Islamophobia Without Muslims? The “Contact Hypothesis” as an Explanation for Anti-Muslim Attitudes – Eastern European Societies in a Comparative Perspective

A specter is haunting Europe – the specter of Islamophobia. At this point, the historically informed reader might question the newsworthiness of our initial statement. Unease, reservation, and even fear and hatred against Islam and Muslims have a long tradition in Europe (Benz 2017). Since Edward Said’s (1978) seminal study on ‘Orientalism’, it is a commonplace to acknowledge that the West has associated Islam with negative images and stereotypes for hundreds of years. The essential novelty of Europe’s old specter of Islamophobia is that Eastern European governments joined the anti-Islam chorus during the so-called ‘refugee crisis’ of 2015. Czech president Miloš Zeman stated that Muslim mass integration into European societies is practically impossible for ‘cultural reasons related to Islam’. He considered Islam as a ‘religion of death’ and argued that the term ‘moderate Muslim’ is as contradictory as referring to ‘moderate Nazis’ (Trait 2016). Jaroslaw Kaczyński, the strong men of the Polish law and justice party, warned that Muslims pose a danger for Europe’s ‘Christian identity’. In his view, Muslims consider ‘churches as toilets’ and their ultimate goals are the ‘establishment of Sharia law’. Furthermore, Kaczyński regarded Muslims as a ‘menace for public health’ as immigrants come along with diseases and parasites (Cienski 2015). Victor Orbán claimed that Hungary is the last ‘bastion against the Islamization of Europe’. In his view, the influx of Muslim immigrants equates a ‘danger for European’s employment and living conditions’ (Boffey 2018; Kokot 2015). To put it bluntly, political leaders in Eastern Europe portray themselves as protectors of their nation’s ‘Christian identity’ that is allegedly endangered by an ‘invasion of Muslims’ (Schenkkan 2016).

But what is the position of the citizens? The perception of Islam and Muslims as an aggressive menace to Western societies gained importance in the aftermath of the Cold War (Huntington 1997) and since the terrorist attacks of 9/11 (Ruf 2014). International surveys show that anti-Muslim resentments are widespread among Western public (Helbling 2012; Pickel and Yendell 2016; PEW Research 2018; Strabac and Listhaug 2008; Zick, Küppers, and Hövermann 2011). Given a climate of fear due to several terrorist attacks committed by the ‘Islamic State’ (e.g., Paris, Marseille, Barcelona, and Berlin), it seems unlikely that the career of Islam as the scapegoat of Europe has reached its final climax yet. However, research on Islamophobia – which deals with public opinion and relates to studies on right-wing extremist and populist parties – has focused predominantly on Western European societies (Allen 2010; Ciftci 2012; Helbling 2012; Kaya 2015; Marfouk 2016; Mudde 2007; Pickel and Yendell 2016; Strabac and Listhaug 2008). Given the hostile reactions of Eastern European politicians during the so-called refugee crisis, one might argue that there is still a considerable research gap about the causes of citizen’s susceptibility for Islamophobia in Eastern Europe.

A few studies reveal a puzzling finding. Nowadays, Islamophobia seems to be more widespread in Eastern Europe than in Western Europe, even though Muslim communities are virtually nonexistent in most Eastern European societies (Strabac and Listhaug 2008; Pickel and Öztürk 2018). This provokes two questions: Is it empirically true that Islamophobia is more widespread in Eastern Europe than in Western Europe? And if yes, how can we explain this puzzling pattern? We hypothesize that Islamophobia is more prevalent in Eastern Europe than in Western Europe and we suppose that this empirical pattern is caused by a mixture of more pronounced ethnocentric sentiments of nationalism and less contacts with Muslims.

This assumption leads us to the contact hypothesis and alternative social– psychological explanations (e.g., Social Identity Theory, Integrated Threat Theory, and Deprivation Theory). We shed light on the differences in the magnitude of Islamophobic attitudes across Europe using these theories, and we present evidence on both the individual level and the societal level. On the individual level, we expect general effects of the social–psychological determinants of anti-Muslim prejudice. In this regard, individuals in Eastern and Western Europe do not differ dramatically. We nevertheless argue that the prime reasons for more elevated levels of Islamophobia in Eastern Europe are rooted on the societal level: The assumed negative individual-level effect of contacts on anti-Muslim prejudice lacks an amplifier in Eastern Europe as Muslim communities are virtually inexistent in these parts of Europe. To put it bluntly, the lack of direct contacts in Eastern Europe fosters citizen’s antipathy toward Muslims.

It is also conceivable that Islamophobic attitudes lead to less contact with Muslims. However, research on the contact hypothesis suggests the causal direction we deploy in our article: Contact is likely to reduce prejudice against out-groups (Pettigrew and Tropp 2006).

