Cross-national survey of 5,500 people across Britain, Netherlands, Norway, and Denmark reveals perceived exclusion predicts radical potential equally everywhere—and most radicals are from ethnic majorities
Summary
A landmark cross-national study published in Terrorism and Political Violence has upended conventional wisdom about radicalisation in Europe. Analysing survey data from over 5,500 respondents across Britain, the Netherlands, Norway, and Denmark, researchers tested whether generous Nordic welfare states protect minorities from radicalisation better than leaner regimes like the UK’s. The answer is a decisive no.
The study, conducted by Tahir Abbas and colleagues from Aston University, the University of Edinburgh, and Open Horizon, reveals what they call the “Uniform Logic of Exclusion”: perceived structural exclusion predicts support for political violence at a strong and statistically indistinguishable rate in all four countries. The institutional buffering hypothesis—that comprehensive welfare states weaken the conversion of grievance into radical potential—finds no support whatsoever.
Three findings carry profound policy implications. First, every country-by-exclusion interaction tested was statistically null. Perceived exclusion is as radicalising in Copenhagen as in Birmingham. Second, while conversion rates are uniform, levels of political alienation follow an institutional gradient: highest in the UK, lowest in Norway and Denmark. The institutional signal appears in grievance levels, not in how quickly exclusion turns into radical views. Third, minority-majority differences in radical potential are modest (Cohen’s d between -0.24 and -0.60), and because majorities are demographically dominant, roughly four in five respondents in the top band of the radical-potential scale in every country are members of the ethnic majority.
Muslim identity appears protective rather than risky, with the Muslim main effect trending negative (b = -0.25, p = .079). Women report significantly lower radical potential than men. And the largest minority-majority gap in radical potential was found in Denmark, not Britain—a surprise given the UK’s securitised counter-extremism apparatus.
The implication is stark: European counter-extremism policies focused on Muslim minorities are targeting the wrong population. Radicalisation is driven by perceived structural exclusion everywhere, and the real threat pool is majority-dominated in all four societies. The authors call for decoupling risk assessment from Muslim identity, targeting perceived structural exclusion itself, and re-allocating counter-extremism attention towards the majority share of radical potential.
Introduction: Testing the Welfare State Buffer
For two decades, European counter-extremism policy has overwhelmingly fixated on minority communities, particularly Muslims, as the primary locus of volatility. This securitised gaze assumes that political disaffection stems from internal deficits—cultural separateness, theological incompatibility, failed assimilation. But what if the driver of radicalisation lies not in the culture of the actor but in the architecture of the environment?
A new study published in Terrorism and Political Violence puts this question to a rigorous cross-national test. Researchers from Aston University, the University of Edinburgh, and Open Horizon surveyed over 5,500 people across Britain, the Netherlands, Norway, and Denmark—four societies sharing democratic institutions and post-war migration histories but differing systematically in welfare-state architecture.
The hypothesis was elegant and theoretically well-motivated: comprehensive Nordic welfare regimes should buffer the experience of exclusion, weakening the conversion of perceived structural disadvantage into support for political violence. The residualist British regime, with its securitised counter-extremism apparatus organised around the Prevent duty, should show a stronger conversion.
The answer, across every specification, is no. Perceived structural exclusion predicts radical potential strongly and uniformly across all four countries. The institutional buffering hypothesis is rejected in its slope-moderation form: every interaction term is null in the country-specific models, the pooled contrasts, and the structural variants.
This finding has profound implications for theory and policy. It suggests that radicalisation is driven by perceived structural exclusion everywhere, and that counter-extremism infrastructure calibrated to minority populations is mis-calibrated against the actual distribution of radical potential.
The Uniform Logic of Exclusion: What the Study Found
Data and Design
The analysis uses the 2023 wave of the DRIVE survey, funded under the European Union’s Horizon 2020 programme. From 6,243 returned questionnaires, the final sample comprises 5,501 respondents: United Kingdom (n=2,535), Netherlands (n=1,963), Norway (n=368), and Denmark (n=635). The sampling design mixed a general-population sample with minority oversamples, ensuring sufficient representation of native-born visible minorities, self-identifying Muslims, and foreign-born working-age respondents.
Intersectional minority status coded respondents as “majority” only if they were both native-born and ethnically dominant; all others were classified as minority. This refinement matters because the lived experience of second- and third-generation minorities differs in kind from that of recent migrants: they are not negotiating the difficulty of arrival but the chronic condition of being positioned as provisional citizens in the country of their birth.
