Freedom of religion, women's agency and banning the face veil: the role of feminist beliefs in shaping women's opinion

Abstract

Several countries have imposed bans on the wearing of face veils, a controversial option considered in Bill 94 by the province of Quebec in 2010. This paper examines non-Muslim women's support for the acceptability of the niqab in public spaces. Analysing the 2010 Quebec Women's Political Participation Survey, we find that key feminist arguments – that wearing the niqab is a woman's free choice, a matter of freedom of religion and a visible symbol of women's oppression – are important drivers of opinion. Their role in shaping opinion, however, is complex and mirrors divisions among feminist groups in the province. Additional attitudinal drivers include generation, exposure to the practice and openness to immigration. Equally important, our findings suggest that being a member of a racial minority, feelings of cultural insecurity and religiosity are of little consequence for thinking on the issue.

Keywords:

• face veils
• feminist beliefs
• threat perception
• inter-group relations
• public opinion
• tolerance

Introduction

Muslim women's head coverings¹ have generated controversy in a number of countries. Many European countries regulate the wearing of headscarves or other head coverings in the public sphere. Examples include France's 2004 ban on any ‘conspicuously worn’ religious symbols in schools, effectively banning Muslim students from wearing any head coverings, and the burqa bans adopted by France and Belgium in 2010.

These bans raise interesting questions from the perspective of intersectionality (Crenshaw 1991; Rottmann and Ferree 2008). How do non-Muslim women view bans on Muslim women's head coverings? Why do some non-Muslim women oppose bans while others applaud them? How do non-Muslim women balance considerations like women's personal agency versus freedom from what are widely portrayed as patriarchal religious practices? While Muslim women's head coverings have elicited a good deal of debate among feminists, surprisingly little attention has been paid to the opinions of non-Muslim women.

We examine non-Muslim women's attitudes towards the niqab using a survey undertaken in the Canadian province of Quebec in the midst of debate on the issue. The issue came to prominence in 2010 when a young woman enrolled in a French-language course refused to remove her niqab to allow the instructor to properly assess her pronunciation because men were present. Despite attempts to accommodate her, her repeated refusal resulted in her removal from the course. Soon after, the Quebec government introduced a bill limiting the right of women wearing face veils to receive or deliver services in nearly every public institution, including childcare centres, school boards and public health facilities, if it limited communication, limited others from identifying the wearer or presented a security risk.² The government defended the bill by invoking the principles of gender equality and the importance of the state's religious neutrality.

Quebec is a particularly interesting case (see Conway 2012). As in many European countries, there is considerable ambivalence about multiculturalism (Sharify-Funk 2010) and the niqab debate surrounding the bill highlighted the tensions between the rights of religious believers and values like secularism and gender equality that are central to the province's political culture. Moreover, the Quebec case makes it possible to extend the study of public opinion on Muslim women's head coverings to the wearing of the niqab. Previous studies have tapped opinion towards headscarves or the Muslim veil (Saroglou et al. 2009; van der Noll 2010). These are rather vague terms that probably connoted the hijab to most respondents. Public support may well differ depending on the type of head covering. The niqab is likely to elicit more opposition since it covers the entire face, save for the eyes.

The province's feminist groups were divided on the issue. The Simone de Beauvoir Institute (2010) issued a statement strongly opposing the bill and expressing its commitment ‘to supporting bodily and personal autonomy for all women, as well as all women's capacity to understand and articulate their experiences of oppression on their own terms’. La Fédération des femmes de Québec (2010), on the other hand, offered qualified support, arguing that the bill struck an important balance between gender equality and reasonable accommodation. Other women's organizations fully supported the ban on the grounds of women's equality (see e.g. AFEAS 2010).

The thinking of non-Muslim women on the issue echoed similar themes. Our findings reveal that the key attitudinal forces driving acceptance of the niqab in public spaces are the perception that the niqab is freely chosen and the belief that it should be viewed as a matter of religious freedom, whereas opposition to the practice is motivated by the belief that it is a visible symbol of women's oppression. However, the tensions between these beliefs and perceptions leave many women conflicted.