What exactly do we mean by Islamophobia? First, it is to state that Islamophobia is a highly contested term (Allen 2010). A great deal of attention in the research field is dedicated to the subtle differences (and flawed boundaries) between Islamophobia, critics of Islam, and hostility toward Islam and its adherers (Bleich 2011; Pfahl-Traughber 2014). Some authors do not even stop short of advocating for the avoidance of the term’s usage in the scientific realm (Kahlweiß and Salzborn 2014). Bleich (2011) argued that social scientists would be ill-advised to do so, as the term labels a social reality, that is to say, “that Islam and Muslims have emerged as objects of aversion, fear, and hostility in contemporary liberal democracies” (Bleich 2011, 1584). In this vein, Bleich (2011) proposed a widely used definition of Islamophobia, which is “indiscriminate negative attitudes or emotions directed at Islam or Muslims” (Bleich 2011, 1585). We adopt this definition, as it is widespread among the scientific communities, but we limit our scope of Islamophobia to indiscriminate negative attitudes directed at Muslims. There are good empirical and normative reasons to do so.

Uenal (2016) presented empirical evidence that attitudes toward Islam and Muslims constitute two different dimensions. From a normative perspective, a too broad understanding of Islamophobia – which encompasses critique of Islam – is problematic as it puts legitimate criticism of religion (e.g., the subordination of women) under the suspicion of prejudice (Alexander and Welzel 2011). As Kaya (2015, 451) pointed out, this narrow understanding of Islamophobia is largely in line with Allport’s (1971) characterization of prejudice – which is “an antipathy based upon a faulty and inflexible generalization”, and “thinking ill of others without sufficient warrant” (Allport 1971, 20).

To make a long story short, Islamophobia describes negative attitudes toward individuals based on their perceived religious background. As nonpracticing Muslims face discrimination because of their ethnocultural characteristics, it is an alleged group identity that drives anti-Muslim prejudice (Cinnirella 2012). On these grounds, there are good reasons to search for the causes of Islamophobia in the light of social–psychological theories of prejudice and stereotyping (Adorno et al. 1950; Allport 1971).

The empirical part of our article focuses on one specific manifestation of Islamophobia: The support for an immigration ban for Muslims. Of course, this policy preference encompasses different phenomena such as ethnocentrism, discomfort with religious people, and anti-immigrant positions in general, but in our opinion, it represents anti-Muslim prejudice in the first instance. The seventh round of the European Social Survey (ESS 2014) incorporates a question that allows us to examine whether Europeans support a Muslim ban by asking whether Muslims from other countries should be allowed to come and live in their countries?. It is a unique opportunity to scrutinize to what extent right-wing populist and extremist parties call for a complete Muslim ban is supported by European citizens. Even if hostility toward Muslim immigration is associated with opposition to immigration in general, it is reasonable to assume that the survey question captures Islamophobic attitudes. For many Europeans, it is not immigrants in general, but Muslims who are perceived as the ‘main problem’. In contemporary Europe, political debates about Muslim immigration come along with populist rhetoric about the ‘Islamization of Europe’. On a regular basis, populist politicians blame governments for their foot-dragging vis-à-vis an alleged ‘Muslim invasion of Europe’. This rhetoric is an attempt to attract prejudiced voters and to portray themselves as the only credible guardian of ‘the people’ (Bremmer 2018).

However, there is a certain mismatch between populist’s heavy rotation of anti-Muslim statements and the factual size of Muslim minorities living in Europe. The scatterplot in Figure 1 shows the association between the factual presence of Muslim minorities in European societies and the average support for a Muslim ban. In line with the previous research on Islamophobia beyond the West, the data highlight that Islamophobia is more widespread in Eastern than in Western Europe (Strabac and Listhaug 2008; Pickel and Öztürk 2018). So far, Eastern European countries have been comparatively unpopular choices for migrants to settle and particularly Muslim communities are virtually absent in most Eastern European societies. Nevertheless, Eastern European citizens are on average more prone to oppose Muslim immigration. The smaller the factual presence of Muslim minorities in European societies, the higher the average support for a Muslim ban. The scatterplot reveals eye-catching evidence for the phenomenon of ‘Islamophobia without Muslims’. At this point, it is important to acknowledge that the ESS (2014) has not been carried out in all member states of the European Union. Furthermore, it does not allow representative statements about Eastern Europe as the respective round of the ESS (2014) only contains six Eastern European societies.

Figure 1

But with a glance on the available data, the anti-Muslim rhetoric of Miloš Zeman, Jaroslaw Kaczyński, and Victor Orbán seems to be in line with the public opinion in their countries. Over one in two respondents in the Czech Republic (56%) and Hungary (50%) and at least one in three in Estonia (41%), Lithuania (37%), and Poland (32%) reject Muslim immigration. Slovenia is rather an outlier among East European societies. Less than one in four of Slovenian respondents (20%) is in favor of a Muslim ban. At this point, it is important to emphasize that the rejection of Muslim immigration has no Eastern European peculiarity. There is a remarkable support for a Muslim ban in countries such as Portugal (31%), Ireland (24%), Austria (20%), and Belgium (18%). Given the fact that citizens in the Netherlands (13%), France (12%), Germany (8%), and Sweden (4%) are less inclined to reject Muslim immigration in general, there is nevertheless a clear pattern that the receptiveness for anti-Muslim prejudice is much higher among citizens in Eastern Europe. How can we explain this finding?

