“Our Christian nation”: the role of Christian religiosity in explaining opposition to Muslim expressive rights

ABSTRACT

This research aimed to investigate to what extent the relationship between endorsing a Christian representation of Dutch national identity and opposition to Muslim expressive rights can be explained by perceived ethnocultural incompatibility (between the Dutch and Muslim ways of life), and to what extent these relationships depend on individuals’ level of Christian religiosity. We hypothesized that for nominal Christians – but not for religious Christians – higher levels of perceived ethnocultural incompatibility could explain the positive relationship between Christian representation endorsement and opposition to Muslim expressive rights. To test this hypothesis, an online survey was conducted among native Dutch adults (N = 275). The results showed that perceived ethnocultural incompatibility fully mediated the relationship between Christian representation endorsement and opposition to Muslim rights for nominal, religious and non-Christians. This illustrates that perceiving the Netherlands as Christian can translate into more opposition to Muslim expressive rights via stronger perceptions of ethnocultural incompatibility.

KEYWORDS:

• Opposition to Muslims
• national identity
• incompatibility
• Christian religiosity
• intergroup relationships
• religious tolerance

Religious people express their religion in diverse ways. Catholic nuns cover their bodies and their hair to show their devotion to God, and in certain orthodox Protestant circles, women place value on dressing modestly and covering their head in church. Thus far, these ways of expressing religion have not led to any controversies in Dutch society. However, Muslim religious garments like the hijab and niqab, as well as other religious expressions of Muslims (e.g. Islamic schools, the building of mosques), have been the subject of many parliamentary and societal debates. The belief that Islam is incompatible with national identity is widespread in the Netherlands and many other European countries (Pew Research Centre 2018). These perceptions of identity incompatibility are fuelled by populist radical-right party leaders, such as Geert Wilders, leader of the Freedom Party (PVV), which argue that Islam and religious expressions of Muslims should be limited because they are incompatible with the Dutch Judeo-Christian nation. Similar discursive strategies highlighting the importance of Christian heritage for national belonging and for excluding Muslims as “religious others” are used by populist radical-right parties in other western countries (Hafez 2014). These political and societal debates highlight the importance of conducting social scientific research on the way in which individuals’ perceptions of Christianity in relation to national identity shape their attitudes towards immigrant out-groups, and Muslims in particular.

Research indicates that while individual Christian religiosity has drastically declined in Western societies as a result of secularization (Voas and Doebler 2011), many native majority members in these societies still consider Christian religion and Christian cultural heritage as essential parts of their national identity (Straughn and Feld 2010; Trittler 2017). Studies in the Dutch context have demonstrated that native majority members who believe that national identity is rooted in Christianity are more opposed to Muslim expressive rights (Smeekes and Verkuyten 2014), but little is known about the mechanisms that can explain this relationship. Research has demonstrated that perceptions of identity incompatibility form an important explanation for the link between individual beliefs about national identity and belonging and out-group attitudes (Sindic and Reicher 2009; Smeekes, Verkuyten, and Poppe 2012). Populist radical-right parties often use a Christian representation of national identity to highlight the incompatibility with Muslims and Islam and the subsequent need for strict policies targeting their religious expression. However, it is unknown whether native majority members’ perceived incompatibility between the national and Muslim way of life forms an explanatory mechanism for the link between endorsing a Christian national identity representation and opposing Muslim expressive rights. In the current study, it is therefore investigated if perceived ethnocultural incompatibility, which is defined as the belief that Islam is incompatible with Dutch national identity, can explain this link.

Moreover, little is known about the role that personal religiosity plays in the relationship between endorsing a Christian representation and negative attitudes towards Muslims. While research demonstrates that Christians more often endorse a Christian representation of national identity than non-Christians (Smeekes and Verkuyten 2014), it has also been shown that higher levels of Christian religiosity can go together with lower levels of opposition to Muslim expressive rights (Carol, Helbling, and Michalowski 2015; Eskelinen et al. 2022). Building on previous research by Storm (2011), which distinguished between non-Christians, religious Christians and nominal Christians (i.e. identifying as Christian but being hardly religious), this paper aims to investigate whether the link between endorsing a Christian representation and negative attitudes towards Muslims (in the form of perceived ethnocultural incompatibility and opposition to Muslim expressive rights) depends on these different levels of Christian religiosity.

In sum, this study extends existing research on endorsing a Christian representation of national identity and attitudes towards Muslims by focusing on perceived ethnocultural incompatibility as an explanatory mechanism and by investigating the role of Christian religiosity in these relationships. The findings of this study can help societies to better understand why and for whom the Christian narrative about national identity results in exclusionary attitudes towards Muslims. As such, this contributes to our understanding of the mobilizing potential of this Christian narrative in relation to support for populist radical-right parties, which have incorporated this narrative in their new master-frame to enlarge their electorate.

Theory

Christian representation of national identity and intergroup relations

Nations are mainly defined as communities that live and move together through time and people often perceive these communities as sharing a culture and identity that is transmitted from generation to generation (Anderson 1983; Sani 2008). Research has shown that people find comfort in the belief that their national group and its culture have historical endurance, because it satisfies their psychological need for identity continuity – i.e. the need to see a sense of connection between one’s past, present and future identity (Smeekes and Verkuyten 2015). In Western societies, politicians often portray rapid socio-cultural changes related to immigration and cultural diversification as a threat to the continuity of traditional national culture and identity (Mols and Jetten 2014). However, the content of this traditional national culture and identity is not self-evident and can be defined and understood in different ways. In the Western context, there are different ideas about the historical roots that provide the basis for understanding “our” shared national heritage and identity. This means that people do not only understand their national identity as having temporal endurance, but also have ideas about how cultural heritage determines the contents of their national identity. This latter aspect is relevant for the study of intergroup relations, because depending on the particular historical content that people believe to be at the heart of their national identity they are likely to position themselves favourably or unfavourably towards the presence of immigrant out-groups in society.

