By Triana McNeil

  A hate crime occurs nearly every hour in the U.S., based on data reported to the FBI. Investigations of recent hate crimes have suggested that exposure to hate speech on the internet may have contributed to the attackers’ biases against their victims. In 2021, the FBI placed hate crimes at the same national threat priority level as preventing domestic violent extremism. GAO was asked to review information on hate crimes and hate speech on the internet. This report examines (1) the extent to which DOJ collects data on hate crimes that occur on the internet, (2) what company data indicate about steps selected companies have taken to remove hate speech and violent extremist speech from their internet platforms, and (3) what is known about users’ experience with, or expression of, hate speech on the internet, and its relationship to hate crimes and domestic violent extremism. GAO analyzed U.S. hate crime data and interviewed DOJ officials. GAO analyzed data and interviewed officials from six selected companies operating internet platforms with publicly available policies prohibiting hate and violent extremist speech. GAO assessed peer reviewed and nonprofit studies that described hate speech on the internet, hate crimes, and domestic violent extremist incidents. What GAO Recommends GAO recommends that BJS explore options to measure bias-related criminal victimization that occurs on the internet through the National Crime Victimization Survey or in a supplemental survey, as appropriate. BJS agreed with this recommendation.

Washington, DC: U.S. Government Accountability Office, 2024. 87p.

By Alex Goldenberg, Joel Finkelstein,

In the predawn hours of September 12, 2001, on board a helicopter heading from Liberty State Park to State Police Headquarters, I had my first opportunity since the terrorist attacks the previous morning to wonder, “how the hell did they pull this off?” It was inconceivable to me, with the trillions of dollars our nation had spent on a global early warning system to prevent another Pearl Harbor surprise attack, that we were unable to prevent the 9/11 attacks or, with the exception of the heroism of the passengers and crew of United 93, to stop them in progress. The question haunted me for the remainder of my term as New Jersey’s Attorney General and beyond until, as Senior Counsel to the 9/11 Commission, I was able to help piece together precisely how the terrorists managed to succeed. At every turn, they hid in plain sight. They traveled openly and freely among the Americans they despised, then disappeared when circumstances warranted. Specifically, on the day of 9/11 itself, the first thing the hijackers did once they secured the cockpits was to turn off the transponders that identified the aircraft to military and civilian controllers. This had the effect of making the planes’ radar signals disappear into the clutter of raw radar data, making it extremely difficult to track the planes. A lot has changed in the years since the 9/11 attacks happened and the 9/11 Commission Report was issued. There were no smart phones then, no Twitter or Instagram, no Google or Snapchat. The revolution in communication technology since has transformed both the way we live and the tactics employed by the extremists who want to kill us. What has remained constant, however, is the extremists’ strategy of using the instrumentalities of freedom recursively in order to destroy it, and the challenge to governments to anticipate the new generations of tactics in order to frustrate their employment. The Report you are about to read, “Cyber Swarming: Memetic Warfare and Viral Insurgency,” represents a breakthrough case study in the capacity to identify cyber swarms and viral insurgencies in nearly real time as they are developing in plain sight. The result of an analysis of over 100 million social media comments, the authors demonstrate how the “boogaloo meme,” “a joke for some, acts as a violent meme that circulates instructions for a violent, viral insurgency for others.” Using it, like turning off the transponders on 9/11, enables the extremists to hide in plain sight, disappearing into the clutter of innocent messages, other data points. It should be of particular concern, the authors note, for the military, for whom “the meme’s emphasis on military language and culture poses a special risk.” Because most of law enforcement and the military remain ignorant of “memetic warfare,” the authors demonstrate, extremists who employ it “possess a distinct advantage over government officials and law enforcement.” As with the 9/11 terrorists, “they already realize that they are at war. Public servants cannot afford to remain ignorant of this subject because as sites, followers, and activists grow in number, memes can reach a critical threshold and tipping point, beyond which they can suddenly saturate and mainstream across entire cultures.” This Report is at once an urgent call to recognize an emerging threat and a prescription for how to counter it. As such, it offers that rarest of opportunities: the chance to stop history from repeating itself.