The feminist debate

Feminist thinking on the issue of Muslim women's head coverings is similarly conflicted. As Rottmann and Ferree (2008, 485) observe: ‘This is difficult theoretical terrain for all feminists.’ The feminist debate over Muslim women's head coverings reflects the larger ‘feminism/multiculturalism’ debate (Saharso 2008). Nussbaum (1999, 30) has summed up the dilemma for Western feminists:

To say that a practice endorsed by tradition is bad is to risk erring by imposing one's own way on others… To say that a practice is all right whenever local tradition endorses it as right and good is to risk erring by withholding critical judgement where real evil and oppression are surely present.

While feminist critics of Muslim women's head coverings do not characterize the practice as evil, they do typically consider it to be a symbol of Muslim women's lack of agency and their subordination to Islamic patriarchal norms (see Rottmann and Ferree 2008). From this perspective, a ban on head coverings is justified on grounds of gender equality and protecting Muslim women from oppression.

Rejecting the notion that the practice reflects Muslim women's subjugation, feminist opponents of a ban view veiling as an authentic choice that must be respected. Phillips (2009, 42) has cautioned against assuming that religious women's choices reflect false consciousness: ‘[R]esistance takes many and subtle forms, and that what looks to an outsider like submission can sometimes be better understood as empowerment or subversion.’ Feminists have also problematized the gender equality frame, arguing that it results in the othering of Muslim women (Sauer 2009) and reinforces negative stereotypes (Phillips and Saharso 2008). Krivenko (2012, 25) goes further, arguing that framing the issue in terms of equality rather than sexuality blinds us to Western cultural attitudes that are just as discriminatory.

Indeed, feminists who support Muslim women's choice to veil charge that feminist critiques of the practice are culturally essentialist: ‘[W]hen we criticize other cultures than our own there is a risk that we use double standards, that we use essentialist notions of culture and that we speak for, and thereby deny the autonomy and agency of, minority women’ (Saharso 2008, 6). These failings are seen in a failure to consult women who wear the veil.

Recognizing this lacuna, a number of studies have turned to focus groups or in-depth interviews with hijab-wearing women. Studies conducted in Canada (Ruby 2006), France (Afshar 2008; Wing, Smith, and Nigh 2005–2006) and the USA (Droogsma 2007) reveal very similar and equally diverse motivations, ranging from religious observance³ and modesty, to avoiding the male gaze, resisting sexual objectification and taking control of their own bodies, to asserting a Muslim identity and resisting assimilation. Far from seeing head covering as oppressive, hijab-wearing women often characterize the practice as empowering, contrasting ‘the pressure on [Western] women to reveal their bodies with their own choice to cover; the first reflects patriarchal oppression while the second reflects conscious resistance to oppression’ (Droogsma 2007, 309); however, there may, of course, be instances where head covering are imposed on women.

Given the typical media representation of Muslim head coverings as epitomizing ‘the veiled and oppressed victim’ (Bullock and Jafri 2000, 36), we expect non-Muslim women to be more likely to see the niqab as a visible symbol of women's oppression than a freely made choice. We further expect that these contrasting views will influence women's opinions about the proposed ban on the niqab. Women who associate the niqab with oppression will be more likely to favour the ban, whereas those who regard it as an expression of the wearer's agency will oppose it.

Threat perception and inter-group relations

Women may oppose a ban on the wearing of the niqab even though they find the practice abhorrent. This is the essence of tolerance: ‘a willingness to “put up with” those things that one rejects… One is tolerant to the extent one is prepared to extend freedoms to those whose ideas one rejects, whatever these might be’ (Sullivan, Piereson, and Marcus 1979, 784). One of the strongest predictors of intolerance is the potential threat associated with a group or practice (Gibson 2006; Golebiowska 1999).

Drawing on integrated threat theory (Stephan and Stephan 1996), van der Noll (2010) found that the perception that Muslims are unwilling to acculturate was one of the most important predictors of support for a headscarf ban in France, the Netherlands and the UK (see also Dekker and van der Noll 2012). This type of symbolic threat is likely to play an important role in non-Muslim women's views about a niqab ban. Symbolic threats arise when practices are perceived to deviate from societal norms and values. The threat that the niqab posed to Quebec values was a prominent theme in briefs submitted to the National Assembly (Conway 2012), even though no more than ninety and perhaps as few as twenty-four women wear the niqab in Quebec (Patriquin and Gillis 2010).

