Challenging the Muslimification of Muslims in research on ‘liberal democratic values’: why culture matters beyond religion

ABSTRACT

This study critiques the use of ‘Muslim’ as an analytic category and overfocus on religiosity as an explanatory variable in studies on conflicts between Muslims and national majorities over ‘liberal democratic values’. We call this tendency the Muslimification of Muslims. We demonstrate how this research reproduces and reinforces stereotypes drawn from dominant resonant public debates. To challenge these assumptions, we turn the research inquiry around so that ‘Muslim’ and religiosity become objects not tools for analysis. Revisiting the EurIslam survey data-set, explicitly designed for studying socio-cultural distances between Muslims and majorities, we examine boundary construction over ‘liberal democratic values’. For Muslims, first, we test for differences between four ethnonational family origin groups – Ex-Yugoslavians, Moroccans, Turks, Pakistanis- and second, for the explanatory power of religiosity compared to non-religious cultural variables. Findings are clear-cut: family ethnonational origin matters and there are different group trajectories of acculturation; religiosity has a very modest impact and much less than self-identification with settlement-country which pushes in the opposite acculturative direction. Simply put, regarding the construction of differences over ‘liberal democratic values’, not all Muslims are the same, and it is not all about practicing Islam. It is time for a re-think and a de-Muslimification of academic research.

KEYWORDS:

• Muslim
• liberal democratic values
• value conflicts
• acculturation
• religiosity

‘Muslim’ is a category towards which one must take a stance: one cannot simply inhabit it in an unreflective manner’ Rogers Brubaker (2013, 5)

Introduction

Several years ago, I was the academic speaker at a public event where people who are Muslims discussed being Muslims in Britain. For a white secular (but Church-of-England-baptised) British-born male academic, it was fascinating to hear the inside story of individuals brought together by a common sense of belonging as ‘Muslim’. Most people were under 40 years, but the superdiversity in the room was striking: by ethnonational family background, race, Islamic faith, phenotype, class background, accent and attire – ranging from suits to hipster, hijab to baseball cap. When people talked more about being Muslims in Britain this strong diversity became even more evident through their personal experiences. A well-spoken black man with family connections in Africa spoke about pursuing a career as a journalist in the BBC, the national cultural broadcasting institution, while a white-convert discussed her feminist activism through a local NGO in South London. For the people present, self-identifying as ‘Muslim’ and the attachment to Islam that implies, mattered a great deal in how they saw themselves in Britain, but it has clearly shaped their lives in radically different ways and to different degrees. Apart from being Muslims, the participants’ other main shared experience was being confronted by dominant stereotypical and often negative stigmatising representations of Muslims and Islam as a religion in British public life. How public institutions, the mainstream media and other people see Muslims and Islam importantly influenced their lived experiences of being Muslims in Britain too. Participating in this event rekindled my concerns about the way social science produces knowledge about and depicts Muslims and Islam, not least in quantitative survey research. This contribution is in part self-critique.

First, to state what really should be a very obvious fact, Muslims living in Europe are a highly heterogeneous category, and for most purposes should be treated as such. Social scientists tend to lump this heterogeneous set of individuals into a single category for analysis, ‘Muslim’, often uncritically without reflexivity.¹ This is problematic because it reifies ‘Muslim’ into a single homogeneous group and treats them as if they are a single solidary group in the social world – an assumption that is most likely false in most contexts, and at very least requires empirical verification.

A second problem arises because ‘Muslim’ is a religious identity that is commonly applied to significant minorities of immigrant origin from Muslim-majority countries in Europe – for example, Pakistanis in Britain, Turks and Moroccans in the Netherlands. Studies that take ‘Muslim’ as the analytic category automatically emphasise religion and religiosity over other cultural identities and social attributes, when looking for possible explanations for ‘integration problems’ of minorities relative to majority populations. This marks all ‘Muslims’ as culturally different from everyone else in the national community on religious grounds, while giving such claims – that are not established facts – a false veneer of a quasi-scientific legitimacy. Indeed, as this paper will argue an overemphasis on religious difference and (assumed) religiosity for ‘Muslims’ has become almost standard, so that alternative non-religious cultural explanations are rarely considered, nor given equal weight.² This overemphasis on religiosity as an über-explanatory variable for Muslim minorities’ attitudes and behaviour has an implicit tendency to depict them as Islamists. Whether intentionally (or probably most often not), researchers applying this lens risk reinforcing dominant resonant stereotypes in public discourses that culturally other Muslims and Islam in inaccurate and stigmatising ways. Such studies provide a source of legitimacy for political actors who mobilise anti-Muslim and anti-Islamic claims publicly but on questionable or flawed scientific evidence. In fact, research shows that only about a third of people who are Muslims of immigrant origin in Europe practice Islam strictly, with the rest non-practicing or agnostic (Dassetto, Ferrari, and Marechal 2007, 8; Joppke 2015, 156). Our own findings for religiosity among the Muslim sample gives a similar ballpark figure of 38.7% (see appendix). As a religious and ethnic identity ‘Muslim’ is especially prone to conflations in the public imagination that see all Muslims as adherents of the fundamentalist forms of Islamic faith that are deemed problematic to liberal democracies.

Third, academic research tends to reinforce unreflective dogmas over the assumed or real conflicts between ‘Muslims’ and ‘national majorities’ over ‘liberal democratic values’ that resonate strongly in dominant public discourses. This becomes especially problematic when studies are based on assumptions that pitch ‘Muslim’ against the national ‘majority’ population over a presumed ‘culture clash’ over adherence to so-called ‘liberal democratic values’ (see especially, Huntington’s (2002) ‘clash of civilisations’). Traditionally, academic discussions over ‘liberal democratic values’ emphasised the need for shared core norms and values that cut across ethnic and religious groups as a basis for democracy in increasingly diverse societies (see e.g. Held 1987). There has always been some dispute over the definition of core democratic values for Western liberal democracies, that leads to some slippage across political and social values but most include some combination of commitments to a separation of religious and secular authority, the rule of law and social pluralism, institutions of representative democracy, protection of individual rights and civil liberties and gender equality (Norris and Inglehart 2004). However, academic opponents of multiculturalism (e.g. Sniderman and Hagendoorn 2007) tend to reduce this broad (civic) understanding of ‘liberal democratic values’ to a set of assumed ethnonational characteristics of majority populations –‘Christian’, ‘British’, ‘Dutch’ values – against which they ‘test’ Muslim minorities of immigrant origin for whether they fit in the liberal democratic national political community, or not. Of course, viewed through this one-eyed lens the cultural and religious value differences between Muslim minorities to majorities automatically constructs them as potential opponents to liberal democracy. As we discuss below, this reductive understanding defines the research field on ‘liberal value conflicts and Muslims’ in a way that reinforces stereotypes about Muslims and Islam.

There is some academic critique of these tendencies. Anthropologists and ethnographers argue that who or what is a Muslim in Europe today is anything but straightforward (Grillo 2004; Bowen 2007). From sociology, Brubaker (2013, 6) raised the methodological problem about the pitfalls of categorising ‘Muslims’ lucidly in a short commentary a decade ago: ‘It is to risk what might be called ‘methodological Islamism’ by focussing too exclusively on the ‘Muslimness’ of Muslims and treating ‘Muslim’ as a master status and a continuously salient self-identification.’ His analogy to the highly influential ‘methodological nationalism’ (Wimmer and Glick Schiller 2003) critique of the nation-state-centredness of migration research, shows how central – yet unacknowledged – he perceives this problem to be. Brubaker makes a compelling call for social scientists to use greater reflexivity and social awareness when they apply ‘Muslim’ as a category of analysis, and to be clear about how this relates to and shapes understandings of ‘Muslim’ in the social world, as a category of social, political and religious practice. Ten years later, it is hard to make the case that lessons have been learned and this has been done at all.

We think that a significant body of research on Muslim minorities in Europe constructs and reproduces a bias and overfocus on: group boundaries showing strong differences between ‘Muslims’ and (national non-Muslim, or Christian/secular) ‘Majorities’, to the exclusion of other boundaries ethnic, race, class, gender, etc.; religiosity to explain Muslims’ behaviour and attitudes, at the expense of other cultural or socio-economic variables; and value conflicts (over gender, religion and freedom of speech, etc.) to explain mutual social relations between Muslims and Majorities, rather than other forms of socio-cultural interaction (such as friendship and neighbourhood ties, and shared identities, etc..) that cut across group boundaries. These general tendencies are what we mean by the Muslimification of Muslims by academic research. It is a sociological imagination deficit and methodological flaw that we critique and try to challenge, empirically. We are not aware of other attempts to do this. The argument is straightforward: lumping all Muslims together as a single group and using religiosity as an über-explanatory variable without consideration of alternative socio-cultural explanations is deeply problematic but common for studies of ‘value conflicts’. Such problems (and the assumptions behind them) are seldom acknowledged. It seems hard for research on Muslims to break out of this dominant mode of interpretation, that is path dependent, and becomes almost a self-fulfilling prophecy.

This paper is written in the spirit of challenging the assumptions and biases that leads to a Muslimification of Muslims. The aim is to challenge some assumptions of survey research on Muslims by undertaking survey research on Muslims. We apply the same method (analysis of survey data) that is very often used to ‘find’ and ‘explain’ supposed Muslim differences to majority populations. To do this, we turn the research inquiry around, so that the ‘Muslim’ category and ‘religiosity’ become the objects rather than tools of analysis. This approach fits the spirit of Brubaker’s suggestion (2013) and is the only research to do so that we are aware of. Revisiting the EurIslam data-set that was explicitly designed to examine socio-cultural gaps between Muslims and non-Muslim Majorities across six European countries (see Statham and Tillie 2016), our research inquiry mimics a well-trodden path by examining differences between Muslims and majorities over ‘liberal democratic values’.