The Network Contagion Research Institute , 2021. 11p.

Edited by Giray SADIK  

According to Strategic Concept 2022, terrorism, which poses the immediate asymmetric and transnational threat, is one of two main threats to NATO. Even though NATO’s focus is currently directed to the other main threat, Russia, particularly after its war against Ukraine, terrorism has remained a major threat across the NATO territory and periphery. Besides, the adaptations of terrorist organizations in response to current wars such as the one in Ukraine are likely to exacerbate terrorist threats for NATO Allies and partners. Therefore, there is an ongoing need for research and learning around global terrorism landscape and its implications for NATO. To this end, this research aims to identify the lessons learned for NATO from the RussiaUkraine war for countering terrorism effectively. Russia’s war on Ukraine since February 2022 has led to dramatic changes in global geopolitics and the all-encompassing domains of security, connectivity, and modern warfare. Although significant and widespread, the effects of this ongoing war on global terrorism have yet to be comprehensively analyzed. To address this gap in a timely manner, this project aims to examine the effects of the Russia-Ukraine war on terrorism through expert discussions on the contemporary trends and lessons learned for global counter-terrorism efforts, and NATO Allies and Partners. To this end, we organized a workshop in hybrid format (in-person and online) with the contributing authors of the edited book. Workshop participants include academics, practitioners, and subject matter experts from various NATO Allies and Partners.   

Ankara, Turkey: Centre of Excellence Defence Against Terrorism (COE-DAT) , 2025.   138p.

By Shai Farber

  This systematic review unveils significant convergence between cybercriminals and terrorists over the past decade, characterized by substantial overlap in tactics, capabilities, and infrastructure exploitation. Key findings illuminate the adoption of cybercriminal methods by terrorist organizations, the engagement of cybercriminal groups in terrorist activities, the proliferation of enabling technologies such as cryptocurrencies and the dark web, and the role of social media in facilitating knowledge exchange between diverse threat actors. While direct terrorist involvement in proftdriven cybercrime appears limited, the escalating convergence underscores urgent imperatives for enhanced cooperation, legal refinement, the establishment of flexible counter-threat authorities, and collaboration with the private sector. Continuous monitoring of this rapidly evolving landscape is crucial to inform effective policies and interventions in an increasingly interconnected digital world.

Security Journal (2025), 23p.

By Nina Wilén and Kwesi Aning

Over the past decade, Western and multilateral security interventions in the Sahel have failed to stabilize the region, while national and local efforts, particularly in Mali and Burkina Faso, have remained largely militarized. These approaches have proven ineffective in addressing the underlying causes of instability, such as food insecurity, lack of basic services, and limited economic opportunities. The key lesson learned is that military action alone cannot resolve the Sahel’s security crisis—political engagement from local, national, and regional actors is crucial.

Konrad-Adenauer-Stiftung e. V. 2 COUNTER EXTREMISM PROJECT, 2025// 28p.

