Proceedings, 2021, GSE 2020

Countering violent extremism (CVE) implies many pressing ethical issues. For policymakers and professionals, it is essential to identify concrete ethical dilemmas and to understand the underlying more abstract ethical issues. The proposed typology of ethical issues consists of four different levels: the structural, the political, the professional, and the personal levels. Ethical issues on the structural level are rooted in the conditions of the world risk society, such as the morality of CVE and counterterrorism in general or its fundamental inconsistencies. On the political level, the phenomenon of the state of emergency, the politicization of CVE and counterterrorism, the legitimacy of state interventions, and the tension between secrecy and transparency all play a role. Professional values can conflict with organizational interests on the professional level. Finally, on the personal level, integrity can come under pressure due to conflicting values. An exploration of the benefits for counterterrorism practitioners of the consequentialist, deontological, and virtue-ethical approaches concludes that none of them offers a sound approach to the practice of counterterrorism in a liquid world risk society. This provides an opportunity to highlight the potential of the philosophical concept of “compromise” (Benjamin) in reconciling opposing principles and underlying values of the key ethical approaches. A major source of guidance and inspiration for compromise is phronesis, or practical wisdom. Another powerful concept is the ethics triangle (Svara). This ethics triangle implies that public administrators should strive towards a balance of virtue, principle, and good consequences, all seen from the perspective of the duty of the public interest. There are several ethics tools available that can inform the design of CVE programs and support dealing with the various types of ethical issues once a CVE program is started. First, it is useful to engage in a closer analysis and value clarification of the ethical issues at stake. Second, the already established tool of privacy impact assessment represents an inspirational point of departure for a broader ethics impact assessment. Third, the institutionalization of ethics support in the field of CVE can be realized by the establishment of ethics committees, the appointment of ethics advisors or ethics rapporteurs, or the installation of reflection groups. Fourth, the implementation of tools of ethics support like moral case deliberation can empower CVE professionals to deal with ethical issues. It can also improve their decision making, support collective learning, and contribute to the development of policies and guidelines in the field of CVE and counterterrorism. Last, but not least, evaluating CVE programs can benefit from taking ethical issues into account. Moving beyond mandatory rituals of ethics checks can lead to an in-depth engagement with professionals and stakeholders within CVE programs and a re-adjustment and fine-tuning of programs if necessary. In the long run, the integration of ethics into the design, implementation, and evaluation of CVE programs can strengthen the legitimacy of these programs among targeted communities, professionals, and societies at large. Full article

P/CVE is a field of research and practice that focuses on complex social programs that, similarly to the field of public health, are most often multidisciplinary in nature. In the proposed program evaluation framework, adapted from the field of public health, we adopt a definition of “program” to include any type of organized activity (i.e., direct service interventions, community mobilization efforts, research initiatives, policy development activities, communication campaigns, etc.). The proposed framework includes three levels of evaluation, six implementation steps, and ten criteria for determining the quality of the evaluation. The three evaluation levels are based on the unit of analysis: system, population, or individual. System-level evaluations are studies conducted at the system level, where data are gathered to understand how an organizational model is developed or implemented (i.e., What type of inter-agency functions need to be developed to enhance the effectiveness of CVE programs?). In population-level evaluations, data are gathered from a segment of the population to understand the impact of the intervention on a group of individuals (i.e., Do school campaigns increase awareness about violent extremism? Do educational interventions directed to law enforcement enhance their ability to respond to cases of violent extremism?). Finally, there are programs executed at the individual level; evaluations of this type may be focused on what worked to reintegrate a specific individual—as such, the unit of analysis is at the individual level (i.e., What type of cognitive therapy interventions work in supporting reintegration efforts? What type of relationships with family and friends are necessary to support the individual in reintegrating into society?). The six implementation steps proposed in the framework are designed to guide the evaluator through the process of evidence gathering: (i) stakeholders’ engagement; (ii) description of the program and the context in which it is implemented; (iii) selection of the study design and framing of how activities are expected to be related to desired outcomes; (iv) collection of data; (v) interpretation of results; and (vi) delivering of results to stakeholders. Finally, the proposed 10 standards are based on questions we should ask to determine the quality of the evaluation. We determined that it is important to understand whether the evaluation is useful, context-specific, ethical, accurate, feasible, independent, culturally sensitive, systemic oriented, and transferable, and to include a counterfactual situation (what-if scenario). Full article

