The Relevance of the Centrality and Content of Religiosity for Explaining Islamophobia in Switzerland

Research on Islamophobia in Switzerland, and on the role of religiosity in relation to Islamophobia, is in its infancy. Against this background, we analyzed data from an online survey conducted in Switzerland on “Xenosophia and Xenophobia in and between Abrahamic religions”. The results of a multivariate analysis revealed that, besides right-wing authoritarianism, social dominance orientation, and political orientation, indicators related to religion play a crucial role. We found that the greater the role of religion, and the more central it is for the individual, the more likely it is that the individual has a positive view of Islam. We claim that a person’s level of religiosity is accompanied by her adoption of religious values, such as neighbourliness and tolerance, and that the more religious individuals are, the more likely they are occupied with different religions, which leads to tolerance as long as it is not accompanied by a dogmatic and fundamentalist religious orientation. Also relevant is that the preference for the state to have a secularized relationship with religion is accompanied by a fear of Islam. We propose that studies on Islamophobia, as well as on other prejudices, should use differentiated measures for religiosity; the Centrality of Religiosity Scale (CRS) turned out to be a reliable instrument of measurement in this regard.


Introduction
Switzerland has a reputation for not being particularly friendly to Islam. The reasons for this reputation lie in the ban on minarets in 2009 and other political demands designed to limit the religious practice of Muslims, including the demands made by the Swiss far right for the state not to recognize Islam, not to allow the funding of mosques, to control what Imams teach, to restrict the import of halal meat, and to ban Islamic religious care (Ademović-Omerčić 2018, p. 650). In this context, it is not surprising that population surveys should reveal that Islam and Muslims are rejected by large sections of the population.
There have also been reports of anti-Muslim crimes in Switzerland, with between 20 and 30 such incidents being reported in 2017. However, despite the high quantity of research on Islamophobia, research on this subject in Switzerland is still in its infancy. There have been a few studies on the discussion of Islam and Muslims in Switzerland (e.g., Cheng 2015;Behloul 2009;Lindemann and Stolz 2014), but only a very few have focused on Islamophobic attitudes in the country (e.g., Stolz 2005;Berger and Berger 2019). Thus, one reason for the present study is the lack of research on the specific case of Switzerland. Another lies in the fact that, despite the long tradition of research on prejudice, it has not been possible to work out adequately the importance of individual religiosity in relation to prejudice, it has not been possible to work out adequately the importance of individual religiosity in relation to Islamophobia. Most studies that test hypotheses use few and often inadequate indicators of religiosity, which produce contradictory results. Against this background, we discuss the intertwining of different explanatory approaches from research in social psychology on prejudices, as well as explanations related to religion. Our argument is that Islamophobia can be explained to a considerable degree by factors related to religion, as well as by prominent and well-proven theories such as authoritarianism and social dominance orientation. However, for this it is necessary to distinguish systematically between the centrality and content of religiosity. To measure the centrality of religiosity, we use the Centrality of Religion Scale (Huber and Huber 2012) and to measure the content of religiosity, we use the constructs of religious fundamentalism and secular threat.

Islamophobia in Switzerland: A Few Figures and the State of Research
To gauge the extent of Islamophobia in Switzerland, it is worth looking at various population surveys ( Figure 1). The overview in Figure 1 shows the level of approval given to several statements in representative population surveys in Switzerland. The results of the Religion Monitor 2013 reveal that more than half the population in Switzerland view religious plurality as a source of conflict. Only about one third believe that Islam fits into the Western world, and half believe that Islam is a threat. At the same time, hostility to Muslims is lower. According to data from the European Values Study, 14% of the population in Switzerland would not like a Muslim as a neighbour (that figure was 12% in 2012), and in the European Social Survey (ESS 2014), 13% said that they would not allow any Muslims to immigrate to Switzerland. The discrepancy between negative attitudes towards Islam and those towards Muslims may not be as relevant as it might seem. The very fact that Islam appears to be problematic for many people in Switzerland can be discriminatory for Muslims if their religion is not accepted and their religious freedom is limited. Only a very few studies have concentrated on the causes of Islamophobia in Switzerland. One example is Jörg Stolz's study, which was based on testing theories to explain Islamophobia among citizens of Zurich (Stolz 2005). The design of this quantitative study conducted between 1994 and 1995 allowed a review of several theories, one of which proved to be particularly fruitful: namely, the theory of traditionalism (Stolz 2000), which claims that there is a link between the desire to conserve Only a very few studies have concentrated on the causes of Islamophobia in Switzerland. One example is Jörg Stolz's study, which was based on testing theories to explain Islamophobia among citizens of Zurich (Stolz 2005). The design of this quantitative study conducted between 1994 and 1995 allowed a review of several theories, one of which proved to be particularly fruitful: namely, the theory of traditionalism (Stolz 2000), which claims that there is a link between the desire to conserve traditions in a rapidly changing world and the rejection of modern developments in society, such as the immigration of Muslims. Another example is a recent study by Roger Berger and Joel Berger, which was based on lost-letter experiments in Zurich (Berger and Berger 2019). The authors discovered that citizens tested in Zurich tended to distinguish between religious and non-religious Muslims, and to discriminate against religious Muslims and communities in the same way that they discriminate against native Swiss who are connected to a Christian sect. They concluded from their findings that it is less about classic xenophobia, and more about the fear that secular principles will not be accepted. This they called the "threat to secularization" explanation.