Theoretically, we work within a broad acculturation/cultural assimilation perspective that is the common terminology for discussions on the cultural adaption of minorities of immigrant origin (on acculturation, see Berry 1997; on the cultural dimension of new assimilation theory, see Alba and Nee 2003; Alba 2020, 2024; Alba and Foner 2015). As Europeans we will use the term acculturation. However, we acknowledge this perspective draws significantly on the work of Richard Alba (see Statham and Foner 2024 this volume for discussion), not least his influential reformulation that acculturation/assimilation should be understood as a two-way process of mutual interaction and cultural adaptation between minority and majority groups, that can lead to a shrinking of the social distance in values and norms and blurring of the group boundaries between them over time. In this view, there is no implication that acculturation is inevitable or occurs in a linear irreversible fashion, while there is still considerable space when acculturation processes occur for the expression of distinctive identities and cultural differences. Importantly, the degree to which minorities and majorities mutually interact across group borders is not a theoretical position, but a subject to be settled empirically (see discussion Ch. 6; Alba 2020).

In our research design, we do not just lump all Muslims together into one category, but also test for important differences by ethnonational family origin -ex-Yugoslavia, Morocco, Turkey, Pakistan – within the Muslim category. If categorisation by ethnonational family origin group is meaningful, then this questions, or at least relativises, the value of treating all Muslims the same over ‘liberal democratic values’. In the next step, we look at explanations for Muslims’ self-perceived differences to national majorities over ‘values.’ First, we test the explanatory power of religiosity, but importantly we also test for a set of non-religious cultural variables drawn from acculturation theories. This allows us to gauge the relative explanatory power of religiosity compared to other cultural factors that are often omitted in studies of Muslims in Europe. Last, we examine which religious and cultural factors are influential in shaping Muslims’ perceived differences to majorities by ethnonational family origin group. This provides comparative insight on the distinctiveness of acculturation trajectories by ethno-national origin group, and again queries how meaningful it is to ‘group’ all Muslims together over ‘liberal democratic values’.

In the next section, we look at the drivers behind ‘Muslim’ becoming so bright a boundary marker for minorities of immigrant origin in Europe. Then we look at how polemics over multiculturalism strongly shaped academic understandings of the presence of Muslims and Islam in Europe. Next, we critique a recent case of research on value conflicts and Muslims, to show an example of Muslimification, that is, how assumptions introduce a cultural bias into academic research. We then briefly present the EurIslam design and data, before presenting our own study. Our analysis is structured around the three queries over Muslims’ perceived differences to the national majority over ‘liberal democratic values’: differences by ethnonational family origin group among Muslims; religiosity vs cultural variables as explanations for Muslims perceived differences; and religiosity vs cultural variables by ethnonational family origin group. Finally, we reassess claims about Muslimfication in light of our findings.

When and why did ‘Muslim’ become the boundary? Public controversies over Muslims in Europe

Today, it is surprising to point out that 25 years ago ‘Muslim’ was a relatively unused category for studying groups of immigrant origin in Europe. Instead, it was routine to categorise minorities from Muslim-majority countries by ethnonational origin, as ‘Pakistanis’, ‘Turks’, ‘Moroccans’ or as ‘black’ or ‘Asian’³ groups in the UK, where the state sponsors racial categorisations. Of course, minorities of immigrant origin have several overlapping, cross-cutting and competing identities, and may at times and within different contexts, hold one of these multiple ethnonational, religious or racial identities as the most important. At the same time, they are confronted by state immigration and integration policies in their settlement societies that sponsor specific status categories – immigrant, foreigner, ethnic minority, racial minority, asylum seeker, etc. – as well as dominant public discourses, that can importantly influence and alter the balance of their self-identification (Koopmans et al. 2005, 113–115). In the social world, how minorities self-identify as groups is not given, but the us and them of social grouping are constructed in their meaningful interactions with members of dominant majority populations, public institutions, and with other minorities, and importantly how these others see them (for discussion, see Jenkins 1997). This is the basic idea of symbolic boundary drawing within studies on cultural assimilation and acculturation that examine socio-cultural distance between groups (Alba and Nee 2003; Alba 2020, 2024).

What happened to make the religious pan-identity ‘Muslim’ so salient, as a self-identification, and as the dominant category by states and public discourses, for the very diverse minorities who are Muslims across Western Europe? The simple answer is that post 9/11 and other terrorist attacks by Islamic extremists, Western military interventions in the so-called War on Terror, and a period of global Islamic resurgence, divisions between Islam and the West became a highly prominent politicised cleavage in all European countries. Islam is now the key site for the demarcation of boundaries between majority populations and individuals of immigrant origin across Europe. Issues around the perceived and real problems of the integration of Muslims and accommodating Islam as a religion resonate fiercely in public debates. Seemingly endless political controversies over teachers wearing headscarves, mosque-building, public funding of Islamic schools and real or perceived incidents of Islamic fundamentalism, resonate in the mass media, reinforcing a strong demarcation and bright boundaries over differences between Muslims and non-Muslims in the public imagination, that impacts on the way that people relate to one another in the social world. The presence of Muslims has been increasingly politicised as a general cleavage over culture, religion, values and belonging to the national community in a way that is consequential for these groups: ‘(T)here have been intensified debates about the need for migrants and their descendants who are Muslims to identify with their countries of residence and to accept their core values and norms, especially in the domains of democracy, separation of church and state, and gender equality’ (Statham and Tillie 2016, 179).

This increasing focus on cultural differences seen through a lens of ‘liberal democratic values’ has transformed basic understandings of ‘integration’ for Muslims in Europe. General policy approaches have shifted from framing minority integration as problems of discrimination and social inequality, towards questions of cultural values and community cohesion. At the same time, the so-called ‘end of multiculturalism’ in the Netherlands and Britain brought in new ‘citizenship tests’ which require applicants to meet a set of constructed ethnonational values, ‘Britishness’ or ‘Dutchness’, assumed to represent the majority national society – a far cry from the explicitly civic duties and values that used to be naturalisation criteria. Some state policies directly target ‘Muslim’ minority populations – for example, PREVENT in the UK (O’Toole 2021) – as the group that poses extremist threats to and finds it hardest to fit in with liberal democratic norms and public institutions. Again, this seems designed to make it harder for minorities from Muslim-majority countries to join the national community. Regarding public representations, dominant mass media discourses often carry stories that focus on ‘value conflicts’ in a way that perpetuates reductive and stigmatising stereotypes of Muslims. This exaggerates the prominence of public controversies over Islam in the public imagination relative to other less controversial social issues that are most likely more important in shaping the lives of Muslims and their social relations with non-Muslims. At the same time, Muslim minorities themselves increasingly self-identify as ‘Muslim’. This can be understood in part as a reaction to being othered by the strong demarcation against ‘Muslims’ – ‘reactive religiosity’ (Voas and Fleischman 2012) – but also has its own dynamic. The increasing salience of being Muslim among second and third generations is part of a global resurgence of Islamic faith beliefs and lifestyles (Berger 1999), even if it is explicitly ‘made in Europe’. In sum, the constructed difference of ‘Muslims’ over values and religion in contrast to national majority populations of liberal democracies has become a public norm – an established ‘social fact’ that is seldom questioned, nor treated with sufficient reflexivity.

Academic scholarship has participated fully in this radical transformation for understanding Muslim integration in Europe. There has been an increasing focus on studying Muslims’ culture and religion in their own right – that is, assumed or real ‘value conflicts’- instead of how culture is a factor within socio-economic integration and structural inequalities, for example, in research on ‘ethnic penalties’ to examine labour market discrimination. Previously discreet research fields on liberal states’ accommodation of minority religions and Islam (see e.g. Fetzer and Soper 2005; Statham et al. 2005; Laurence 2012) and moral value differences (see e.g. Diehl, Koenig, and Ruckdeschel 2009; Röder 2014) became increasingly mainstream and used as a lens for interpreting assumed general divisions in society (see e.g. Koopmans 2015 who compares Muslims and Christians on religious fundamentalism and hostility to outgroups). This shift was driven strongly by resonant controversies over multiculturalism, that in Europe are almost exclusively about Muslims. Scholarly ‘culture wars’ replicate and legitimate equivalent political polemics in the public domain.

Politicised ‘culture wars’ over multiculturalism: the shift towards ‘liberal democratic values’ as a test for ‘Muslims’

The cleavage over multiculturalism in Europe is primarily about the relationship between Muslim minorities and Islam and the liberal democratic nation-state and society. From one side, some advocates of multiculturalism argue that Islam and being Muslim is just another cultural attribute and that state actions can encourage beneficial interactions and acculturation processes between Muslim minorities and other members of society. This is Modood’s stance (2000, 188): ‘The political demands of Muslims  …  are for some degree of Islamicisation of the civic; not for getting the state out of the sphere of cultural identities, but in some small way for an inclusion of Muslims into the sphere of state-supported culture.’ In this view, Muslims’ demands for Islamic group rights are not incommensurable with the values of liberal democracies but can be accommodated, especially if the state actively supports acculturation processes and counteracts discrimination by majorities. It needs noting that some multicultural theorists see Islam as less culturally adaptable and benign than Modood.⁴ But for the most part, they still emphasise that liberal democracies can incorporate Muslims relatively unproblematically, and even if Islam remains a problematic issue, it is a contained and manageable one.

