organisations or individuals, and the resource-intensive nature of many of these collaborations (Atkinson, 2019; Bond & Gittell, 2010; Crawford, 1999; McCarthy & O'Neill, 2014; Yousanamouth, 2019). There is also the possibility that multiagency approaches could lead to adverse outcomes (Galloway, 2017; Norton, 2018). For example, multiagency responses that have poor levels of coordination and communication could lead to cases falling through the cracks where no one agency responds under the misguided assumption that another partner agency is taking the lead (Richards, 2017; Smith, Laszlo, Ayers, & Smith, 1992). Ransley (2016) also raises the possibility of increased coercion from multiagency responses. Therefore, when assessing the effectiveness and the intended outcomes of multiagency interventions, it is also important to consider the context, potential backfire effects and quality of the processes underpinning multiagency collaboration. A broad range of agencies and experts can be involved in multiagency approaches for reducing radicalisation to violence or violent extremism (Crawford, 1999). Nevertheless, the public police are often one of the first points of contact with individuals who have radicalised to extremism. The public police are also the first point of call for those who are concerned about or report known associates, friends or family members as being at risk of radicalisation. As such, police are important partners for identifying, reducing and building resilience to radicalisation (Cherney, 2015). This review will, therefore, focus on the effectiveness of police-involved multiagency interventions for reducing radicalisation to violence and improving multiagency collaboration. Multiagency interventions are generally characterised by two or more entities partnering to solve a shared problem. These entities may be government agencies (such as education, immigration, customs, home affairs, employment, housing, health), or nongovernmental agencies, including: local councils, businesses, community organisations (such as churches, mosques and other houses of worship) and service providers (such as resettlement agencies, local health providers). This review will include any multiagency intervention, where at least one of those partners is the public police and where the intervention explicitly aims to address terrorism, violent extremism, or radicalisation to violence. This type of intervention can include a range of approaches, including: police engaging with different community and agency stakeholders to help identify terrorist threats (Innes, Roberts, & Innes, 2011; Ramiriz, Quinlan, Malloy, & Shutt, 2013); police working with other agencies to refer, assess, or case-manage individuals convicted of terrorism or identified as at-risk for radicalisation (Cherney & Belton, 2019); or police forming task forces or partaking in regular structured meetings with other agencies to problem-solve issues pertaining to radicalisation or extremism (Koehler, 2016). Some observe that there is a great deal of heterogeneity in the risk factors and triggers for radicalisation (Dalgaard-Nielsen, 2010, 2018; Horgan 2009). This means there is a variety of risk and background factors that may lead an individual (or a group of individuals) to radicalise to violent extremism (Campelo et al., 2018; Carlsson et al., 2020; Vergani, Iqbal, Ilbahar, & Barton, 2018). Research also demonstrates the complex nature of different progression pathways from radicalisation to violence (Horgan, 2008; Kruglanski et al., 2019; Kruglanski, Webber, & Koehler, 2020). The literature is, therefore, in agreement that the complex nature of radicalisation risk and pathway processes to violence makes it impossible for any single agency, organisation or entity to address the problem alone (Dalgaard-Nielsen, 2018). As such, interventions to address the problem of radicalisation to violence are often characterised by multiagency partnerships, working across different service delivery sectors (see e.g., Cherney & Belton, 2019; Crawford, 1999; Innes et al., 2011). Multiagency partnerships may disrupt pathways from radicalisation to violence by collectively addressing multiple risk factors in a holistic and coordinated manner (Butt & Tuck, 2014). The multiagency approach to tackling violent extremism may be effective because it fosters a coordination of effort (Crawford, 1999; Kelman et al., 2013), draws from a broad range of expertise (Crawford, 1999), allows for information and intelligence sharing (Cherney, 2018; Murphy, 2008; Slayton, 2000) and enables the pooling of resources (Crawford, 1999; Sestoft, Hansen, & Christensen, 2017). El-Said (2015) describes a range of different ways that multiagency partnerships are created: by formal and informal arrangements, such as legislative or regulatory frameworks, memoranda of understanding or policy standards stipulating channels for information sharing or better interpersonal relations between agencies (see also Koehler, 2016). These arrangements create opportunities for referrals being made from various sources (Koehler, 2017), increasing the capacities for partnerships to detect and respond to those at early pathways to radicalisation and violence. The capacities of multiagency partnerships to better detect and respond to problems over and above what is possible by