Tolerance of Muslim minority identity enactment: The roles of social context, type of action and cultural continuity concern.

There have been strong debates in many European countries about religious identity enactment of Muslims, with the wearing of the headscarf in public places being a central symbolic topic. This study investigated the importance of the context (private versus three public contexts) for tolerance of Muslim identity enactment (e.g., the wearing of headscarves) among a national sample of Dutch majority group members. Using an experimental design, it was found that tolerance was highest in the private context, followed by the street context and then the contexts of work. Furthermore and in all contexts, tolerance of Muslims persuading others to start enacting their religious identity in a similar way was lower than tolerance of identity enactment itself. Moreover, both types of tolerance were found to differ by context only for majority members who were highly concerned about the continuity of their ingroup's cultural identity (i.e., cultural continuity). It is concluded that context-related and action-related variance, as well as cultural continuity, are important for majority members' tolerance of Muslim minority identity enactment.

In many European countries, societal debates evolve around various forms of Muslim identity enactment, such as the wearing of religious clothing that is often strongly connected with one's sense of identity. Muslim minorities face various social constraints that may compromise their ability to act upon their religious belief. For example, several countries have discussed, or adopted, legislative measures to ban teachers and civil servants from wearing a headscarf (Human Rights Watch, 2009). Furthermore, a majority of the Dutch, German and French population has been found to favour a ban on headscarves in public places (Pew Research Center, 2005; Van der Noll, 2010).

This negative attitude towards Muslim identity enactment is often considered to result from negative feelings towards Muslims (Van der Noll, 2014). However, people can have various other reasons for opposing, for example, the wearing of headscarves (e.g., secular beliefs), and their opposition is likely to depend on the situation (e.g., more public or more private contexts). Starting with the seminal work of Stouffer (1955), research on tolerance of the enactment of dissenting beliefs indicates that context‐related variance in tolerance is common, as it involves a situational convergence of various considerations (e.g., political orientation, situationally salient values). The classical conceptualization of tolerance implies the acceptance of practices and beliefs that one disapproves of, and depending on the situation different considerations for showing forbearance can be relevant (Verkuyten & Yogeeswaran, 2017). People are likely to allow the enactment of minority religious identities in some circumstances for reasons of religious freedom, while disallowing the same enactment in other contexts for secular reasons (Stouffer, 1955; Verkuyten & Slooter, 2007). For instance, people may accept the wearing of the headscarf in general but not in public institutions (e.g., at work as a civil servant).

Furthermore, context‐related variance may be especially likely for majority members who are concerned about the continuity of their ingroup's cultural identity, as wearing a headscarf in public places, for example, may be perceived as undermining this identity. Yet, although various studies have investigated tolerance of the headscarf (e.g., Helbling, 2014; Van der Noll, 2010), to our knowledge, no empirical research has considered tolerance across several public as well as private contexts, and whether the impact of the context on majority group members' tolerance of Muslim identity enactment depends on their cultural continuity concern.

The central theoretical proposition that we test in this study is that tolerance is lower when minority identity enactment is considered to have more negative sociocultural consequences (Capelos & Van Troost, 2012; Chanley, 1994). Using a national sample of Dutch majority members and an experimental design, we tested this proposition in three ways. First, we investigated tolerance of Muslim women enacting their religious identity, such as wearing a headscarf, in several contexts, reasoning that Muslim identity enactment has more negative sociocultural consequences in public contexts (street, work) as compared to private contexts. Second, we considered both people's tolerance of Muslims enacting their religious identity and tolerance of persuading others to enact their religious identity in a similar way. Trying to persuade co‐believers to also wear a headscarf might be perceived as having more negative sociocultural consequences and might therefore elicit lower tolerance compared to the wearing itself. Third, we examined whether the effect of context on tolerance of identity enactment is especially strong for majority group members who are concerned about the continuity of their group's cultural identity.