By Noémie Bouhana | Caitlin Clemmow | With contributions from Philip Doherty

Post 9/11, research on radicalisation and involvement in terrorism and violent extremism (henceforth, ITVE) has focussed on the role of individual-level attributes, such as the psychological characteristics, socio-economic background, demographic features, experiences, beliefs, past behaviours and criminality, and the social embeddedness of radicalised individuals and people involved in acts of terrorism (Borum, 2011; Stern, 2016; McCauley & Moskalenko, 2017; Sageman, 2008; Wiktorowicz, 2005). Likewise, existing processes, methods and tools to identify and evaluate the risk of ITVE operate largely at the individual level. The Vulnerability Assessment Framework (VAF) used in the context of Channel is the best example of this. Comparatively less attention has been devoted to identifying and evaluating the characteristics of the socio-physical contexts which may contribute to the emergence of the risk of ITVE, or, conversely, to its suppression. Yet research on crime, arguably a closely related problem space, makes a strong case that many of the causes and enabling conditions of criminal development (of which ITVE is an instantiation) are found in the environment: crime and criminals concentrate in space and time. Modifying the environment in which crime occurs and the specific settings in which criminals develop and act has been shown at length to be an effective approach to crime prevention and disruption (Weisburd et al., 2012). In crime prevention, as in other areas of behavioural change, one size does not fit all: to design effective programs, the ability to tailor interventions to the socio-physical context is paramount, which means that the effect of key contextual features on the outcome of interest must be understood (Sampson et al., 2013); notably, the effect of exposing (susceptible) individuals with particular kinds of characteristics to (criminogenic) places with particular kinds of features. Situational Action Theory [SAT] is a theory of crime causation, which, as its name suggests, puts context at the centre of the explanation of crime (P.- O. H. Wikström et al., 2009). In 2010, ahead of the revision of the Prevent Strategy, the UK Home Office commissioned a Rapid Evidence Assessment of the cause of al-Qaeda-influenced radicalisation seen through the lens of SAT (Bouhana & Wikström, 2011), which identified the key levels of analysis in a causal model of radicalisation as relating to individual vulnerability to radicalising influence, exposure to radicalising settings, and the mechanisms of emergence of said settings. In the decade since, this theoretical model has been refined and now underpins a research programme concerned chiefly with going beyond "who" and "why" to the "where" of terrorism involvement. It has inspired, notably, the EU FP7-funded international PRIME project on lone actor terrorism, which, among other findings, has shown that individual-level risk indicators are multifinal (i.e., their meaning determined by the context in which they arise), and has provided support for a formulation of individual vulnerability which encompasses susceptibility to exposure to radicalising settings, as well as cognitive susceptibility to moral change with significant implications for prevention efforts (Corner et al., 2018). It has also produced evidence for distinct configurations of individual/ context interaction patterns leading to the emergence of lone actor risk, with further implications for risk assessment (Clemmow et al., 2019). To better support policy strategy and counter-terrorism practice, the theoretical SAT model of radicalisation was reformulated and expanded into a risk analysis framework for lone actor terrorism [PRIME RAF] (Bouhana et al., 2016, 2018) and a systemic inference framework to inform counter-extremism strategies [S5], commissioned by the UK Counter-Extremism Commission (Bouhana, 2019). Both of these integrate SAT with the broader behavioural and socio-cognitive knowledge-base to refine our understanding of personenvironment interaction. While the RAF is concerned specifically with the effect of situational interaction on the motivation and capability of lone actors to act (behaviour), S5 is more broadly focused on explaining the acquisition of extremist propensities (beliefs). Both frameworks are analytically related and can be logically integrated. However, these remain general frameworks. Inherent in their interactionist logic is the fact that, while environmental processes of risk can be set out in generic terms, specific risk (or protective) factors and – crucially – their observable indicators must be elicited with regards to a particular context. This is because indicators are subject to change over space and time (P. O. H. Wikström & Bouhana, 2017) and because risk assessment is most effective when calibrated to context (Hamilton et al., 2021). A recently completed, Home Office-funded project guided by S5 (State of the Union [SOTU]; Bouhana and Schumann, 2021), which aimed to assess the observability of key analytical concepts of mainstream models of terrorism involvement, demonstrated the difficulty, yet also the necessity of translating said concepts into observable indicators, inasmuch as any model is intended to guide risk assessment or intervention design and evaluation. It also suggested that environmental indicators may be more reliably observed than susceptibility-related indicators using commonly available sources of information, and it tentatively provided further support for the notion that susceptibility and exposure interact in specific ways, leading to the emergence of risk. This project investigated the feasibility of developing a framework to assess the risk of involvement in terrorism and violent extremism at the level of place, with a view towards – further down the line – designing an assessment tool along the lines of the Vulnerability Assessment Framework, intended to assess environments instead of individuals. As a first step towards establishing this feasibility, this project investigated whether the mechanisms articulated in the S5 framework could be operationalised; in other words, to what extent they could be associated with a set of observable indicators relevant to the UK ITVE context. Given the relative scarcity of empirical research into the characteristics of extremist social ecologies and settings in the UK (and largely elsewhere) – which would be prerequisite to the conduct of a systematic review – the researchers elected to carry out an e-Delphi exercise with U.K. Prevent practitioners, the results of which are presented at length in this report.