The notion that the content and meaning that people ascribe to their group membership is crucial for understanding intergroup relations is a key premise of self-categorization theory (SCT; Turner et al. 1987). The theory proposes that group behaviour is dependent on the contents and meanings that people attribute to their group identity, such as shared norms, values and beliefs. More specifically, SCT argues that self-categorization in terms of a particular group membership, such as national identity, results in a process of depersonalization in which individuals stereotype themselves in terms of what defines their own group compared to relevant out-groups. Through this process of self-stereotyping, the norms, values and beliefs of one’s group become part of the psychological self and thereby provide the guidelines for appropriate intergroup behaviour. In line with SCT, there is a large body of research that has shown that the attitudes of national majority members towards ethnic and immigrant out-groups depend on whether national identity is defined in terms of ethnic (i.e. ancestry and descent) or civic (i.e. citizenship and community engagement) characteristics (Meeus et al. 2010; Pehrson, Vignoles, and Brown 2009; Wakefield et al. 2011). While the ethnic understanding translates into more negative attitudes towards these groups, a civic understanding is often found to go together with more positive attitudes (Reijerse et al. 2013).

A more recent line of research departing from SCT has focussed on how historical representations of national identity affect attitudes towards Muslims in Western European contexts (e.g. Badea, Bender, and Korda 2020; Smeekes, Verkuyten, and Poppe 2011; 2012; Smeekes and Verkuyten 2014). One historical representation that figures prominently in European discourses on national identity and cultural diversity is the idea of being a Christian nation (Straughn and Feld 2010; Trittler 2017). This Christian representation refers to the belief that national identity is rooted in Christian heritage, customs and traditions. In increasingly secular Western societies where Christianity is still the dominant cultural religion, Muslims are often portrayed as the most visible “others”. The Christian representation of national identity is often evoked in public debates to argue for strict measures on public expressions of Islam (e.g. wearing a hijab or niqab, Islamic schools, mosques), as the increasing presence of Muslims is portrayed as a threat to the continuation of “our” Christian national heritage and traditions.

Experimental and survey research has shown that when a Christian representation of national identity is salient or endorsed by national majority members, this results in more opposition to Muslim expressive rights (Badea, Bender, and Korda 2020; Smeekes, Verkuyten, and Poppe 2011; Smeekes and Verkuyten 2014). A possible reason for this relationship could be that native majority members think that the Muslim way of life is incompatible with the original national way of life (Pew Research Center 2018). When native majority members believe that national identity and culture are strongly rooted in Christian heritage and traditions, the increasing presence and public visibility of Islam is likely to be perceived as eroding this traditional Christian way of life and therefore as incompatible with national identity. This perceived incompatibility is likely to result in negative attitudes towards public expressions of Islam. This prediction is in line with studies showing that perceived incompatibility relates to negative attitudes towards relevant out-groups (e.g. Martinovic and Verkuyten 2012; Sindic and Reicher 2009). In this study, the term “ethnocultural incompatibility” will be used in order to emphasize the nature of the incompatibility that Dutch natives perceive: Muslims are seen as incompatible with Dutch society because of the ethnocultural differences between their culture and original Dutch culture.

Christian religiosity and its influence on opposition to Muslim expressive rights

Studies have demonstrated that even though religious Christians are more likely to endorse the idea that national identity is rooted in Christian traditions than non-Christians (Smeekes and Verkuyten 2014), Christian religiosity does not necessarily go together with negative attitudes towards Muslims. For example, Eskelinen et al. (2022) found that stronger religious identification among national majority members in historically Christian nations (Austria, Finland, Germany and Norway) was related to greater support for Muslim expressive rights. Moreover, research by Carol, Helbling, and Michalowski (2015) has demonstrated that religious Christians were more supportive of Muslim expressive rights compared to non-religious individuals in six European countries. Taken together, this pattern of findings suggests that the relationship between endorsing a Christian representation and opposing expressive rights for Muslims could be dependent on the level of Christian religiosity.

This idea is linked to previous research on different ways of being affiliated with Christian religion. For example, Storm (2011) distinguishes between religious and nominal Christians. Where religious Christians consider their religiosity as an important part of themselves, nominal Christians identify as Christian but report low levels of religiosity. The idea of nominal Christians is also in line with work by Demerath (2000) on the phenomenon “cultural religion” that can be observed throughout Christian Europe. According to Demerath, cultural religion is “ … a style of religion that resides ‘in the culture’ without compelling active belief or participation” (2000, 136). Unlike religious Christians, adherents of cultural religion do not attend church frequently, nor do they subscribe to certain religious beliefs. They may identify as Christian, but report low levels of religiosity, because personal religiosity is not important to them (Demerath 2000). For cultural religionists, the dominant religious tradition of the nation is important because it signifies historical continuity. Moreover, they see the religion of the nation as an important part of national identity, that can also be used to separate natives from non-natives.

Storm (2011) found that in Great Britain, nominal Christians have a more ethnic understanding of national identity, compared to religious Christians and non-Christians. This means that nominal Christians define their national identity more in terms of ethnicity and having a shared history and culture, whereas religious Christians and non-Christians more strongly emphasize the civic and symbolic aspects of national identity. In addition, the study by Storm showed that nominal Christians see religious belonging as tightly connected to the nation: being a native for them means also being a Christian. Thus, for nominal Christians, Christianity has ethnocultural significance rather than existential significance. This means that Christians with a low level of individual religiosity use the concept of Christianity as an ethnic identity marker, in order to distinguish between natives (i.e. Christians) and non-natives (i.e. Muslims).