When evaluating the effectiveness of public health, and behavioral and medical interventions, various methods can be used to systematically judge the merit, worth, value, and significance of the program’s outcomes. This presentation will describe what is meant by “evaluation approaches” by providing examples of the characteristics of several different types of approaches, such as, the appreciative inquiry, beneficiary assessment and case study approaches. The presentation will then focus on introducing the developmental evaluation approach by comparing it to the traditional evaluation approach in terms of its purpose, the evaluator’s role, accountability, measurement, and results. Examples of the causal inference pathways underlying the developmental and traditional approaches will illustrate how developmental approaches can accommodate multivariable and multivariate models, this can be used to evaluate the intricacies of complex real-world problems and programs. Full article

The nominal group technique (NGT) was developed in the 1970s as a structured brainstorming and multi-stage consensus-building process to solicit feedback from a group of stakeholders on a given topic. The approach was intended as an evaluation method to provide semi-quantitative rank-ordered feedback from group participants. This lecture presents a variation of this technique created by the Harvard team, namely “scenario-based NGT”. The format proposed includes elements of tabletop exercises (scenario with a timeline and discussion phase) as well as elements of traditional NGTs (silent brainstorming and ranking). The technique we developed was based on a case-study approach (“scenario”) which we then tested in two countries (USA and Sweden) with existing P/CVE initiatives at different stages of development. We conducted scenario-based NGT sessions in both locations and then systematically analyzed the results using iterative qualitative coding based on a common framework. Results were analyzed to achieve consensus on the most common system-level challenges and system-level functions, necessary to overcome those challenges, in each location. Full article

In the field of counter-terrorism (CT) and countering violent extremism (CVE), policymakers are in constant need of accurate data to make informed decisions to support existing programs and develop new approaches to prevent radicalization to violence. The goal of the comparative analysis in this presentation is to identify the types of data needed to assess the impact of CT and CVE programs based on each country’s policy goals. A comparative analysis of the five countries’ specific CT/CVE policies was conducted to identify common themes and data needs. The first most widely discussed theme is the need to maintain and expand collaborations and information sharing across countries—all five policies strongly emphasize the importance of such collaborative efforts. All policies address the need for strengthening collaborations at the local level, considering the important role civil society plays in the frontline response to violent extremism. In particular, the North Macedonian policy recognizes the need to fully engage in multidisciplinary interagency efforts that include civil society in the process for reconciliation of ethnic and cultural divides, educate and promote democratic values in schools and faith based communities. According to the policy documents, it can be found that there is a need for a better understanding of what types of collaborative efforts and partnerships are needed to establish effective CT and CVE programs. All policies stress the need to address a range of extremist ideologies including Jihadist, Far Left, and Far Right groups to address radicalization in the online space as well as through in-person interventions. In terms of interventions, there is a need to understand what type of training is most effective to equip frontline professionals with the knowledge and skills to intervene when individuals engaged in VE come to their attention. The United States policy is innovative with respect to the others because it introduces the concept of targeted violence. By doing so, it recognizes the importance of including situations where ideology is not a motivating factor or the motivations are unknown behind the acts of violence. The Swedish policy is distinguished by its detailed legislation supporting the prevention of terrorist acts. The UK policy emphasizes the need to contrast ideologies and views that are not aligned with UK values. All policies recognize the need for evidence on strategic efficacy and recognize the fact that programs and policies have been widely implemented without scientific proof of their effectiveness. In particular, the Canadian policy points to the need for identifying best practices that can be transferred from case to case or country to country. As an area of policy improvement across countries, there is certainly a lack of clarity on the roles and responsibilities of the many agencies that may be potentially involved in prevention efforts, still leaving a nebulous space in terms of when and how security intercepts social work and public health. Full article

This presentation introduces workshop participants to the role of developmental evaluation in multi-sector collaborative efforts to detect and displace vulnerabilities to the risk of violent extremism. Our discussion shares the journey that the Ottawa Police Service took to mobilize a wide variety of human service partners under a shared commitment to work upstream—where there is an opportunity to support individuals and families during early signs of vulnerability to radicalization. The result of this collaboration is the Ottawa Vulnerability Reduction Protocol: a multi-sector prevention tool aimed at identifying opportunities to support individuals before vulnerability for risk of radicalization to violence occurs. The protocol helps human service professionals, who may not have expertise in violent extremism, to develop the capacity, opportunity, and confidence to contribute towards upstream solutions. Participants of test simulations describe the protocol as allowing for support to be mobilized without labeling and stigmatizing an individual (i.e., radical, extremist), while also requiring the types of support that most human service professionals are comfortable providing. Construction of the protocol and accompanying resources (e.g., Vulnerability Detection Tool, Shared Needs Assessment, Integrated Support Plan) was supported through a developmental evaluation framework. By incorporating the continuous collection of data, ongoing stakeholder feedback, and insight from the preventing and countering violent extremism field, the evaluator assisted community partners in determining a structure and approach to building the protocol. This developmental evaluative approach helped diverse human sectors to navigate multiple perspectives and identify a shared pathway forward. Full article