Secularism is widely considered a core Quebec value. Where freedom of religion implies a right to worship according to the tenets of one's religion, secularism asserts a freedom from public religion, and insists that the state be neutral on religious questions. Although Quebec does not have the same constitutional commitment to secularism as France, it is widely perceived as a valuable legacy of the province's ‘Quiet Revolution’ in the 1960s, which challenged the predominant role of the Catholic Church (Sharify-Funk 2010, 542). With the proposed niqab ban, secularism clearly trumped freedom of religion. Accordingly, we can expect women's views about wearing the niqab in public spaces to reflect the importance that they assign to freedom of religion.

The niqab debate also played into cultural insecurities. Multiculturalism, understood as ‘an obligation to accord the history, language and culture of non-dominant groups the same recognition and accommodation that is accorded to the dominant group’ (Kymlicka 2003, 150), suggests that face veils, as a symbol of cultural identity, should be tolerated in the public sphere. However, Quebec has never officially endorsed multiculturalism, although Canada is formally committed to multiculturalism as state policy. As a French-speaking minority within a largely English-speaking continent, Quebec's policies have concentrated on protecting the French language and culture from assimilation. Accordingly, the working policy is one of interculturalism, which holds that immigrants should be integrated into a common public culture through the medium of the French language (Rocher and Labelle 2010). As Sharify-Funk (2010, 537) observes, a ‘preoccupation with pure laine (literally, “pure wool”) Québecers' insecure minority status within Canada has… generated considerable ambivalence about the broader implications of multicultural policy and has at times reduced empathy for other Canadian minority groups'. Minority group practices that are perceived as threats to this ‘common public culture’ should generate more negative attitudes and be less tolerated than others. We should therefore expect a strong Quebec identity to be associated with support for a niqab ban.

Van der Noll (2010) found that perceptions of threat were associated with support for a headscarf ban among the prejudiced and non-prejudiced alike in France, Germany, the Netherlands and the UK. At the same time, prejudice – as measured by negative attitudes towards Muslims – had an independent and powerful effect. Similarly, Saroglou et al. (2009) found that subtle prejudice towards immigrants in general partially explained Belgians' anti-veil attitudes. We can expect negative attitudes towards immigration to have a similar effect on women's attitudes towards the wearing of the niqab in public spaces.

The challenge of integrating large numbers of immigrants coming from very different cultures necessarily raises questions about how far host societies should go in accommodating minority religious and cultural practices. The debate in Quebec did not begin with Bill 94. In 2007, the government established a commission on ‘Accommodation Practices Related to Cultural Differences’, or what came to be known as ‘reasonable accommodation’. Reasonable accommodation refers to the obligation of private and public institutions to accommodate diversity in their staff and clientele, so long as the accommodation does not cause excessive disruption (Marois 2005, 2). The niqab issue cuts to the heart of what constitutes reasonable accommodation (Conway 2012) and so women's attitudes regarding reasonable accommodation are likely to be linked to opinion about the niqab.

Inter-group anxiety is also likely to influence opinion. This type of threat refers to anxiety about social interactions with members of an out-group (Stephan and Stephan 1996). Women who would feel uncomfortable seeing a niqab-clad woman could be expected to have more negative views about allowing the practice in public spaces.

A sense of unease towards Muslim groups and practices might be mitigated, however, by exposure to ethno-religious diversity. There are at least two theoretical bases for this expectation. First, the cognitive empowerment hypothesis suggests that being exposed to ethnic and religious diversity can foster tolerance (Harell 2010). The key idea is that exposure to diversity fosters cognitive skills, such as perspective taking and seeing the different sides of an issue, that are essential to tolerance. Second, the inter-group contact hypothesis predicts that contact between majorities and minorities can reduce prejudice by increasing empathy and reducing inter-group anxiety and perceptions of threat on the part of the majority (Dekker and van der Noll 2012; Pettigrew 2008; Stephan and Stephan 1996). Importantly, there is evidence that these positive effects generalize beyond the groups involved in the contact (Pettigrew 1997; Reich and Purbhoo 1975). Thus, even though the odds of friendship or even acquaintance with a veil-wearing woman are very low, both hypotheses predict that exposure to ethno-religious diversity per se will encourage acceptance of the practice. Accordingly, opposition to the wearing of the niqab should be weakest on the Island of Montreal, where Quebec's immigrant population is concentrated (Chui, Tran, and Maheux 2009). We would also expect to see a generational pattern in support of the practice: young women who have grown to adulthood in a time of increasing ethno-religious diversity should be more accepting than older women who were socialized when society was more homogeneous (Saroglou et al. 2009). Similarly, women who are in the workforce or pursuing post-secondary education should be more supportive than women who are less exposed to diversity on a day-to-day basis as a result of their confinement to the domestic sphere (see Nunn, Crockett, and Williams 1978).

