The accommodation of Muslim body practices in German public swimming pools

ABSTRACT

The renegotiation of secularity for Muslims has reached sports facilities such as swimming pools, where the human body is central and an element of cross-cultural contestation. The present study identifies factors that could explain intra-organizational variations in religious accommodation. Data from a survey of geo-located swimming pools in Germany (n = 339) shows that conflict among users becomes more likely if the variation in body practices, the share of right-wing populist votes and the population of immigrant background increases. On the organizational level, we find that 75 per cent of the pools in our sample allow the burqini while only 10 per cent provide separate swimming hours for women. Higher shares of right-wing populist (and to a lesser extent of conservative) votes in the organization’s environment correlate with a higher likelihood of burqini bans. Our findings suggest that the relationship between organizational change and citizens’ attitudes is loose at best.

KEYWORDS:

• Organization
• religious accommodation
• swimming pools
• immigrant integration
• Muslims

Why should those interested in secularity be interested in religious accommodation for Muslims in German swimming pools? Following Gorski and Altınordu (2008) we assume that current regimes of state-religion relationship as they exist in Western Europe and elsewhere are the result of a historic and still ongoing process. Historically negotiated arrangements in Western Europe are mainly (though not exclusively) geared towards Christians and Jews. They are currently re-negotiated for growing secular as well as Muslim shares of the population. As has been argued elsewhere (Cadge et al. 2017), this process of re-negotiation differs across regimes of state-religion relationship and across types of organizations within such nationally defined regimes. The current study holds the type of organization (swimming pools) and the country (Germany) constant and instead seeks to understand what (local) factors may explain differences between individual organizations of the same type in terms of accommodation outcomes and user reaction towards accommodation.

In Germany, swimming pools are public organizations, mainly run by municipal authorities. School swimming lessons take place here and swimming pools facilitate sport and leisure activities for the local public. Swimming pools are secular spaces – at least in Germany, religious communities do not run swimming pools and have not done so in the past. Swimming pools are also places where the encounter of different body practices, possibly marked by religion, can be observed. Body practices can range from very prude to very permissive. Their degree of permissiveness can vary for example across individuals, social groups, religions or national cultures. Very prude body practices involve the total avoidance of swimming pools followed by swimming in gender-separated spaces only. For example, as late as 1959 some German Catholic priests intervened successfully against gender-mixed swimming (Pfaffenzeller 2019). Prude body practices comprise swimwear covering most of the body such as the Islamic burqini. Historically, one may think of the conflicts when women started to cut down on the length of their swimwear in the 1920s and when the bikini became popular in the 1960s. On the more permissive end of the scale, there is particularly revealing swimwear such as string-bikinis worn by women and men in mixed-gender pools. On this end of the scale, we also see body practices such as changing one’s clothes in areas visible to the opposite sex, women sunbathing topless and the free-body-culture (FKK) of naked swimming and sunbathing traditionally prominent in some German regions.

Swimming pools are often presented as extraordinary in the sense that few other places exist where the sexes meet in such revealing clothing. This is particularly true in Germany where body practices in swimming pools are rather permissive or at least more permissive than in several other European countries. In internet sources,¹ European visitors and newcomers to Germany frequently express their bewilderment and shame after encounters with Germans showering naked and people sunbathing naked in public places. Thus, differences in body practices are easily observable across European nation-states – and they are even more pronounced when relatively conservative Muslims in Germany who do attend gender-mixed public swimming pools but only in body-covering swimwear are included in the comparison.

This article explores how and why swimming pools differ in their accommodation of conservative Muslim body practices and what factors (including the accommodation for Muslims provided by pools) might explain conflict between different types of users. Our data shows that the users and the broader environment were reluctant towards religious accommodation, drawing bright boundaries between Muslim and non-Muslim populations. The data also shows that the current state of the renegotiation of secularity in German pools is partially accommodating for observant Muslims. We discuss the nature of the claims as well as the mobilization of right-wing populist parties as potential explanations.

Theoretical perspectives on religious accommodation in organisations and body practices in swimming pools

A growing body of literature is following Bender et al.’s (2013) call to study religion outside of congregations and thus focuses on the negotiation of the role of religion in public organizations. Many of these studies, be they on prisons (Beckford and Gilliat 1998; Becci 2011; Jahn 2015; Martínez-Ariño et al. 2015; Harms-Dalibon 2017; de Galembert 2020), on the military (Bertossi and Wihtol de Wenden 2007; Hansen 2012; Bertossi 2014; Michalowski 2015), on hospitals (Bertossi and Prud’homme 2011; Cadge 2013; Bernardo 2018) or on the police (Thériault 2009; Gauthier 2011; Lillevik 2019) contain a reflection on the negotiation of secularity and religious pluralism in a specific organizational context. They show that there are important differences in the accommodation of religious pluralism not only across national but also across organizational contexts. For example, studies (for many see Griera et al. 2015; de Galembert 2020) discuss how a dominant professional group in the organization frames religion as functional or unfunctional in that particular organizational context thereby deciding whether or not it should be accommodated (e.g. religion may be presented as beneficial to the rehabilitation of prisoners but as an obstacle to scientific medical care in hospitals). Similarly, Adam and Rea (2018) who pool a broad range of public and private organizations under the label of “workplace organizations” find that organizational feasibility, continuous service delivery or the non-hindrance of customers are legitimate arguments used for or against religious accommodation in that particular context (for a comparative framework of organizations including such arguments about functional aspects of religion see Cadge et al. 2017, 5).