Against this, anti-multiculturalists find the religion of Muslims an incommensurable historic barrier, that is, a set of values that challenges liberal democracy and prevents social cohesion. This is Islam as a barrier to acculturation. Huntington’s (2002) infamous ‘clash of civilisations’ thesis set the agenda: it depicts Western societies pulling themselves apart along religious divisions between Muslims and Christians as a direct result of immigration. Scholars who reject multiculturalism focus on the apparently irresolvable ‘value conflicts’ between Muslims and Western liberal democratic values often grounded in religious difference. Today most versions of similar arguments are less explicit and more sanitised than the ‘clash of civilisations’. Sniderman and Hagendoorn’s (2007) ‘collision of values’ argues that the problem arises from Muslim minorities’ behaviour and attitudes, while the blame lies with the political elite’s alleged support for multiculturalism:

‘The fundamental issue, it turns out, is not diversity but loyalty. Do Muslim minorities want to adopt the country that they have come to and its core values as theirs? Or do they want to live in it, but not be part of it, reserving their fundamental loyalty for the country they came from and its culture and institutions? They are questions that cut deep … Among the many ironies of our story, this is perhaps the most gratuitous. Multiculturalism encouraged an ambiguity of commitment’ (2007, 11).

Note how seamlessly the authors shift cultural diversity into a question of Muslim minorities’ assumed disloyalty to the Netherlands, something which only they, that is, Muslims, not the Dutch majority population, can affect and have responsibility for addressing. Sniderman and Hagendoorn’s (2007) thesis requires that Muslims undergo a one-sided cultural assimilation to adopt the core values of a culturally ‘thick’ ethnonational ‘Dutch’ liberal democracy. It requires that they reject the culture and institutions of the country they came from -ignoring that 2nd and 3rd generations are born, grew up and socialised in the Netherlands – otherwise they are not part of Dutch society. This is all presented as if it is a question of individual choice for Muslims. However, the cultural baggage and religious faith someone is born into cannot be viewed as an individual choice in a meaningful sense. Family background is culturally ‘sticky’, especially on religion because liberal states uphold freedom of religion, so make few efforts to change minority religious beliefs. This idea of one-sided acculturation is outdated and largely discredited, not least since Alba and Nee’s (2003) new assimilationism made clear it is a two-way process of adaptation by majorities and minorities. Likewise, transnational perspectives have long established that being committed to ‘here’ or ‘there’ is not a zero-sum binary individual choice (see Bivand Erdal and Oeppen 2013). Last even the authors’ claim that multicultural policies created an ambiguity of commitment among Muslims in the Netherlands is political claim that lacks empirical verification. They have no data or analyses on Muslims’ attitudes, and no data that relates ‘Dutch’ attitudes to policy decisions, or over time.⁵

Sniderman and Hagendoorn’s (2007) ‘collision of values’ research was intentionally provocative. It aimed to provide legitimacy to the resonant anti-multicultural and anti-Muslim sentiments in the Netherlands at a time when politics was turning against multiculturalism. The problem is that they shoehorn ‘liberal democratic values’ into the basis of an assumed ethnocultural cleavage between Muslim immigrants, on one side, and Dutch Majorities and liberal democracy, on the other, then raise this to a general theory of society. Viewed through their lens, ‘liberal democratic values’ constructs a bar so high that people with foreign family origins in Muslim countries can never reach it – that seems to be the intention.

Of course, many academic studies try to examine value differences between Muslim minorities and majority populations over religion, gender, sexuality and moral issues, within context and without raising this to a general theory of society.⁶ This research challenges some of the exaggerations, assumptions and stereotypes about ‘value conflicts’ that appear in the polemics on multiculturalism and Islam, empirically. In their influential study on values between Muslim and Western countries based on World Values data, Norris and Inglehart (2004, 155) conclude that, ‘The most basic cultural fault line between the West and Islam does not concern democracy – it involves issues of gender equality and sexual liberalization.’ Similarly, Joppke (2015, 150) finds, ‘most importantly, the real integration deficit seems to be with respect not to political but to moral-ethical values.’ These findings counter the idea of conflicts over political values and point to cultural differences in sexual mores and gender attitudes, that are less of a challenge to liberal democracy and more open to acculturation processes. Regarding the overfocus on religiosity as an explanatory variable, estimates show that only a third of Muslims in Europe practice their religion strictly, with the rest non-practicing or agnostic (Dassetto, Ferrari, and Marechal 2007, 8; Joppke 2015, 156). This places into question the common conflation of being Muslim as an ‘Islamic’ rather than an ‘ethnic’ identity. It counters the assumption that all Muslims of immigrant origin practice Islamic faith and relativises the idea that all people who identify with Islam present a fundamentalist challenge to liberal democratic societies. Similarly, but when discussing majority discrimination against Muslim minorities of immigrant background, Adida, Laitin, and Valfont (2016, 5), concur that, ‘research to date has assumed, rather than shown, that religion is the source.’

Survey-based research on ‘value conflicts’: reproducing Muslim stereotypes?

The main point that we want to emphasise is about the normative assumptions behind the standard survey research that depicts ‘value conflicts’ as a general feature of society. Notwithstanding the high technical quality and accuracy of their analyses, seemingly neutral studies on ‘value conflicts’ and ‘Muslims’, can often do little more than emphasise the ‘otherness’ of Muslims as cultural aliens, because that is the premiss on which the idea of value conflicts is based. A recent example presented at the same conference as this paper -a gathering explicitly focussed on ‘culture’ in integration- is illustrative and underlines the continued salience of this type of academic thinking.⁷

Ivarsflaten et al. (2022) offer an interesting study based on majority opinions on Muslims, that tries to show that some seemingly closed ‘value conflicts’ (i.e. irresolvable ones) can be made open to negotiation, through innovative acts and cultural practices. The authors select ‘handshaking controversies’ as an example that ‘appear(s) to pit gender equality against religious freedom in a closed value conflict logic’ (2022, 2). The basic idea is that Muslims refuse to shake hands with (non-Muslim) officials in public settings, especially when the official is a woman. This is tested by presenting respondents from the majority population with a range of fictious scenarios of handshaking controversies, varied by gender and across settings. Then the authors ask, if the majority becomes more tolerant, if Muslims are prepared to offer another gesture, that is, ‘placing the hand on the heart’ rather than shaking hands. The experimental study is embedded within a representative sample of the German population. The basic finding is that that majorities insist that Muslims should shake hands, especially with female officials, but this becomes less so, when Muslims offer the alternative gesture of placing hand on heart. From this, the authors conclude that ‘conflicts over handshaking are thus open to resolution through cultural ingenuity’ (2022, 2).

The study’s technical application is a sound piece of social science. However, the assumptions merit scrutiny. First, the selection of a controversial event (in this case constructed hypothetically) to pursue the research question, normalises the exceptional and makes it appear representative of events in the social world. This is highly questionable. ‘Handshaking’ is not a common controversy of this type. The constructed scenarios are most likely of miniscule relevance to social relations between Muslims and non-Muslims, although they claim to mimic everyday life. Second, a fundamental problem is that ‘handshaking’ is not a specified problem within Islamic faith, nor for most women and men with family origins in Muslim-majority countries. This means that it is only a good indicator for pitting ‘gender equality against religious freedom’ (2022, 2) if one starts from stereotypical (and mostly false) assumptions that not shaking hands is a general requirement of Islamic faith, that all Muslims are the same, and that religiosity shapes their behaviour. While these might be common assumptions among majority populations and in media representations, we think the academy can do better. Here we run again into the problem of reification, when all Muslims are lumped together and presented as if they are a solidary social group with distinctive cultural/religious characteristics. Third, the findings suggest that if Muslims would place their hand on their heart instead of refusing to shake hands, then a significant portion of the majority population would accept this, something the authors take as indicative of potential ‘inclusive tolerance’ by majorities. However, the authors see the cultural ingenuity required to open up this closed value conflict as a one-sided (not two-way) acculturation process. In other words, Muslims have to culturally adapt -in this case by inventing and adopting a custom that has no meaningful prior cultural heritage- so that majorities can be considered to be ‘inclusively tolerant’. Finally, even if the research remains focussed on majority opinions (and they choose not ask Muslims what they think), why could respondents not be presented with a scenario where a white supremacist or member of the extreme right (i.e. non-Muslim people who explicitly challenge and reject ‘liberal democratic values’) refuses to shake hands with an official? That would at least extend the frame of reference for understanding ‘democratic value conflicts’ beyond ethnocentric stereotypes about Muslims.

This paper: using survey research to critique assumptions about Muslims

This article mimics the standard approach of survey research on Muslims and ‘liberal democratic values’ to test and challenge some of the cultural assumptions behind such perspectives. Specifically, we examine the similarities and differences in acculturation over ‘liberal democratic values’ between Muslim minorities of immigrant origin from four different family countries of origin (Ex-Yugoslavia; Morocco, Turkey, Pakistan) in four countries of settlement (NL, UK, F, D).

Research Question: How do individuals who are Muslims, but have different family ethnic national origins (Turkey, Morocco, Pakistan and Former-Yugoslavia), vary in their self-perceptions of similarities/differences relative to majority populations over ‘liberal democratic values’ in their respective countries of settlement (The Netherlands, Britain, France, and Germany)?