On the other hand, religious Christians share their religiousness with Muslims and this may result in a sense of religious solidarity. In other words, religious Christians and Muslims have a common superordinate “religious” identity which sets them apart from the majority of secular natives in Western Europe. Accordingly, religious Christians and Muslims are likely to attach importance to similar rights when it comes to freedom of religion (e.g. right to start religious schools and expressing one’s religion in public). This could also explain why Christian religiosity has been linked to more positive attitudes towards Muslim expressive rights (Carol, Helbling, and Michalowski 2015; Eskelinen et al. 2022).

Based on the theoretical reasoning above, we predict that the positive effect of endorsing a Christian representation of Dutch national identity on perceived ethnocultural incompatibility and opposition to Muslim expressive rights will be stronger for nominal Christians than for religious Christians. In addition, we expect that for nominal Christians – but not for religious Christians – higher levels of perceived ethnocultural incompatibility can explain the positive relationship between endorsing a Christian representation of Dutch national identity and opposition to Muslim expressive rights (see Figure 1). For non-Christians, no specific predictions are made, because this group has not yet been studied apart from Christians in relevant literature. Therefore, for non-Christians these relationships will be analysed exploratively.

Figure 1. Conceptual model.

Methods

Participants and design

In order to investigate the relationships hypothesized above, an online self-administered survey was conducted, in which attitudes towards Muslims and Dutch identity among native Dutch majority members were measured. The survey was written in Dutch. Ethical approval for conducting the survey was granted by the Ethical Review Board of the Faculty of Social and Behavioural Sciences of Utrecht University (approval number 22-1343). Participation in this study was on a voluntary basis and was not rewarded. Since the aim of this study is to explain native Dutch majority opposition to the minority outgroup of Muslims, Dutch adults who identify as Muslim were not included in the target population because of their outgroup membership. Because of the limited time and resources of the researchers, a convenience sample was used to obtain the data. Participants were approached via the recruitment website Survey Circle and the link to the survey was shared with the personal network of the first author via the social media platforms WhatsApp, Instagram and LinkedIn. The main disadvantage of using this sampling strategy is that the results are likely not fully generalizable to the larger population, since the sample used in this study most likely is not representative of the general population of native Dutch adults. In the discussion section, the limitations of this approach and the eventual composition of the sample are discussed in more detail.

In total, 284 respondents participated in the study. Nine of these respondents reported to have at least one non-native Dutch parent. Because the target population of this study consist of adults with a native Dutch background, which is defined as having two parents of Dutch descent, these respondents were excluded from the analysis. Thus, eventually 275 respondents were included in the final sample. The ages of the respondents ranged from 18 to 78 (M = 34.06, SD = 14.78). Of all participants, 122 identified as male (44.4 per cent), and 151 as female (54.9 per cent). One person identified as non-binary and another person indicated that they would rather not disclose their gender. These two persons have been coded as missing with respect to the variable gender. The sample was relatively highly educated: 198 respondents (72 per cent) of the respondents had completed or were enrolled in higher tertiary education at the time, whereas 77 respondents (28 per cent) reported to have completed or to be enrolled in primary school, secondary education or lower tertiary education (see Table 1 for a complete overview).

Table 1. Frequency table of educational level.

	Frequency	Percentage	Cumulative Percentage
Primary education	1	0.4	0.4
Lower secondary education	15	5.5	5.8
Upper secondary education	39	14.2	20.0
Post-secondary non-tertiary education	22	8.0	28.0
Short-cycle tertiary education	22	8.0	36.0
Bachelor’s degree	101	36.7	72.7
Master’s degree	67	24.4	97.1
Post-graduate education (PhD)	8	2.9	100
Total	275	100

Note: N = 275.

Procedure

Participants completed the survey in Qualtrics, an online survey programme. The survey could be completed on a desktop computer as well as a mobile phone and usually took five to ten minutes. Before starting the survey, an information letter about the topic of the study and the questionnaire itself was shown to the participants, who could give their informed consent to partake in the study and to allow the use of their data by ticking a box. The first part of the survey consisted of questions about demographic characteristics and questions regarding the religiosity of the participant, after which the key variables were measured in the following order: endorsement of a Christian representation of Dutch national identity, perceived ethnocultural incompatibility, and opposition to Muslim expressive rights.

Participants were able to withdraw from the survey at any time by simply closing the screen. At the end of the survey, participants had to actively allow the use of their data and they received an email address for questions about the survey. No personal data such as names or IP-addresses were stored; the survey was completely anonymous. Data were treated confidentially and were only used for analysis as is described below. Furthermore, the data were stored safely in accordance with the guidelines of Utrecht University.

This research is about a potentially invasive topic, namely attitudes towards a minority outgroup. Participants could have experienced discomfort when they filled out the survey questions about their opposition to Muslim expressive rights or their perceived ethnocultural incompatibility with Muslims, because they may have thought that their opinion is not in line with what is socially desirable. As a result, social desirability bias could have occurred. The researchers have attempted to limit this by assuring the participants at the start that: (a) the survey was fully anonymous, (b) the researchers were interested in their personal opinion, and (c) that there were no right or wrong answers to the questions. In addition, (d) participants were able to skip questions they preferred not to answer.

Measures

Opposition to Muslim expressive rights

Opposition to Muslim expressive rights was measured using six statements drawn up by Verkuyten (2007). This six-item scale has proven to be internally consistent and has previously been used in research in the Dutch context (e.g. Smeekes, Verkuyten, and Poppe 2011; Smeekes and Verkuyten 2014). Participants had to indicate to which extent they agreed or disagreed with every statement, on a scale ranging from 1 (strongly disagree) to 7 (strongly agree). Examples of statements that are used, are: “In the Netherlands, Muslims should have the right to establish Islamic schools”, and “In the Netherlands, Muslim women should have the right to wear a headscarf anywhere they want” (all items can be found in Appendix 1). The items were recoded so that a higher score indicates larger opposition to Muslim expressive rights, and averaged into a scale (α = .91).