Advocates of the contact hypothesis would argue that the virtual absence of Muslim communities is a relevant explanation for anti-Muslim prejudice in its own right. The contact hypothesis states that an individual’s contact with out-groups is conducive to teardown existing prejudice (Allport 1971; Pettigrew 1998).

Pettigrew and Tropp’s (2006) meta-analysis presented compelling evidence that contact with out-groups is likely to refute existing prejudice and stereotypes. In doing so, the authors invalidate three frequent points of criticism. The first criticism was put forward by Allport (1971). He argued that there are certain preconditions that determine whether contacts reduce prejudice. However, contacts with out-groups have an overall tendency to reduce prejudice (Harmon-Jones and Allen 2001; Lee 2001). The second criticism relates to the direction of causality: Is it really contact that reduces prejudice or do prejudiced individuals simply avoid contacts with out-groups? There is evidence for both causal directions. Prejudiced people indeed avoid contacts with out-groups (Herek and Capitanio 1996), but across the board the causal path from contacts to reduced prejudice appears to be much stronger (Pettigrew 1997). The third criticism states that the reduction in prejudice is restricted to the individuals who are directly involved in the contact situation. However, Pettigrew and Tropp’s (2006) findings indicated that contact’s tendency to reduce prejudice comes along with less reservation toward the whole out-groups (Rhodes, Halberstadt, and Brajkovich 2001).

In line with the contact hypothesis – and bearing in mind its shortcomings – we assume that individuals who stay in contact with out-groups, make friends with people from other ethnicities, and perceive these contacts as convenient are less likely to feel prejudice toward Muslims (Freitag and Rapp 2013). However, we argue that it is of tremendous importance to distinguish between the individual and the societal level. On the individual level, we expect similar effects among individuals in Eastern and Western Europe.

This microlevel effect, however, lacks an amplifier in Eastern European societies. To put it bluntly, the lower extent of contact possibilities is the prime reason why anti-Muslim prejudice is more prevalent in Eastern European societies. We argue that the sheer absence of Muslims can even intensify anti-Muslim prejudice. Most European citizens have a rather restricted knowledge about Islam and seldomly meet Muslims as ‘normal citizens’ (Pickel 2018). This pattern particularly applies to Eastern European citizens as they tend to have ‘para-social contacts’ with Muslims (Horton and Wohl 1956). In the absence of direct contact possibilities, it seems likely that media consumption (e.g., TV and Internet) comes along with an illusion of direct contacts. ‘Para-social contacts’ give rise to anti-Muslim prejudice for two reasons. First, mass media in general leans toward a negative news bias. Second, news coverage of terrorist attacks committed by Islamists shapes the prevailing image of all Muslims (Ahmed and Matthes 2016; European Islamophobia Report 2018; Frindte 2013; Pickel and Yendell 2016, Saeed 2007). In doing so, media facilitates prejudice against Muslim immigrants. This overall tendency is likely to be intensified by political communication strategies of right-wing populist parties that use different media platform to spread fear about Muslims (Wodak 2015). This leads us to our two societal-level hypotheses:

Even though we consider the contact hypothesis as a powerful explanation for the rejection of Muslim immigrants, it is of paramount importance to account for alternative explanations. One good reason to do so is the remarkable support of a Muslim ban in Western European societies (e.g., Austria, Belgium, the Netherlands, and the United Kingdom) with substantial Muslim communities. On a basic level, the support for a Muslim ban arises from a process of categorization. Citizens attribute negative characteristics toward a group of individuals based on their (perceived) religious background. As this process of categorization comes along with the construction of collective identities, we consider Social Identity Theory (Tajfel 1969; Tajfel 1982; Tajfel and Turner 1979) as a good starting point to identify factors that promote the rejection of Muslim immigrants. From the perspective of Social Identity Theory, prejudice is the result of an “actor’s identification of themselves and the others belonging to different social categories” (Tajfel and Turner 1979, 40). One precondition for anti-Muslim attitudes to unfold out of this categorization process is that individuals “have internalized their group membership as an aspect of their self-concept: they must be subjectively identified with the relevant ingroup” (Tajfel and Turner 1979, 41). At this point, nationalism comes into play, because it might be considered as the most powerful source of an individual’s collective identity (Gat 2012). We define national identity as an individual’s “subjective or internalized sense of belonging to the nation” (Huddy and Khatib 2007, 65). It derives from an individual’s “knowledge of his/her membership of a social (...) group together with the value and emotional significance attached to it” (Latcheva 2014, 3941). As nations represent ‘imagined communities’ (Anderson 1991; Brubaker 1996), an individual’s national identity manifests itself in positive attachments toward collective symbols, a common langue, history, and tradition. Nationalism is likely to be a source of anti-Muslim attitudes. Nationalism establishes differences between groups of people and fosters dividing lines between in-groups and out-groups (Gat 2012; Hjerm 1998; Latcheva 2014).