Against the idea that being Muslim and Islam is automatically a barrier to being part of a liberal democratic society, we think that the degree to which minority groups who are ‘Muslim’ are socially distant in norms, values, and behaviours to other groups in society, or not, is a question to be resolved empirically. This draws on the perspective of acculturation or cultural assimilation as a two-way process of socio-cultural interaction between minorities and majorities, that can lead to a shrinking of the social distance between them over norms and values over time (see Alba 2020, 2024). For research design, it requires including non-religious cultural variables and giving them equal weight to religious variables. Mainstream acculturation literatures include several non-religious factors that can potentially reduce social distance between minorities and majorities, leading to more cohesive relations across group boundaries (see Statham and Tillie 2016), including: sharing a language (language competencies of minorities in the dominant language are important in this respect); mutual identification and acceptance that promotes common interests and solidarity across ethnic and religious groups; social networks and trust, especially through ‘bridging capital’ that reaches out to the outgroups; intermarriage that leads to greater family engagement across group boundaries; and shared norms and values as the basis for democracy in a diverse society, that is, ‘liberal democratic values’, although as we have noted the definition of what is contained with this is disputed.

Our inquiry into the social distance and boundaries between ‘Muslims’ and ‘Majorities’ over ‘liberal democratic values’ proceeds in three distinct steps.

First, we study self-perceptions of difference to the outgroup over ‘liberal democratic values’ between ‘majority’ and ‘Muslims’. Instead of lumping all Muslims together, we use the EurIslam data-set to examine variations between ethno-national groups of Muslims from the four different family countries of origin. Simply put, this allows us to test whether there are discernible patterns by family country of origin for Muslims that suggest distinct acculturation trajectories by ethno-national group – that is, that all Muslims are not the same because they are Muslims, but that ethno-national group matters for views on ‘liberal democratic values’.

Second, we look at the explanatory power of variables that would account for this self-perceived difference by Muslims relative to the ‘majority’ over liberal democratic values. Specifically, we aim to test the explanatory power of religiosity relative to non-religious cultural variables, including self-identification with family origin country, and with settlement country, settlement country language proficiency, media use in settlement country language, ‘bridging’ social capital with majority outgroup, and intermarriage with the majority in the family. Here our objective is to gain insight into the specific interplay of cultural factors that can increase or decrease Muslims’ perceived social distance to the ‘majority’ over liberal democratic values. This allows us to unpick the degree to which religiosity is a bright boundary and incommensurable barrier for Muslims to self-identify with ‘liberal democratic values’, as it is often depicted in academia and the public domain, or whether other cultural factors play a relatively more or less influential role than religiosity in increasing or decreasing the possibilities for acculturation. In other words, is it all about Islam as a religion, or does culture matter in other ways, that could bring other more acculturative outcomes?

The third step combines the previous two by examining the relative explanatory power of the religious and non-religious cultural variables for the four ethno-national groups by family country of origin. Here we ask whether the same cultural factors hold for explaining how Muslims who are Turks, Moroccans, Pakistanis and from former Yugoslavia, self-perceive their difference from the ‘majority’ over ‘liberal democratic values’? Does one-size fit all Muslims, or are distinct acculturation trajectories discernible by ethno-national group?

Data source and sampling: EurIslam survey design features

We use a publicly accessible large-scale survey data-set (7000+ interviews) generated by the EurIslam project.⁸ This source has been validated by many published articles.⁹ The data-set’s design makes it especially useful for our purposes, not least because questions aimed explicitly, ‘to retrieve information on variables relating to socio-cultural integration, including language competences, identification and belonging, core norms and values, and social capital’ (Statham and Tillie 2016, 185). Importantly, the questions retrieve the cultural information that matters for our inquiry by including religion alongside a range of other cultural variables. Also, the project oversamples people who are Muslims of immigrant descent residents in six Western European countries. This allows ‘Muslim’ sample sizes that are meaningful and large enough to work with, especially when compared to those retrieved from general population surveys designed for other purposes.

EurIslam’s sampling strategy for Muslims has another key characteristic: it selects people whose families come from four origin countries – Turkey, Morocco, Pakistan and former-Yugoslavia. The sample includes Muslims resident in one of the settlement countries, who were born in one of the origin countries, or who have at least one parent born there. This ties the sample to people with family immigrant backgrounds and excludes native Muslims and ‘white’ converts. Each family origin country represents a distinctive linguistic and cultural regional source of immigration to Western Europe – Central and Eastern Europe (former-Yugoslavia), the Maghreb (Morocco), Asia (Turkey) and South Asia (Pakistan). The selection includes at least one important source of postwar immigration from a predominantly Muslim country for each of the four settlement countries we use in this paper – the Netherlands (Moroccans, Turks), France (Moroccans), Germany (Turks) and Britain (Pakistanis). The EurIslam data on Britain, Germany, France and the Netherlands is used for this study, because these are the countries with the largest Muslim minority populations, and they are the cases that the author is most familiar with as a researcher making interpretations. This excludes Switzerland and Belgium, two complex multi-lingual multi-ethnic countries.

Moroccans, Turks and Pakistanis are identifiable in their settlement countries as ethnonational groups and as minorities of people of colour. By contrast, Muslims from former Yugoslavia arrived as refugees from the conflict in the early 1990s. Mostly, they have ethnonational identities as Bosniak and Kosovo-Albanian, though a very small minority could be Macedonian, Montenegrin or Serb. Due to this complexity, we refer to ‘Ex-Yugoslavian’ in the analyses here, while recognising this means for the most part Bosniak and Kosovo Albanian as ethnic or ethnonational group. Including this ‘homeland’ national culture dimension is important for our inquiry, partly because the attitudes and practices often ascribed to Islam, for example, moral conservatism, might actually result from non-religious factors specific to the cultural heritage of an origin country or region. Of course, Morocco, Turkey and Pakistan are Muslim-majority countries, whose distinctive national cultures, are importantly historically shaped by Islam, albeit in different ways and by different branches of Islamic faith, just as Western liberal democratic countries are shaped by Christian faiths. Still, this does not mean that all aspects of Moroccan, Turkish or Pakistani culture, whether in the ‘homeland’ or present as the characteristics of an ethnonational group in the settlement country, ought to be considered religious and ‘Islamic’.

EurIslam used CATI techniques (computer-assisted telephone interviewing) to retrieve samples of four Muslim groups in each settlement country using an onomastic (name-based) sampling frame that identified typical Turkish, Moroccan, Pakistani, Bosnian-Muslim and Kosovo-Albanian first and family names from electronic phone directories (landline and mobile). Professional polling agencies conducted the interviews using the questionnaire in each country in 2010–2011. Bi-lingual interviewers were used so that interviewees could choose to be interviewed in the dominant language of the settlement country or in Turkish, Moroccan-Arabic, Urdu, Bosnian or Albanian. Often surveys, and especially general surveys, do not offer a possibility for the choice of language to minorities in interviews. This is problematic for research on minorities, and particularly for questions on sensitive cultural issues because it biases samples to more acculturated respondents.

A last important feature is that the data-set includes a sample of non-Muslim ‘nationals’ for each settlement country – that is, people born in one of the settlement countries, who also have two parents born there. These national non-Muslim ‘majority’ samples are used as a baseline against which to measure the Muslims groups’ similarities and differences in attitudes and behaviour – that is, the relative social-cultural distance or gap to the national ‘majority’. From a methodological perspective, it is essential to have data from both sides (ingroup/outgroup) to study a symbolic boundary-marker, because acculturation is a two-way process. Without data on ‘Muslim’ and ‘majority’ sides of the equation, it becomes impossible to measure the socio-cultural distance between them meaningfully. The risk is falling back on assumptions and stereotypes about national cultural values or ‘Muslim’ differences as a proxy for data on the missing group.

Analysis and findings: making ‘Muslim’ and ‘Religiosity’ research objects and not just tools for studying ‘liberal democratic values’

Query 1: Muslims’ self-perceived differences to majorities over ‘liberal democratic values’: variations by family ethno-national origin

In the first step, we study how people who are Muslims perceive their own similarities and differences over ‘liberal democratic values’ relative to the national majority (as an outgroup) in their respective settlement country. Our specific aim is to see if there are distinctive patterns within the group of Muslims, when we distinguish between their family ethnonational backgrounds in specific origin countries, former-Yugoslavia, Morocco, Turkey and Pakistan. This speaks to our core inquiry about the validity of lumping all Muslim minorities of immigrant origin together under a single pan-identity ‘Muslim’. Does family ethnonational origin, another type of distinctive symbolic boundary marker, make sense as a category of analysis?

We noted earlier that there is not a single definition for Western ‘liberal democratic values.’ It is a contested field that leads to some slippage across political and social values, but most scholars include some combination of commitment to: a separation of religious and secular authority, the rule of law and social pluralism, institutions of representative democracy, protection of individual rights and civil liberties, and gender equality (Norris and Inglehart 2004). To measure self-perceived distances to outgroup over ‘liberal democratic values’, we use survey questions that ask how similar or different respondents think most people from their respective outgroup are to themselves over three issues: freedom of speech, gender divisions and the role of religion in society. This issue selection represents the standard selection applied for discussions of ‘liberal democratic values’ with respect to Muslims.

For Muslim respondents, the outgroup is the majority from their respective country of settlement, that is, British, Dutch, French or German. Majority respondents are simply asked about Muslim as an outgroup. Responses score on a four-point scale, ranging from very similar to very different.

Survey Question:

How different or similar do you think most people of (French/British/German/Dutch – Muslim) origin are compared to you on specific topics? Are they very similar (1), quite similar (2), quite different (3) or very different (4) to you?