Endorsement of a Christian representation of Dutch national identity

Whether participants endorsed a Christian representation of Dutch national identity was measured using the same five items as Smeekes and Verkuyten (2014). Sample items are: “The Dutch have passed on their Christian traditions to next generations”, and “The Dutch identity is rooted in Christian traditions” (see Appendix 1). Participants indicated their (dis)agreement with these statements using a seven-point scale which ranged from 1 (strongly disagree) to 7 (strongly agree). The scores were averaged into one coherent scale (α = .82).

Perceived ethnocultural incompatibility

Perceived ethnocultural incompatibility was assessed with five items based on Sindic and Reicher (2009) and adjusted to the context of the present study. Item examples are “The fact that the Dutch and Muslim ways of life differ does not mean that they are necessarily in opposition”, and “Muslims cannot be seen as truly Dutch” (see Appendix 1 for all items). The answer options ranged from 1 (strongly disagree) to 7 (strongly agree). Again, the scores were averaged into a scale (α = .84).

Christian religiosity

Christian religiosity was measured using two questions. First, participants were asked which religion they most affiliated with. Secondly, they had to indicate how important their religion is to them. Answer options ranged from not important at all (1) to very important (7). 98 respondents indicated that they did not affiliate with a religion (35.6 per cent). Hence, these respondents were coded as non-religious (1). 176 respondents affiliated with Christianity (64 per cent). These respondents were put in one of two categories based on their score on the religiosity item: those who indicated that their religion is not very important to them (3 or lower) were coded as nominal Christians (2), and those with a score of 4 or higher on religiosity were coded as religious Christians (3). Of the 176 Christians in the sample, 145 were coded as religious Christians (52.7 per cent of the total sample) and 31 as nominal Christians (11.3 per cent of the total sample). One respondent did not disclose their religious affiliation and has been treated as missing with regard to this variable.

Control variables

Education. Several studies have shown that people with lower levels of education are more strongly opposed to Muslim expressive rights compared to higher educated people (Smeekes, Verkuyten, and Poppe 2012; Smeekes and Verkuyten 2014; Velasco González et al. 2008). Therefore, educational level has been controlled for in the analysis. Educational level was measured using an ordinal scale ranging from primary education (1) to PhD (8), based on the ISCED guidelines (CBS 2011). Participants were asked to indicate the level of the highest educational degree they have completed or are currently enrolled in.

Age. It has been found that acceptance of Muslims decreases as age increases (Smeekes, Verkuyten, and Poppe 2012). Moreover, in the Netherlands, older people are more often religious than younger people (CBS 2020), which may cause stronger endorsement of a Christian representation of Dutch national identity among older people. Therefore, age was measured by asking participants to state their age in years, so that age could be controlled for as a continuous variable in the analysis.

Gender. Various studies that have been conducted in the European context have demonstrated that men generally hold more negative views of Islam and Muslims than women (see Helbling 2013). Therefore, the respondents were asked to indicate their gender by choosing one of the given options (male, female, non-binary, or other). Subsequently, gender was controlled for in the analysis.

Analysis

SPSS version 28 was used to screen and analyse the data. As participants were not forced to answer every question of the questionnaire, some answers were missing. However, this only concerned three participants: one person did not state their religious affiliation, another person did not provide their education level, and yet another person did not want to disclose their gender. These responses were coded as missing and the data from these respondents could hence still be used for the analysis.

Before the analysis, the assumptions of linearity, homoscedasticity and normality of residuals were checked, and it was concluded that none of these assumptions were violated. Low VIF values indicated that multicollinearity was absent. By graphing boxplots and calculating leverage values, six outliers were found. The main analysis has been conducted with and without outliers. The absence of the outliers did not change the outcome of the main analysis, and therefore, only the results of the analysis that includes the outliers will be reported below. The results of the analysis without the outliers can be found in Appendix 2.

Each scale that is used in this study was meant to measure one underlying factor. In order to test whether this was the case, exploratory factor analysis with oblique rotation was performed for every scale individually. According to the Kaiser criterion, the eigenvalues of these analyses showed that all three scales measured their intended constructs. Subsequently, exploratory factor analysis was performed again for all items at once. This analysis revealed that two items of the scale that was intended to measure perceived ethnocultural incompatibility loaded equally strong on the factor for opposition to Muslim expressive rights. These items were “Muslims cannot be seen as truly Dutch” and “Muslims undermine the original Dutch way of life”. These items were therefore removed from the scale. The new perceived ethnocultural incompatibility scale consisting of three items proved to be sufficiently reliable (Cronbach’s α = .74).

The main analysis was conducted using the PROCESS-macro extension in SPSS. First, a mediation model was tested to investigate whether perceived ethnocultural incompatibility mediated the relationship between the Christian representation and opposition to Muslim expressive rights. Age, education and gender were included as covariates in this model. Second, we tested a moderated mediation model to investigate whether the effect of the Christian representation on perceived ethnocultural incompatibility and opposition to Muslim expressive rights was moderated by levels of Christian religiosity (see Figure 1). Specifically, this model included an interaction between Christian religiosity and Christian representation in order to test if the effects in the model differed between non-religious people, nominal Christians and religious Christians. Again, age, education and gender were added as covariates.

Results

Preliminary analyses

In order to gain a descriptive understanding of the differences between non-religious people, nominal Christians and religious Christians in relation to the key constructs in the model, Table 2 reports the means and standard deviations for the three Christian religiosity groups separately. Table 2 also shows whether the mean scores of the three groups significantly differ from each other on the three main variables. The results show that nominal and religious Christians displayed an equal amount of opposition to Muslim expressive rights. Non-Christians displayed significantly lower opposition to Muslim expressive rights compared to these two groups. Non-Christians also least endorsed a Christian representation of Dutch national identity. Nominal Christians showed more Christian representation endorsement than non-Christians, but less than religious Christians. Lastly, Table 2 also indicates that religious Christians perceived significantly more ethnocultural incompatibility than non-Christian and nominal Christians. The latter two groups did not significantly differ from each other in this regard.