Many existing programs for countering violent extremism focus on either end of the radicalization spectrum. On one hand are prevention programs aimed at deterring individuals from starting down the path to violent extremism. On the other hand are disengagement/de-radicalization programs designed for assisting individuals who have been fully radicalized. Conspicuously absent are programs for those who fall in-between, into what might be referred to as the pre-criminal space: individuals who have begun to exhibit signs of radicalization, but for whom radicalization is not yet complete. The British Columbia Shift (BC Shift) initiative was created to assist individuals determined to be in this pre-criminal space; that is, those deemed to be in danger of radicalizing. The goal of BC Shift is to stop individuals from traveling further down the path of radicalization, and, ideally, to turn individuals away from the path. BC Shift operates as a navigational model, connecting at-risk individuals with services and supports in the community. BC Shift is a government initiative supported by the Canada Centre for Community Engagement and Prevention of Violence. It is a civilian organization that partners very closely with, but is separate from, law enforcement. In addition to its primary CRVE mandate, BC Shift has rapidly evolved and expanded into several other responsibilities, including coordination on national CVE standards; liaising with other CVE programs across Canada; maintaining stakeholder relationships; and helping create capacity through dialog and training. Although the program only began accepting referrals in 2019, its operation has already revealed many important lessons for CRVE programs. First, it is critically important to have the right people in the room. There has to be buy-in from the highest levels of partner agencies and stakeholders, particularly early on. Second, programs of this sort should leverage existing resources wherever possible. BC Shift has been lucky enough to coordinate with situation tables, such as the CHART program in Surrey. There are already many organizations doing excellent work in their respective communities; it is very helpful to plug into those resources. Third, even though BC Shift operates as a navigational hub, it has benefitted greatly from having a social worker as part of the team. This skill set is important in helping referred individuals feel comfortable with the process of accessing services and supports. Finally, marketing matters! CRVE programs such as BC Shift have to navigate a complex reality. The very concept of violent extremism is disconcerting to a lot of people in the community; these fears have to be addressed, and difficulties related to differences in perspective and language have to be overcome. BC Shift’s first year-and-a-half of operation has also highlighted several issues that have not yet been satisfactorily resolved. There is, for example, the “low hanging fruit” problem; agencies are typically referring less severe cases. Trying to get agencies to refer more serious cases has proved challenging. We hope that, by outlining these lessons and issues, this presentation proves to be useful to other CRVE initiatives. Full article

Since the 1970s, the Dutch anti-radicalization strategy has been characterized by a close collaboration between police, welfare providers, community groups, and national and local governments. This comprehensive approach aims to identify individuals at risk of radicalization at an early stage, offering them alternative life paths and robustly responding to imminent threats. The comprehensive approach generally enjoys considerable support among professionals, politicians, and the populace at large. However, a swelling tide is asking for more evidence that this combination of “soft” and “hard” interventions is working. Parliament is asking the government to take more measures against emerging extremist threats, but also demands more proof that these measures are effective. Liberal parties in parliament are critical of the financial and societal costs of counter-terrorism and anti-radicalization measures, while right-wing parties worry the government is not tough enough. Over the past five years, the Dutch government has commissioned multiple evaluations. Specific incidents, such as the shooting in Utrecht on 18 March 2019, have been extensively evaluated. A more general review in 2015 examined the counter-terrorism policy as a whole, finding that the ideas for the comprehensive approach were sound; yet, that the investment of ministries in this approach fell apart in the perceived low-threat period, making the intended “comprehensive” strategy become a pick-and-choice approach. Currently, the national and local governments involved in anti-radicalization efforts keep asking for more evaluations. The Ministry of Social Affairs created an evaluation toolkit, which local governments can use to evaluate interventions. The City of Arnhem commissioned a review of its anti-radicalization efforts, which is mainly able to assess the plans and processes, but not yet the outputs of the interventions. Moving forward, more collaboration may be needed within or even between European countries to amass sufficient data to enable full effect evaluations. Full article