While these studies provide clues to individual factors to explain Islamophobia, there is a lack of broader research that tests classical theories from the mostly socio-psychological research on prejudices, as well as religion-related theories for the population in Switzerland.

The Impact of Religiosity on Prejudices: The State of Research
As already mentioned, research on Islamophobia has rarely dealt with religiosity. One exception is the study conducted at the Cluster of Excellence "Religion and Politics" at the University of Münster on "Perception and Acceptance of Religious Diversity" , which, using a multivariate model analyzing attitudes towards Muslims in five countries, revealed that religious exclusivism, i.e., a person's belief that it is only her own religion that contains truth, has a negative impact on attitudes to Muslims (Pollack 2014, p. 52). In addition, a syncretic religiosity has a positive but weak effect on attitudes towards Muslims in Germany and the Netherlands. The authors also revealed for West Germany and the Netherlands that whether a person has a denomination or not plays a role, too. Those with a denomination are slightly more positive towards Muslims than those without. Using the same data, Yendell found that, besides religious dogmatism and pluralism, the frequency of church attendance is positively correlated with tolerance towards Muslims (Yendell 2014) Drawing on data from the German General Social Survey (ALLBUS 2012), Rebenstorf (2018) found that approval for the statement "no religion contains truths" correlates with Islamophobia, which indicates that atheism is positively correlated with Islamophobia. In addition, the likelihood of a person rejecting Islam increases with the importance of believing in God and an exclusivist claim to religion, while church attendance tends to lead to less rejection of Islam. Drawing on the Leipzig Authoritarianism Study 2018, Pickel and Yendell found that Protestants and Catholics are more open to authoritarian attitudes than non-Christians, but there is no statistical difference between the three groups in terms of negative attitudes towards Muslims (Pickel and Yendell 2018, pp. 221-23).
In analyzing data from the Portraits of American Life Study 2006, Jung (2012) discovered that Christians are more likely to have a low level of respect for Islam. Also, the image of a God who punishes his followers correlates positively with the derogation of Islam. This study also considered contact with other religious groups. While higher frequency of contact with Muslims was accompanied by a higher probability of respecting Islam, this did not apply to Evangelical and black Protestants, with increased contact with Muslims leading to a lower level of respect for Islam in these two groups. One explanation could be the high levels of exclusivity in these groups (Jung 2012, p. 123). Doebler (2014) discussed the relationships between religiosity and intolerance among Europeans living in non-Muslim majority countries towards Muslims and immigrants. In her analysis of the European Values Study (2010), she discovered that religiosity has an effect regarding attitudes towards Muslims, and that traditional and modern beliefs in a higher being are strongly and negatively correlated with intolerance towards Muslims, whereas fundamentalism is positively correlated with the derogation of Muslims.