Table 2. Means and SDs by type of Christian religiosity.

	Min	Max	Type of Christian religiosity	M	SD
1. Opposition to Muslim expressive rights	1.00	7.00	Non-religious	2.86b	1.34
			Nominal	3.45a	1.60
			Religious	3.66a	1.63
2. Christian representation of Dutch national identity	1.20	7.00	Non-religious	3.86a	1.09
			Nominal	4.39b	1.12
			Religious	5.58c	0.96
3. Perceived ethnocultural incompatibility	1.00	7.00	Non-religious	2.95a	1.16
			Nominal	3.08a	1.26
			Religious	3.91b	1.42

Note: Comparisons with differing subscripts are significantly different at p < .05 (two-sided).

Overall, respondents scored significantly below the neutral mid-point of the scale measuring opposition to Muslim expressive rights (M = 3.35, SD = 1.56, t(275) = −6.87, p = <.001), meaning that respondents showed relatively low opposition. Respondents scored significantly above the neutral mid-point of the Christian representation scale, which means that respondents in general endorsed a Christian representation of Dutch national identity (M = 4.83, SD = 1.31, t(275) = 10.55, p = <.001). When it comes to perceived ethnocultural incompatibility, respondents scored significantly below the neutral mid-point of the scale (M = 3.49, SD = 1.40, t(275) = −6.07, p = <.001), meaning that on average, respondents did not perceive much ethnocultural incompatibility with Muslims.

Bivariate correlations

Bivariate Pearson correlations (Table 3) indicate that opposition to Muslim expressive rights was significantly positively correlated with perceived ethnocultural incompatibility (r = .652, p < .001), which means that perceiving incompatibility with regards to Dutch and Muslim culture and lifestyle goes together with stronger opposition towards Muslim expressive rights. Moreover, endorsing a Christian representation of Dutch national identity went together with higher levels of opposition to Muslim expressive rights (r = .356, p < .001), and higher levels of perceived ethnocultural incompatibility (r = .344, p < .001). Thus, people who endorsed a Christian representation of Dutch national identity were more likely to experience ethnocultural incompatibility and were on average more strongly opposed to Muslim expressive rights.

Table 3. Means, SD’s and correlations (Pearson) between the variables.

	Min	Max	M	SD	1	2	3
1. Christian representation of Dutch national identity	1.20	7.00	4.83	1.31	–
2. Perceived ethnocultural incompatibility	1.00	7.00	3.49	1.40	.344***	–
3. Opposition to Muslim expressive rights	1.00	7.00	3.35	1.56	.356***	.652***	–
Covariates
4. Age	18	78	34.06	14.78	.339***	.117	.306***
5. Level of education	1	8	5.40	1.63	-.178**	-.125*	-.285***
6. Gender (reference = male)	-	-	-	-	-.206***	-.277***	-.221***

*p < .05. **p < .01. ***p < .001.

Moreover, Table 3 shows that as respondents’ age increased, they endorsed a Christian representation of Dutch national identity more strongly (r = .339, p < .001), and were more strongly opposed to Muslim expressive rights (r = .306, p < .001). Furthermore, both educational level and being female were negatively associated with all three main variables. This means that higher educated respondents in general displayed less opposition to Muslim expressive rights (r = -.285, p < .001), perceived less ethnocultural incompatibility (r = -.125, p = .038), and were less likely to endorse a Christian representation of Dutch national identity (r = -.178, p = .003), compared to lower educated respondents. Similarly, women reported less opposition to Muslim expressive rights (r = -.221, p = <.001), perceived less ethnocultural incompatibility (r = -.277, p < .001), and were less likely to endorse a Christian representation (r = -.206, p < .001), compared to men.

Main analysis

Mediation Model. A mediation analysis was conducted (model 4 of the PROCESS macro), in which endorsement of a Christian representation of Dutch national identity was included as the independent variable, opposition to Muslim expressive rights as the dependent variable, and perceived ethnocultural incompatibility as a mediator of this relationship. The covariates age, education and gender were also included in the model. This analysis (N = 273; see Table 4) revealed a significant total effect of Christian representation endorsement on opposition to Muslim expressive rights (B = .32, p = <.001). The direct effect was not significant (B = .08, p = .231), but the indirect effect was (B = .24, 95 percent CI [.13, .35]). The model explained 23 per cent of the total variance (R² = .23; see Figure 2). According to Mathieu and Taylor (2006), a full mediation is present if both the indirect effect and total effect of the independent on the dependent variable are significant, while the direct effect is close to zero. Therefore, these results indicate that there was a positive effect of Christian representation endorsement on opposition to Muslim expressive rights, and that this effect was fully mediated by perceived ethnocultural incompatibility. This means that the more Dutch natives endorsed a Christian representation of national identity, the more likely they were to oppose Muslim expressive rights, because of stronger perceptions of ethnocultural incompatibility between the Dutch and Muslim ways of life.

Figure 2. Mediation Model Including Age, Education and Gender as Covariates.

Note: N = 273, ***p<.001. Unstandardized coefficients. In parentheses, the direct effect without inclusion of the mediator in the model is shown.

Table 4. Results of the mediation analysis (models with and without covariates).