Experiencing diversity, however, could have the opposite effect. According to the competition hypothesis, diversity leads groups to compete for power and scarce resources (Bobo 1999), which can be perceived as a threat to the majority group's position of dominance (Jackson et al. 2001). The result is in-group solidarity and hostility towards out-groups. This is why lower socio-economic status has been linked to greater intolerance: the disadvantaged may have more to fear from the competition for jobs as a result of immigration (Esses et al. 2001). This is especially true of those with less education. Accordingly, we can expect increasing levels of education to make for greater acceptance of the niqab. Not only do educated women face less competition from immigrants in the job market, but they also tend to have larger and more diverse social networks (Erickson 2004). As a result, they are more likely to come into contact with women from a wide range of backgrounds, which may also make them more accepting of difference (Erickson 2009; Harell 2010).

Finally, Saroglou et al. (2009) have emphasized the importance of taking account of religious attitudes when examining the role of inter-group relations. They found that religiousness appeared to have ‘a “protective effect” against negative attitudes towards and representations of the Islamic veil’ in Belgium (426). This runs counter to research indicating a link between religiosity and prejudice. Since Stouffer's ground-breaking research in the 1950s (Stouffer 1955), religiosity has typically been found to be related to intolerance for out-groups and the rejection of universalist values (see e.g. Golebiowska 2004; Wald and Calhoun-Brown 2010). However, there is also evidence to suggest that the relationship between religiosity and prejudice is contingent (Hunsberger and Jackson 2005). In a secularized setting like Quebec, religiosity may be associated with greater tolerance of the practices of other religions. Religious women may share a sense of solidarity with other religious groups bred from the need to protect their own religious practices from state interference.⁴ For example, Bill 60, introduced in 2013 by the Quebec government, proposes banning civil servants from wearing any visible religious symbols.

Data and methods

To assess the various factors influencing attitudes toward the niqab, we draw on data from the 2010 Quebec Women's Political Participation Survey (2010 QWPPS). The twenty-three-minute telephone survey was undertaken by CROP Inc. from 2 June to 2 July 2010. A total of 1,201 interviews were completed with French-speaking women aged eighteen years and over.⁵

The survey asked respondents their attitudes on the wearing of the niqab in four increasingly challenging contexts: while (1) shopping; (2) working as a pharmacist; (3) teaching in a public school; and (4) voting in a Quebec election. The dependent variable combines responses to these four questions (coefficient α = .71). Since the dependent variable is ordinal, all models are estimated using ordered logistic regression. We estimate a series of models. The first model includes social background characteristics, the most distal predictors of attitudes. In addition to age, education, residency on the Island of Montreal and focus on the domestic sphere, controls for racial minority status and for working in the public sector were included. The variables are dummy-coded with the named categories coded ‘1’: age (under thirty-five, and fifty-five and over); education (completed college or university); resident of the Island of Montreal; neither a student, nor employed or self-employed; non-European ancestry (excluding Aboriginal peoples); and public sector worker.

The second model adds basic identities and beliefs that may influence women's opinions on the niqab. Feminist identity taps whether women considered themselves very feminist (scored as 1), somewhat feminist (scored as 0.5) or not feminist at all (scored as 0). National identity was captured by a question asking whether women identified primarily or exclusively with Quebec (scored as 1, and 0 otherwise). Religiosity was measured by asking about the importance of religion in the woman's life, with respondents indicating it was very or somewhat important scored as 1. Finally, views about immigration were captured by a question on the desired level of immigration. Respondents who indicated that immigration should be reduced were given a score of 1; those who supported existing or increased levels were scored 0.