London: CREST - Centre for Research and Evidence in Security Threats (CREST), 2025. 50p.

By Remadji Hoinathy, Malik Samuel and Akinola Olojo

  Some Lake Chad Basin countries (Cameroon, Chad, Niger and Nigeria) have been dealing with violent extremism for over a decade. Disarmament, demobilisation, repatriation, reintegration and resettlement processes in these countries may offer useful lessons for other West African contexts, including Mali and Burkina Faso, or more recently affected countries such as Benin, Côte d’Ivoire and Togo. Such lessons include incentivising defections, coordinating at national and regional levels, gender sensitivity, appropriate legal frameworks and community engagement.

Key findings Understanding the circumstances that lead people to join and leave violent extremist groups, and their experiences in those groups, is key to crafting incentives for disengagement. The willingness to leave violent extremist groups is often clouded by uncertainty, as well as long waiting periods between disengagement and enrolment in programmes for disarmament, demobilisation, repatriation, reintegration and resettlement (DDRRR). Waves of defection from Boko Haram caught Lake Chad Basin (LCB) states unprepared. DDRRR processes were thus implemented under pressure, hindering early coordination at a regional level. Women are treated mainly as victims, despite some voluntarily joining and playing active roles in Boko Haram, including volunteering as suicide bombers. DDRRR implementation in the LCB has revealed gaps in legal frameworks that require revisions. DDRRR lacks public support in some LCB countries, as many people view it as blanket amnesty for Boko Haram members. 

Recommendations Non-military means of countering violent extremism should incorporate incentives and opportunities for associates to leave these armed groups. By depleting these groups’ human resources, their fighting capacity is reduced. To encourage defection, clear processes for screening, prosecution and integration are needed. National ownership of these processes, inclusivity, adaptation to local context and adequate resourcing are key, from inception to implementation. National ownership should not lead to isolated approaches, but rather create bridges between countries to enable a cohesive regional approach. Organisations such as the Economic Community of West African States, the Liptako-Gourma States Integrated Development Authority, the G5 Sahel and the Accra Initiative could offer relevant regional frameworks for this. Authorities in charge of DDRRR programmes should consider the diverse needs and backgrounds of ex-associates. Distinguishing individuals according to why they joined and their roles within the groups is important for providing appropriate treatment. The specific needs of women and children should be taken into account. Affected West African countries should proactively formulate relevant legal and institutional frameworks. International and regional provisions and standards should be taken into account, including the need for transparent and predictable screening, prosecution and rehabilitation processes. Due to their position connecting the LCB and West Africa, Niger and Nigeria could play a key role in sharing lessons learned from DDRRR implementation. Community organisations, platforms, and media should be used by the DDRRR authorities to raise public awareness of the reintegration process. This would help prevent perceptions of general amnesty for Boko Haram members and preconceived ideas about the disengaged from undermining the process. 

Pretoria: Institute for Security Studies, 2023. 12p.

By Irfan Idris and Kusumasari Ayuningtyas

The ideological influence of Jemaah Islamiyah continues to be a threat to security even after its disbandment in June 2024. Some of its former members may align with other extremist groups, while others may seek to operate as lone-wolf terrorists. The possibility of JI’s former members transitioning into lone wolf actors stems from the enduring power of its ideology and indoctrination process.