• In the way, they think about freedom of speech

• In the way, roles are divided between men and women in households

• In how they think about the role of religion in society

We start by discussing the broad binary categories Majority and ‘Muslim’. This effectively mimics survey research that just selects ‘Muslim’ as an analytic category, by lumping the four ethnonational minority groups together, then evaluating them against a measure for majority values. After controlling for age, gender and education, we conduct a one-way analysis of variance that allows comparison across the groups by estimated marginal mean score. Table 1 shows findings for respondents’ perceived similarities and differences over ‘liberal democratic values’ to outgroup. ‘Muslim’ simply combines the four groups by family ethnonational origin (ex-Yugoslavian, Moroccan, Turkish, Pakistani), while ‘Majority’ is a broad category for non-Muslim nationals in the settlement countries. The columns show means for freedom of speech, gender and religion, as well as all three issues combined, that we take as ‘liberal democratic values’. Put simply, a higher mean score (range from low 1 to high 4) indicates a group’s perception of stronger differences, that is, they see a greater socio-cultural distance to the outgroup.

Table 1. Self-perceived socio-cultural distance to outgroup over ‘liberal democratic values’ in respective settlement country estimated marginal means: controlled for age, gender and education (Four point scale 1–4: 1 very similar; 2 quite similar; 3 quite different; 4 very different).

	Issues combined	Freedom of speech	Gender in the household	Role of religion in society
Muslim^+	2.70	2.47	2.69	2.93
Majority	3.14	2.91	2.91	3.27

⁺ four family ethnonational origin groups combined to make ‘Muslim’.

A first finding is the very strong (statistically significant) degree to which national majorities see people who are Muslim to be different than themselves over ‘liberal democratic values’. Symbolic boundaries are constructed from both sides, and the ‘Muslim’ group also sees their respective national majority as (statistically significantly) different from themselves. Taken at face value, this confirms that a bright symbolic boundary between Muslim/Majority exists over ‘liberal democratic values’, at least in people’s imaginations. But importantly, the data also gives a hint to what drives this relationship, by showing that the degree to which national majorities see Muslims to be different from themselves, is much more than the other way round, that is, the degree to which Muslims see their national majorities as different (see e.g. for all issues combined: Muslim 2.70; Majority 3.14). In other words, the bright boundary over ‘liberal democratic values’ is especially strongly constructed from the side of Majorities, compared to Muslim minorities.¹⁰

Dominant public discourses seem to have left their mark on how British, Dutch, French and German majorities see these issues. When they see the world through a lens of ‘liberal democratic values’ majority perceptions reinforce high barriers to acculturation for Muslim minorities. Indeed, if significant socio-cultural distances over ‘liberal democratic values’ are to be defined as a problem, then majority perceptions of Muslims are a ‘social fact’ that importantly contributes to this problem. Put differently, any possibility for Muslim minorities to acculturate to ‘liberal democratic values’ is highly dependent on significant perception shifts by people from national majority populations too.

To move beyond just categorising Muslims as ‘Muslim’, we disaggregate the data by family origin country, which allows comparison by ethnonational minority group. Table 2 shows the estimated marginal means, again after conducting a one-way analysis of variance with controls for age, gender and education. The subscript letters on means are when one ethnonational group is highly statistically similar (p > .05) to another group, but statistically different from the others – for example, Turks and Pakistanis (subscript a) are statistically similar to each other over the three issues combined, but different to ethnonational minorities with origins in Morocco and Ex-Yugoslavia.

Table 2. Self-perceived socio-cultural distance to outgroup over ‘liberal democratic values’ in respective settlement country estimated marginal means: controlled for age, gender and education (Four point scale 1–4: 1 very similar; 2 quite similar; 3 quite different; 4 very different)

Family ethnonational origin	Issues combined				Freedom of Speech				Gender in the household				Role of religion in society
	M	SE	95% CI	Rank	M	SE	95% CI	Rank	M	SE	95% CI	Rank	M	SE	95% CI	Rank
Ex-Yugoslavian	2.46	.03	2.40/2.51	1	2.27a	.04	2.19/2.35	1=	2.53a	.04	2.46/2.60	1=	2.56	.04	2.49/2.63	1
Moroccan	2.62	.03	2.57/2.66	2	2.36a	.04	2.29/2.43	1=	2.52a	.03	2.46/2.59	1=	2.96b	.03	2.90/3.02	2
Turkish	2.87a	.02	2.83/2.91	3=	2.64b	.03	2.59/2.71	3=	2.85b	.03	2.80/2.92	3=	3.11a	.03	3.06/3.17	3=
Pakistani	2.85a	.03	2.80/2.90	3=	2.62b	.04	2.54/2.69	3=	2.85b	.04	2.78/2.92	3=	3.08ab	.03	3.02/3.15	3=
Majority	3.14	.02	3.09/3.18		2.91	.03	2.84/2.96		3.24	.03	3.18/3.29		3.27	.03	3.21/3.32

Note: Rank reflects perceived similarity to national majority (Rank 1 = highest perceived similarity).

Same subscripts within a column indicate that the estimated means do not differ significantly, p > .05.

Turning to variations by ethnonational group, it is striking how differences between the groups are not random but follow a clear structure and pattern across the issues. Muslims with family origins in former Yugoslavia see themselves as most similar to their national majority over ‘liberal democratic values’. This holds across the issues. Muslims with Moroccan heritage are close to the ex-Yugoslavians, and highly significantly similar in their perceived differences to the majority over freedom of speech and gender issues. But Moroccans differ significantly from ex-Yugoslavians over the role of religion, seeing much greater differences to the majority. By contrast, Muslims with family origins in Turkey and Pakistan are highly significantly similar to each other and see themselves as more different from the national majority to a significant degree compared to ex-Yugoslavians and Moroccans. This holds across all three issues, except for Moroccans over religion. Importantly, this clear structuring of similarities and differences between ethnonational minority groups provides evidence that there are distinctive views over the possibilities for acculturation over ‘liberal democratic values.’

In other words, when looking through a lens of ‘liberal democratic values’ not all Muslims are the same in the way they see their relationships to the majority. On the contrary, there still seems to be important analytic traction in using ethnonational categories of analysis, which unlike ‘Muslim’ do not automatically emphasise religion, and were after all the dominant categories for studying these same populations twenty years ago, prior to the salience of culture wars over Islam.

A last point to make in this discussion is methodological: it concerns the use of survey data for studying acculturation processes. Some survey-based studies on ‘value conflicts’ in Europe place the obligation for cultural adaptation -implicitly or explicitly- one-sidedly on the shoulders of Muslim minorities (Sniderman and Hagendoorn 2007 is an explicit example, see earlier discussion). This should be rejected as theoretically flawed since acculturation processes clearly need to be conceptualised and operationalised as a two-way process of adaptation between majorities and minorities. However, a common problem that remains in the research design of ‘value conflict’ studies is that they often use survey data on only one side of the equation, either on majorities, or on Muslims. Sometimes this is because of data limitations. But if acculturation is a two-way process, then it is methodologically necessary to collect and use data on majorities and minorities. In this study, without the (admittedly broadbrush) data we present on national majorities’ perceptions of difference, we would have no baseline against which to assess the chances for acculturation by minorities, nor would the strength of majority perceptions as a barrier be visible. When studying assumed or real ‘value conflicts’ the interpretive risks of having data on only one side of the boundary are heightened. Such studies tend fall back on stereotypical assumptions of what a national culture consists of -often rehashed from resonant media debates on ethnocentric ideas of ‘Britishness’ or ‘Dutchness’- as a baseline to which minorities are supposed to acculturate towards.

Query 2: Explaining Muslims’ self-perceived differences to national majorities over 'liberal democratic values': religion or culture?

In a second step, we look for possible reasons that account for Muslims’ self-perceived differences relative to their respective national majorities over ‘liberal democratic values’. Our specific focus is on culture: we test the explanatory power of religiosity relative to non-religious cultural variables. Non-religious cultural practices feature prominently in acculturation theories, including self-identification with settlement and origin countries, settlement country language proficiency and use, ‘strong’ and ‘weak’ social ties with the majority, intermarriage, etc.. (see earlier discussion). These potentially acculturative factors can be pitched against practicing Islam as a supposed barrier for Muslims over ‘liberal democratic values’. Our objective is to gain insight into which specific cultural factors -religious and non-religious- significantly increase or decrease Muslims’ perceived social distance to the majorities. This addresses the important question over the degree to which Muslims’ religiosity constitutes a barrier to their acculturation towards ‘liberal democratic values’ – that is, the master narrative in dominant public discourses on Muslim integration. This narrative is also an assumption and claim of many social scientific studies, not least those that do not test for non-religious cultural variables. Is it really all about religion? Or do Muslims’ other cultural practices matter too, perhaps even offering the possibility of acculturative outcomes? When studying Muslims as ‘Muslim’ there has been virtually no research within the ‘value conflicts’ field that looks for alternative explanations.

To assess the degree to which religiosity and cultural practices account for Muslims’ perceived differences to majorities over ‘liberal democratic values’, we conduct multivariate regression analyses with the sample of Muslims (grouped together) and for the three issues combined.¹¹ In the first model, we include socio-demographic, human capital and status variables (age, gender, level of education, country of settlement, country of family origin, immigrant generation, citizenship of settlement country, citizenship of family origin country). Taken together these factors provide a basic background picture for predictors that increase or decrease Muslims’ perceived differences to majorities. There is not space to discuss these findings in detail. Instead, we focus on the explanatory power of variables for Muslims’ religiosity (second model), and then their cultural practices (third model), and the additional explanatory effect that these have as predictors, net of the socio-demographic, human capital and status variables, for perceived difference to national majorities. Table 3 shows the findings. Operationalisation of the variables from the survey questions is presented in Appendix 1.