agencies or entities working alone are enhanced through better information sharing and referral processes (Cherney, 2018; Murphy, 2008; Slayton, 2000). Partnerships can enhance programme planning and design so that counter radicalisation strategies address the required risks and vulnerabilities among individuals and groups (Koehler, 2017). The range of expertise across multiagency partners also help to enhance programme implementation by ensuring the all required components of a strategy are delivered (Crawford, 1999). They are likewise important in relation to programme evaluation by enabling the sharing of data that can be used to assess programme effectiveness (Cherney, 2018). Police cannot tackle the problem of radicalisation, violent extremism and terrorism on their own (Cherney & Hartley, 2017). Many of the risk factors for radicalisation and violent extremism are complex (Dawson, Amarasingam, & Bain, 2016; Hafez & Mullins, 2015; Kruglanski et al., 2019, 2020). Research suggests that is not just the presence of risk factors, but rather the accumulation of risk factors (Campelo et al, 2018; Carlsson et al., 2020; Simi, Sporer, & Bubolz, 2016) and what Vergani et al. (2018) describe as the push, pull and personal nature of the radicalisation process. The complexity of the process, therefore, can trigger a range of different vulnerabilities. Some of these vulnerabilities relate to a lack of sense of belonging (Harris-Hogan, 2014), which requires different institutional responses spanning the family, educational and work context, all of which contribute to the formation of a sense of identity (Kruglanski et al., 2019). The complexity and variability of the radicalisation process provides an opportunity for police to partner with various agencies and community groups to tackle radicalisation in a multifaceted manner. As such, multiagency interventions have become an important approach to tackle the problem of radicalisation and violent extremism (Butt & Tuck, 2014; Mucha, 2017; Sestoft et al., 2017). Existing evidence, however, does not provide a clear understanding of the effectiveness of police-involved, multiagency approaches to radicalisation (Cherney & Hartley, 2017; Koehler, 2017; MacDonald, 2002). In addition, there are no existing reviews of multiagency programmes, with police as partner, for addressing radicalisation to violence.1 Given the cost of forming multiagency interventions and the organisational complexities of managing and maintaining these types of responses, it is imperative to know whether current multiagency approaches that include police partners are effective for reducing radicalisation to violence and enhancing multiagency collaboration. Policy-makers, practitioners and researchers also need to understand not only whether the intervention works, but also how the intervention works (mechanisms), under what conditions or contexts (moderators), and what the implementation considerations and cost implications are. This review will fill a significant gap in the evidence-based literature for countering violent extremism in two ways. First, by quantitatively synthesising the existing evidence for the impact of multiagency police-involved programmes on reducing violent radicalisation or enhancing multiagency collaboration. Second, by qualitatively synthesising research that reports on the mechanisms, moderators, implementation considerations and economic information pertaining to police-involved multiagency programmes to counter radicalisation to violence. The results from this review will inform future decision-making regarding the design and evaluation of multiagency programmes by synthesising the evidence for their effectiveness, identifying potential gaps in the evidence-base and providing insight into what level of investment is required for the implementation and evaluation of primary studies. The first objective of this review (Objective 1) is to answer the question: how effective are multiagency interventions with police as a partner at reducing radicalisation to violence or improving multiagency collaboration? If there is sufficient data, the review will also examine whether the effectiveness of these interventions varies by the following factors: geographical location, target population, nature of the intervention approach (e.g., number of components, specific intervention techniques), and number and type of multiagency partners. The second objective of this review (Objective 2) is to qualitatively synthesise pertinent information about how the intervention might work (mechanisms), under what context or conditions the intervention operates (moderators), the implementation issues and economic considerations. To fulfil the objectives of this review, two types of studies will be included. The specific type of studies used to address each review objective may overlap, and are detailed in the subsections below. To be included in the review of effectiveness (Objective 1), a study must be a quantitative impact evaluation that employs a randomised experimental (e.g., randomised controlled trials) or a quasiexperimental design with a comparison group that does not receive the intervention. Eligible comparison groups are: “business-as-usual” treatment, no intervention, or an alternative intervention (treatment-treatment designs). Cross-over designs Regression