Research on mental health and violence among marginalized communities has identified the importance of engaging communities, diminishing stigma, addressing multiple outcomes including strengths, and building social connections. Within the United States, Countering Violent Extremism (CVE) policies and programs have been criticized for failing in these areas. Recent efforts have sought to build multidisciplinary teams for the prevention of targeted violence and terrorism that explicitly seek to address these critiques and work to build the capacity of multidisciplinary providers to work with youth at risk for targeted violence and terrorism. Community Connect was a Boston, US-based community-based program that worked with youth at risk of violence, including ideologically-based violence. This program achieved broad community buy-in and successfully linked referred youth to a broad range of services in their communities. To bring the program to scale, an adaptation of Community Connect was developed that accepted referrals from a regional federally-convened threat assessment team, the Massachusetts Bay Threat Assessment Team (MassBayTAT). This multidisciplinary services team (MDST) maintains several essential functions from Community Connect, such as providing a thorough psychosocial assessment and maintaining regular contact and coordination between diverse providers, as well as making key changes to accommodate a regional scope. Given the nascent state of the field, both formative as well as summative evaluations play important roles in shaping and evaluating multidisciplinary violence prevention teams, as is evident in the iterative adaptation of the above-described multidisciplinary approaches. Evaluation of a multidisciplinary team for VE should assess both team development as well as case outcomes. Building trust within a community of diverse providers and disciplines and achieving a ‘whole of society’ approach to violence prevention is in and of itself an outcome that should be sought, as well as a reduction in violence at the individual level. Mixed-methods evaluations are needed to capture both the process and outcomes that are central to an effective multidisciplinary team for the prevention of terrorism and targeted violence. Full article

The place of clinical, medical, or health professional interventions in addressing violent radicalization is a topic of ongoing debate. Although violent radicalization is primarily a social phenomenon with significant psychological dimensions, the high prevalence of mental health “issues” and past psychiatric diagnosis in lone actors suggests that it may be useful to distinguish socialized actors who have strong ties to structured extremist organizations from relatively socially isolated actors who claim, and even boast about, virtual affiliation to extremist groups. For the latter, the potential efficacy of mental health interventions should be considered. However, because of the risk of profiling, stigmatization of minorities, pathologizing social dissent, and resistance, clinical intervention may cause harm and should be carefully evaluated. Until the effectiveness of clinical interventions in reducing radical violence is improved through evaluative research, exchanges about existing clinical models can be useful to support practitioners in the field and provide initial insights about good and potentially harmful practices. The Quebec model of clinical services to mitigate violent radicalization (secondary and tertiary prevention) is structured around three pillars: multiple access points to facilitate outreach and decrease stigma; specialized teams to assess and formulate treatment plans based on existing best evidence in forensic, social, and cultural psychiatry; and collaborative involvement with primary care services, such as community mental health, education, and youth protection institutions, which are in charge of social integration and long-term management. Beyond the initial assessment, the program offers psychotherapy and/or psychiatric interventions services, including mentorship to foster clients’ social integration and life-skill development. Artistic programs offering a semi-structured, nonjudgmental environment, thus fostering self-expression and creativity, are very well received by youth. A multimedia pilot program involving young artists has been shown to provide them with alternative means of expressing their dissent. Three years on from its inception, the preliminary evaluation of the Quebec clinical model by its partners and clinicians suggests that it could be considered a promising approach to address the specific challenges of individuals who present as potential lone actors at high risk of violent radicalization. The model does not, however, appear to reach many members of extremist groups who do not present individual vulnerabilities. While initial signs are positive, a rigorous evaluation is warranted to establish the short, medium, and long-term efficacy of the model, and to eventually identify the key elements which may be transferable to other clinical settings. In 2020, a five-year evaluative research project began to examine these questions. It is important to consider that any intervention can be harmful if due attention is not paid to structural discrimination and violence stemming from associated marginalization and exclusion. Clinical care can in no way replace social justice, equity, and human rights—all key pillars in primary prevention against violent radicalization. In the meantime, however, providing empathy and care in the face of despair and rage may prove most beneficial in decreasing the risk of violent acts. Full article