Table 3. Muslims’ self-perceived differences to the majority population over ‘liberal democratic values’ (issues combined): results of regression analyses

	Model 1:Demographic, human capital and status variables			Model 2: + religiosity			Model 3: + cultural (non-religious) variables
	b	(SE)	Beta	b	(SE)	Beta	b	(SE)	Beta
Intercept	2.48	.05	***	2.45	.05	***	3.00	.11	***
Gender (ref. men)	0.12	.02	.08***	0.10	.02	.07***	0.07	.03	.05*
Age (centred, in years)	−0.00	.00	−.01	−0.00	.00	−.02	−0.00	.00	−.03
Education (ref. primary or less)
Secondary education	−0.07	.03	−.05*	−0.06	.03	−.04*	0.00	.03	.00
Tertiary education	−0.19	.04	−.11***	−0.17	.04	.10***	−0.08	.04	−.05*
Family country of origin (ref. Ex-Yugoslavia)
Turkey	0.41	.04	.26***	0.39	.04	.24***	0.29	.04	.18***
Morocco	0.18	.04	.11***	0.12	.04	.07**	0.14	.04	.08***
Pakistan	0.40	.04	.22***	0.33	.04	.18***	0.36	.04	.20***
Country of settlement (ref. Germany)
France	−0.10	.04	−.06**	−0.11	.04	-.06**	−0.04	.04	−.02
United Kingdom	0.26	.03	.16***	0.29	.03	.18***	0.23	.04	.14***
Netherlands	0.12	.04	.07**	0.11	.04	.06**	.19	.04	.10***
Generation (ref. 1st generation)
1.5 generation	−0.11	.03	−.07**	−0.10	.03	−.06**	0.04	.04	.02
2nd generation	−0.07	.04	−.04	−0.07	.04	−.04	0.06	.04	.04
Citizenship country of settlement	−0.15	.03	−.09***	−−0.15	.03	−.09***	−0.06	.03	−.04
Citizenship country of Origin	0.10	.03	.06**	0.09	.03	.05**	0.06	.03	.04
Practicing religiosity				0.20	.03	.12***	0.10	.03	.06**
Self-identification with family country of origin							0.04	.01	.05**
Self-identification with country of settlement							0.11	.01	−.18***
Language proficiency country of settlement							−0.05	.02	−.07**
Media consumption in language of country of settlement							−0.04	.01	−.07**
Acquainted with members of majority population							−0.01	.01	−.01
Friends with members of majority population							-.03	.01	−.05*
Intermarriage in family							−.07	.03	−.05**
Explained variance (R²)		.119***			.132***			.19***

*** p < .001; ** p < .01; * p < .05

Briefly, the analysis of socio-demographic, human capital and status variables mostly follow expectations from the extent literature. Muslim women perceive differences to majorities over liberal democratic values significantly more than men. Age does not make a difference. The better educated see fewer differences to majorities. The 1.5 generation is less likely to see differences to the majority than the first generation, but this is less so for the second generation. Possessing citizenship of the settlement country reduces Muslims’ perceived differences to majorities while holding that of the family origin country increases perceived differences, but to a lesser degree. Finally, we see that family country of origin matters significantly, as does settlement country, but to a lesser degree. This finding – that is, that where your family comes from (ethnonational origin) matters more than where it settled (settlement country)- is relevant to this paper’s inquiry. It underlines earlier findings that family ethnonational origin matters.

Moving to the second step, we use two survey questions to operationalise religiosity as a cultural practice:

Survey questions:

Q. To what extent do you see yourself as Muslim? (Very strongly; Strongly; Somewhat; Hardly; Not at all)

Q. How often do you pray? (several times a day; once a day; once or a few times a week; only on special occasions; never; refused; don’t know)

One question asks how strongly they identify as ‘Muslim’ on a five-point scale (ranging from ‘very strongly’ to ‘not at all’), while the other asks about the frequency of praying on a five-point scale (ranging from ‘several times a day’ to ‘never’). We consider people to be religious, if they identify strongly or very strongly as ‘Muslim’ and pray at least once a day.

The second column shows what happens when we add religiosity into the mix. Indeed, we do find that there is a significant unique effect (0.12) of religiosity, meaning that being religious and practicing Islam increases perceived differences to majorities over ‘liberal democratic values’. However, the degree to which religiosity increases explained variance on top of the socio-demographic, human capital and status variables, is significant, but actually relatively small (an additional 1.3% – R² from .119 to .132). This finding about the modest impact of religiosity matters. It goes against most dominant public and political discourses on ‘liberal democratic values’ that assume that it is Muslims’ religiosity and practicing Islam that is the crucial (incommensurable) barrier that prevents their acculturation. Likewise, it questions the centrality of religiosity as an über-explanatory variable in many social scientific studies on Muslims and ‘value conflicts’. Of course, there is a correlation, and if researchers stop the inquiry there -as many do- then they have religiosity as an answer. But what happens when we bring other cultural variables into the equation?

The third model in Table 3 adds a set of non-religious cultural variables drawn from acculturation theories. Following these, if individuals from Muslim minorities have, a high self-identification with their settlement country, a high proficiency in the settlement country language, a high consumption of media in the settlement country language, many acquaintances with majority population (weak ties), many friendships with majority population (strong ties) and family experiences of intermarriage with the majority, then we expect them to see fewer differences between the national majority and themselves over liberal democratic values. Conversely, if they strongly self-identify with their family country of origin, this leads to expectations of greater self-perceived differences to the national majority.

The findings are striking. These cultural variables have a significant impact -explained variance R² increases by another .058, an additional 5.8% – that is greater than religiosity and pushes overall towards acculturation. Importantly, adding these cultural practices to the story even reduces the significant unique effect of religiosity (from .12*** to .06**) for explaining perceived differences to the majority. High self-identification with the country of settlement has the strongest acculturative effect – it is clearly stronger than religiosity and pushes in the opposite direction. Proficiency and media consumption in settlement-country language, plus having family experiences of intermarriage, have an acculturative effect roughly equivalent to religiosity, but again push in the opposite direction. By contrast, high self-identification with family origin country increases perceived differences to the majority, like religiosity, and to a similar degree. Social ties (weak and strong) and bridging capital with majorities do not seem to have an effect.

Overall, the picture that comes from these findings is that religiosity should be seen as one cultural factor among several, and that it is not the cultural factor that drives the relationship over Muslims’ perceived differences to majorities over ‘liberal democratic values’. On the contrary, the degree to which Muslims identify with their nation of settlement matters more, which opens up the possibility of a pathway towards acculturation and downplays practicing Islam as a barrier to social cohesion.

Query 3: Accounting for self-perceived difference to majority over ‘liberal democratic values’, variations by family ethnonational origin group

The findings in Query 2 still involved lumping all the Muslims together as a single group. Here we disaggregate and follow the same line of inquiry by ethnonational minority group. This allows to see if similar or different factors are at work that account for their perceived differences to the majority over ‘liberal democratic values’ across the groups of Muslims. Generally, strong differences between ethnonational groups would indicate distinctive acculturation trajectories, while similarities suggest that it makes sense after all to categorise them together into a single Muslim category for ‘liberal democratic values’ studies. Table 4 shows standardised regression coefficients based on the full model per group for Muslims with Ex-Yugoslavian, Moroccan, Turkish and Pakistani family origins.

Table 4. Self-perceived differences to majority over ‘liberal democratic values’ by family ethnonational origin: Full model per group (standardised regression coefficients).

	Former-Yugoslavian	Moroccan	Turkish	Pakistani
Gender (ref. men)	.10*	.06	.02	.01
Age (centred, in years)	.01	.11	−.22***	−.01
Education (ref. primary or less)
Secondary education	.09	.01	−.04	.03
Tertiary education	.03		−.15***	−.03
Country of settlement (ref. Germany)
France	−.04	.03	−.09*	−.05
United Kingdom	.13**	.22***	−.01	−.27***
Netherlands	.10*	.25***	−.06	.15**
Generation (ref. 1st generation)
1.5 generation	.09	.09	−.09*	−.02
2nd generation	.03	.14*	−.13*	−.00
Citizenship country of settlement	−.10*	−.00	.03	−.03
Citizenship country of origin	−.01	.06	.08	.05
Practicing religiosity	−.02	.08	.05	.11**
Identification with country of settlement	−.27***	−.10*	−.20***	−.10*
Identification with country of origin	.09*	.02	.03	.05
Language proficiency country of settlement	−.04	−.07	−.07	−.11*
Media consumption in language of country of settlement	.03	−.07	−.05	−.11*
Acquainted with members of majority population	.02	−.04	−.05	.06
Friends with members of majority population	−.13**	−.13**	.01	−.01
Intermarriage in family	−.01	.04	−.06	−.03
Explained variance (R²)	.202***	.185***	.171***	.167***

*** p < .001; ** p < .01; * p < .05

The striking overall picture is one of clear differences in the factors that account for ex-Yugoslavian, Moroccan, Turkish and Pakistani Muslims’ perceived differences to the majority over ‘liberal democratic values.’ Contrary to the commonly applied research assumptions, one size does not fit all. Providing detailed explanations for how each ethnonational minority group sees differences from the majority would require digging deeper into the data and a different research design. However, even this general comparison points towards differences, that become especially interesting regarding our focus on cultural variables. The first striking takeaway is that religiosity is a statistically significant factor that increases perceived differences to the majority only for Pakistani Muslims, and not for ex-Yugoslavian, Moroccan and Turkish Muslims. Again, this finding challenges the central status that is often given to religion and practising Islam as an über-explanatory variable in discussions over ‘liberal democratic values’ conflicts. Practising Islam might be a factor for how some Muslim people see themselves as different from the dominant values in the settlement society, but this is not widespread, and is more the case for those with family ethnonational heritages in some origin countries (e.g. Pakistan) than others. Again, this finding points to the need for researchers to look inside ‘Muslim’ as a category of analysis, and to look beyond religiosity as a cultural explanation for how Muslim people see they fit in, or not, with the dominant political and social values of their settlement societies.