In his book America and the Challenges of Religious Diversity, the American sociologist of religion, Robert Wuthnow (2005), distinguished between three groups which differ regarding their religiosity and their attitudes towards Islam. He called them "Spiritual Shoppers", "Christian Inclusivists", and "Christian Exclusivists". These three types differ regarding their attitudes towards Islam as follows: while, for instance, "only" 37% of Spiritual Shoppers agree that Islam is fanatical, that figure is 47% for Christian Inclusivists, and 55% for Christian Exclusivists (Wuthnow 2005, p. 216). The three groups also differ regarding the role that they wish for Muslims: 71% of Spiritual Shoppers support the idea of Muslims gaining a stronger presence, while that figure is only 50% for Inclusivists and 36% for Exclusivists. Moreover, 46% of Exclusivists agree that it should be more difficult for Muslims to settle in the US (Inclusivists: 42%; Spiritual Shoppers: 24%) (Wuthnow 2005, p. 217 One of the most well-known studies on the relationship between religiosity and prejudices is the study by Allport and Ross (1967), who explored this relationship by differentiating between intrinsic and extrinsic religiosity, the latter being defined as a means to achieve a self-serving end, and the former as a means in itself. They found that intrinsically-oriented people have less ethnic prejudice than people with extrinsic religiosity (Allport and Ross 1967, p. 441), and came to the conclusion that those who embody intrinsic and extrinsic motivation ("indiscriminately pro-religious people") are particularly intolerant because they have a "dogmatic mind" that leads them to adopt undifferentiated perceptions (see Allport and Ross 1967, p. 441).
The difference between intrinsic and extrinsic religiosity also played a role in a survey by Batson et al. (1978) on racial prejudices, which also introduced a third category: namely, quest religiosity. They found that controlling for the effects of social desirability either diminishes or erases the negative correlation between intrinsic religion and prejudice. However, quest religiosity correlated negatively with racial prejudice even when social desirability was controlled. Scheepers et al. (2002) examined the relationship between different dimensions of religiosity and prejudice in eleven cross-national European samples. Their analysis revealed that Catholics and Protestants are more prejudiced towards ethnic minorities than non-religious people; that frequency of church attendance is positively correlated with prejudices; that subscribing to doctrinal beliefs reduces prejudice; and that the salience of religiosity and spirituality is negatively correlated with prejudice, whereas religious particularism is positively correlated.
Doctór emphasized the importance of authoritarianism, dogmatism, and fundamentalism, which in his study were the strongest predictors for people to reject new religious movements (Doktór 2002, pp. 558-59). Jelen and Wilcox (1991) found that dogmatism and belonging to a Protestant denomination affect political intolerance. Merino also noted in a recent study of the US that a religiously exclusive position reduces a person's willingness to grant the same rights to members of non-Christian religious communities in American society (see Merino 2010, p. 243). In a study on how religious orientations influence the acceptance of homosexuality, Herek found that the intrinsically oriented were even more intolerant than the extrinsically oriented. In his opinion, an intrinsic religious orientation only increases tolerance for other groups if the religious doctrine followed calls for tolerance of the relevant group (see Herek 1987, p. 35). According to Herek, religious fundamentalism can explain prejudices towards particular groups better than extrinsic or intrinsic religiosity (Herek 1987, p. 40).
Drawing on the Leipzig Authoritarianism Study, Pickel and Yendell (2018) found no significant differences between Catholics, Protestants, and those without a denomination when it comes to extreme right-wing attitudes, as well as anti-Muslim sentiments, antiziganism, xenophobia, and anti-Semitism (Pickel and Yendell 2018, pp. 221-23). Previous waves of the same study did not show a consistent picture (see also Rebenstorf 2018). In the 2010 wave, it was Protestants who had the highest values for different dimensions of right-wing extremism (which are accompanied by prejudices) (Decker et al. 2010); in 2012, those without a denomination (Decker et al. 2012); and, in 2014 and 2016, Catholics (Decker et al. 2014. Zick and Küpper revealed in a study based on "Deutsche Zustände" data that Protestants are the most likely to have racist views, while those without a denomination are the least likely (Küpper and Zick 2006). Other studies by the same authors showed only small effects, and the presence of prejudice towards only a few minorities (for instance, not Muslims) (Zick and Küpper 2014). In a comparative country study, the authors concluded that there actually is a correlation between religiosity and prejudices, since the more religious people are, the more prejudiced they are likely to be. However, they also concluded that the effect of religiosity is very weak, and that not every form of religiosity is problematic (Küpper and Zick 2010). Ekici and Yucel (2015) analyzed data from the European Values Study (EVS) 2008, and discussed the effects of religiosity on religious and racial prejudice in Europe. The latter was measured by the question of which neighbours people would not like to have (Christians, Muslims, Jews), and subsumed under one variable. What the answers revealed is that religious particularism is accompanied by more religious and racial prejudice, whereas doctrinal belief and individual spirituality are both correlated with less religious prejudice. They also found that nonreligious individuals have a higher level of religious prejudice than members of religious denominations.