The final model adds five attitudinal variables that are more proximate to opinions about the acceptability of the niqab. Two relate directly to the feminist debate: first, that the niqab represents women's visible oppression; and second, that women who wear the niqab do so ‘freely’, exercising individual agency rather than accepting the dictates of a patriarchal religious institution or succumbing to family pressure. A third variable taps into the possibility that a belief in the freedom of religion motivates support for the right of women to wear the niqab, while a fourth tests whether opposition to the niqab stems from a general feeling of discomfort with an unfamiliar religious and cultural practice. These four variables have a common scoring: strongly disagree (0), somewhat disagree (0.33), somewhat agree (0.66) and strongly agree (1.0). The final variable captures views about whether reasonable accommodation has ‘gone too far’, coded 1 for strongly disagree and 0 for strongly agree.

Findings

Regardless of the scenario, a majority of respondents view the niqab as unacceptable (see Table 1). The near-consensus on the final two scenarios is especially striking: only 8% of the women sampled considered wearing the niqab while teaching in a public school to be acceptable and a mere 6% found it acceptable when voting. More respondents accepted the right to wear the niqab while shopping (35%), although only 12% considered it acceptable while working as a pharmacist.

Table 1. Acceptability of the niqab.

		Shopping	Working as a pharmacist	Teaching in a public school	Voting
Wearing the niqab is acceptable when:		35.1	12.1	7.7	6.0
	No scenarios	One scenario	Two scenarios	Three scenarios	All four scenarios
Wearing the niqab is acceptable under:	63.9	22.6	6.4	4.0	3.0

Note: Entries are percentages.

Overall, almost two-thirds (64%) of respondents considered the niqab unacceptable under any scenario and almost a quarter (23%) found it to be acceptable under only one, almost always when the woman was shopping. When it came to the more challenging scenarios, the numbers dropped off precipitously. Few believed that the niqab was acceptable in more than one scenario, and only 3% of women considered it acceptable under all four scenarios.⁶

There was considerable variation in beliefs about the practice (see Table 2). Nonetheless, a majority of respondents rejected the notion that the women are exercising freedom of choice when they wear the niqab and that freedom of religion gives them a right to wear the face covering. Similarly, a majority considered the niqab to be a symbol of oppression, admitted to feeling uncomfortable seeing a women wearing one, and believed that reasonable accommodation has gone too far. As such, it is very likely that these beliefs are at least partly driving attitudes on the acceptability of the niqab.

Table 2. Beliefs underpinning attitudes on the acceptability of the niqab.

	Strongly agree	Somewhat agree	Somewhat disagree	Strongly disagree	Don't know	n
Wearing the niqab is a free choice	6.0	21.2	34.9	34.3	3.5	1,196
The niqab is a visible symbol of women'soppression	51.0	23.2	13.1	8.7	4.0	1,195
Freedom of religion gives women the right to wear the niqab	7.9	23.3	28.2	36.5	4.1	1,191
Seeing a woman in a niqab makes me uncomfortable	30.4	20.3	26.8	20.5	2.0	1,196
We have gone too far with reasonable accommodation	52.3	24.4	12.2	6.9	4.2	1,198

Note: Entries are percentages (except n).

Our first model looks at the impact of social background characteristics (see Table 3). The fault lines are clear. There are marked generational differences: other things being equal, the estimated probability of considering the niqab to be acceptable under at least one scenario is thirty-eight points higher for a woman under the age of thirty-five compared with a woman aged fifty-five or over.⁷ This lends support to our expectation that experiencing greater diversity during the formative years has made younger generations of Quebec women more accepting of practices such as wearing the niqab.

Table 3. The impact of social background characteristics, basic identities and beliefs, and proximate attitudes on acceptance of the niqab.