The outcome variable, Radical Potential, measured attitudinal permissibility of violence (“Sometimes people need to use violence to achieve their goals,” 1-7). This sits deliberately on the opinion side, capturing the breadth of the pool from which mobilisation can draw rather than behavioural involvement.
Key Findings
Uniform Conversion Across Regimes. The central result is a theoretically informative null. Perceived structural exclusion predicts the attitudinal endorsement of political violence at a strong rate—approximately 0.45 scale points per unit of perceived exclusion, β ≈ 0.24-0.31 standardised in the pooled models. Country-stratified slopes were virtually identical: UK 0.45, Netherlands 0.45, Norway 0.52, Denmark 0.53. The four regression lines climb in near-parallel, and their 95% confidence bands overlap throughout the observed range.
Minority-Majority Differences Are Modest. Minorities report slightly higher radical potential everywhere: Cohen’s d = -0.24 in the UK, -0.28 in the Netherlands, -0.36 in Norway, and -0.60 in Denmark. The largest gap is Danish, not British—contrary to the intuition that the most securitised national context would show the largest minority surplus.
The Majority Holds Most Radical Potential. Because majorities are demographically dominant, roughly four in five respondents at the top of the radical-potential scale (score 5 or above) in every country are members of the ethnic majority. In the UK, approximately 293 majority respondents sit at the top against roughly 77 minority respondents—a four-to-one ratio that holds across all four countries.
Muslim Identity Is Protective. Muslim identification was not a positive predictor of radical potential in any specification. It trended protective (b = -0.25, p = .079), consistent with previous research showing a 78.4% reduction in the odds of violence-support associated with strong Muslim identity. The empirical case for Muslim identity as a risk marker in instruments such as the Prevent duty is, on this evidence, inverted.
Political Alienation Follows an Institutional Gradient. While conversion rates are uniform, levels of political alienation vary systematically. Alienation is highest in the UK (roughly +0.15 SD above the pooled mean), close to the pooled mean in the Netherlands, and lowest in Norway and Denmark (roughly -0.33 SD below the pooled mean). The institutional signal is in levels, not slopes.
Radical Potential by Country and Status
| Country | Ethnic Majority Mean (1-7 scale) | Ethnic Minority Mean (1-7 scale) | Difference (Cohen’s d) | Interpretation |
|---|---|---|---|---|
| United Kingdom | 2.48 | 2.89 | -0.24 | Small difference |
| Netherlands | 2.62 | 3.12 | -0.28 | Small difference |
| Norway | 2.30 | 2.94 | -0.36 | Small-to-moderate difference |
| Denmark | 2.43 | 3.44 | -0.60 | Moderate difference |
What this tells us: Minorities report slightly higher radical potential in all four countries, but the difference is modest. The largest gap is in Denmark, not the UK—surprising given the UK’s securitised counter-extremism apparatus.
Who Holds the Most Radical Potential? (Score 5 or Above)
| Country | Ethnic Majority (n, %) | Ethnic Minority (n, %) | Ratio (Majority:Minority) |
|---|---|---|---|
| United Kingdom | ~293 (79%) | ~77 (21%) | 4:1 |
| Netherlands | ~261 (79%) | ~62 (21%) | 4:1 |
| Norway | ~39 (77%) | ~11 (23%) | 4:1 |
| Denmark | ~86 (79%) | ~20 (21%) | 4:1 |
What this tells us: Despite slightly higher average radical potential among minorities, the sheer demographic dominance of ethnic majorities means that roughly four out of five people at the highest levels of radical potential in every country are members of the ethnic majority.
What Predicts Radical Potential? (Minority Sub-Sample)
| Predictor | Effect Size | Statistical Significance | Direction |
|---|---|---|---|
| Perceived Structural Exclusion | b = 0.43-0.46 | p < 0.001 | Strong positive |
| Age | b = -0.024 | p < 0.001 | Negative (younger = higher) |
| Female Gender | b = -0.29 | p = 0.013 | Negative (women lower) |
| Muslim Identity | b = -0.25 | p = 0.079 (trend) | Protective (trend) |
| Benefits Receipt | b = 0.54 | p < 0.001 | Positive (financial vulnerability) |
| Partnership Status | b = 0.26 | p = 0.032 | Positive |
| Country × Exclusion Interactions | All null | All non-significant | No country differences |
What this tells us: Perceived structural exclusion is the strongest predictor of radical potential. Age, gender, and Muslim identity are protective. Financial vulnerability increases radical potential. Crucially, country interactions are all null—the exclusion-radicalisation link is uniform across all four countries.