In recent years, the fight against terrorism and political violence has focused more on anticipating the threats that they pose. Therefore, early detection of ideas by local professionals has become an important part of the preventive approach in countering radicalization. Frontline workers who operate in the arteries of society are encouraged to identify processes towards violent behavior at an early stage. To date, however, little is known about how these professionals take on this screening task at their own discretion. Research from the Netherlands suggests that subjective assessment appears to exist. This is due to the absence of a clear norm for preliminary judgments. However, such an approach affects prejudice or administrative arbitrariness, which may cause side effects due to unjustified profiling. The publications about the Dutch case are inspired by the concept of “performativity”, (de Graaf, B., & de Graaff, B. G. J. (2010). Bringing politics back in: The introduction of the ‘performative power’ of counterterrorism. Critical Studies on Terrorism, 3(2), pp. 261–275.) which points to a distinct relationship between the performative power of counterterrorism instruments and the effectiveness of the local approach. Performativity contends that the overall effect of the policy in question is not necessarily determined by the policy measures and their intended results, as such, but more by the way in which they are presented and perceived. This means that, in order to create an equitable approach, governments, whether local or national, should focus more on the actual practice performed by frontline practitioners. The focus on practices is part of a larger project, entitled ‘Gatekeepers of Justice’ (See: https://www.internationalhu.com/research/access-to-justice), by the Research Group Access2Justice (Research Centre of Social Innovation at Utrecht University of Applied Science), led by professor Quirine Eijkman, Deputy President of the Netherlands Institute for Human Rights. Full article

In the Nordic countries, prevention of radicalization and violent extremism is based on an existing crime prevention collaboration. The core tenet of crime prevention approaches is that early radicalization prevention is best organized as a joint effort, where individual cases are assessed holistically and relevant information shared. This presentation is based on Nordic comparative research that provides a critical analysis of policies, perceptions, and practices regarding multiagency approaches. At the city level, we systematically explore how core components of a Nordic governance model contribute to and might be preconditions for effective multiagency collaboration and secure societies. Variations in these city-level approaches lead to an important question: do these various multiagency approaches to countering violent extremisms and radicalization constitute a unified model? The project provides an informed platform for spreading Nordic experiences and models of governance to other countries in the area of radicalization prevention and countering violent extremism. The multiagency collaboration in Nordic cities tends to be composed of a smaller number of professionals, e.g., three to eight, representing social and health services, a coordinator, and the police. Their positions vary; some represent leadership in the municipality, others are positioned at more mid-level management, and others are frontline professionals handling clients on a daily basis. The origin of these collaborative efforts is based on the School, Social and health services, and Police framework (SSP) in Denmark, Norway, and the Ankkuri group in Finland. We assert that future efforts can also be more need-based, comprising municipal/local initiatives with an element of the guiding national policies, and include the legal frameworks to guide professionals on ethical issues like confidentiality. This presentation is part of a larger project entitled “Nordic Multiagency Approaches to Handling Extremism: Policies, Perceptions, and Practices”, led by professor and center director Tore Bjørgo (University of Oslo) and is funded by NordForsk (Nordic Council of Minister), 2018–2021. Full article

This presentation outlines the results of the primary programmatic evaluation efforts the Emergency Preparedness Research Evaluation and Practice (EPREP) Program has conducted since 2016. The presentation begins with an overview of the methodology of selecting outcome measures to evaluate program efficacy, as well as a description of the evaluation framework. Results of the longitudinal and quasi-experimental 2017 evaluation of the Online4Good Academy—on of the training events at the focus of the Boston CVE Pilot Program—are presented and discussed. In 2018, the EPREP Program utilized a longitudinal and quasi-experimental design to evaluate the efficacy of the Peer2Peer antihatred campaign Kombat with Kindness. Results and implications from this study are discussed. The final portion of the presentation describes the more recent activities of the EPREP Program—an evaluation of the online safety program Operation 250 (OP250). This portion of the lecture describes the psychological framework and theory of change under which OP250 implements their initiative. During the final segment we also present the preliminary results of a randomized controlled trial conducted at two different study sites in Massachusetts designed to evaluate the programs’ efficacy. Full article