	Social background characteristics	Basic identities and beliefs	Proximate attitudes
Under 35 years of age	0.77 (.20)***	0.71 (.20)***	0.31 (.22)
55 years or older	−1.00 (.23)***	−0.89 (.24)***	−0.48 (.28)†
College or university graduate	0.44 (.17)*	0.30 (.18)	0.53 (.22)*
Resident of the Island of Montreal	0.36 (.17)*	0.32 (.17)†	0.41 (.18)*
Not working or studying	−0.40 (.26)	−0.45 (.27)†	−0.24 (.30)
Racial minority	0.09 (.25)	0.12 (.24)	0.24 (.25)
Public sector worker	−0.14 (.18)	−0.14 (.18)	0.10 (.20)
Primary/exclusive Quebec identity		0.12 (.17)	0.24 (.19)
Admit fewer immigrants		−0.77 (.20)***	−0.37 (.22)†
Religiosity		−0.12 (.17)	−0.47 (.19)*
Feminist identity		−0.47 (.27)†	0.08 (.29)
Wearing the niqab is a free choice			1.80 (.35)***
The niqab is a visible symbol of oppression			−0.64 (.31)*
Freedom of religion gives the right to wear niqab			2.75 (.30)***
Niqab-wearing woman causes me unease			−1.21 (.23)***
Reasonable accommodation has gone too far			−0.04 (.28)
Nagelkerke pseudo-R^2	.16	.19	.45
AIC	1.99	1.98	1.66
Number of cases	952	952	952

Note: The column entries are ordered logistic regression coefficients, with robust standard errors shown in parentheses.

*** p < .001, ** p < .01, * p < .05, † p < .10.

The other key fault line relates to education. As we predicted, education has a significant impact on views about the wearing of the niqab in public spaces. Women who have completed college or university are nine points more likely to find the wearing of the niqab acceptable under at least one scenario than women who lack a college or university diploma. Education is a powerful marker of socio-economic status in post-industrial societies, but just why women's opinions are divided along these lines requires further study. It may reflect the fact that women with less education are more likely to perceive immigrants as a threat in the competition for employment (Esses et al. 2001), or that more highly educated women have a wider network of friends and acquaintances and occupy positions that bring them into contact with people from different backgrounds (Erickson 2004, 2009; Harell 2010).

The potential importance of contact is highlighted by the positive effect of living on the Island of Montreal, where most of Quebec's immigrant population is concentrated. Living there boosts the acceptance of niqab wearing in public under at least one scenario by an estimated eight points. However, contrary to expectations, being confined to the domestic sphere does not significantly diminish support. Women who are full-time homemakers or retired have an eight-point lower probability of considering the wearing of the niqab in public to be acceptable in at least one public setting, compared with women who are working or studying outside the home, but the effect falls just short of conventional levels of statistical significance (p = .12).

We might have expected women who belong to a racial minority to be more sympathetic towards the cultural and religious practices of other minority groups. However, this is not the case. Belonging to a racial minority makes little or no difference to a woman's opinion about the niqab.⁸ Working in the public sector is also unrelated to opinion: apparently neither the prospect of serving a niqab-wearing woman nor working alongside her influences public sector workers' opinion on the issue.

We had expected that cultural insecurity might lead women with a strong Quebec identity to be less accepting of practices that are foreign to the traditional French-speaking mainstream, but there is no evidence to support this expectation (see Table 3).⁹ It may be that this is much more of a concern in elite-level discourse about the niqab (Conway 2012). Similarly, there is little to suggest that being religious is associated with negative opinions about the wearing of the niqab or that solidarity among religious women makes for greater openness to the practice. What does matter is openness to immigration. The effect is both statistically robust and substantial. This is consistent with Saroglou et al.' s (2009) finding that anti-immigrant attitudes drive some of the anti-veil sentiment. The probability of accepting the niqab in at least one public setting is fifteen points lower for women who favour reducing levels of immigration, compared with those who would keep the level as it is or who would accept more immigrants. However, only a little over a third (35%) of women wanted cuts to immigration. The modal preference was to maintain the current level.

Feminist identity also matters. Women who identify as strong feminists were significantly less likely, by nine percentage points, to find the niqab acceptable in public places than women who reject the label altogether. These strong feminists were much more likely to view the niqab as a symbol of women's oppression than a reflection of the wearer's agency or a matter of religious freedom. Sixty-one per cent strongly agreed with the first argument, compared with only 46% of women who did not consider themselves feminists. Conversely, 45% strongly disagreed with the statement that women wear the niqab by choice and 50% strongly disagreed with the argument about religious freedom. The comparable figures for non-feminist women were 34% and 31%. Like the feminist organizations, feminists themselves were clearly divided on these matters.