CONTEXT OF RELIGIOUS IDENTITY ENACTMENT

Empirical research on attitudes towards (Muslim) minority identity enactment has investigated these attitudes in general (Van der Noll, 2014), or in a specified social context such as in school (Helbling, 2014), in politics (Simon et al., 2018) or at work (Van der Noll, Rohmann, & Saroglou, 2018). The findings of these studies are difficult to compare because there tend to be differences between tolerance in a general sense and tolerance in specific contexts, as well as between various contexts (Chanley, 1994). To our knowledge, no research has systematically examined situation‐related variance in tolerance of Muslim identity enactment across private and several public contexts. Yet, such variance would indicate that people take the context into account and do not base their tolerance of Muslim identity enactments only on, for example, anti‐Muslim sentiments or principled considerations (e.g., secular beliefs).

One important reason to expect that the context matters for people's tolerance is that religious enactment in the private sphere, compared to various public contexts, has much less sociocultural consequences and therefore is easier to tolerate. Research has shown that people are less tolerant when a particular practice is considered to have negative implications for themselves, their group or society more generally (Bannister & Kearns, 2009; Chanley, 1994). For instance, people may fear that religious minority enactment in public contexts may have negative consequences for social cohesion, invokes societal and cultural change or threatens the secular nature of the state and its public institutions. Thus, identity enactment in public contexts can be expected to be tolerated less than in the private sphere (e.g., home environment, with family members).

However, there are different sorts of public contexts, and we therefore additionally examined tolerance of Muslim identity enactment in three public contexts: on the street, in a general work context and working as a civil servant. First, people are likely to be less tolerant of minority expression in the open space of the street than in private contexts, as private expressions are of little consequence for society at large, and people generally feel that one should not interfere in the private domain (Slater, 1998). Religious expression in the street is a visible form of minority identity enactment, and research in the United States has shown that minority groups are construed as less American when expressing their minority identity publicly versus privately (Yogeeswaran, Dasgupta, Adelman, Eccleston, & Parker, 2011). Therefore, we expected to find lower tolerance for religious identity enactment in the street context than in the private sphere (Hypothesis 1a).

Second, we examined the difference between the street context versus the two work contexts (i.e., the general work context and working as a civil servant). Although the wearing of a headscarf is visible in the street, the consequences for majority members are less direct and obvious compared to the work context or in social interactions with civil servants. People may perceive the latter two settings as having to be ‘colour‐blind’ situations in which minority identity enactment is less appropriate, and research has found that such a perception can lead to lower minority group acceptance (Dovidio, Gaertner, & Saguy, 2015). Thus, it is expected that tolerance for religious identity enactment in both work contexts is lower than in the street context (Hypothesis 1b).

Lastly, we compared the level of tolerance in the work context and as a civil servant. The latter implies a role as state representative for which religious identity enactment can be considered as going against the principle of state neutrality. For example, employees in Dutch companies are generally allowed to wear a headscarf, while public officials at the court or police officers are not allowed to wear headscarves for reasons of state neutrality (Saharso & Lettinga, 2008). In line with the principle of state neutrality, we expected tolerance for religious identity enactment as a civil servant to be lower than in the work context (Hypothesis 1c).

TOLERANCE OF ENACTMENT AND OF PERSUASION

In addition to context‐related variance in tolerance, there might be act‐related variance: variation in what people are asked to tolerate. Research has shown that tolerance levels differ for dissenting beliefs, the behavioural expression of these beliefs and the mobilization of others to engage in the same practice (Gieling, Thijs, & Verkuyten, 2010). For example, majority members in the Netherlands were found to be more tolerant of Muslims involved in practices such as the wearing of religious clothing and the refusal to shake hands with someone of the opposite sex (i.e., tolerance of enactment), than of Muslims trying to persuade other Muslims to engage in these practices (i.e., tolerance of persuasion) (Gieling et al., 2010).