This presentation summarizes a register-based study on women who have been identified as belonging to three violent extremist milieus in Sweden: violent Islamic, violent far-right, and violent far-left extremism. We studied the women in these milieus along a number of analytical dimensions, ranging from demographic and educational to criminal background and network relationships, and compared them to three reference groups: (i) non-extremist biological sisters to female extremists in the study population; (ii) men in the respective extremist milieus; and (iii) female members of other antagonistic milieus such as organized crime. Our results showed that there are both similarities and differences between groups. In some cases, like age and region of birth, there are commonalities between violent far-right and violent far-left women. Regarding region of birth and migration background, women affiliated to violent far-right and violent far-left extremism are predominantly born in Sweden. Women affiliated to violent Islamic extremism tend to be born in Sweden to a greater extent than men in the same milieu, but to a much lesser degree than women in the violent far-right and violent far-left. When it comes to education, women in the violent Islamic milieu are closer to women in violent far-right extremism. Women in violent far-left extremism perform best at school, with consistently higher grades. The average score of women in violent far-left extremism is identical to that of their sisters, and women in violent far-left extremism perform on average substantially better than men in the same milieu. Women in violent Islamic extremism, in contrast, perform on average similarly to men in violent far-left extremism, and they perform better than their biological sisters. Regarding labor market attachment, violent Islamic extremists have the weakest attachment and the highest dependency upon financial assistance as well as a low employment share (36 percent in 2016), but also a relatively high share of individuals with a high number of unemployment days, suggesting that women in violent Islamic extremism experience higher social exclusion. We find the highest employment share among women in violent far-left extremism, where 89 percent are gainfully employed in 2016 (80 percent for at least three of the last five years) and about a 20 percent unemployment share. Men in violent far-left extremism have an employment share around 10 percent below that of the women in far-left extremism for 2016. The highest fractions of individuals that have not been in contact with the health system due to mental disorders are among violent Islamic extremism, with the women’s fraction at 84 percent, compared to their non-extremist sisters and men in the same milieu that are just above 79 percent. Women in violent far-left extremism have the highest share of in-patient major mental disorders among the extremist milieus (3 percent), higher than men in the same milieu (less than 1 percent) as well as than women and their sisters in the other categories. During the period 2007–2016, 68 percent of individuals in the extremist milieus are covered by the register of suspected individuals. The coverage is substantially higher for men, 72 percent than for women, 43 percent. Compared to their sisters, women in all three milieus are criminally active to a much higher extent. However, women in all three milieus are less criminally active than women in other antagonistic milieus, among whom 67 percent have been suspected at least once. In all three milieus, the share of men with a criminal record is about twice as large as that of women. As far as the gender aspect is concerned, we know that extremist milieus generally have a conservative view of the role of women in society. In our results, this is reflected in the low rates of crime in women compared to men, and relatively marginal positions in the co-offending networks. The fact that women in violent far-left extremism have stronger positions in their networks than the other women in the study population is expected, given that the ideology of this milieu allows for greater equality. This means that women in violent far-left extremism participate more often than, e.g., women in violent far-right extremism, in political actions where violence is common. This pattern of gender roles and criminal involvement also holds concerning women in violent Islamic extremism. This milieu has a more traditional view of the role of women than views among even violent far-right extremists. Women in violent Islamic extremism are less involved in crime and, in particular, violent crime. Full article

In this presentation, we draw a parallel between public health screening processes of asymptomatic individuals to detect a disease (i.e., cancer) and the attempt of identifying individuals who may be classified as being “at risk” of conducting a violent act. We discuss concepts related to the early detection of a condition through screening. In particular, we focus on the underlying history of the disease over time, as it moves through different stages: onset, start of asymptomatic detectability, start of symptomatic detectability, and death. Notably, such description refers to the latent (possible) insurgence of the disease and to its would-be progression if left undetected and untreated. As part of this framework, we discuss the relevant quantities that are involved when one performs a screening test. These include the probabilities of true and false positive and negative test results and the area under the curve (AUC) index. A positive test result is typically defined through the crossing of a threshold by some measured index. The behavior of the probabilities that describe the features of the test procedure as one changes the threshold that defines a positive result can be illustrated through an online application that was developed for this presentation, available at https://marcobonetti.shinyapps.io/shinyapp/. Other quantities that are discussed are the negative predictive value (NPV) and the positive predictive value (PPV). The latter quantity represents the probability that a subject is truly a case when the test result is positive. The PPV may be surprisingly low, especially if the disease has a low prevalence in the population. We then discuss some modern methods for screening through the analysis of online data streams, either through active or passive data collection, the difference being that in the first kind, the subject is active in the data collection process; in the second kind, secondary data streams are used, such as Twitter tweets, Facebook posts, or Google searches. Machine learning algorithms allow the study of phenomena at the population level (e.g., trends in a phenomenon of interest), but also the prediction of features for individual subjects. Two specific examples are discussed, both based on the use of Facebook data. The first example deals with the prediction of a depression diagnosis from Facebook language, with a comparison of the results of the automatic classification algorithm and the true (known) depression diagnoses of a set of consenting subjects. The second example is a Facebook’s suicide prevention program, which has been in place for over 10 years. The program identifies individuals that, based on their posts, comments and live videos, may be at risk of committing suicide. Full article