Beyond nationalism, religion can be a source of collective identities. Eastern European politicians regularly stress the ‘Christian heritage’ of their nations to justify their refusal to host Muslim refugees. As Muslim integration turns out to be complicated in Christian heritage societies (Adida, Laitin, and Valfort 2016), one might ask if and how religiosity is related to anti-Muslim attitudes. Across the board, the potential effects of religiosity on prejudice are considered as highly ambivalent (Allport 1971, Allport and Ross 1967). On the one hand, almost all religious teachings demand the devoutness to feel compassion and to supply care for those in need. Consequently, one might expect positive effects of religiosity toward the acceptance of Muslim immigrants. On the other hand, many authors argue that religious citizens direct their generousness only to their coreligionists (Adorno et al. 1950; Ben Nun Bloom, Arikan, and Courtemance 2015; Djupe and Calfano 2013; Gibson 2010; Norenzayan 2014). If true, religiosity drives antipathy toward Muslim immigrants. We argue that an individual’s religiosity comes along with anti-Muslim prejudice for two reasons. First, an individual’s sense of belonging to a religious group is a source of identification with the in-group beyond nationalism (Hunsberger and Jackson 2005; Jackson and Hunsberger 1999). Second, the very existence of Muslims and their sheer religious ‘otherness’ can induce a threat among the in-groups and may aggravate the ‘we-versus-them-divide’ (Helbing 2014; Pickel 2018; van der Noll 2010).

As the categorization process is kept alive by an allegation of negative out-group characteristics, the emergence of feelings of anxiety and threat becomes a realistic scenario (Pickel and Yendell 2016). This assumption is highly plausible as we deal with Islamophobia. As mentioned earlier, there is a long tradition of negative images of Muslims in European history (Benz 2017, Said 1978), and one might argue that terror attacks consolidate such feelings toward Muslims (Cinnirella 2012). The impact of fear occupies a central position in the Integrated Threat Theory (Gonzalez et al. 2008; Jonas and Fritsche 2013; McLaren 2003; Stephan and Stephan 1996; Quillian 1995). In a nutshell, the theory states that in-groups’ perceived realistic and symbolic threats are a powerful source of prejudice toward out-groups (Stephan and Stephan 1996). When Victor Orbán warned his compatriots that Muslims pose a danger for their employments and current living conditions and may turn out to be terrorists, he tried to produce (or to strengthen) realistic threats among the Hungarian public (Kokot 2015). A realistic threat perception is in place when the in-groups consider the very existence of the out-groups as a danger for their physical and material well-being (Stephan and Stephan 1996).

Much of the anti-Muslim rhetoric we described earlier aims to encourage symbolic treats among Eastern European citizens. Symbolic fear is a likely outcome if in-groups are concerned about “group differences in morals, values, standards, beliefs, and attitudes” (Stephan and Stephan 1996, 418). Miloš Zeman’s characterization of Islam as a ‘religion of death’ insinuated sharp differences between the violent and brutish culture of Muslims and the peaceful and civilized characteristics of the Czech population. On top of that, buzzword like the alleged ‘Islamization of Europe’ gives the impression that Muslim immigrants’ main goal is to change the nature of European’s native cultures. If Eastern European citizens are susceptible for this kind of rhetoric, rejective attitudes toward Muslim immigrants become a likely outcome.

In the long run, the combination of collective identities (e.g., nationalism and religion) and widespread threat perceptions (e.g., realistic and symbolic fears) is likely to trigger more rigid forms of “in-group favoritism and discrimination against the out-group” (Tajfel and Turner 1979, 39). In-group favoritism and hostility toward out-groups come close to Levinson’s (1949) description of ethnocentrism. For ethnocentric individuals, the distinction between in-groups (those groups with which the individual identifies himself) and out-groups (with which the individual does not have a sense of belonging and which is regarded as antithetical to the in-groups) is of paramount importance (Levinson 1949, 20). The perception of a sharp divide between ‘we’ and the ‘others’ is exaggerated to the extent that “outgroups are the objects of negative opinions and hostile attitudes; ingroups are the objects of positive opinions and uncritically supportive attitudes; and outgroups are regarded as properly subordinate to ingroups” (Levinson 1949, 20). We consider ethnocentrism to be a more valid explanation for Islamophobia than nationalism. National identification in terms of constitutional patriotism may fulfill positive functions for a democratic political community (and does not necessarily breed Islamophobia) (Habermas 1990, 147-156).

Ethnocentrism, however, is likely to come along with anti-Muslim attitudes as it “involves blind attachment to certain national cultural variables, uncritical conformity with the prevailing group ways, and rejection of other nations as outgroups” (Levinson 1949, 20; and see: Garner and Selod 2014; Meer and Modood 2009). Right-wing populist leaders gain from these ethnocentric sentiments in their electoral campaigns. In this vein, the scapegoating of Muslims comes along with severe criticism of liberal democracies’ political elites (Mudde and Kaltwasser 2017). The so-called refugee crisis turned out to be an opportunity window for populist leaders. It allows them to portray the political elites as unwilling to defend ‘the people’ against an ‘invasion of Muslims’. Evidence of that recipe’s payoff is provided by the electoral success of right-wing populist parties in the last years.