Individual religious expression and persuading others to act in a similar way have different sociocultural implications. It is one thing to tolerate religious identity enactment, but another to accept that others are persuaded to express their identity in a similar way. Research has shown that the perception of minority group size is related to perceived threat (McLaren, 2003; Outten, Schmitt, Miller, & Garcia, 2011) and that ‘threat in numbers’ predicts negative attitudes towards minorities (Earle & Hodson, 2019). As such, trying to persuade others implies mobilizing fellow Muslims to engage in these practices, which majority group members may consider as threatening their identity and culture. Therefore, we expected tolerance of persuasion to be lower than tolerance of identity enactment (Hypothesis 2). This is particularly likely in public contexts but might also extend to the private sphere because the mobilization of others to, for example, also start wearing a headscarf may be perceived as increasing the number of practicing Muslims in society.

CULTURAL CONTINUITY CONCERN

Apart from context‐based and act‐based variance in tolerance, tolerance might also depend on the extent to which majority members perceive Muslim minority identity enactment as undermining the continuity of their normative way of life. A sense of continuity has been conceptualized as an important identity need (Vignoles, 2011), and people will more strongly strive for cultural identity continuity when they perceive that it is challenged or threatened (see Smeekes & Verkuyten, 2015). Research has demonstrated that concern for cultural continuity can have negative implications for outgroup attitudes (e.g., Jetten & Wohl, 2012; Smeekes & Verkuyten, 2015) and for the acceptance of Muslim minority practices (Velthuis, Verkuyten, & Smeekes, 2020). In contrast, majority members who are little concerned about the continuation of their normative way of life tend to support cultural diversity and societal change, with the related recognition and acceptance of minority identities in various domains of life (Verkuyten, Yogeeswaran, Mepham, & Sprong, 2020). Thus, we expected that stronger cultural continuity concern is related to lower tolerance (Hypothesis 3a) and, more importantly, that the predicted context‐related differences in tolerance are especially likely for majority members who emphasize the importance of ingroup cultural continuity (Hypothesis 3b). Participants with higher continuity concern are more likely to consider situational differences in sociocultural consequences of religious identity enactment, and therefore, the situation in which it is enacted is likely to matter more for their tolerance judgements (both persuasion and enactment tolerance).

In testing the different predictions, we additionally examined tolerance among Dutch majority members who are religiously affiliated and those who are not. The reason is that religiously affiliated people, compared to non‐religious people, may for example consider the specific context less relevant for the ability of religious groups to act upon their religious beliefs (e.g., Sleijpen, Verkuyten, & Adelman, 2020).

METHOD

Data and participants

Initially, 850 respondents completed an online survey in 2019 after being drawn from a gross sample of 1,640 panel members from the GfK/Ipsos panel of over 80.000 Dutch citizens. The response rate was 52%, which is common in the Netherlands (Stoop, Billiet, Koch, & Fitzgerald, 2010). All respondents participated with consent, and the data collection was approved by the ethics committee of the Faculty of Social and Behavioral Sciences at Utrecht University (FETC18‐063). Eighteen respondents were removed by research agency GfK/Ipsos to assure data quality. We additionally excluded five participants who indicated that both of their parents were not Dutch, and one participant who indicated they were Muslim, resulting in N = 826.

In total, 51% of the respondents was female. The mean age was 55 years (M = 54.89, SD = 16.23, range 18–88 years), and slightly less than half of the sample (43%) was religiously affiliated. In total, 26% had obtained lower levels of education (primary school or lower secondary education), 27% average levels (lower tertiary or higher secondary education) and 47% had obtained higher levels of education (higher tertiary education). The sample closely matches the general Dutch population in terms of demographic characteristics but with a slight overrepresentation of older people and higher educated.