In a study on extreme right-wing attitudes, Anonymous analyzed data from a German population survey (ALLBUS 2018), and differentiated between commitment to church, attachment to God, and belief in supernatural powers. Their findings revealed that commitment to church is negatively correlated, and belief in supernatural powers positively correlated, with extreme right-wing views, whereas attachment to God plays no significant role.
Results of the very few studies that considered religiosity as an indicator of prejudices, racism, and Islamophobia give the impression that, if taken into account at all, religious indicators often play a minor role. In general, it seems that salience of religiosity, intrinsic religiosity, and doctrinal beliefs tend to be negatively correlated with prejudice, whereas religious fundamentalism and particularism are always positively correlated. Findings on religious affiliation and church attendance are contradictory. This unclear picture overall is not necessarily due to the fact that religiosity offers no explanatory potential; it is also due perhaps to the fact that previous studies have not sufficiently considered certain forms of religiosity. Thus, religiosity is a black box that needs further examination.

Religiosity and Prejudices in a Model of Religiosity
As previous surveys on the relationship between religion and prejudice have not yet produced reliable results, Rebenstorf (2018) claimed that the Centrality of Religiosity Scale (CRS) (Anonymous, Anonymous) is particularly fruitful when it comes to examining the relationship between religiosity and prejudices, since it exceeds the usual characteristics of denomination, church attendance, and self-assessment of religiosity. The CRS is the backbone of a two-parameter model of religiosity, and understands religiosity as a function of the centrality and content of the personal religious construct system of the individual (Huber 2003(Huber , 2007. The centrality parameter depicts the strength and influence of the personal religious construct system in the individual. It is measured by the CRS. The CRS consists of items that measure how strongly five core dimensions of religiosity are developed in a person: namely, intellect, ideology, public practice, private practice, and experience. Following Charles Glock (1962), Stark andGlock (1968), Anonymous (2003) argued that these dimensions are the most important modes expressing religiosity in a person. Therefore, measuring the general intensity of these dimensions can yield a representative cross-section of a person's religious expressions of life. His second argument was that the amount of a person's religious expressions of life increases the likelihood that religiosity will be central in the individual's personality, and will influence her norms, experiences, and behaviour.
On the other hand, the content parameter depicts the direction in which the individual is led by her personal religious construct system. There are, of course, many possible religious contents that may be relevant for the individual, e.g., religious emotions (Huber and Richard 2010), or the experience of forgiveness by God (Huber et al. 2011). The CRS and the two-parameter model of religiosity was the basis for Bertelsmann Stiftung's international Religion Monitor (Huber 2009;Huber et al. 2011;Huber and Krech 2009), which therefore included measures of many possible contents of religiosity besides the CRS, e.g., religious reflexivity, concepts of God, religious fundamentalism, and religious emotions (Huber 2009). Both parameters (centrality and content) are necessary to gain a clearer picture of the relevance of religiosity for the experiences and behaviour of the individual. In the context of prejudice and Islamophobia, we expect two contents of personal religious construct systems to be particularly relevant: religious fundamentalism and secular threat.
Like Stark and Bainbridge (1996), we claim that church members gain the same pro-social norms as their religious companions in the moral community of a church. Engaged church members (for instance, those who regularly attend church services) are therefore more likely than others to adopt religious values such as neighbourliness and tolerance, and are thus more likely to be tolerant towards other religious communities and their members (Doebler 2014, p. 64). Secondly, we claim that the combined dimensions of private practice, experience, ideology, and intellect have much in common with Allport and Ross's type of intrinsic religiosity. We argue that the more present the experience of transcendence is to something divine or a higher power, the more likely this experience is to develop a universalizing dynamic. This could happen in such a way that the individual internalizes human values such as goodness and love for her neighbour, while at the same time promoting tolerance and openness towards other religious communities. In addition, we claim that religious knowledge may concern a person's knowledge not only of her own religion, but also of other religions, which can lead to a more nuanced image of Islam than the one propagated in the media of an aggressive and backward Islam. We also claim that trust in God is accompanied by interpersonal trust, which itself reduces prejudice. All in all, we expect religiosity to increase tolerance towards Islam.