Beyond collective identities, fear, and ethnocentrism, authors refer to the relative deprivation theory (Runciman 1966; Stouffer et al. 1949) to explain prejudice toward out-groups. Deprivation theory expects a relationship between unfavorable socioeconomic conditions and pejorative attitudes toward out-groups. In this vein, individuals who lack action resources (e.g., lower strata, unemployed, and less educated persons) are assumed to be the most prejudiced citizens of the social strata (McCutcheon 2000). Having said, it is nevertheless important to distinguish between objective and relative deprivation. For the emergence of prejudice, it is relative deprivation that matters (Runciman 1966; Stouffer et al. 1949). Even if there is no objective deprivation in place, it is citizens’ perceived socioeconomic disadvantage that makes them more prone to disapprove out-groups. Furthermore, it is a customary practice to differentiate between individuals (e.g., the perception of personal disadvantage) and collective deprivation (e.g., the perception that the whole in-groups are disadvantaged vis-à-vis out-groups). Across the board, the authors argue that collective deprivation breeds prejudice (Vanneman and Pettigrew 1972; Walker and Mann 1987; Walker and Smith 2001). The rhetoric of right-wing populist parties and politicians is likely to intensify these effects. Again and again, populists such as Victor Orbán portrayed Muslims as a burden for their host countries’ welfare systems (Kokot 2015). In line with these theoretical arguments, we assume that collective deprivation matters for the support of a Muslim ban.

At this point, we want to recall that we assume invariant causal effects of the outlined social–psychological drivers of Islamophobia among citizens in Europe. To put it bluntly, the causes of citizens’ susceptibility for Islamophobic attitudes are rather uniform across Europe. It is societal-level factors that matter. The absence of Muslim communities in Eastern Europe leads to less contacts, less friendships with immigrants, and more skeptical perceptions of intergroup contacts. These factors strengthen a social climate in which anti-Muslim prejudice prevails. We therefore expect clear societal-level effects that rest upon comparable individual-level underpinnings.

Theories need to be tested against empirical evidence. For this purpose, we exploit the European Social Survey (2014). Since 2002, the European Social Survey Program conducts public opinion polls every 2 years. Every round of the ESS includes repeated questions and up to three topic-orientated modules. The seventh round of the European Social Survey (2014) sheds light on questions that relate to immigration and asylum issues.

Therefore, it offers a unique opportunity to study the social–psychological determinants of rejective attitudes toward Muslim immigrants in a comparative perspective. The survey contains acknowledged indicators that enable us to empirically measure constructs that are integral parts of theoretical explanations for the formation of prejudiced attitudes.

See https://www.europeansocialsurvey.org/.html for additional information.

It is important to note that the European Social Survey (2014) was conducted on the verge of the so-called refugee crisis. We do not consider the timing of the survey as a serious drawback for our analysis. The data allow us to show that there is a certain continuity of Islamophobic sentiments and its enabling social–psychological personality traits – which also indicates that

Islamophobia is not entirely a by-product of the polarized debates that came along with the so-called refugee crisis (Helbing 2012; Pollack et al. 2014). As there have been several terrorist attacks since 2015, the figures we present might nevertheless turn out as rather conservative estimations.

The respective survey contains data for 18-member states of the European Union, Switzerland, Norway, and Israel. Questions that relate to the management of the so-called refugee crisis caused serious political conflicts within the European Union. Therefore, we decided to restrict our case selection to its member states. Even though the ESS (2014) offers a unique opportunity to shed light on the extent and causes of derogatory attitudes toward Muslims beyond Western Europe, we must raise a caveat at this point: There is a certain lack of Eastern European countries in our sample. When we talk about Eastern Europe, in the following, we can only refer to six Eastern European cases (Czech Republic, Estonia, Hungary, Lithuania, Slovenia, and Poland). Of course, this a clear selection bias for comparative research. As our comparative study relies upon clear theoretical and conceptual assumptions, we nevertheless argue that an analysis of the ESS (2014) is suitable to give answers to our research questions. This goes without saying that more detailed research is necessary to scrutinize whether our empirical results travel to other Eastern European societies.

The operationalization of our theoretical constructs is listed in Table 1. We decided to use a coding scheme that allows us to bring every item into the same scale range. We standardized every item into a scale from 0 to 1.0. Due to this procedure, different response schemes have been transformed to a normalized range with minimum 0, maximum 1.0, and fractions of 0.10 for intermediate positions. Among other things, this procedure simplifies the interpretation of coefficients in regression analyses (Welzel 2013). As a robustness check, we control for respondent’s educational level, gender, and age. A priori – and in line with previous research – we expect that the less educated, the elderly, and female respondents are more likely to support a Muslim ban. It is Islam’s alleged misogynistic nature that makes women more likely to oppose Muslim’s immigration (Marfouk 2016). Due to this theoretical consideration, we treat males as the reference category in our statistical models.