A second finding concerns identification with country of settlement. This is the only factor that is statistically relevant for all four ethnonational groups -stronger for ex-Yugoslavian and Turkish, compared to Moroccan and Pakistani- and is acculturative, it reduces perceived differences to national majorities. In other words, the more Muslims self-identify as British, Dutch, French or German – that is, the collective national identities of their respective settlement countries- the less they see ‘value conflict’ issues as a boundary between themselves and non-Muslim people.

At least as far as Muslims’ own perceptions are concerned these findings offer a different narrative than the contentious European debates over Muslim minorities that emphasise Muslims’ assumed religiosity and the supposed incommensurability of Islam as a religion with ‘liberal democratic values’ as a barrier to integration. Instead, it seems that more efforts to make Muslims feel part of the national political community and identify with the country would produce acculturative dividends. For example, policies that stimulate Muslim self-identification with the settlement nation and seek to legitimate this inclusion within the dominant understandings of majorities too. According to our findings, religious attitudes and practicing Islam do not present a significant obstacle. The dominant public narrative on Muslims and ‘value conflicts’ greatly exaggerates this factor. Indeed, the political intent behind mobilising claims over ‘liberal democratic values’ is often precisely to stimulate widespread negative public attitudes over the real or assumed cultural and religious differences of immigrant minorities who are Muslims. When focussed on specific groups of immigrant origin this can become a proxy for quasi racist sentiments.

Of course, our analyses of Muslims’ perceived differences to the majority over ‘liberal democratic values’ tell us nothing about the other side of the equation in this two-way acculturation process – that is, factors that can lead majorities to reduce their perceptions of Muslims’ differences over ‘liberal democratic values’. It could be that majority perceptions continue to construct high boundaries against Muslims by focussing on assumed or real ‘liberal democratic values’ differences independently from Muslims’ own perceptions, attitudes and behaviours. In such a scenario, the degree to which Muslims self-identify with the settlement nation, or not, will have no impact in diminishing ‘value conflicts’, because they are driven by majority public attitudes. Again, this points to the strong impact of majority public attitudes in constructing ‘liberal democratic values’ as a barrier for Muslims. This factor is also separate from liberal state’ efforts to extend group rights for Muslims, leading to a situation where as Joppke and Torpey (2013, 141–142) point out, ‘Islam, may still be more ‘barrier’ than ‘bridge’ to including immigrants in Europe but only as a matter of mentalities, not of institutions.’

Concluding discussion – the need for a new narrative and de-Muslimification

This study was designed in the first instance for purposes of critique: to demonstrate important limitations in the commonly applied academic thinking on this topic – what we call the Muslimification of Muslims. We are unaware of any other attempts to turn the question around and make the analytic usefulness of ‘Muslim’ and ‘religiosity’ objects rather than the tools for analysis. So, does it really make sense for research to lump all these people together as a single (religious) group ‘Muslims’?

Our empirical findings are clear-cut. First, not all ‘Muslim’ groups are the same. Family ethnonational origin -as ‘Pakistani’, ‘Turk’ etc..– matters too: it has significant analytic traction for how people who are Muslim see boundaries to national majorities over ‘liberal democratic values’. Second, it is not all about religiosity and Islam. In fact, we find religiosity has a very modest impact in shaping Muslims’ perceptions of their differences to majorities over ‘liberal democratic values’. Identifying with the settlement country matters much more and pushes in the opposite acculturative direction. Indeed, the modest degree to which religiosity increases Muslims’ perceived social distance to the majority is reduced when (non-religious) cultural factors are included. Finally, when we split the Muslims by family ethnonational original group, religiosity was statistically significant only for Pakistanis.

Thinking back to the heterogeneity and superdiversity of participants at the being Muslims in Britain meeting in the opening anecdote, we see a different social reality to that depicted by a rigid binary division between Muslim and non-Muslim (Christian) nationals that is the dominant master narrative in social science, and pretty much the only narrative in political and public discourses.¹² In the social world there is an increasing diversification and heterogeneity of ethnic and religious identities and social belongings among people who are Muslims (and also non-Muslims). This is carried by processes of demographic and intergenerational change, not least increasing mixing across ethnic and religious boundaries, and the onset of ageing societies where the predominantly white ‘baby boomer’ generation is being replaced (Alba and Foner 2016). The analytic logic of survey-based research on Muslims seems outdated and has not really adapted to the factual superdiversity of today’s European populations. We should also note that today’s mainstream or majority is also not a homogenous mass against which minorities can simply be measured against. In some cases, this restricted viewpoint can be because of data limitations that are seldom acknowledged. Most general population surveys do not have minority samples that are large enough to differentiate, so it is not possible to say anything about differences between groups who are Muslims. Many surveys do not include questions that allow to study the range of cultural variables included in this study. Ultimately, however, the persistence of this dominant one-eyed approach is largely the result of a path-dependent thinking.

Generally, there needs to be more sociological imagination in defining academic research questions about Muslim minorities. More reflexivity is required. There needs to be more intersectional thinking in the way that social scientific studies define research questions, apply perspectives, and design surveys and research projects on this topic. This would allow research to start acknowledging the diversity and hybridity of group boundaries in today’s societies, and in some cases even try to capture the social processes that carry them forward. It requires thinking out of the box and not simply following the dominant masterframe narrative that resonates in public debates focussing on controversies over Islam. Religiosity has too often been an easy answer plucked from political debates, often by those who do not look for, or consider using other cultural variables commonly used in acculturation studies. Scholars need to acknowledge the highly politicised context of using the ‘Muslim’ category in academic research. They should not shy from using ‘Muslim’, which remains a valid and salient category of social practice. However, there needs to be greater reflexivity and awareness when using ‘Muslim’ as an analytic category, and this rationale needs detailing explicitly when presenting the research. In short, it is high time for a de-Muslimification of academic thinking.

Reaching definitive answers on the substantive issues raised by our findings would require further research and detailed fine-grained analyses. However, we still think that the clear-cut direction is a sufficient basis to offer some speculative thoughts on policy implications.

Generally, our findings imply new directions for policy thinking over Muslim minority ‘integration’ too. Policy thinking could be de-Muslimified and tailored more towards the specific (nonreligious) social needs of different ethnonational minorities, who seem to have distinctive trajectories of acculturation. Religiosity and Islam matters only to a very limited degree in the construction of symbolic boundaries over ‘liberal democratic values’, at least from the Muslim side. Practicing Islam is not an unsurmountable barrier to acculturation processes, but needs understanding within context, as a specific, sometimes problematic, but ultimately manageable issue of religious accommodation within liberal democracies. More than religion as a barrier, our substantive findings point to the importance of self-identifying with the national identity and proficiency in the language of the settlement country as important acculturative factors that influence Muslim minorities’ perceptions of their socio-cultural distance to the majority. This suggests that the dominant public master narrative over Muslims in liberal democratic societies is wrong to depict incommensurable value cleavages between Muslims and non-Muslim national majorities. Perhaps if political elites and state policies focussed more on supporting general (non-religious) acculturation processes for Muslim minorities and legitimated such ideas so that they resonated in the minds of majorities too, instead of targeting Muslims and Islam as a ‘problem’, then more progress could be made in bringing people together.

Such policies would not necessarily equate with an extension of religious group rights to Muslims, as commonly assumed in the literature (see Joppke 2015). For example, they could be policy initiatives explicitly including Muslim minorities within understandings of Britishness, Dutchness or Germanness, etc.., and supporting language skills, etc … This could look something like the initiatives advocated by Putnam (2007) to enhance bridging capital across ethnic and religious group boundaries. Of course, such piece-meal incremental, evolutionary ideas of acculturation processes lack the drama of cleavages, culture clashes and value conflicts. They would also need a new supportive public narrative advanced by political elites to legitimate their existence and potential success. Here symbolic inclusion within the national community matters too. One is reminded of Putnam’s anecdote about when he was invited to the UK Royal Household to advise on ‘integration’ and community cohesion. According to his version, he suggested that the Queen (and Head of State) should encourage one of her grandchildren to marry a Muslim, thereby legitimating the presence of Muslims in the core national symbolic institution the monarchy – after which he was never invited back!¹³

The serious point is that any shift in policy thinking requires a sharp about turn in the established master narrative on Muslims of immigrant origin, of which there currently appears little prospect. Anti-Muslim and anti-Islamic sentiments directed towards minorities of immigrant origin have played a significant part in the rise of right-wing populists across Europe. Even in Sweden, often touted as Europe’s most open and tolerant country, today there are increasing incidents of (white) individuals burning the Qur’an in public to assert their liberal right to freedom of expression.¹⁴ Of course, far from being neutral, such claims make ‘liberalism’ an ethno-religious barrier – they are intended to provoke, incite, harm and exclude Swedish people of immigrant origin who are Muslim. It shows whose foot the boot is firmly on in Europe’s controversies over Islam, Muslims and ‘liberal democratic values’. The paradox here is that far from being a measure for Muslim acculturation, ‘liberal democratic values’ and ‘freedom of speech’ is a construction that is used as a stick to beat Muslims of immigrant origin. Used this way, claims made in the name of ‘freedom of speech’ legitimate forms of cultural racism that would otherwise be unacceptable.