S. Rajaratnam School of International Studies, NTU Singapore, 2025, 3p.

By Noor Huda Ismail

The concept of asabiyya – group solidarity – originally articulated by Ibn Khaldun, has been hijacked by extremist groups in Southeast Asia to create division and promote radical ideologies. Reclaiming asabiya is crucial, redefining it as a tool for social cohesion and unity by drawing on cultural values that foster mutual respect and collective responsibility. Regional cooperation and digital counter-narratives are essential in countering extremism and fostering long-term peace in the region.

S. Rajaratnam School of International Studies, NTU Singapore, 2025. 4p.

By Alexander Palmer, Skyeler Jackson, and Daniel Byman

Despite the recent attack in New Orleans, longer-term trends suggest that jihadist terrorism in the United States is not resurgent. The number of jihadist plots and attacks in the United States has been low since the territorial defeat of ISIS. The average lethality of jihadist terrorist attacks has also fallen since the peak of the caliphate. International groups such as the Islamic State and al Qaeda have filled more of an inspirational role rather than directly orchestrating attacks on the United States. Although formal links between attackers and plotters in the United States appear to be rare, this has not prevented individuals from carrying out lethal “lone wolf” attacks. Overall, jihadist terrorism in the United States does not merit additional resources at this time; however, international terrorism organizations continue to merit counterterrorism efforts.

Washington, DC: The Center for Strategic and International Studies (CSIS) , 2025. 12p.

By Iftekharul Bashar

In 2024, Bangladesh faces a growing threat of radicalism fuelled by political instability, security vacuums, and weakened counterterrorism infrastructure. Unless the interim government adopts a zero-tolerance policy and strengthens law enforcement and intelligence efforts, extremist groups may gain further ground, leading to long-term instability. COMMENTARY Bangladesh is facing an increasing threat of radicalism and terrorism. The political violence and instability of 2024 have worsened the situation, with extremist elements exploiting the security vacuum created by the regime change. Weakened law enforcement and intelligence agencies, along with the rise of armed groups and radical ideologies, pose a serious threat to the country’s stability. Bangladesh risks long-term insecurity with broader regional implications without swift and decisive action.

Singapore: S. Rajaratnam School of International Studies (RSIS), NTU , 2024. 4p.

By Yannick Deepen and Sabine Kurtenbach

Non-state armed actors (NSAAs) come in a variety of shapes and sizes, including warlordled groups, insurgencies, militias, and organised-crime syndicates to name just the most prominent examples hereof. In war or lower-level armed conflict, as well as violence-prone contexts, these groups pose acute problems for peacebuilding, democratic governance, and sustainable development. They control resources and territories, as well as compete, cooperate, or align with political and economic elites both within and beyond the state to promote the unstable status quo that serves their interests. Various challenges arise when dealing with NSAAs. Prevailing strategies vary between repression, co-optation, and submission to the rule of law. This analysis maps cross-regional trends regarding NSAAs based on available datasets. On this basis we develop a typology for NSAAs and a conceptual frame for their transformation. We then apply this to three paradigmatic case studies – Colombia, Mali, the Philippines – and formulate recommendations for the designing of improved strategies vis-à-vis development cooperation and other external actors.

GIGA Working Papers, No. 337,

By Omar Ashour

How did ISIS – a widely hated, massively outnumbered and ludicrously outgunned organisation – manage to occupy over 120 cities, towns and villages from the Southern Philippines to Western Libya? Seeking to understand ISIS’s combat effectiveness, Omar Ashour analyses the military and tactical innovations of ISIS and their predecessors in Iraq, Syria, Libya and Egypt. He shows how their capacity to mix conventional military tactics with innovative guerrilla warfare and urban terrorism strategies allowed ISIS to expand and endure beyond expectations.

Edinburgh: Edinburgh University Press, 2021.

By Rohan Gunaratna

The conflicts in Israel, Gaza, Lebanon, Yemen and Syria will have a generational impact. The 7 October 2023 attack against Israel in Gaza and its consequences have set back the prospects of a two-state solution to the Palestinian issue and may attract a new wave of foreign fighters to embark on terrorist activities like what was seen following the military debacles in Afghanistan, Iraq and Syria. What can be done to stem another rise of the terrorist fighters?