Design and measures

A between‐subjects experimental design with four randomly assigned conditions was used to manipulate the context in which the identity enactment occurred. Vignettes with concrete and realistic situations were used in order to increase the ecological validity of the experiment (Steiner, Atzmüller, & Su, 2016). Based on previous research in the Netherlands (e.g., Hindriks, Verkuyten, & Coenders, 2017), the vignettes introduced a fictitious interview about ‘being Muslim in the Netherlands’ that was ‘recently published in a well‐known newspaper’. The alleged interview was with a thirty‐year‐old Muslim woman called Fatma who was born in The Netherlands. In the interview, Fatma was first asked whether she finds it important that Muslims in the Netherlands can enact their religious beliefs. In the three public contexts and after her affirmative answer (‘yes of course’), she was asked for an explanation (‘why?’). For making religious identity enactment salient, she explained her answer by stating, ‘because your religion is who you are, it is your identity which you should always be able to show, like with a headscarf and in your behavior’. The headscarf was mentioned because of being often perceived as emblematic of Muslim identity. Subsequently and for manipulating the three public contexts, she gave an affirmative answer after being asked ‘so for example also if one is a civil servant at the municipality or as a police officer?’ (civil servant condition, n = 202); ‘so for example also at work?’ (work context condition, n = 207); ‘so also when one, for instance, goes shopping?’ (street condition, n = 208). In the private context condition (n = 209), Fatma answered that she ‘does not really’ find it important that Muslims in the Netherlands publicly enact their religious beliefs because ‘one's religious belief is something private, that you experience when you are with your family or pray to God. You do not need to show that everywhere to other people’ (see full text of vignettes in Appendix A in Data S1).

Manipulation check. The four contexts were expected to vary in the degree to which they evoke negative feelings because of the perceived sociocultural consequences. Therefore and following previous research (Adelman, Verkuyten, & Yogeeswaran, 2021; Sleijpen et al., 2020), we asked participants to indicate on a 7‐point scale from very negative (1) to very positive (7) how they felt towards Muslims like Fatma. Results of an analysis of variance (ANOVA) demonstrated that there were indeed significant differences in feelings, F(3,822) = 30.25, p < .001, ηp  = .10, with a pattern of increasingly more negative feelings from the private context to working as a civil servant (see Table A1 in Appendix B in Data S1).

Tolerance of identity enactment was assessed with two items on 7‐point scales. The first item explicitly considered acceptance in spite of a negative attitude: ‘To what extent do you think the way of life of Muslims like Fatma should be accepted, despite one being negative about it?’. The second item was: ‘Do you think it is OK that Muslims like Fatma enact their religious identity this way?’. The latter was recoded in such a way that a higher score represented higher tolerance (r = .48 between the two items). Tolerance of persuasion was also measured with two items (7‐point scales; Verkuyten & Slooter, 2007) and counterbalanced with the other two items (see Appendix C in Data S1): ‘Do you think it is OK if Muslims like Fatma try to persuade other Muslims to engage with their religion in the same way?’ and ‘Do you think it is OK if Muslims like Fatma organize religious assemblies to spread their views?’. Both had an answer scale from 1 (totally OK) to 7 (totally not OK) and were recoded so a higher score meant higher tolerance levels of the mobilization of others (r = .56).

Confirmatory factor analysis (CFA) was performed in Mplus (version 7.3) to empirically examine whether enactment and persuasion tolerance were empirically separate constructs. Results demonstrated that the proposed two‐factor structure had an acceptable fit to the data, χ 2 (1) = 18.06, p < .001; CFI = 0.98; TLI = 0.88; RMSEA = 0.15 [0.09–0.21]; SRMR = 0.02, with standardized factor loadings above .64 (see Kline, 2016). Importantly, an alternative one‐factor model had a worse fit, Δχ 2 (1) = 25.37, p < .001, indicating empirical support for examining enactment and persuasion tolerance as two separate dimensions. However, since it may be argued that RMSEA is relatively high and TLI relatively low (Hu & Bentler, 1999) – which is common in models with low degrees of freedom (Kenny, Kanisan, & McCoach, 2014) – we also performed additional robustness checks with single‐item measures in assessing the context‐effects (see Table A2 in Appendix D in Data S1).