Beliefs about the wearing of the niqab clearly matter (see Table 3). Women who strongly agree that it represents a free choice are significantly more likely – by twenty-nine percentage points – to consider the practice acceptable under at least one scenario than those who strongly disagree. Conversely, construing the niqab to be a visible manifestation of women's oppression reduces the likelihood of considering the practice to be acceptable in any of the scenarios. However, the effect is much smaller (ten points). By far the most important attitudinal correlate of opinion is the belief that the issue is one of freedom of religion. The effect on opinion is both statistically robust and substantively strong: strongly agreeing with the freedom of religion interpretation increases the estimated probability of being willing to allow the practice under at least one scenario by forty-eight points, compared with strongly disagreeing. Clearly, freedom of religion trumps all other determinants when it comes to explaining opinion on this question.

At the same time, there is evidence that discomfort with an unfamiliar practice may be driving some of the opposition to the niqab. Half of the women agreed that seeing a woman wearing the niqab made them uneasy and this sense of unease influences their openness to it. A nineteen-point difference exists in the estimated probability of finding the practice acceptable in at least one scenario between those who admit to a strong sense of unease and those who strongly deny it. Despite the widespread belief that reasonable accommodation has gone too far and the prominence of the issue, opinion about reasonable accommodation is not a significant factor in women's views about the acceptability of the niqab. The effect was trivially small and dwarfed by its standard error.

Controlling for the impact of more proximate beliefs reduces the impact of age on attitudes towards the niqab, suggesting that these attitudes help to explain generational differences. On the other hand, once more proximate attitudes are taken into account, a significant negative relationship emerges between the importance that a woman assigns to religion in her life and acceptance of the niqab. Further investigation reveals that religiosity is associated with a stronger belief that freedom of religion includes the right to wear the niqab (not shown). Once this belief is taken into account, the impact of religiosity is more in line with previous research, although relatively small (seven points).

The key attitudinal forces driving opinion on the niqab are clearly the perception that the practice is freely chosen and the belief that it should be viewed as a question of religious freedom. However, these forces are not necessarily mutually reinforcing. Some women are clearly conflicted. Almost half (47%) of the women who believe that freedom of religion means that women have the right to wear the niqab do not believe that women freely choose to wear the face covering, and fully two thirds (69%) consider it to be a visible symbol of women's oppression (not shown). Moreover, a third (36%) of these women admit that seeing a woman wearing the niqab makes them feel uneasy, despite their belief that the wearer is exercising her right to freedom of religion. This sense of unease is also found among strong feminists, but many more (88%) of those who agree that freedom of religion gives women the right to wear the niqab consider it a visible symbol of their oppression.

Conclusion

Non-Muslim women in Quebec were clearly divided on the acceptability of the niqab in public spaces, especially along the lines of age and socio-economic status. The key social fault line is generational: younger women are much more open to the practice, at least in some circumstances. The inference is that this reflects greater exposure to ethno-religious diversity during their formative years, which, working through more proximate attitudes, reduces their opposition to the niqab. This is not the only evidence to show that exposure may matter. Women who live on the Island of Montreal – which is likely to be associated with greater exposure to members of ethno-religious minorities – are also more accepting of the practice. Both the cognitive empowerment hypothesis and the generalized inter-group contact hypothesis suggest that exposure to diversity can make for a greater willingness to accept a practice that is disliked. The results are certainly suggestive, but much more work is required to sort out which, if either, of these interpretations is correct. This will require detailed data on women's social networks, as well as their work environments and neighbourhoods.

The ban on the niqab was hypothesized to cue a number of perceived threats. One perceived threat relates to appropriate levels of immigration. It is impossible with the data at hand to determine whether this reflects a perceived threat to the majority culture, subtle prejudice or a general desire for conformity. Each may be at play. What is clear is that a desire to reduce immigration goes hand in hand with a reluctance to accept the niqab in public settings. Yet identifying strongly with the province appears to have no discernible impact on Quebec women's views about the wearing of the niqab, suggesting that the particular political, social and historical context within which the debate took place may have minimized the link between two. And while women's personal religiosity has a negative impact on their views on the public display of a religious symbol, mirroring findings elsewhere, this link is only apparent once the impact of more proximate attitudes are controlled.