	With covariates (N = 273)				Without covariates (robustness check; N = 275)
	95% Confidence interval				95% Confidence interval
	B	SE	Lower	Upper	B	SE	Lower	Upper
Effect of CRE on PEI	.19***	.04	.11	.26	.22***	.04	.15	.29
Effect of PEI on OMER	1.28***	.10	1.07	1.49	1.34***	.11	1.13	1.55
Direct effect of CRE on OMER	.08	.07	-.05	.22	.21**	.07	.08	.35
Indirect effect via PEI	.24*	.06	.13	.35	.30*	.06	.19	.41
Total effect of CRE on OMER	.32***	.08	.16	.48	.51***	.08	.35	.67
	R^2 of the total effect model = .23, p = <.001				R^2 of the total effect model = .13, p = <.001

Note: CRE: Christian representation endorsement; PEI: perceived ethnocultural incompatibility; OMER: opposition to Muslim expressive rights. *p < .05, **p < .01, ***p < .001. For the indirect effects, no p-values were obtained, so the asterisk indicates that the confidence interval does not include zero.

Moderated Mediation Model. According to Hayes (2015, 3), a moderated mediation occurs “ … if the proposed moderator variable has a nonzero weight in the function linking the indirect effect of X on Y through M to the moderator”. In simpler terms, this means that the proposed mediation is dependent on the level of a moderator (see Preacher, Rucker, and Hayes 2007). Since we expected that the strength of the effect of Christian representation endorsement on ethnocultural incompatibility (Figure 1, path a) and opposition to Muslim expressive rights (Figure 1, path c’) would be dependent on Christian religiosity, we conducted a moderated mediation analysis using model 8 of the PROCESS macro extension in SPSS. This means that Christian religiosity was added to the mediation model as a moderator of the effects of the Christian representation on perceived ethnocultural incompatibility and opposition to Muslim expressive rights (see Figure 1). Age, education and gender were included as covariates in this model.

The results of the moderated mediation analysis (N = 272; see Table 5) showed that the hypothesized conceptual model (see Figure 1) was not supported by the data. Christian religiosity (with the categories non-religious, nominal Christian (M1) and religious Christian (M2)) did not moderate the relationship between the Christian representation and perceived ethnocultural incompatibility (Figure 1, path a; M1: B = .14, p = .311; M2: B = -.02, p = .812). The R² change of this model was .004, p = .488. Christian religiosity also did not moderate the relationship between the Christian representation and opposition to Muslim expressive rights (Figure 1, path c’; M1: B = -.01, p = .977; M2: B = -.20, p = .235). The R² change of this model was .003, p = .453. Thus, neither of the hypothesized interactions of Christian religiosity was significant.

Table 5. Results of the Moderated Mediation Analysis (hypothesized models with and without covariates).

		Hypothesized model (N = 272)				Hypothesized model without covariates
						(robustness check; N = 274)
			95% Confidence interval				95% Confidence interval
	B	SE	Lower	Upper	B	SE	Lower	Upper
Effect of CRE on PEI	.09	.07	-.05	.23	.12	.08	-.03	.26
Effect of interaction 1 (CRE x nominal Christians) on PEI	.14	.14	-.14	.42	.11	.15	-.18	.40
Effect of interaction 2 (CRE x religious Christians) on PEI	-.02	.10	-.21	.17	.02	.10	-.18	.21
Effect of PEI on OMER	1.36***	.11	1.15	1.57	1.41***	.11	1.19	1.62
Indirect effect via PEI for non-Christians	.13	.09	-.05	.31	.16	.10	-.03	.35
Indirect effect via PEI for nominal Christians	.32	.17	-.07	.63	.32	.17	-.01	.65
Indirect effect via PEI for religious Christians	.09	.12	-.14	.33	.18	.12	-.03	.43
	R^2 change = .004, p = .488				R^2 change = .002, p = .752
Direct effect of CRE on OMER	.29*	.12	.05	.54	.38**	.13	.12	.63
Effect of interaction 1 (CRE x nominal Christians) on OMER	-.01	.24	-.49	.47	-.04	.26	-.55	.47
Effect of interaction 1 (CRE x religious Christians) on OMER	-.20	.17	-.53	.13	-.13	.18	-.48	.22
Direct effect of CRE on OMER for non-Christians	.29*	.12	.05	.54	.38**	.13	.12	.63
Direct effect of CRE on OMER for nominal Christians	.29	.21	-.13	.71	.34	.23	-.11	.78
Direct effect of CRE on OMER for religious Christians	.10	.12	-.14	.33	.25*	.12	.01	.49
	R^2 -change = .003, p = .453				R^2 -change = .001, p = .759
Index of moderated mediation			95% Confidence Interval				95% Confidence Interval
	Index	SE	Lower	Higher	Index	SE	Lower	Upper
Nominal Christians	.20	.19	-.23	.53	.15	.19	-.21	.53
Religious Christians	-.03	.15	-.32	.26	.02	.15	-.26	.33

Note: CRE: Christian representation endorsement; PEI: perceived ethnocultural incompatibility; OMER: opposition to Muslim expressive rights. *p < .05, **p < .01, ***p < .001.

In short, these findings mean that for all three groups the effect of the Christian representation on opposition to Muslim rights works in a similar fashion. That is, the more non-religious, nominal Christians and religious Christians endorse a Christian representation of national identity, the more likely they are to oppose Muslim expressive rights, because of stronger perceptions of ethnocultural incompatibility.

Additional analyses

In order to check the robustness of the results of the mediation and moderated mediation analyses described above, both analyses were repeated without adding the covariates to the model. The general pattern of results did not differ much from that of the main analysis (see Figure 3). The main difference between the two models was that the mediation model without covariates showed a significant direct effect of the Christian representation endorsement on opposition to Muslim expressive rights (B = .21, p = .002), whereas the model which included covariates did not show this direct effect. In both models (with and without covariates), the indirect effect is significant, which means that the effect of Christian representation endorsement on opposition to Muslim expressive rights is mediated by perceived ethnocultural incompatibility. Thus, the finding that perceived ethnocultural incompatibility mediates the relationship between Christian representation endorsement and opposition to Muslim expressive rights can be seen as robust. However, the fact that the direct effect of the Christian representation on opposition remained significant in the mediation model without covariates means that we observed partial mediation instead of full mediation (Mathieu and Taylor 2006).