Cultural continuity concern was measured with three items (7‐point scales), all starting with the introductory sentence ‘immigrants in the Netherlands can live as they wish’ (Velthuis et al., 2020). This was followed by three items that emphasized the importance of national cultural identity: ‘…as long as Dutch culture is preserved’, ‘if Dutch traditions continue to exist’ and ‘as long as Dutch identity is not undermined’. The three items were combined into a reliable scale (α = .91), with a higher score indicating a stronger concern about the majority's cultural continuity.

Additionally, some variables were measured which we subsequently used to check whether the randomization of the experimental conditions was successful. We examined whether there were differences between the experimental conditions in terms of gender, age, level of education (seven categories comparable to the international ISCED measure), religious affiliation (0 = no, 1 = yes), political orientation (7‐point self‐placement question) and national identification (10‐point single‐item measure, Postmes, Haslam, & Jans, 2013). The randomization was successful because there were no significant differences (p  > .515) across the experimental conditions for these variables, and also not for cultural continuity concern.

Analytic strategy

Tolerance of enactment and of persuasion were investigated as multiple dependent variables using a one‐way multivariate analysis of variance (MANOVA) in SPSS (version 24.0), with context as a between‐subjects factor. There were no missing values on the constructs of interest, and assumptions for multivariate analyses of variance were met.

The multivariate analysis was followed up with univariate analyses of variance with planned contrasts to test our hypotheses H1a–H1c. A first contrast compared the private to the street context; a second contrast compared the street with both work contexts together and the last contrast involved the work versus the civil servant condition. Since the contrasts were three correlated comparisons, we applied the Bonferroni correction in order to have control over the (cumulative) Type I error, using a more conservative α of .017 (α divided by the number of comparisons, see Field, 2009). Moreover, since the contrasts tested directional hypotheses, one‐tailed p‐values were considered (Field, 2009).

RESULTS

Descriptive findings

Descriptive statistics for the main variables are shown in Table 1. As indicated by the overall mean scores across conditions, participants were relatively tolerant of identity enactment (significantly above the neutral midpoint of the scale, t[825] = 8.58, p < .001, d = 0.31), while they were not so tolerant of trying to persuade others to engage in similar identity enactment (significantly below the neutral midpoint of the scale, t[825] = −10.53, p < .001, d = 0.37). Moreover, cultural continuity concern was endorsed relatively strongly (significantly above the neutral midpoint, t[825] = 40.86, p < .001, d = 1.42). The two types of tolerance were positively associated and negatively related to cultural continuity concern.

TABLE 1

Means, standard deviations and correlations of main variables (N = 826)

	1	2	M (SD)
1. Enactment tolerance	—	—	4.42 (1.39)
2. Persuasion tolerance	.57***	—	3.46 (1.47)
3. Cultural continuity	−.23***	−.16***	5.73 (1.22)

Note: *** p < .001. All scales range 1–‐7.

Context and tolerance

Figure 1 displays the means for both types of tolerance per experimental condition. Findings of the MANOVA demonstrated that there was a significant multivariate effect of context on enactment tolerance and persuasion tolerance, Pillai's V = .10, F(6, 1,644) = 14.91, p < .001, η 2 = .05. Separate univariate analyses of variance revealed significant effects (p s < .001) of context on enactment tolerance as well as on persuasion tolerance, with similar, medium effect sizes (see Table 2).

FIGURE 1

Means of enactment and persuasion tolerance per context

TABLE 2

Means, standard deviations and results of analyses of variance for enactment and persuasion tolerance of the four contexts (N = 826).

	1. Civil servant	2. Work	3. Street	4. Private	F	η p 2
	M (SD)	M (SD)	M (SD)	M (SD)
Enactment tolerance	4.03 (1.35)	4.16 (1.37)	4.36 (1.35)	5.10 (1.26)	26.52***	.09
Persuasion tolerance	3.21 (1.39)	3.13 (1.38)	3.38 (1.45)	4.10 (1.48)	20.07***	.07

Note: *** p < .001.. All scales range 1–7. Multivariate analysis of variance demonstrated a significant effect of context on enactment and persuasion tolerance, Pillai's V = .10, F(6, 1,644) = 14.91, p < .001, η 2 = .05.