S. Rajaratnam School of International Studies, NTU Singapore, 2025. 3p.

Edited by Katharina Karcher, Yordanka Dimcheva ,· Mireya Toribio Medina. Mia Parkes

This open access book sheds light on collective practices of remembering, imagining and anticipating in relation to recent acts of urban terrorism in Europe. Analysing a range of personal and collective responses to urban terrorism in contemporary Europe, this book shows that current debates on this issue are shaped by multiple co-existing and intersecting memories of political violence in the past. Moreover, despite public declarations of unity and solidarity, collective memories of urban terror in contemporary Europe are far from consensual - memory can be both a catalyst for and an impediment to social and political change. Drawing on case studies from a range of European countries and creative responses by survivors, artists, and poets, this interdisciplinary volume introduces readers to key methods (e.g. discourse analysis and (auto-)ethnography) and concepts (e.g. Lieux de Mémoire and ‘grassroots memorials’) for the study of the memorialization of terror attacks.

Cham: Palgrave Macmillan, 2024. 317p.

By Michael Jensen, Sheehan Kane, and Elena Akers

Profiles of Individual Radicalization in the United States (PIRUS) is a database of 3,203 Islamist, far-left, far-right, and single-issue extremists who have radicalized to violent and non-violent extremism in the United States from 1948 through 2021. This update adds 955 subjects to the database who were charged with committing extremist crimes in the United States from 2019-2021.1 The database is freely available for download on START’s webpage at https://www.start.umd.edu/profiles-individual-radicalization-united-states-pirus-keshif. The analysis for this research brief draws on the full dataset and illustrates the important differences that exist across ideological groups and highlights recent trends in the data.

College Park, MD: START , 2023. 10p.

By Michael Jensen, Sheehan Kane, and Elena Akers

Included with the current update to the Profiles of Individual Radicalization in the United States (PIRUS) data are a suite of variables that allow users to investigate extremist offenders who plotted to commit, or committed, mass casualty attacks in the United States from 1990-2021. A mass casualty plot is defined as an event in which the perpetrator(s) intended to kill or injure four or more victims. Unlike other data resources on mass casualty events, PIRUS allows users to investigate successful mass casualty attacks as well as unsuccessful plots where the perpetrators clearly intended to kill or injure four or more people but failed. The inclusion of failed and foiled mass casualty plots in PIRUS gives users a more accurate estimate of how often extremists plan to cause significant harm. It also allows users to compare ideological groups and offenders on measures of plot success. This research brief provides an overview of the mass casualty crimes in PIRUS, including their frequency, links to ideological and sub-ideological groups, and a comparison of their perpetrators.

College Park, MD: START (March). 2023, 4p.

By Rohan Gunaratna

SYNOPSIS As we enter the new year, armed conflicts continue to rage across the globe, destabilising nations, regions, and even the global order. Terrorist attacks and state responses create ripple effects, which not only shape the battlefields but redefine states, fragment societies, and radicalise communities. What does the terrorism threat landscape look like in 2025?

COMMENTARY Global terrorism will remain a persistent and pervasive threat to the stability and peace of the world. The Islamic State, Al Qaeda, and Iran-sponsored Shiite and Sunni militia have exacerbated instability around the globe, while geopolitical competitions have polarised and fragmented the community of nations. Unless governments work together to mitigate common security challenges, threat actors will exploit the gaps, loopholes, and weaknesses in global security systems. All nations need to be steadfast in preventing and preempting threats from taking root and facing up to threats from state and non-state actors including lone wolf attackers. In addition to disinformation operations, threat actors will attack information infrastructure. Nations will need to secure the online domain as hostile states increase their cyberattacks, sabotage infrastructure, assassinate public officials, and engage in espionage.