discontinuity designs Designs using multivariate controls (e.g., multiple regression) Matched control group designs with or without preintervention baseline measures (propensity or statistically matched) Unmatched control group designs without preintervention measures where the control group has face validity Unmatched control group designs with prepost intervention measures which allow for difference-in-difference analysis Short interrupted time-series designs with control group (<25 pre- and 25 postintervention observations (Glass, 1997) Long interrupted time-series designs with or without a control group (≥25 pre- and postintervention observations (Glass, 1997) Less rigorous quasiexperimental designs can be used to illustrate the magnitude of the relationship between an intervention and an outcome, yet have limitations for establishing causality. Therefore, we will exclude the following weaker quasiexperimental designs in the synthesis of intervention effectiveness: Raw unadjusted correlational designs where the variation in the level of the intervention is compared to the variation in the level of the outcome; and Single group designs with pre- and postintervention measures. To be included in the qualitative synthesis of the potential mechanisms, moderators, implementation issues and economic considerations related to the intervention (Objective 2), each study must either be (a) already included in the quantitative synthesis impact evaluations (see above for review Objective 1); or (b) be an empirical study reporting on an eligible intervention. To be an empirical study, the authors must either be reporting on primary quantitative or qualitative data or conducting secondary analysis of primary quantitative or qualitative data. We acknowledge that qualitative studies may not be present “data” per se, but report on empirical work such as textual themes from key-informant interviews or focus groups, or information gathered by observational methods (e.g., participant-observers). Purely theoretical work, opinion pieces or research reports that only summarise, reference or describe previous intervention studies will not be used for the qualitative synthesis. However, this type of document may be harvested for relevant references. Individuals of any age, gender or ethnicity Micro places (e.g., street corners, buildings, police beats, street segments) Macro places (e.g., neighbourhoods, communities, police districts) Report on a multiagency intervention where police are a partner, defined as some kind of a strategy, technique, approach, activity, campaign, training, programme, directive or funding/organisational change that involves police and at least one other agency (Higginson, Eggins, Mazerolle, & Stanko, 2015). Police involvement is broadly defined as: Police initiation, development or leadership; Police are recipients of the intervention or the intervention is related, focused or targeted to police practices; or Delivery or implementation of the intervention by police. The other agencies or entities involved in the intervention may be government or nongovernmental agencies, including government agencies (such as education, immigration, customs, home affairs, employment, housing, health), local councils, businesses, communities (such as churches, mosques and other houses of worship), and services providers (such as resettlement agencies, local health providers). AND Report on a multiagency intervention with police as a partner that aims to address terrorism, violent extremism, or radicalisation to violence, as defined or specified by study authors. Police being trained OR police training or educating partner(s), to improve recognition, referral and responses to radicalisation, including guiding at risk populations towards numerous forms of support services offered by various partnerships, such as life skills mentoring, anger management sessions and cognitive/behavioural therapy (Home Office, 2015a). Community awareness programmes or training delivered to police OR police delivering community awareness training or programmes to partner(s) to help partner(s) identify someone who may already be engaged in illegal terrorist-related activity and are referred to the police (Home Office, 2009). Police working in partnership with universities to train, engage, intervene and consult on action plans to reduce at-risk youth to extremist messaging (Angus, 2016). Approaches that involve police working with other agencies to refer, assess, or case-manage individuals convicted of terrorism or identified as at risk of radicalisation (Cherney & Belton, 2019). Police partnering with other agencies to address radicalisation or extremism through regular structured/unstructured focus groups or meetings that may or may not be formalised (e.g., memoranda of understanding) or by forming task forces or multiagency intervention teams. Police working with external agencies to divert an individual away from violent extremism (e.g., UK Channel programme; Home Office, 2015b). Police officers undertaking various forms of engagement with different community and agency stakeholders to help identify terrorist threats (Innes et al., 2011; Ramiriz et al., 2013). For the review of effectiveness (Objective 1), we will include studies with two main categories of outcomes. The first is radicalisation to