In the last decade, advances in statistical modeling and computer science have boosted the production of machine-produced contents in different fields: from language to image generation, the quality of the generated outputs is remarkably high, sometimes better than those produced by a human being. Modern technological advances such as OpenAI’s GPT-2 (and recently GPT-3) permit automated systems to dramatically alter reality with synthetic outputs so that humans are not able to distinguish the real copy from its counteracts. An example is given by an article entirely written by GPT-2, but many other examples exist. In the field of computer vision, Nvidia’s Generative Adversarial Network, commonly known as StyleGAN (Karras et al. 2018), has become the de facto reference point for the production of a huge amount of fake human face portraits; additionally, recent algorithms were developed to create both musical scores and mathematical formulas. This presentation aims to stimulate participants on the state-of-the-art results in this field: we will cover both GANs and language modeling with recent applications. The novelty here is that we apply a transformer-based machine learning technique, namely RoBerta (Liu et al. 2019), to the detection of human-produced versus machine-produced text concerning fake news detection. RoBerta is a recent algorithm that is based on the well-known Bidirectional Encoder Representations from Transformers algorithm, known as BERT (Devlin et al. 2018); this is a bi-directional transformer used for natural language processing developed by Google and pre-trained over a huge amount of unlabeled textual data to learn embeddings. We will then use these representations as an input of our classifier to detect real vs. machine-produced text. The application is demonstrated in the presentation. Full article

Academics and policymakers recognize the absence of empirically grounded research to support the suppositions on which terrorist focused policies are based. (Sageman, Marc. 2014. “The Stagnation in Terrorism Research”. Terrorism and Political Violence 26 (4): 565–80) We developed our project, Mapping Networks and Narratives of Online Right-Wing Extremists in New South Wales, (Department of Security Studies and Criminology. 2020. Mapping Networks and Narratives of Online Right-Wing Extremists in New South Wales. https://0-doi-org.brum.beds.ac.uk/10.5281/zenodo.4071472) to illuminate this space. Using the analysis of large-scale online data to generate evidence-based insights into online Right-Wing Extremism (RWE) across the state, our research focused on four key questions: (1) What is the nature of the online RWE environment in New South Wales, Australia (NSW)? (2) How is this movement distributed across NSW? (3) How are themes and narratives framed in different online contexts to mobilize support? (4) What level of risk does the online right-wing environment pose? These questions were left purposely broad to facilitate an exploratory project into what was, in 2018–2019, still a relatively little studied milieu. We combined expertise from computational science, security studies, and behavioral science. We were funded by the Department of Communities and Justice, NSW. We identified two distinct—yet connected—levels of risk. The first was a creeping threat to democracy fueled by networks and content that challenged the fundamental principles of pluralistic liberal democracy. The second was a risk of violence perpetrated by individuals and/or groups that advocate and/or support the use of violence as a tactic to achieve an ideological end. The communities we examined were primarily characterized by networks of individuals as opposed to formal groups. The role played by individual influencers has important ramifications for policy communities: attention should be paid to issues of proscription and moderation. While this milieu engaged with Australian issues and events, it was notably far more obsessed with American issues: particularly those focused on populist narratives and Trumpism. Despite being hateful and extreme, online RWE communities are, firstly, spaces of sociability for users, where social networks are maintained by shared values and norms. For those involved, these spaces engender positive experiences: individuals might share an image of their dinner cooking in their kitchen interspersed with “shitposting” and virulent hate speech. While we identified a variety of narratives that focused on the delegitimization of government and dehumanization of others, the central theme was that of “white identity under threat”. We observed five distinct stages of moderation approach and echo chamber strength. A series of issues for future consideration were identified from the analysis: (1) Awareness raising for key stakeholders across different levels of government and civil society about the revolutionary and anti-social agenda of RWE communities. (2) Building awareness about the civic underpinnings of representative liberal democracy and the threat that RWE poses. (3) Expanding current Countering Violent Extremism infrastructure provided by the NSW government to individuals and communities vulnerable to right wing extremism. (4) The local government is well positioned to deliver programs in rural communities impacted by RWE. (5) Upskilling front-line workers to recognize the risks associated with RWE, and providing pathways into CVE intervention programs for individuals identified as being at-risk. Full article