Figure 3. Mediation Model Without Covariates.

Note: N = 275, **p <.01 ***p <.001. Unstandardized coefficients. In parentheses, the direct effect without inclusion of the mediator in the model is shown.

The moderated-mediation analysis without the covariates of age, education and gender (N = 274) did not yield results that differed from the analysis which included the covariates. Christian religiosity neither moderated the relationship between Christian representation endorsement and perceived ethnocultural incompatibility (Figure 1, path a; M1: B = .11, p = .746; M2: B = .02, p = .874; R² change = .002, p = .752) nor the relationship between Christian representation endorsement and opposition to Muslim expressive rights (Figure 1, path c’; M1: B = -.04, p = .873; M2: B = -.13, p = .464; R² change = .001, p = .759). These results show that the conclusions of the main analysis also hold when the covariates are removed from the model. Therefore, it can be concluded that the results of the main analysis are robust.

Discussion and conclusion

While Christian religiosity has drastically declined in Western societies as a result of secularization, many native majority members still view Christian religion and Christian cultural heritage as essential parts of their national identity (Straughn and Feld 2010; Trittler 2017). Populist radical-right parties in these contexts often strategically evoke a Christian representation of national identity to highlight incompatibility between national and Muslim ways of life and to argue for stricter policies on the religious expression of Muslims, such as the building of mosques and Islamic schools.

The aim of this study was twofold. First, we wanted investigate to what extent feelings of ethnocultural incompatibility can explain the relationship between endorsing a Christian representation of Dutch national identity and opposition to Muslim expressive rights. While previous research has looked at the relationship between endorsing a Christian representation and opposition to Muslim expressive rights (e.g, Badea, Bender, and Korda 2020; Smeekes and Verkuyten 2014), it is unknown whether perceived incompatibility between the national and Muslim ways of life can explain this relationship. Second, this study aimed to clarify whether the relationships between the Christian representation and perceived ethnocultural incompatibility and opposition to Muslim rights are dependent on individuals’ level of Christian religiosity. Based on previous research by Storm (2011) and Demerath (2000), we distinguished between non-religious people, nominal Christians (i.e. people who identify as Christian but who indicate that their religion is not very important to them) and religious Christians (i.e. people who identify as Christian and indicate that their religion is important to them). We investigated these aims using an online survey among a broad sample of native Dutch majority members.

Our results revealed that there was a positive relationship between endorsing the Christian representation and opposition to Muslim expressive rights and that this relationship was fully mediated by perceived ethnocultural incompatibility. This finding extends previous work which observed a positive relationship between endorsing a Christian representation of national identity and opposition to Muslim expressive rights (Badea, Bender, and Korda 2020; Smeekes and Verkuyten 2014). Our findings show that stronger perceptions of ethnocultural incompatibility between the Dutch and Muslim way of life form an explanation for why endorsing a Christian representation is related to more opposition to Muslim expressive rights.

However, the findings revealed that the effects of Christian representation endorsement were not dependent on the level of Christian religiosity. In other words, the positive relationship between endorsing a Christian representation and opposition to Muslim expressive rights via stronger perceptions of ethnocultural incompatibility were present for nominal Christians, religious Christians and non-Christians. This latter finding goes against our prediction that the proposed positive links between the Christian representation and perceived ethnocultural incompatibility and opposition to Muslim rights would only be present for nominal Christians but not for religious Christians. We based this prediction on previous work by Storm (2011) which showed that nominal Christians have more exclusionary understandings of their national identity compared to religious Christians and non-Christians.

It is possible that the inability to find support for the moderating role of Christian religiosity is caused by a lack of power, given the small sample of nominal Christians (N = 31) in this study. Keeping this limitation in mind, comparison of the means nevertheless revealed some interesting differences between the groups. First, we observed that religious Christians on average perceived more ethnocultural incompatibility than nominal and non-Christians, and that compared to non-Christians, both nominal and religious Christians expressed stronger (and similar levels of) opposition to Muslim expressive rights. On the one hand, these findings are in contrast with studies showing that religious Christians are more supportive of Muslim expressive rights compared to non-Christians (Carol, Helbling, and Michalowski 2015; Eskelinen et al. 2022), based on the idea that both Muslims and Christians share a common identity of “being religious”. Instead, our results align with theory and research on group threat, which proposes that stronger group identification makes people more likely to perceive out-groups as threatening, which, in turn, fuels out-group negativity (Stephan 2014).

On the other hand, these findings are in line with the work of Storm (2011) and Demerath (2000), because they show that people who identify as Christian but are hardly religious (i.e. nominal Christians) are more opposed to Muslim expressive rights than non-Christians. This means that viewing Christian religion as something one belongs to, rather than something one believes in, can also result in stronger opposition to the presence of Muslims. Taken together, these findings highlight the relevance of distinguishing between different ways of affiliating with Christian religion for understanding its relationship with attitudes towards religious out-groups.

One alternative explanation for not finding support for our prediction is that we did not distinguish between Catholic and Protestant Christians. Protestant Christians are most likely overrepresented in the sample, considering the fact that the personal network of the first author mostly consists of Protestant Christians. A report by De Hart and Van Houwelingen (2018) shows that when Catholics cease to be religiously active, they will still hold onto their Catholic identity – a phenomenon that is much less visible among Protestants. This indicates that the idea that the Netherlands is and should remain a Christian nation resonates more with nominal Catholics. Some evidence for this can be found by looking at recent election outcomes. In the (predominantly) Catholic south of the Netherlands, political parties that emphasize the importance of protecting the “Dutch Judeo-Christian culture” are electorally more successful than elsewhere in the Netherlands (Kiesraad n.d.). Therefore, there is reason to assume that both Christian representation endorsement and perceived ethnocultural incompatibility with regards to Muslims will be more strongly present among nominal Christians with a Catholic background compared to nominal Christians with a Protestant background. As the sample most likely includes very little Catholic Christians, it is possible that the presence of ethnocultural incompatibility among nominal Christians in the Netherlands is underestimated in this study. For this reason, future studies could try to investigate our hypothesis using a sample that is broader with respect to Christian denominational affiliation. Another related suggestion for future research is to investigate whether our theoretical model replicates when using somewhat different operationalizations of Christian religiosity and exclusionary attitudes towards Muslims. For example, studies could try to more directly measure whether people identify as nominal or religious Christians and/or focus on measures of prejudice or social distance towards Muslims.