Planned contrasts showed, as expected (H1a), that tolerance was significantly lower for religious expression in the street than in the private context, for both enactment (t[822] = −5.62, p < .001, d = 0.56), as well as persuasion tolerance (t[822] = −5.17, p < .001, d = 0.49). Also as expected (H1b), religious expression at work and as a civil servant (average of both conditions M = 4.09) elicited significantly lower enactment tolerance as compared to the street context (t[822] = −2.37, p = .009, d = 0.20), but only marginally for persuasion tolerance (average of both work conditions M = 3.17), t(822) = −1.73, p = .043, d = 0.14. Lastly (H1c), religious expression as a civil servant did not elicit lower tolerance than in the more general work context, neither for identity enactment (t(822) = −1.06, p = .145, d = 0.10) nor for persuasion tolerance (t(822) = 0.57, p = .285, d = 0.06).

We performed a robustness check with religious affiliation as additional factor (in a two‐way MANOVA), which yielded similar results to the main analyses (see Table A3 in Appendix E in Data S1). Thus, the effect of context on tolerance of religious expression was not different for religiously affiliated and non‐religiously affiliated majority group members.

Differences between two types of tolerance

In order to assess H2 about the mean difference between the two types of tolerance, paired‐samples t‐tests were performed, taking into account that both forms of tolerance were measured within the same individual (Field, 2009). As expected, the findings showed that overall, tolerance of persuasion was significantly lower than tolerance of enactment, t(825) = 20.67, p < .001, r = .58 (see Table 1). This was found in all four experimental conditions: as a civil servant (t[201] = 9.59, p < .001, r = .56), in the work context (t[206] = 10.63, p < .001, r = .60), the street (t[207] = 10.65, p < .001, r = .59) as well as the private context (t[208] = 10.47, p < .001, r = .59). Furthermore, results of an ANOVA showed that there were no significant differences in the difference between enactment and persuasion tolerance (Δ enactment‐persuasion) between the four contexts, F(3, 822) = 1.09, p = .351. Thus, persuasion tolerance was lower than enactment tolerance in all four contexts.

Cultural continuity concern

Multivariate analysis of covariance (MANCOVA) was used to investigate H3a and H3b about the direct and moderating role of cultural continuity concern (mean‐centered variable). Results demonstrated that cultural continuity concern was significantly and negatively associated with enactment and persuasion tolerance, Pillai's V = .06, F(2, 820) = 23.69, p < .001, ηp 2 = .06. Separate univariate analyses revealed significant negative relations between continuity concern and enactment tolerance, F(1, 821) = 45.89, p < .001, ηp 2 = .05, as well as with persuasion tolerance, F(1, 821) = 21.12, p < .001, ηp 2 = .03. In line with H3a, majority members who were more concerned about ingroup cultural continuity were less tolerant of Muslim women enacting their religious identity and of persuading others to engage in similar practices.

Importantly, there also was a significant multivariate interaction effect between context and continuity concern on tolerance, Pillai's V = .03, F(6, 1,636) = 3.56, p = .002, ηp 2 = .01. Separate univariate analyses revealed significant interaction effects on both enactment tolerance, F(3, 818) = 4.42, p = .004, η 2 = .02 and persuasion tolerance, F(3, 818) = 4.12, p = .006, η 2 = .02. Thus, as expected, the effect of context on tolerance depended on individual differences in concerns about cultural continuity.