Before turning to the statistical methods, we used to test our assumption; it is important to keep in mind that we will present evidence for the individual and the societal level. On both level, we scrutinize the explanatory power of the contact hypothesis and control for the alternative social–psychological determinants we discussed in the theoretical section of our article. On the individual level, we run several ordinary least-square regressions to test our hypotheses (Pickel and Pickel 2017). We are convinced that only separated individual-level regression models for each country are detailed enough to ensure that we have comparable causal patterns of Islamophobia within countries that have been surveyed by the ESS (2014).

In this vein, our empirical approach resembles the idea of a Most Different System Design: We compare cases that vary considerably about their societal-level features (e.g., Muslim minorities’ population size, experience with democracy, and socioeconomic development), but assume that the support for a Muslim ban is caused by similar social–psychological determinant patterns (Pickel 2016).

For the societal level, we used population means of the scales that we constructed for our individual-level regressions (Table 1). The European Social Survey collects its data by means of a random selection scheme. An aggregation of individual-level data therefore allows us to describe the average frequency of contacts and prevalence of anti-Muslim attitudes in European societies (Pickel 2009). As the number of cases is rather restricted, we rely upon visualizations (e.g., scatterplots) and report results from bivariate regressions. To shed light on the explanatory power of contacts on the societal level, we show how strongly each of the alternative explanations (which is the factual share of Muslims and a social climate of nationalism, religiosity, threat perceptions, ethnocentrism, and collective deprivation) affects the support for a Muslim ban, when we control for the average frequency of contacts with migrants.

What causes anti-Muslim prejudice? Does Islamophobia rest upon similar social–psychological patterns in European societies? In view of our regression results (Table 2), there is empirical evidence for the contact hypothesis. Except for Hungary, there is a clear pattern: Respondents who stay in contact with migrants, make friends with people from other ethnicities, and perceive these contacts as convenient are on average less inclined to support a Muslim ban (the effects range from b = -.07 in Estonia to b= -.25 in Belgium).

There is much less empirical support for our assumption that the identification with an in-group – via religion or nationalism – is a fundamental precondition for prejudice against Muslims. Both factors show varying effects in different national contexts. In Denmark (b =.06), Germany (b =.04), and Spain (b =.07), there are citizens with nationalistic sentiments who are more likely to prefer an immigration ban for Muslims. In contrast to our hypothesis, we observe that Irish citizens with a strong sense of national pride are less likely

to support a Muslim ban (b = -.09). In Eastern Europe, nationalism turns out to be a nonsignificant parameter of Islamophobia. Beyond nationalism, our results confirm Allport’s and Ross’ (1967) old wisdom that religion has ambivalent repercussions for prejudiced attitudes toward out-groups. While devout people in Finland (b =.04), Austria (b =.08), Ireland (b =.06), the Netherlands (b =.06), Portugal (b =.11), and Spain (b =.06) tend to support restrictive immigration policies vis-à-vis Muslims, the opposite effect of religiosity is observable in the Czech Republic (b = -.05) and Germany (b = -.03). In these cases, religious individuals take the biblical commandment of love thy neighbor in the literal sense and express less hostility toward Muslims than their nonreligious fellow citizens.

Beside these peculiarities of national contexts, our results show that Islamophobia rests upon quite similar social–psychological underpinnings in European societies. Essentially, it is a mixture of threat perceptions and ethnocentrism that drives anxiety toward Muslims. In view of Muslim’s extensive career as scapegoats of the West, it is hardly surprising that European’s threat perceptions turn out to be the core driver of Islamophobia. The rejection of Muslim immigration is caused by both realistic and symbolic threat perceptions. Right-wing populist’s ‘Islamization of Europe’ rhetoric is likely to fall on fertile grounds among European citizens, as symbolic fear perceptions in general exceed the effect size of realistic threat perceptions.

The perception that the influx of immigrants endangers ‘the culture of European nations’ turns out to be the core driver of Islamophobia. However, symbolic and realistic threat perceptions should not be seen in isolation of each other as fear of terrorism and the perception of sharp cultural differences vis-à-vis Muslims are closely interlinked in the mind-sets of European citizens (Pickel 2018, 26). Hungary and Finland are the exceptions to this pattern; in these countries, it is realistic threat perception that trumps the explanatory power of symbolic fears.

As mentioned earlier, there is no strong effect of religion and nationalism. Thus, it would be premature to conclude that in-group favoritism and pejorative attitudes toward out-groups do not matter at all for Islamophobia. Our empirical results unambiguously reveal that ethnocentric worldviews drive hostility toward Muslim immigrants. Europe’s legacy of ethnocentrism (and even outright racism) still matters for prejudice toward out-groups (regression coefficients vary between b = .05 in Austria and b =.16 in Poland), and one might conclude that Islamophobia is hardly a new trend among citizens in Europe.