Limitations

This study is subject to several limitations. First of all, because of limited time and means, this study used a convenience sample, which in this case means that respondents were recruited via the personal network of the first author. Therefore, the sample used in this study is not representative of the general population of native Dutch adults. As the descriptive analysis shows, the sample is relatively young and relatively highly educated. Because younger and higher educated people generally show less opposition to Muslim expressive rights compared to older and lower educated people (Smeekes, Verkuyten, and Poppe 2012), the amount of opposition to Muslim expressive rights that is present among native Dutch adults is most likely underestimated in this study, which limits the generalizability of the results to the broader population. In addition, the order in which the main concepts were addressed in the survey may have caused order effects. Since the respondents were confronted with questions about Christianity first, this may have caused their Christian identity and Christian national heritage to be salient and this may have affected their responses to the questions about perceived ethnocultural incompatibility and Muslim expressive rights. Future research could test these predictions among a representative sample of the native Dutch adult population and could use question randomization to limit order effects.

Conclusion

In sum, the present research provides evidence that among native Dutch majority members, endorsement of a Christian representation of Dutch national identity was positively related to opposition to Muslim expressive rights. Furthermore, this study is the first to show that endorsing the view that Muslim culture is incompatible with Dutch culture can explain this relationship, and that this explanation holds irrespective of whether the individual is non-religious, a nominal Christian, or a religious Christian. This is a relevant finding, because it helps us to understand why populist radical-right politicians who use the Christian representation of national identity in their rhetoric are successful in mobilizing broad segments of society against Muslims and Islam. In addition, the finding that both religious and nominal Christians show stronger opposition to Muslim expressive rights than non-Christians shows that rather than mere religiousness, identification as a Christian is associated with stronger opposition to Muslims. This finding illustrates that different ways of affiliating with Christianity may affect attitudes towards Muslims, which implies that researchers should take both “believing” and “belonging” into account in investigating intergroup and/or interreligious relationships.

Disclosure statement

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

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Appendices

Appendix 1: Items Used to Measure Key Constructs

Opposition to Muslim Expressive Rights

• In the Netherlands, Muslims should have the right to show and practice their faith in public.

• In the Netherlands, Muslim women should have the right to wear a headscarf anywhere they want.

• In the Netherlands, Muslims should have the right to not only celebrate their Islamic holidays at home, but also in public.

• In the Netherlands, Muslims should have the right to establish Islamic schools.

• In the Netherlands, Muslims should have the right to build mosques.

• On Dutch television, there should be a place for programmes by and for Muslims.

Endorsement of a Christian Representation of Dutch National Identity

• The Dutch have passed on their Christian traditions to next generations.

• Christian norms and values are an important part of the Dutch identity.

• Christian norms and values have always been important in Dutch society.

• The Dutch identity is rooted in Christian traditions.

• The Netherlands originally is a Christian country and should remain one.

Perceived Ethnocultural Incompatibility

• The Islamic way of life is incompatible with the Dutch way of life.

• Muslims and the Dutch are like a jigsaw puzzle. They may differ, but they fit together well.

• The fact that the Dutch and Muslim ways of life differ does not mean that they are necessarily in opposition.

• Muslims cannot be seen as truly Dutch.*

• Muslims in the Netherlands undermine the traditional Dutch way of life.*

*This item was eventually removed from the scale based on the factor loadings of the exploratory factor analysis.

Christian Religiosity

• Which religion do you feel most connected to?

• How important is your religion to you?

Appendix 2: Results of the analysis without outliers

Main Analysis Without Outliers

Preliminary analysis revealed that there were six outliers present in the data. The main analysis has been repeated without the outliers in order to see if removal of the outliers influences the results.

Mediation Model. The main effects model analysis was repeated, and this analysis (N = 267) revealed a significant total effect of Christian representation endorsement on opposition to Muslim expressive rights (β = .24, p = <.001). The direct effect was not significant (β = .07, p = .242), but the indirect effect was (β = .17, 95 per cent CI [.04, .09]). The model explains 22 per cent of the total variance (R² = .22). Similarly to the main analysis which included outliers, these results indicate that there is an effect of Christian representation endorsement on opposition to Muslim expressive rights, and that this effect is completely mediated by perceived ethnocultural incompatibility.

Moderated Mediation Model. Subsequently, the hypothesized model was tested after removing the outliers from the dataset. A moderated mediation analysis was conducted using model 8 of the PROCESS-macro extension in SPSS. The results of this analysis (N = 266) showed that the hypothesized conceptual model (see Figure 1) was not supported by the data. Christian religiosity (with the categories non-religious, nominal Christian (M1) and religious Christian (M2)) did not moderate the relationship between Christian representation endorsement and perceived ethnocultural incompatibility (Figure 1, path a; M1: B = .02, p = .951; M2: B = -.08, p = .628). The R² change of this model was .001, p = .869. Christian religiosity also did not moderate the relationship between Christian representation endorsement and opposition to Muslim expressive rights (Figure 1; path c’; M1: B = -.13, p = .593; M2: B = -.17, p = .229). The R² change of this model was .003, p = .477. Thus, similarly to the analysis that included outliers, neither of the hypothesized interactions of Christian religiosity was significant.

In sum, it can be concluded that removing the outliers did not lead to different results.