Our own findings underline that majority perceptions are an important factor that reinforces high barriers to acculturation for Muslim minorities. Dominant public discourses about Muslims and Islam seem to have left their mark on how populations view these issues. Any possibility for acculturation by Muslims is highly dependent on significant perception shifts by people from national majority populations too. An important point here is that Muslim acculturation and ‘integration’ to their settlement societies is a distinct social process to cultural racism and discrimination by majority populations. This means that majority opposition to Muslims may persist independently from the degree to which Muslims acculturate or ‘integrate’, not least when such sentiments are stimulated and legitimated by resonant public discourses where anti-multiculturalists -including so-called public intellectuals- find it easy to mobilise their claims. There is also a methodological point: studies of national populations’ views tend to lump them together in a way that reifies them into a homogeneous group. This hides the important dividing lines and divisions within majorities. It matters because the existence of white supremacists or anti-Muslim racists can become lost in the category of ‘national majority’, while their existence in the real world is a challenge to the values of liberal democracy and the wellbeing of minorities of immigrant origin.

Acknowledgements

This paper has gone through several iterations over several years. Jolanda van der Noll deserves very special thanks and acknowledgement for being involved earlier in the project and undertaking the technical analyses. The later versions of the paper benefitted from airing to specialist audiences at the WZB, Berlin, in September 2022, and at MIM in Malmö, in March 2023. Sarah Scuzzarello offered detailed critical comments on a late version. I originally designed this paper as a lecture for the Migration Masters students at Sussex to demonstrate the limits of research on Muslims and was encouraged to write it up by their enthusiasm and support for the perspective.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The EurIslam project and data were supported by the European Commission Seventh Framework Programme [grant number SSH-2007-3.3.1 215863].

Notes

1 For example, there is no discussion of what or who is a Muslim in the actually detailed and valuable survey research by Lewis and Kashyap (2013) on the British case. Muslims are simply lumped together as Muslims and compared to ‘Britons of other religions’.

2 In many cases, the failure to test for non-religious cultural variables is due to data inadequacies, that is, that such questions simply do not exist in the survey. However, this is hardly ever discussed or acknowledged as a research limitation. Indeed the strong salience of religiosity as an explanatory variable for Muslims to the exclusion of others is demonstrated by a paper that like this one uses the EurIslam data-set -which was designed explicitly to include religious and non-religious cultural variables – but which still plucks religiosity out of the air to make general causal statements on religiosity as the basis for Muslims’ values on sexuality and democracy: ‘Across the countries, and thus independently of the specific historical and political context, Muslims were found to be more religious and to have a lower sense of national belonging and these accounted for their lower support of democratic government and of liberal sexual mores’ (Eskelinen and Verkuyten 2020, 2358).

3 In this UK categorisation ‘Asian’ refers to people with origins in the Indian Subcontinent.

4 Taylor is at the opposite end of the spectrum from Modood on this among pro-multiculturalists, he excludes Islam from his politics of recognition, because (1994, 62), ‘[f]or mainstream Islam, there is no question of separating politics and religion’ and ‘western liberalism is … (an) organic outgrowth of Christianity.’ Interestingly, Taylor’s critique exposes that far from being neutral as its ideology claims, liberalism has an ethnocentric cultural basis derived historically from Christian values. Conversely, anti-multiculturalists present liberal democracy as neutral but attribute it with core underpinning values that have a ‘thick’ fixed cultural imprint – as a specific ‘Christian’ ‘ethnic’ national identity derived from the assumed cultural of the national majority population.

5 Their survey is based on a sample size of 2007 ‘Dutch’ adults (16 and older), selected randomly by telephone numbers, with a response rate of 37%, at one time-point, collected Jan-April 1998 (Sniderman and Hagendoorn 2007, 139).

6 See for example, Koenig (2007), Diehl, Koenig, and Ruckdeschel (2009), Röder (2014), Carol and Milewski (2017).

7 This paper was presented at a forum for scholars in the US and Europe working on culture and immigration/integration where Richard Alba was the keynote. 8th Annual Conference on Migration and Diversity Does Culture Matter For Integration? Empirical Patterns and Regulation of Difference, WZB, Berlin, 14–16 September 2022.

8 For a theoretical and conceptual outline of the aims and design of the EurIslam project, see Statham and Tillie (2016). For further information on EurIslam sampling frame, codebook and data, as well as response rates and representativeness checks, information is collated and accessible on DANS Data Station Social Sciences and Humanities. For the EurIsam data-set see, https://ssh.datastations.nl/dataset.xhtml?persistentId=doi:10.17026/dans-xx7-5x27 (accessed 14/9/23).

9 Among others, see specific contributions using the survey data by EurIslam team members to Statham and Tillie eds (2018).

10 Similarly, in a study – using EurIslam data – that compares majority and Muslim perceptions of religious group rights provisions for Muslims and Christians, Statham (2016) finds that it is important shifts in majority population perceptions that drives the relationship and produces a significant ‘gap’ between Muslims and Majorities over such issues, when the question moves from state provision for Christians to Muslims. While Majorities become much more opposed to religious group rights when the provision is for Muslims, this is not the case for Muslims, who hold similar views over rights provisions for Christians and their own religion. Again, this research emphasises the importance of majority perceptions – stimulated by resonant public debates- as a contextual barrier towards acculturation processes over religious group rights.

11 For the multivariate regression analyses the answers to the three questions are averaged into one scale of perceived differences (α = 62 in the total sample), in which a higher value reflects more perceived differences.

12 It should be made clear that this overemphasis on constructed binary divisions between Muslims and non-Muslims is not only by academic opponents of Islam and multiculturalism, but also by scholars who see themselves as socially progressive and multiculturalists (see e.g. Carens 2000 Ch. 6; Lewis and Kashyap 2013). They also end up using interpretative frameworks that can only speak to and thereby emphasise and reproduce the same stereotypes about binary divisions.

13 Author’s recollection of Putnam’s speech at Centre for the Study of Ethnicity and Citizenship, Bristol.

14 ‘It’s a racism crisis’: call for action on Qur’an burnings in Sweden. The Guardian, Miranda Bryant 3/08/23. https://www.theguardian.com/world/2023/aug/03/muslims-react-quran-burnings-stockholm-sweden(accessed 22/08/23).

Appendix. Additional Notes on Sample and Measures (not already in the main text)

Sample

The analyses are from samples collected in the Netherlands, Britain, France and Germany by the EurIslam project. Respondents from the four family ethnonational background groups identify as ‘Muslim’. The samples of the majority population include people who identify as Christian or not religious. This results in a total sample of n = 5087 respondents across all countries and groups. Age of the respondents ranges from 17 to 92 years old (M = 41, SD = 15.5) and 50 percent are female. Around one-fifth of the respondents have completed primary education or less (19 percent), a half (48 percent) secondary education or vocational training, and a quarter (27 percent) tertiary education. For the sample of people who are Muslims from one of the four family ethno-national groups (n = 3609), age ranges from 17 to 86 years old (M = 38, SD = 13.6), and half (49 percent) are female. A quarter (25 percent) completed primary education or less, a half (47 percent) secondary education or vocational training, and a quarter (24 percent) tertiary education. The number of people who are Muslims who met our criterion for ‘religiosity’ – pray at least once a day and identify strongly or very strongly as Muslim – is n = 1396, that is, 38.7 percent of Muslim sample.

Demographic, human capital and status variables

We constructed dummy variables for country of settlement and family country of origin. Gender (male = 0 or female = 1) was noted by the interviewer. Respondents were asked to indicate their year of birth, based on which we calculated their age at the time of data collection. The highest achieved level of education was coded according to the International Standard Classification of Education (ISCED) into primary education or less (1; reference category), secondary education or vocational training (2), tertiary education (3). To assess generation, respondents were asked when they arrived in their settlement country. Second-generation immigrants are those who were born in their country of settlement. People who born abroad but who arrived in the country of settlement when they were at most 16 years old were categorised as the 1.5 generation, and those who arrived in the country of settlement after the age of 16 years old were counted towards the first generation immigrants. To assess citizenship, respondents indicated all countries in which they held citizenship. We used this information to create two dummy variables indicating whether respondents had citizenship of their country of origin and whether they had citizenship of their country of settlement (1 = yes, 0 = no).

Cultural (non-religious) variables

To assess self-identification with the family country of origin and country of settlement respondents were asked to what extent they see themselves as a person of these respective countries. The answer categories offered ranged from 1 = very strongly to 5 = not at all, we reverse coded these items so that a higher score reflects a stronger identification with these countries. To measure language proficiency respondents were asked how often they have problems with a conversation in the national language of their country of settlement. The answer categories ranged from 1 = never to 5 = always. In addition, for language of media consumption respondents were asked in which language they mostly used media (such as reading newspapers or watching television), ranging from 1 = always in the language of the country of settlement to 5 = always in their parents’ mother tongue. We reverse coded both items so that a higher score reflects a better proficiency and higher media use in the language of the country of settlement. Contact with members of the majority population (‘bridging’ social capital) was measured with three questions. First, focusing on ‘weak ties’, respondents were asked to indicate how many of their acquaintances from their neighbourhood are members of the majority population. Focusing on ‘strong ties’, respondents were furthermore asked to indicate how many of their good friends are members of the majority population. For both questions, the answer categories offered ranged from 1 = nearly all; 5 = almost none up to one, and were recoded in order for a higher score reflecting more contact with members of the majority population. Finally, for intermarriage in family respondents were asked whether they had family members who are married or cohabitating with members of the majority population. We created a dummy variable distinguishing between one or several (1) and none (0).