However, as many studies have revealed, if religion becomes fundamentalist, then it can also be associated with intolerance (Doebler 2014;Eisinga et al. 1995;Laythe et al. 2002). Nevertheless, it is difficult to provide a conceptual definition of religious fundamentalism . It is sometimes defined as radical and reflexive traditionalism (Riesebrodt 1990), as authoritarianism (Altemeyer andHunsberger 1992, 2004), as a militant reaction to modernity (Almond et al. 2003), as a self-immunizing form of thought (Meyer 1995(Meyer , 2011, and as an intratextual type of religious search for meaning (Hood et al. 2005). Because the approaches are so diverse and contradictory, we propose using a definition from the Chicago "Fundamentalism project" and from the research by Hood et al.: Religious fundamentalism is a religious attitude that is characterized by a syndrome of certain characteristics. At its core, an exclusivist understanding of one's own religion and a strictly intratextual search for absolute religious truth come into effect. Religious exclusivity and absolutism do not in themselves constitute a fundamentalist attitude. Additional characteristics come into play. Dualistic constructions which make a clear distinction between an area of good and salvation on the one hand and an area of evil and misery on the other are essential. After all, one should only speak of a fundamentalist-religious attitude if there is a strong social cohesion and a high level of commitment to one's own religious group (Huber in print).
As we know from Social Identity Theory (Tajfel and Turner 1986), such religious fundamentalism is accompanied by socio-psychological dynamics. In order to achieve a higher level of self-esteem, individuals tend to increase the value of their social in-group, while at the same time devaluing social out-groups. This dynamic becomes stronger when the social in-group is exclusive, as is the case with fundamentalist religious groups.
Another fruitful theoretical approach is the "threat to secularization approach" (Berger and Berger 2019). For Berger and Berger, one major dimension of Swiss identity is secularism: Switzerland has a long tradition of secularism, and the state controls and places strong restrictions on the Catholic and Protestant churches. They also pointed to a number of historical events that show how strongly in favor of secularism the Swiss are, such as the ban on the Jesuit order in 1874, on new monasteries and dioceses, and on kosher butchering in 1873, the exclusion of clerics from the Swiss parliament until 1999, and the ban on minarets in 2009. It was therefore not surprising for Berger and Berger that Switzerland has a comparatively high degree of secularism and regulation of religious issues (Berger and Berger 2019;Fox 2008). From this perspective, Islamophobia is not the "result of general xenophobia", but rather a rejection of "religious groups outside the publicly recognized churches".

Political and Social-Psychological Dimension of Prejudices: "the Usual Suspects"
Another potentially useful approach is provided by the theory of authoritarian personality, which has a long tradition in sociology and social psychology. It was mainly used by Wilhelm Reich (1933) and Erich Fromm (1941) to explain the rise of national socialism at the end of the 1920s and early 1930s, and gained prominence through the work of Adorno and his co-authors (Adorno et al. 1950). It is now very important in research on prejudice and right-wing extremism (Decker and Brähler 2018). Typical of the authoritarian personality is the belief in absolute obedience or submission to authority, and the oppression of those considered weak. Authoritarian individuals have an authoritative, strict, or oppressive character, particularly towards subordinates. Psychological projection, as described in psychoanalysis, also plays a role, with a person's own inferiority being transferred to foreigners, who are accused, for example, of being aggressive and dishonest. An authoritarian mindset can be accompanied by insecurity and low self-esteem (Larsen and Schwendiman 1969), as well as by low self-acceptance (Kornyeyeva and Boehnke 2013). Adorno's original theory was based on the fact that German society long had an ideal of upbringing that consisted of punishment. Although modern society has become more open and the physical punishment of children is now forbidden and less common, this approach still sees authoritarianism as being omnipresent, and points to the fact that people are also exposed today to coercion and (performance) pressure (as exerted, for example, by the economic system) (Decker and Brähler 2016), and often lack recognition. Drawing on the concept of an authoritarian personality, Adorno and his colleagues developed the f-scale (f stands for fascism), which quantitative studies on prejudices and right-wing extremism have since made broad use of, although there has been wide criticism of the scale, particularly because its wording creates acquiescent response styles. In response, Altemeyer has since developed a common instrument to measure authoritarianism (see Altemeyer 1981): namely, the RWA (Right-Wing Authoritarianism) scale, which has an equal number of pro-and anti-authoritarian statements. Altemeyer reduced the nine subdimensions of authoritarianism outlined by Adorno et al. (1950) to three: authoritarian aggression (authority-sanctioned general aggression towards others), authoritarian subservience (subservience to established authorities and general acceptance of their statements and actions), and conventionalism (strong adherence to established social conventions) (see Beierlein et al. 2014). According to Altemeyer (1981Altemeyer ( , 1996, these three dimensions suffice to describe and measure the RWA personality traits. Unlike the psychodynamic approach of Adorno et al. (1950), which sees the cause of the authoritarian personality as lying in early childhood, Altemeyer sees RWA as a personality trait that is socialized in early youth.