Recent interest in terrorist risk assessment and rehabilitation reveals that the likelihood and risk factors for terrorist disengagement, re-engagement, and recidivism are poorly understood. In this presentation, I review related literature on criminal desistance, disaffiliation from new religious movements, commitment, and turnover in traditional work organizations, role exit, and the investment model to develop a series of theoretical starting points for gauging the likelihood and predictors of risk, which can help inform evaluation efforts. I then highlight key findings from the existing literature on terrorist disengagement and re-engagement/recidivism as well as key differences across samples and the methodological challenges associated with such research—mainly the absence of control groups, relatively small sample sizes, the need for a lengthy time horizon, and inconsistencies in what constitutes re-engagement and recidivism. Then, using data collected on 185 terrorist engagement events for 85 individuals representing over 70 unique terrorist groups, I present my and my colleagues’ findings on the drivers of terrorist disengagement and re-engagement. We find that terrorist disengagement is a lengthy process more commonly driven by “push” rather than “pull” factors, specifically disillusionment with the strategy or actions of the terrorist group, disillusionment with leaders or other members, disillusionment with one’s day-to-day tasks, burnout, difficulty living a clandestine lifestyle, difficulty coping with attacks, and psychological distress. Importantly, “de-radicalization” is only cited as playing a “large role” in just 16% of disengagement events in our sample. I then discuss how one’s role within a terrorist group offers insight into the disengagement process. Our research shows that leaders and violent operatives have a harder time disengaging than those in logistical or support roles because of the sunk costs associated with their involvement and/or the fewer opportunities available to them. We also find that individuals in certain roles are more/less likely to experience certain push/pull factors for disengagement. I conclude by discussing our research on terrorist re-engagement, which shows that in the short term, a deep commitment to the ideology, maintaining ties to individuals still involved in terrorism, and being young increase the likelihood one will return to terrorism. Full article

In recent years, there has been a proliferation of programs aimed at preventing radicalization and disengaging known violent extremists. Some programs have targeted individuals through the use of case management approaches and the development of individual intervention plans (e.g., the Desistance and Disengagement Program and the Channel program in the UK; the Australian New South Wales Corrections Proactive Integrated Support Model—PRISM—and state-based division initiatives in Australia). There is a broad consensus in the literature that the evaluation of such initiatives has been neglected. However, the evaluation of case-managed interventions to counter violent extremism (CVE) is challenging. They can have small caseloads which makes it difficult to have any comparison or control group. Client participation can vary over time, with no single intervention plan being alike. This can make it hard to untangle the relative influence of different components of the intervention on indicators of radicalization and disengagement. In this presentation, results from primary research that set out to evaluate case-managed CVE interventions in Australia and develop evaluation metrics are presented. This research involves the examination of interventions implemented by New South Wales corrections and state police. The effectiveness of these interventions was assessed against a five-point metric of client change. Client change overtime was analyzed using case note information collected by the various interventions on client participation. Results show that client change is not a linear process and that the longer an individual is engaged in a case-managed intervention, the more likely they are to demonstrate change relating to disengagement. Specific case studies are used to illustrate trajectories and turning points related to radicalization and to highlight the role of case-managed interventions in facilitating disengagement. Key elements of effective interventions include the provision of ongoing informal support. Investment in capturing case note information should be a priority of intervention providers. Different challenges confronted by case-managed CVE interventions are highlighted. Full article

While much time and effort in terrorism studies has gone into finding an answer to the question of why people radicalise, the question of how to rehabilitate individuals has received renewed attention with the return of many citizens who had joined the Islamic State. This has led to an increase in both programs and interventions aimed at ‘deradicalising’ or ‘disengaging’ individuals as well as studies that aim to assess the effectiveness of such efforts. This article/presentation provides an overview of the main elements of rehabilitation and reintegration programs for violent extremist offenders based on international research assessing 34 rehabilitation and reintegration programs. In the second half of the article, an evaluation of the Dutch approach to reintegrating violent extremist offenders in and after prison is presented. Finally, the main lessons learned from these programs as well as the academic literature are discussed. Full article