Theory vs. Reality—Testing the Buffering Hypothesis
| Hypothesis | Prediction | What the Data Found | Supported? |
|---|---|---|---|
| H3a: Buffered Nordic Conversion | In Norway and Denmark, the exclusion-radicalisation link is weaker than in the UK | Nordic slopes: 0.52 and 0.53; UK slope: 0.45 | NO—slopes are virtually identical |
| H3b: Unbuffered British Conversion | In the UK, exclusion exerts a stronger direct effect | UK direct effect in SEM: β = 0.178; Nordic-pooled: β = 0.174 | NO—effects are virtually identical |
| Institutional Buffering | Comprehensive welfare states weaken conversion of exclusion into radical potential | All country × Exclusion interactions are null | NO—conversion is uniform |
| Levels, Not Slopes | Institutions affect how much alienation people feel, not how quickly exclusion converts | Alienation highest in UK, lowest in Nordic countries | YES—institutional signal is in levels, not slopes |
What this tells us: The data reject the idea that generous welfare states buffer the conversion of exclusion into radical potential. The conversion rate is uniform everywhere. What varies is how much political alienation people feel overall—and that follows an institutional gradient.
Policy Implications: What This Means for Counter-Extremism
Decouple Risk Assessment from Muslim Identity
The study’s findings directly challenge the operational logic of counter-extremism programmes like the UK’s Prevent duty. Muslim identification is not a positive predictor of radical potential in any specification. It trends protective. The empirical case for Muslim identity as a risk marker is, on this evidence, inverted.
A concrete first step, short of wholesale reform, is the removal of religious-identity markers from referral rubrics and their replacement with structural-exclusion markers—repeated adverse institutional contact, housing insecurity, perceived discriminatory treatment—which the models identify as the operative variables.
Target Perceived Structural Exclusion Itself
Because the exclusion-to-radical-potential conversion is uniform, the policy lever is exposure rather than conversion: the volume of group-differentiated unfairness citizens perceive. Three exposure channels are documented within the DRIVE programme: media framing of minorities as security problems; discriminatory institutional encounters in everyday settings, particularly education; and the urban institutional environment, where proactive and inclusive city administrations measurably outperform fragmented ones.
The Danish pattern is a standing caution against the assumption that high-trust welfare states are self-correcting on this margin. A generous regime that renders group-targeted treatment legible—as the “ghetto” legislation does—generates exactly the perceived exclusion the uniform logic then converts, at the common rate, into radical potential.
Re-Allocate Counter-Extremism Attention
This is not a call to securitise majorities; it is a call to de-securitise the framework altogether and to treat radical potential as a population-level property tracked by exclusion and alienation—in which majorities, by weight of numbers, hold roughly four-fifths of the mass in every country observed.
Concretely, de-securitisation means relocating Prevent-style duties from the security frame into safeguarding and welfare frames whose performance is evaluated by procedural-justice outcomes rather than referral volumes. It means resisting the export of identity-indexed referral architectures into universalist systems that currently lack them. And it means investing in the repair of procedural fairness and local civic infrastructure as counter-extremism policy in its own right.
Conclusion
This study set out to test whether the institutional architecture of the state conditions the conversion of perceived structural exclusion into radical potential, comparing the residualist British welfare regime with the comprehensive Nordic regimes and the hybrid Dutch case. The hypothesis was theoretically well-motivated, pre-stated, and cleanly rejected.
Perceived exclusion predicts the attitudinal endorsement of political violence strongly, and at a statistically indistinguishable rate, in all four societies. What varies across regimes is the stock of political alienation, not the exchange rate between exclusion and radicalism. Minority-majority differences are modest and largest in Denmark; Muslim identification trends protective; and in every country the large majority of respondents at the top of the radical-potential scale belong to the ethnic majority.
The implication for theory is a re-specification: institutional accounts of radicalisation should be formulated as claims about grievance levels and exposure, not about grievance conversion. The implication for policy is a re-orientation: because the logic of exclusion is uniform, no regime can outsource the problem to its institutional reputation; and because the distribution of radical potential is majority-dominated everywhere, no counter-extremism architecture organised around minority identity can be calibrated to the phenomenon it claims to govern.
The antidote to the present volatility is not more scrutiny of the minority but less exclusion for everyone—a conclusion that the data of four very different welfare states, for once, state in unison.