Another relevant theory to explain prejudices is Social Dominance Orientation (SDO), which is a measure of the individual level of acceptance of group-based hierarchies and the corresponding inequalities (Pratto et al. 1994;Sidanius and Pratto 2001).
"SDO is defined as a very general individual differences orientation expressing the value that people place on nonegalitarian and hierarchically structured relationships among social groups. It expresses general support for the domination of certain socially constructed groups over other socially constructed groups, regardless of the manner in which these groups are defined. (...) Individuals differ in the degree to which they desire group-based inequality and dominance for any number of reasons" (Sidanius and Pratto 1999, p. 61).
SDO has often been shown to have a high level of explanatory power when it comes to different kinds of prejudices or political attitudes (e.g., Newman et al. 2014;Dru 2007;Cohrs and Asbrock 2009;on Islamophobia, see Uenal 2016).
SDO emerged from social dominance theory (Sidanius and Pratto 1999), which is a multilevel theory that focuses on the retention and stability of group-based social hierarchies. These hierarchies bestow privileges on dominant groups and are present in nearly all stable societies. According to Sidanius and Pratto, hierarchies consist of three systems: (1) Age (adults are more privileged than children); (2) Gender (men usually have more power then women); (3) An arbitrary system (culturally defined group-based hierarchies).
Another relevant concept is that of the political spectrum. Generally, the left is characterized by an emphasis on "ideas such as freedom, equality, fraternity, rights, progress, reform and internationalism", while the right is associated with "notions such as authority, hierarchy, order, duty, tradition, reaction and nationalism". Many studies have revealed that a person's political position is linked to Islamophobia, with people who view themselves as left-wing generally being less Islamophobic than those who view themselves as right-wing (Anonymous; Anonymous).

Method and Data
We analyzed data from a subsample of people of other faiths than Islam, of a non-representative online survey conducted in the research project "Xenosophia and Xenophobia in and between Abrahamic religions" (XENO 2018) funded by the Swiss National Science Foundation (SNF).
The sampling was oriented towards the aim of the project, namely to investigate structures of prejudice in the mutual perception of members of Abrahamic religions as well as of non-denominational and secularists. By secularists we mean non-denominationalists who belong to an association that is critical of religion and advocates the separation of church and state. Within Christianity, not only members of the two large national churches but also members of free churches and of Orthodox churches were supposed to be examined. These project goals could hardly be achieved with a representative sample, since Jews, Muslims, free church, and Orthodox Christians, as well as secularists would not be present in a sufficient size. Therefore, samples were taken from secularist organizations (e.g., freethinkers) as well as from churches and religious communities. All Jewish communities and associations in Switzerland were contacted and asked for help. Furthermore, numerous Muslim mosque communities and associations as well as Orthodox churches and free churches were written to and asked for help. The same procedure was followed with congregations of the two large national churches in the Bern and Zurich areas. All the organizations contacted were provided with text modules for their newsletters as well as further information on the study-e.g., the positive decision of the ethics committee at the University of Anonymous. The online questionnaire was accessible via a website that also contained background information on the study. In addition to the national languages German and French, the questionnaire could also be completed in Albanian, Bosnian, English, Croatian, Serbian, and Turkish. The translations were checked by translation and retranslation. The online survey was open from November 2015 to December 2018. A total of 1093 people took part in the survey (age: m = 47.1, SD = 17.4; female 51.4%; people with migrant background: 31%; people with a university degree: 63%). Of these, 213 people professed their faith in the Reformed Church, 223 in the Catholic Church, and 132 in a free church, 77 in Judaism, and 156 in Islam. Due to the fact that we were interested in attitudes towards Islam in this research Muslims were excluded from the following analysis. In addition, 205 non-denominational participants took part, including 46 people from secularist associations. Furthermore, there were 87 people who belonged to another church or religion or who did not expressed their religious affiliation.