Beyond factors that relate to Social Identity Theory and the Integrated Threat Theory, our results show that distributional conflicts and perception of collective deprivation feed citizen’s demand for more restrictive immigration policies. If citizens compare the government’s treatment of the autochthonous population vis-à-vis new immigrants and reason that the government privileges foreigners, they are more likely to support a Muslim ban (the effects range from b =.07 in Sweden to b=.19 in Austria). These effects turn out to be more relevant for respondents in Western Europe. As significant immigrant communities are part of Western European’s living conditions, this empirical pattern is hardly surprising. However, we observe similar effects in Eastern Europe as relative collective deprivation feeds support for a Muslim ban among respondents in Hungary (b =.13) and Poland (b =.16).

Summing up, we want to emphasize that it is not distributional issues, but the perception of clear-cut cultural differences vis-à-vis Muslims that drives Islamophobia. The effects of the control variables match our a priori assumptions. Respondents with higher educational achievements turn out to be less critical vis-à-vis Muslim immigrants (the regression coefficients vary between b = -.05 in Germany and b= -.10 in Lithuania). As the education level of respondents turns out to be a rather weak parameter, one might conclude that Islamophobia is a phenomenon that shapes the entire range of the social stratum. The gender effect appears to be rather weak as well, but in line with our assumption women are more likely to oppose Muslim immigration in the Czech Republic (b =.03) and in Belgium (b =.04). Furthermore, the elderly is more prone to support a Muslim ban. In Estonia, it is the age of respondents (b =.44) which turns out to be the decisive driver of anti-Muslim prejudice.

Overall, threat perceptions turned out to be the root cause of Islamophobic attitudes. In this vein, it is fear of terrorism and the notion of sharp cultural differences vis-à-vis Muslims that come along with realistic and symbolic threat perception. Is this finding enough to reason that the contact hypothesis is only playing a minor role for the existence of anti-Muslim prejudice? This would be a rash decision. As the explanatory power of our individual-level hypotheses differs considerably between Western (R² ranging from .28 in Ireland to .44 in Germany) and Eastern European societies (R² ranging from .23 in the Czech Republic to .38 in Slovenia), one might ask whether varying societal-level features of European societies matter in their own right?

Our individual-level results reveal that there is a good reason to consider contacts with migrants as an antivenom of Islamophobic attitudes, and respondents in Eastern and Western Europe do not differ too much in this respect. But for all, it is important to acknowledge that the antivenom of contacts is unequally distributed among European societies. Up to now,Eastern European countries have been comparatively unpopular choices for migrants to settle. As Muslim communities are all but existent in most Eastern European societies, Eastern European citizens have less opportunities to adjust their negative stereotypes about Muslims by face-to-face encounter. We argue that this societal-level feature is the prime reason why anti-Muslim prejudice is more prevalent in Eastern European societies. Figure 2 shows the association between the prevalence of contacts with migrants and the average support for a Muslim ban in European societies. Again, there is a clear pattern (now on the societal level) which is in line with our theoretical assumption: Less contact with migrants comes along with a higher support for a Muslim ban.

Figure 2

Do we have good reasons to consider the prevalence of contact with migrants as the decisive factor that drives the rejection of Muslim immigration? Table 3 summarizes beta-coefficients from bivariate regressions, showing how strongly each societal-level predictor influences the support for a Muslim ban. Regarding the alternative explanatory factors of Islamophobia, it is the factual size of Muslim communities (b = -.77), realistic (b =.66) and symbolic threat perceptions (b =.63), ethnocentrism (b =.52), and the perception of collective deprivation (b =.54) that impact on the support for a Muslim ban.

However, none of these factors show an impact as strong as the prevalence of contacts with migrants (b = -.88). The middle column of Table 2 summarizes how strongly each of the alternative explanations affects Islamophobia when we control for the prevalence of contacts. The right-hand column shows the results from the same regression models but indicates the impact of the prevalence of contacts when we control for the alternative explanations. The middle column reveals that contacts with migrants absorb the effects of realistic threat perceptions and ethnocentrism. Furthermore, it is noteworthy that the effects of the factual size of Muslim communities (b = -.30), symbolic threat perception (b =.29), and collective deprivation (b =.31) drop drastically when we control for contacts with migrants. In a nutshell, the effect of contacts with migrants only drops to a marginal extent and remains significantly stronger than that of the alternative explanations. We interpret these results as the confirmation of our guiding hypothesis: The absence of Muslim communities in Eastern Europe leads to less contacts, less friendships with immigrants, and more skeptical perceptions of intergroup contacts. These factors, however, strengthen a social climate in which anti-Muslim prejudice prevails. Less contact with migrants is the core driver of ‘Islamophobia without Muslims’ in Eastern European societies.

Our empirical results provide compelling evidence that Islamophobia is more widespread in Eastern Europe than in Western Europe. ‘Islamophobia without Muslims’ is an appropriate headline for this pattern as Muslim communities are virtually nonexistent in most Eastern European societies. How to account for this finding?