Our primary purpose was to identify the role that feminist thinking plays in driving attitudes. Our findings suggest that its role ought not be ignored. The importance of feminist arguments in women's attitudes on the question is unequivocal. Some feminist organizations within the province were emphasizing women's bodily integrity and right to choose; accepting that women who wear the niqab are doing so by choice clearly makes a woman more open to the practice. Other feminist organizations countered with arguments that highlighted gender equality and freedom from patriarchal practices; these arguments are reflected in the negative impact of the belief that the niqab is a visible sign of women's oppression. What remains to be determined is how these attitudes are transmitted. Are women taking cues from the organizations themselves, or is it a two-step process that is mediated through elite discourse and media coverage?

It is perhaps not surprising that thinking on the issue is complex. Discussions on the face veil elicit passionate debate that is not always easily tied to feminist identity and religious belief. And it pits equally salient values against each other: women's bodily integrity and their right to be free from religious oppression. On this, our findings are likely not unique to Quebec. As immigrant flows become ever more diverse, many host societies have to grapple with the challenges of reconciling freedom of religion with gender equality. Our finding that younger women are much more open to a practice that makes many uncomfortable suggests that addressing the challenge might get easier over time. Such a prediction is tempered, however, by the reality that generational differences are less important to thinking on the issue of the face veil than more proximate attitudes on women's bodily autonomy and their religious freedoms, attitudes that are contradictory rather than reinforcing for many women.

Finally, the complexity of the issue underscores perhaps why feminist self-identification plays only a modest role in shaping attitudes on the niqab. Like feminist groups themselves, self-identified feminist women were divided on the justification for the ban on the niqab. Identifying as a feminist means that while the issue is likely to be particularly salient, looking to feminist groups for cues on the appropriate position to take is unlikely to offer any simple solutions to women's thinking on the issue.

Acknowledgements

The authors thank Katrine Beauregard and Nicole Leonard for their assistance.

Funding

This work was supported by the Social Sciences and Humanities Research Council of Canada [Grant #410-2009-0285].

Additional information

Funding

Funding: This work was supported by the Social Sciences and Humanities Research Council of Canada [Grant #410-2009-0285].

Notes on contributors

Brenda O'Neill

BRENDA O'NEILL is Associate Professor in the Department of Political Science at the University of Calgary.

Elisabeth Gidengil

ELISABETH GIDENGIL is Hiram Mills Professor in the Department of Political Science at McGill University.

Catherine Côté

CATHERINE CÔTÉ is Associate Professor in the École de politique appliquée at Université de Sherbrooke.

Lisa Young

LISA YOUNG is Professor in the Department of Political Science at the University of Calgary.

Notes

1. These coverings include: the hijab, a scarf that covers the head and neck but not the face; the niqab, which covers the face but leaves the eyes exposed; and the burqa, covering the whole body, with netting obscuring the eyes from view.

2. Bill 94 died with the dissolution of the thirty-ninth sitting of the Quebec Assembly in August 2012 but the proposed Quebec Charter of Values (Bill 60), introduced in September 2013, has reignited the debate.

3. There is a good deal of debate, however, about what the Qur'an requires. For useful summaries, see Ruby (2006) and Wing et al. (2005–2006).

4. We thank one of the anonymous reviewers for this insight. The impact of religion may also vary with affiliation, with adherents of minority religions being the most open to a practice like wearing the niqab. However, the province's religious composition makes it difficult to assess this possibility empirically. Quebec is (at least nominally) overwhelmingly Catholic: 82% of our sample identified as Catholic; non-Christians make up only 2%.

5. The response rate was 34%. The data are weighted to adjust for differences in the probability of selection based on household size, as well as the over- and under-sampling of regions that were part of the sample design.

6. Although interesting, a reliable analysis of differences between non-Muslim and Muslim women is impossible due to the small number of Muslim women (n = 18) in the sample.

7. All of the estimated probabilities were obtained using the margins option in Stata.

8. Substituting foreign born for racial minority does not change the result.

9. This finding continues to hold when opinion about immigration is dropped from the model.