In comparison to the population statistic (Bundesamt für Statistik (BFS) 2019), the following deviations can be found. According to official statistics, the average age of the Swiss population is 42.2. In our sample, the average age was 48.3, which means it differed by six years. According to official statistics, 50.4% of the Swiss population is female. In our sample, the proportion was 50.7%, which is only slightly different. A total of 75% of the population in Swittzerland has Swiss citenship, in our sample the proportion was a bit higher (83%). The biggest deviation from the population statistics concerns education. While according to official statistics 28.8% of the Swiss population has a university degree, this figure was significantly higher, at 65%. This is not a surprising deviation and typical for online surveys. As we did not make any statements about the distribution in the population on the basis of our data, we refrained from weighting, as a non-representative sample was sufficient for testing the hypotheses.

Variables
There was a question in the study on general attitudes towards religions: "If you think about how many religions there are in this world, how threatening or rewarding do you perceive the following religions?" (answers: very enriching, quite enriching, quite threatening, very threatening. Figure 2 shows the proportion of those respondents who viewed different religions as threatening). More or less half of the respondents said that they feel threatened by Islam, whereas only a minority of respondents said that they feel threatened by Christianity (12%) and Judaism (15.7%). This corresponds with the findings of other studies on attitudes towards religious plurality in Christian countries (see Pollack et al. 2014;Pickel and Yendell 2018). The dependent variable in the following was the question concerning the perception of Islam.
We next describe the variables that were included in the statistical analysis against the background of theoretical considerations. The precise formulations of the individual statements and questions are in Appendix A.


Political orientation was measured on a scale from 0 (left) to 100 (right). Left-wing individuals (0 to 34) felt less threatened than those in the middle (35 to 65) and those on the right (66 to 100) (35% versus 63% versus 84%);  The full RWA scale is created from the mean of six variables. People with a high mean (≥ 1.5) felt more threatened by Islam than those with a low mean (< 1.5) (75% versus 44%);  Social Dominance Orientation is created from the mean of seven variables. People with a high level (≥ 50) felt more threatened than those with a low level (< 50) (73% versus 48%);  Religious Fundamentalism is created from the mean of six variables. People with a high mean (≥ 1.5) felt more threatened than those with a low mean (< 1.5) (55% versus 43%); 1  Secular threat is created from the mean of two variables. People with a high mean (≥ 2.5) felt more threatened than those with a low mean (< 2.5) (78% versus 45%);  The centrality of religiosity scale (CRSi-7) is created from seven variables. People who were not religious (1 to 2) felt more threatened by Islam than those who were religious (2.2 to 3.8) and than those who were very religious (4 to 5) (59% versus 48% versus 51%);  Age as a metric variable. There was no linear direction between age groups (below 20: 50% felt threatened; 20 to 34: 53%; 35 to 49: 51%; 50 to 64: 51%; 65 and above: 47%); More or less half of the respondents said that they feel threatened by Islam, whereas only a minority of respondents said that they feel threatened by Christianity (12%) and Judaism (15.7%). This corresponds with the findings of other studies on attitudes towards religious plurality in Christian countries (see Pollack et al. 2014;Pickel and Yendell 2018). The dependent variable in the following was the question concerning the perception of Islam.
We next describe the variables that were included in the statistical analysis against the background of theoretical considerations. The precise formulations of the individual statements and questions are in Appendix A.

•
Political orientation was measured on a scale from 0 (left) to 100 (right). Left-wing individuals (0 to 34) felt less threatened than those in the middle (35 to 65) and those on the right (66 to 100) (35% versus 63% versus 84%); • The full RWA scale is created from the mean of six variables. People with a high mean (≥ 1.5) felt more threatened by Islam than those with a low mean (< 1.5) (75% versus 44%); • Social Dominance Orientation is created from the mean of seven variables. People with a high level (≥ 50) felt more threatened than those with a low level (< 50) (73% versus 48%); • Religious Fundamentalism is created from the mean of six variables. People with a high mean (≥ 1.5) felt more threatened than those with a low mean (< 1.5) (55% versus 43%); 1 • Secular threat is created from the mean of two variables. People with a high mean (≥ 2.5) felt more threatened than those with a low mean (< 2.5) (78% versus 45%); • The centrality of religiosity scale (CRSi-7) is created from seven variables. People who were not religious (1 to 2) felt more threatened by Islam than those who were religious (2.2 to 3.8) and than those who were very religious (4 to 5) (59% versus 48% versus 51%); • Age as a metric variable. There was no linear direction between age groups (below 20: 50% felt threatened; 20 to 34: 53%; 35 to 49: 51%; 50 to 64: 51%; 65 and above: 47%); • Gender: male 0/female 1: Males felt more threatened by Islam than females (57% versus 44%); • Education (highest level of educational achievement): secondary school (58% felt threatened); middle school and grammar school (39% felt threatened); vocational school (59% felt threatened); university or college degree (48% felt threatened).