Follow‐up analyses at high continuity concern (at +1 SD above the mean) and low continuity concern (at −1 SD; Aiken, West, & Reno, 1991) demonstrated that the effect of context on enactment tolerance was only significant for high continuity concern, F(3, 198) = 8.84, p < .001, η 2 = .12 (medium‐sized effect), and not for low continuity concern, F(3, 120) = 1.91, p = .132. Similarly, the effect of context on persuasion tolerance was significant for high continuity concern, F(3, 198) = 9.55, p < .001, η 2 = .13 (medium‐sized effect), whereas it was not significant for low continuity concern, F(3, 120) = 0.47, p = .706. Specifically, when people were highly concerned about ingroup cultural continuity, the pattern of tolerance differences between the contexts was similar to the findings for the whole sample, but with lower tolerance for enactment across all contexts (M civil = 3.61, M work = 3.63, M street = 4.26, M private = 4.79) and for persuasion across the three public contexts (M civil = 2.96, M work = 2.75, M street = 3.23, M private = 4.17). Participants who were not much concerned about cultural continuity were more tolerant overall and for them, no significant context‐differences in tolerance were found.

DISCUSSION

Increasing cultural diversity in western societies has made the question of tolerance of minority practices relevant and urgent. In particular the accommodation and acceptance of Muslim identity enactment, such as the wearing of a headscarf, in public and in work‐related settings, has become a much debated issue in many European countries (Human Rights Watch, 2009). The aim of the current research was to examine the context‐related and act‐related variance in majority members' tolerance of Muslims enacting their religious identity, and whether situational variance in tolerance depends on people's concerns about the continuity of the majority's normative way of life. Our overall reasoning was that majority members are less tolerant if Muslim identity enactment is perceived to have a more negative sociocultural impact. We examined the related proposition in three ways: a comparison between four contexts, a focus on two types of tolerance and by considering individual differences in concern about ingroup cultural continuity.

First, using an experimental design, we found that majority members' tolerance of Muslim identity enactment was context‐specific. This indicates that tolerance of Muslim minority practices is not only based on, for instance, general prejudice or secular beliefs. The context matters for tolerance judgements, with lower tolerance in contexts that have stronger sociocultural implications, such as on the street compared to private settings (e.g., home environment) and at work compared to the street context. However, tolerance of identity enactment at work was not significantly different from working as a civil servant. This was not in line with our expectations, but a similar finding was found in another study in the Netherlands (Sleijpen et al., 2020). One explanation is that practices such as wearing a headscarf was not seen as having different sociocultural implications in these two contexts, which corresponds to the finding that the manipulation check question did not elicit a clearly different response between these two experimental conditions (see Data S1). The two work contexts might not be distinct enough and trigger similar feelings about the sociocultural impact of Muslims enacting their religious identity working with colleagues or serving clients. As tolerance was lowest in the work contexts, one implication is that primarily the work context is an area to focus on when trying to improve negative attitudes towards Muslim minority identity enactment.

Second, the findings show that majority members were more tolerant of the identity enactment (e.g., wearing of a headscarf) than of trying to persuade other people to also express their religion in this way (e.g., also start wearing a headscarf). This finding corresponds with the reasoning that the mobilization of other Muslims is considered to have broader sociocultural consequences (Gieling et al., 2010; Verkuyten & Slooter, 2007). The overall mean score for tolerance of persuasion was below the neutral midpoint of the scale, which further indicates that participants were rather intolerant of trying to persuade other Muslims to express their religion in similar ways. Furthermore, lower tolerance for persuasion compared to enactment was found in all four contexts, including the private sphere. This might indicate that trying to persuade fellow Muslims to enact their religious identity is perceived by majority members as increasing the number of Muslims in society who enact their religion, which might be considered threatening to the ingroup cultural identity (McLaren, 2003; Outten et al., 2011).