In line with the contact hypothesis, our results reveal that there are good reasons to consider contacts with migrants as an antivenom of Islamophobic attitudes. We observe a uniform individual-level effect in all countries that have been surveyed by the ESS (2014): individuals who stay in contact with out-groups, make friends with people from other ethnicities, and perceive these contacts as convenient are less likely to feel prejudice toward Muslims. Thus, it is important to keep in mind that the antivenom of contacts is unequally distributed among European societies. As Muslim communities are all but existent in most Eastern European societies, most citizens have no opportunities to adjust their negative stereotypes about Muslims by real-life experience. Our results confirm the core assumptions of the contact hypothesis: Less contact with migrant breeds ‘Islamophobia without Muslims’ in Eastern European societies. The fact that Muslims evoke feelings of fear and hatred among citizens in Eastern Europe is a result of para-social contacts, and the rhetoric of right-wing populist leaders is likely to aggravate these emotions (Wodak 2015). In general, Eastern European citizens rely on media information to build their opinion about Muslims. Shocking news dominates the headlines: As Muslims are frequently portrayed as fanatical fundamentalists, that are prone to violence, many Europeans associate Muslims with terrorism (Pickel 2018). Populist leaders are the major beneficiaries of this trend. Anti-Muslim rhetoric is an uncomplicated way to attract prejudiced voters. To avoid a false impression, there is an element of truth in these stereotypes. Religious fundamentalism is not a marginal phenomenon among Muslims in Europe, and this worldview indeed contradicts core democratic values such as freedom, equality, and tolerance (Koopmans 2015). The sad truth is Muslim extremists justify terrorist attacks on behalf of Islam.

However, there is a dangerous trend that terrorist attacks and media’s tendency to portray Islam as an ‘unintegratable religion’ are likely to shape the prevailing image of all Muslims – no matter if they are refugees, immigrants, or born and raised in the countries where they live. In contrast to members of the other world religions, it is Muslims and Islam who are perceived as the primary threat by European citizens. Nevertheless, there is a major difference: Eastern European citizens have less opportunities to adjust their negative stereotypes about Muslims by face-to-face encounter. We argue that this societal-level feature is the prime reason why anti-Muslim prejudice is more prevalent in Eastern Europe. Right-wing populist parties and politicians see how the wind blows. They take advantage of citizen’s unease with Muslims. Miloš Zeman, Jaroslaw Kaczyński, and Victor Orbán reinforced popular stereotypes as they presented Muslim immigrants as a socioeconomic burden, carriers of diseases, and terrorists. Apparently, Islamophobia comes along with xenophobic nationalism as right-wing populist leaders portray themselves as protectors of their nation’s ‘Christian identity’ against an ‘invasion of Muslims’.

This political strategy works out. Since the so-called refugee crisis of 2015, right-wing populist parties have been highly successful in elections all over Europe. Citizens increasingly support and accept anti-Muslim political positions, speeches, and political activities. Social–psychological theories contribute to a better understanding of this trend. It is not an accident that right-wing populist parties make use of nationalistic rhetoric vis-à-vis Muslim immigrants (Brubaker 1996). Our empirical results reveal that latent ethnocentric worldviews still matter for Islamophobia (Pickel and Yendell 2016; Pollack et al. 2014).

However, it is European’s threat perceptions that turned out to be the core driver of the recent upsurge of Islamophobia. In this vein, the rejection of Muslim immigration is caused by both realistic and symbolic threat perceptions. If right-wing populist leaders equate Muslims with terrorists and portray them as a burden for European’s employments and living conditions, then this is a clear attempt to produce (or to strengthen) realistic threat perceptions among European citizens. The more diffuse ‘Islamization of Europe’ rhetoric appears to be even more problematic. This rhetoric is likely to fall on fertile grounds among European citizens, as symbolic fear perceptions are the decisive individual-level factor that drives Islamophobia. Symbolic threats relate to the perception of sharp cultural and identitarian boundaries between in-groups and out-groups. While political systems can respond to realistic threat perceptions by a multitude of policy measures (e.g., economic policies and security policies), things look more difficult about symbolic threat perceptions. The perception of collective, cultural differences between in-groups and out-groups is much more emotional and less debatable than questions that relate to economic or security issues.

On balance, we argue that the joint effect of varying societal-level characteristics of European societies (e.g., a diverging share of Muslim population) and social–psychological factors (e.g., less contact with migrants which helps to sustain symbolic fears) gives rise to the phenomenon of ‘Islamophobia without Muslim’. In the long run, the xenophobic nationalism of Eastern European governments is likely to evoke political conflicts within the European Union. A first indication for this trend is the Visegrad group’s (e.g., Poland, Czech Republic, Hungary, and Slovakia) declaration to act jointly against binding quotas for the allocation of refugees. This trend implies tough times for the European Union. Eastern European governments owe their popularity to anti-European, nationalistic, xenophobic, and anti-Muslim positions. Miloš Zeman, Jaroslaw Kaczyński, and Victor Orbán again and again agitated against an ‘invasion of Muslims’ that was allegedly orchestrated by ‘Brussels’. Against the backdrop of these developments, we wonder whether Islamophobia enhances Euroscepticism in Eastern Europe. Future research is therefore warranted to shed light on the political consequences of Islamophobia.