Results
To analyze how strong the explanatory power of individual theories is in a more complex hypothesis model, we conducted a hierarchical regression with listwise deletion with each step considering statistically significant indicators only. The step-by-step approach also enabled us to discover whether explanatory variables retained their explanatory power when new indicators were added, and to describe the change in variance by increasing other variables. The regression analysis revealed the extent to which the feeling of threat of Islam was affected by social-structural characteristics such as age, gender, and education (step 1), political orientation, right-wing authoritarianism, social dominance orientation (step 2), religious fundamentalism, centrality of religion, and secular threat (step 3) (Table 1). The first model contained the social-demographic variables of age, sex, and education. With a corrected R 2 of 0.023 there was a predictably low explained variance. While age did not play a significant role, gender did (beta = 0.148), with men feeling more threatened by Islam than women. 1 The fundamentalism scale that was used here was used successfully in the Bertelsmann Foundation's international Religion Monitor (see Huber 2009, p. 28;Huber and Krech 2009, pp. 74-77).
Education also had a significant, but weak, effect (beta = 0.070*), with those who were better educated feeling less threatened by Islam.
The second step added political position, right-wing authoritarianism, and social dominance orientation, which increased the corrected R 2 by 0.161. With a beta of −0.227***, political position was comparatively strongly correlated with the feeling of threat of Islam: the more to the right that people positioned themselves, the more likely they were to feel threatened by Islam.
Not surprisingly, RWA (beta = −1.150***) and SDO (beta = −0.109***) correlated significantly with the feeling of threat of Islam. Education was no longer a statistically relevant factor when these three variables were considered. Gender still played a role, though, since men felt more threatened than women. With a beta of 0.075*, the effect was weaker than in the first model.
The third model is particularly interesting, as we added the scales related to religion. Again, the change to the corrected R 2 was considerable (0.167), showing a value of 0.351. All three variables related to religion-CRS, religious fundamentalism, and secular threat-played a significant role. While CRS was positively correlated with attitudes towards Islam (beta = 0.262***), religious fundamentalism (beta = −0.364***) and secular threat (−0.245***) were negatively correlated, which is not surprising. Interestingly, the roles of RWA and SDO were more or less the same as in the previous model (beta = −0.155*** and −0.105*), whereas the strength of the political position was lower (beta = −0.116***). We conclude that, besides the "usual suspects" of RWA, SDO, and political position, variables related to religion also play a crucial role in explaining Islamophobia: the more central religiosity is to the individual, the less likely she is to feel threatened by Islam. Fundamentalist religious attitudes, as well as the preference for a secular society, make people more likely to feel threatened by Islam. In the last model, none of the socio-demographic variables played a role any longer.

Discussion
The aim of this study was to analyze the relevance of different theories formulated to explain Islamophobia in Switzerland, with special regard to indicators related to religion. The study revealed that, besides political position, right-wing authoritarian attitudes, and social dominance orientation, explanations related to religion also play a crucial role. This is interesting because research on Islamophobia and prejudices either often neglects or does not sufficently operationalize factors related to religion. The results of the multivariate analysis are also revealing, since, in the context of Islamophobia, religiosity brings forth its own dimension besides right-wing orientation, social dominance orientation, and political position. We found that attitudes towards Islam were correlated with religiosity, since the more religion plays a role for the individual and becomes central, the more likely she is to have a positive view of Islam. We assumed that the level of religiosity is accompanied by the adoption of religious values, such as neighborliness and tolerance, and that the more religious individuals are, the more likely they are to be confronted with and occupied by different religions, which leads, for instance, to their having a nuanced picture of Islam and to their being more tolerant. Being strongly religious promotes tolerance towards Islam, as long as this is not accompanied by a dogmatic and fundamentalist religious orientation. Another relevant factor is that the preference for the state to have a secularized relationship with religion is accompanied by a fear of Islam, since many consider Islam to be an anomaly in a secularized world. Finally, we argue that studies on Islamophobia and other prejudices should use differentiated measures for religiosity; the CRS turned out to be a sufficient measure in this regard. Funding: This research was funded by the Swiss National Science Foundation (SNF), project number 100017_1535278).

Conflicts of Interest:
The authors declare no conflict of interest.