violence. For the purposes of this review, radicalisation to violence is defined as the process by which a person adopts extremist views and moves towards committing a violent act (Hardy, 2018; Jensen et al., 2018). It is important to note that “radicalisation” remains inconclusively defined in the literature (Heath-Kelly, 2013) and violence is just one potential outcome of radicalisation (Angus, 2016; Hafez & Mullins, 2015; Schmid, 2013). We also recognise that terminology (e.g., radicalisation and extremism) in the extant literature is often used interchangeably (Borum, 2012), and that outcomes may not be labelled explicitly as “radicalisation to violence”. Other labels that may be used include: radicalisation (Horgan, 2009), extremism, violent extremism (Khalil & Zeuthen, 2016), political violence, ideologically motivated violence, political extremism (Lafree, Jensen, James, & Safer-Lichtenstein, 2018), violent radicalisation (Bartlet & Miller, 2012) and terrorism (Christmann, 2012). Violent Extremist Risk Assessment-2: a risk assessment of the likelihood of violence by an offender who has been convicted of ideologically motivated violence (Pressman & Flockton, 2012). Extremist Risk Guidance Factors (ERG 22+): assesses the needs and risks of offenders who have either been convicted of an extremist offence or have shown behaviours or attitudes that raise concerns about their potential to commit extremist offences (Knudsen, 2020). IAT-8: Assesses the effectiveness of a current intervention at reducing or altering the level of vulnerability to radicalisation (RTI International, 2018). RADAR assessments: identifies individuals who would benefit from services to help them disengage from violent extremism by assessing a variety of observations including religious understanding and knowledge, radicalisation source, intervention goals and progress undertaken to achieve these goals (Cherney & Belton, 2019) Terrorist Radicalisation Assessment Protocol (TRAP-18): a professional judgement instrument for risk and threat assessment of individuals who may engage in lone-actor terrorism (Meloy, 2018). Information sharing (e.g., frequency, quality); Perceptions of trust, respect, or legitimacy within multiagency collaborations; or Degree of shared goals and understanding between multiagency partners. To be included in the qualitative synthesis of the potential mechanisms, moderators, implementation issues and economic considerations (Objective 2), no specific outcome measures are required. Specifically, any empirical study of a police-involved multiagency programme that aims to address terrorism, violent extremism or radicalisation to violence will be examined for qualitative or quantitative data pertaining to potential mechanisms, moderators, implementation or economic considerations (see Supporting Information Appendix C for definitions). We note the differences in the conceptualisation of “outcomes” for quantitative and qualitative studies, whereby qualitative studies may not distinguish between different types of variables such as independent, predictor, outcome, moderator, or mediator variables. Rather, qualitative studies are likely to present thematic textual data drawn from interviews, focus groups or observational methods. In addition, study authors may use mechanism, moderator, implementation, and economic variables as outcome variables or they may use data within these domains as mediators or moderators to explore their impact on study outcomes. To provide a comprehensive synthesis of potential mechanisms, moderators, implementation issues, and economic considerations, we will include empirical studies that report on data in any of these domains, regardless of whether the data is conceptualised as an “outcome variable”. For both the review of effectiveness (Objective 1) and the review of mechanisms, moderators, implementation and economic considerations (Objective 2), we will include studies with follow-up periods of any length. If there is variation in the length of follow-up across studies, we will group and synthesise studies with comparable follow-up durations. For example, short (e.g., 0–3 months postintervention), medium (>3, <6 months), and long-term follow-up (>6 months postintervention). For both the review of effectiveness (Objective 1) and the review of mechanisms, moderators, implementation and economic considerations (Objective 2), we will include studies that report on an impact evaluation of an eligible intervention using eligible participants, outcome(s) and an eligible research design in any setting. All geographical regions will be included in the review (i.e., high-, middle- and low-income countries). If we identify a range of conceptually distinct settings, we will synthesise the studies within the settings separately. We will include studies written in any language. At the title and abstract screening stage, we will use Google Translate to identify whether a non-English language study is potentially eligible for review. If eligible after title and abstract screening, we will again use Google Translate using multiple sections of the document to determine its eligibility for the review. If eligibility cannot be unequivocally determined with this approach, we will call upon our international network of colleagues for assistance with full-text screening and coding