This interpretation is further supported by the role of the ingroup cultural continuity concern which was found to matter for tolerance and the context‐related differences in tolerance. People low in continuity concern were generally tolerant and did not seem to consider the different sociocultural implications within the various contexts. In contrast, participants with higher continuity concern did consider the context of Muslim identity expression and demonstrated the discussed pattern of context differences in their enactment and persuasion tolerance. These findings indicate that cultural continuity concern is an important factor to consider when empirically examining and trying to improve tolerance of majority group members towards Muslim minorities. Concerns about being able to maintain one's cultural identity can be an important reason for the limits of tolerance (Verkuyten & Yogeeswaran, 2017).

Limitations and future directions

Despite its novel contribution to the tolerance research and the understanding of majority members' acceptance of Muslim identity enactment, there are several limitations that provide directions for future research. First, the vignettes were designed and presented in such a way that there was a clear demarcation between the four contexts, to ensure that the conditions did not overlap, which would have made the experimental manipulation unclear. However, as a consequence, the tolerance questions might not match all four vignettes equally well. Specifically, the private context might be more ambiguous (positive about enactment in private and/or negative about enactment in public) than the other three vignettes (positive about enactment in public). Thus, in the private context, the questions asked might be interpreted more in terms of ‘only in private’ or rather in terms of ‘not in public’, which is complementary but not exactly the same. Although it does not seem very likely that this possible ambiguity is responsible for the difference that was found between private and public contexts, it might be relevant for the interpretation of this difference.

Second, we investigated three public contexts that are not all ‘public’ to the same extent. Some work contexts might be considered ‘semi‐public’, although empirically no differences between the two assessed work contexts were found. Further, a sharp distinction between private and public has been criticized because both domains are often interdependent (e.g., Modood, 2015). However, a distinction between private and public contexts and between, for example, situations on the street and at work are common in people's thinking (Slater, 1998). We focussed on people's evaluation of Muslim identity enactment in these different contexts, and our theoretical reasoning was based on the perceived negative sociocultural implications of the same enactments in different contexts. The pattern of findings is in line with this reasoning, but it should be noted that we did not directly assess the perceived sociocultural implications and possible related feelings of threat. Future research could examine whether Muslim identity enactment has a different impact on perceived sociocultural consequences in different contexts, or whether it triggers different levels of threat in various contexts.

A third limitation relates to the national context in which the study was conducted. It remains to be seen to what extent the findings generalize to other national contexts than the Netherlands. On the one hand, many Western European countries are quite secular, which could mean that similar results will be found in other European countries. On the other hand, each national context has its own history and specific rules and regulations, which might influence social norms and attitudes towards, for example, the wearing of headscarves, especially in the public sphere. For instance, France is a strictly secular (‘laïcité’) society with the majority of the French supporting a ban on headscarves in public places, and the French are the least tolerant of the headscarf compared to other Western European countries (Helbling, 2014; Pew Research Center, 2005). In contrast, in countries with a more multicultural tradition such as the United Kingdom, there are quite liberal regulations and attitudes with regard to the wearing of headscarves in public institutions and by civil servants (Van der Noll, 2010). However, these country differences in the level of tolerance do not have to imply that majority members who are concerned about their ingroup cultural continuity do not differ in their tolerance in private and various public contexts or do not show lower tolerance for persuading fellow Muslims to also express their religion.

CONCLUSION

Using a well‐powered survey‐embedded experiment and a national sample, we found that tolerance of Muslim identity enactment shows context‐related and act‐related variance, with lower tolerance in public contexts and for trying to persuade fellow Muslims. Furthermore, the context of religious identity expression was found to have a medium‐sized effect on tolerance among majority group members who endorsed high (versus low) levels of concern about the continuity of their ingroup's cultural identity. These findings demonstrate that tolerance of Muslim identity enactment (e.g., headscarf) depends on the context and on what people are asked to tolerate. The findings provide a more detailed and nuanced understanding of the complexity of people's attitudes towards Muslim minorities. Such an understanding is important for trying to improve intergroup relations and accommodating minority rights in culturally diverse societies.

CONFLICT OF INTEREST

The authors have no conflict of interest to declare.

Data S1 Supporting information.

Click here for additional data file.