S. Rajaratnam School of International Studies, NTU Singapore 2025. 5p.

By Heidi Ellis , Alisa B. Miller , Georgios Sideridis , Rochelle Frounfelker , Diana Miconi , Saida Abdi , Farah Aw-Osman and Cecile Rousseau

Objectives: We examine the association between perceived discrimination, mental health, social support, and support for violent radicalization (VR) in young adults from three locations across two countries: Montréal and Toronto, Canada, and Boston, United States. A secondary goal is to test the moderating role of location. Methods: A total of 791 young adults between the ages of 18 and 30, drawn from the Somali Youth longitudinal study and a Canada-based study of college students, participated in the study. We used multivariate linear regression to assess the association between scores on the Radical Intentions Scale (RIS) with demographic characteristics, anxiety, depression, social support, and discrimination. Results: In the full sample, discrimination, age, and gender were associated with RIS scores. When we examined moderation effects by location, RIS scores were associated with depression only in Montréal, and with social support (negatively) and discrimination in Toronto. None of the variables were significant in Boston. Conclusion: These findings suggest that an understanding of risk and protective factors for support of VR may be context-dependent. Further research should take into consideration local/regional differences

International Journal of Public Health published: 29 March 2021, 10 p.

By Michael Wolfowicz, Yael Litmanovitz, David Weisburd, and Badi Hasisi

Abstract Background: Two of the most central questions in radicalization research are, (1)why do some individuals radicalize when most of those from the same groups or exposed to similar conditions do not? and (2) why do radicalized individuals turn to radical violence while the majority remain inert? It has been suggested that the answer to both questions lie in the cumulative and interactive effects of a range of risk factors. While risk assessment and counter‐radicalization take a risk‐protectivefactor approach, there is widespread debate as to what these factors are and which are most important.Objectives: This review has two primary objectives.1) To identify what the putative risk and protective factors for different radica-lization outcomes are, without any predeterminations.2) To synthesize the evidence and identify the relative magnitude of the effects of different factors.The review's secondary objectives are to:1) Identify consistencies in the estimates of factors across different radicalizationoutcomes.2) Identify whether any significant heterogeneity exists within factors between(a) geographic regions, and (b) strains of radicalizing ideologies.Search Methods: Over 20 databases were searched for both published and gray literature. In order to provide a more comprehensive review, supplementary sear-ches were conducted in two German and one Dutch database. Reference harvesting was conducted from previous reviews and contact was made with leading researchers to identify and acquire missing or unpublished studies.Selection Criteria: The review included observational studies assessing the out-comes of radical attitudes, intentions, and/or radical behaviors in OECD countries and which provided sufficient data to calculate effect sizes for individual‐level risk and protective factors.Data Collection and Analysis: One‐hundred and twenty‐seven studies, containing 206 samples met the inclusion criteria and provided 1302 effect sizes pertaining to over 100 different factors. Random effects meta‐analyses were carried out for each factor, and meta‐regression and moderator analysis were used to explore differ-ences across studies.Results: Studies were primarily cross‐sectional, with samples representing 20 OECD countries. Most studies examined no specific radicalizing ideology, while others focussed on specific ideologies (e.g., Islamist, right‐wing, and left‐wing ideologies). The studies generally demonstrated low risk of bias and utilized validated or widely acceptable measures for both indicators and outcomes.With some exceptions, sociodemographic factors tend to have the smallest estimates, with larger estimates for experiential and attitudinal factors, followed by traditional criminogenic and psychological factors.Authors' Conclusions: While sociodemographic factors are the most commonly examined factors (selective availability), they also tend to have the smallest es-timates. So too, attitudinal and even experiential factors, do not have effect sizes of the magnitude that could lead to significant reductions in risk through tar-geting by interventions. Conversely, traditional criminogenic factors, as well as psychological factors tend to display the largest estimates. These findings suggest the need to broaden the scope of factors considered in both risk assessment and intervention, and this review provides much needed evidence for guiding the selection of factors

Campbell Systematic Reviews: Volume 17, Issue 3 September 2021, 90 p.