if required. We will include studies disseminated between 2002 and 2018 in the review. We will use a common search strategy for the review of effectiveness (Objective 1) and the review of mechanisms, moderators, implementation and economic considerations (Objective 2). The search for this review will be led by the Global Policing Database (GPD) research team at the University of Queensland (Elizabeth Eggins and Lorraine Mazerolle) and Queensland University of Technology (Angela Higginson). The University of Queensland is home to the GPD (see http://www.gpd.uq.edu.au), which will serve as the main search location for this review. The GPD is a web-based and searchable database designed to capture all published and unpublished experimental and quasiexperimental evaluations of policing interventions conducted since 1950. There are no restrictions on the type of policing technique, type of outcome measure or language of the research (Higginson et al., 2015). The GPD is compiled using systematic search and screening techniques, which are reported in Higginson et al. (2015) and summarised in Supporting Information Appendices A and B. Broadly, the GPD search protocol includes an extensive range of search locations to ensure that both published and unpublished research is captured across criminology and allied disciplines. The GPD systematic search uses a broad range of policing and research search terms and systematically progresses the screening of the captured research in sequential stages with increasing specificity. At the initial title and abstract screening stage, records identified by the systematic search are screened on whether they are broadly about police or policing (see Higginson et al., 2015). At subsequent full-text screening stages, documents for records screened as being potentially about police or policing are then screened on whether they report on a quantitative impact evaluation of an intervention relating to police or policing, with no limits on outcome measures. As a result, refined corpuses of policing research can be searched and extracted from the GPD without the need to use policing search terms. Because our review will capture both quantitative and qualitative studies of eligible interventions, we will extract data from the GPD from the point of title and abstract eligibility (i.e., is the document broadly about police or policing). We will search the title and abstracts within this corpus between 2002 and 2018 using the following search terms: *terror* OR extrem* OR *radical*. Searching trial registries (as listed by WHO https://www.who.int/ictrp/network/primary/en/) and Office for Human Research Protections (https://www.hhs.gov/ohrp/international/clinical-trial-registries/index.html); Searching counterterrorism organisation websites (see Table 1); Conducting reference harvesting on both the corpus of eligible documents and existing reviews; Forward citation searching for all eligible documents; Liaising with the Five Country Research and Development Network (5RD), and the Department of Homeland Security Advisory Board network for the Campbell Collaboration grants, to enquire about eligible studies that may not be publicly available; Personally contacting prominent scholars in the field and authors of eligible studies to enquire about eligible studies not yet disseminated or published; and Hand-searching the 12-months prior to December 2018 for the following journals to identify eligible documents yet to be indexed in academic databases: Critical Studies on Terrorism Dynamics of Asymmetric Conflict Intelligence and Counter Terrorism International Journal of Conflict and Violence Journal for Deradicalization Journal of Policing Perspectives on Terrorism Police Quarterly Policing—An International Journal of Police Strategies and Management, Policing and Society Sciences of Terrorism and Political Aggression Studies in Conflict and Terrorism Terrorism and Political Violence Existing literature highlights a growing range of police multiagency approaches that address terrorism, radicalisation, and/or extremism (e.g., see Butt & Tuck, 2014; Wise, Roberts, Formosa, & Chan, 2018). We anticipate that studies captured under the review of intervention effectiveness (Objective 1) will utilise quasiexperimental research designs (e.g., Williams, Horgan, & Evans, 2016). We anticipate that studies captured under the review of mechanisms, moderators, implementation and economic considerations (Objective 2) will utilise quantitative and qualitative data collected and presented in a range of formats, such as analysis of implementation documentation, interviews or focus groups, case studies or cost analyses (e.g., Butt & Tuck, 2014; Mabrey, Hepner, & Ward, 2006; Mastroe, 2016). For studies included in the synthesis of intervention effectiveness (Objective 1), risk of bias will be evaluated using either the Cochrane randomised or nonrandomised risk of bias tools, depending on the study methodology (Higgins et al., 2019). Using these tools, studies will be rated across domains as having high, low or unclear risk of bias. Study authors will be approached to obtain missing data where a domain is rated as “unclear”. Results of the risk of bias assessment will be presented in summary tables and in a risk of bias summary figure. Depending on available data, sensitivity analysis w