Yet, while cross-national and cross-organizational differences in religious accommodation are considered, differences across organizations of the same type located in one country have not been studied systematically, partially because of the qualitative nature of the existing case studies. When discussing the particularities of local strategic action fields (Crozier and Friedberg 1977; Fligstein and McAdam 2013), some studies argue that organizational decision-making on religious accommodation is based on pragmatism (Bertossi 2014; Harms-Dalibon 2017; Adam and Rea 2018) and highly dependent on how individual actors translate expectations that are addressed to them by the environment into organizational practices (de Galembert 2020; also c.f. Harms-Dalibon 2017, 17). The current study acknowledges such idiosyncracies but still looks for more systematic factors that could at least partially explain local variation. Here, a study on the accommodation of religious pluralism in the rooms of silence of three Scandinavian universities, proposes local variations in the composition of users as one potential explanatory factor for variation in accommodation across local contexts (Christensen et al. 2018). Another study by Koopmans, Michalowski, and Waibel (2012) argues that citizenship rights for immigrants are more likely to be restricted in context with higher electoral success of right wing populist parties. We will pick up these ideas in our analyses of survey data on 339 geo-located German swimming pools. In particular, we ask what aspects of the organizational environment may have an influence on organizational decision-making with regard to Muslim accommodation and how the composition of the users may relate to conflict around organizational rules.

Next to theories on cross-national and cross-organizational differences, the sociology of migration produced theories explaining minority religious accommodation by focusing on the claims: Koenig for example (2005) differentiates Muslim minority claims according to their level of demandingness for the liberal state, speaking of claims for toleration, for autonomy, for tolerance and for equal participation in the organizational centre of the state. He argues that equal participation is the most demanding type of claim because it requires the state to consider religious and cultural identities in all organizational processes. Building on Statham et al. (2005), Carol and Koopmans (2013) classify claims for religious accommodation according to their level of obtrusiveness. The authors argue that claims become more obtrusive if they are not claims for parity but for special treatment, if they concern public institutions and in particular the personnel of these institutions, if they concern the religious practices of a small minority within the minority and if they require the rest of society to adapt their behaviour to accommodate the minority. Carol and Koopmans find that in countries which already provide more minority accommodation, claims tend to cover ever more obtrusive issues, thus following a logic of further differentiation and specification. In complementation of these theories one could, in view of the cross-organizational comparison, argue that claims for accommodation vary across organizations, in terms of content as well quantity: while studies of total institutions (Goffman 1961) such as prisons and the military have shown that these organizations are confronted with claims for accommodation ranging from Ramadan to dress and breaks, leisure organizations such as swimming pools receive more specific claims that relate – just like in hospitals and schools (the two other organizations studied in the larger project) – mainly to the question how the human body is displayed in public and regulated by the organization.

Thus, body practices are a central element of the negotiations over religious accommodation in public swimming pools because religions partially find their expression in body practices. Comparing Islam and Christianity, Mellor and Shilling point out “[what … .] these two religions aim to achieve is the initiation of a process whereby bodies are made ‘Christian’ or ‘Muslim’” (Mellor and Shilling 2010, 32). Examples that they give on the Muslim side are the body movements during prayer, the veiling of women and the use of prayer caps by men (Mellor and Shilling 2010, 31). Referring to other authors, Mellor and Shilling (2010, 34) state that “Islam is a religion whose resilience has in the past rested upon ‘its essence as a way of life’ embedded in physical habits and practices rather than contained within a cognitive ‘belief system’”. The more Weberian perspective of Philip Gorski (2003) assumes that religions vary in their degree of putting emphasis on and believing in the human capacity to control and discipline the body. Compared to Protestant Christianity, this belief in the capacity to control the body is weaker in Catholicism and even weaker in Sunni and Shia Islam which is why men and women are required to cover their awra.² Hard-line Muslim jurists insist that Muslim women should only swim in the presence of other Muslim women while comparatively more liberal jurists tolerate burqinis worn in mixed-gender pools (Shavit and Winter 2011, 270–1). Those women who wear the burqini can therefore be seen as social innovators from the perspective of conservative Islam. Some pool visitors in Germany, however, find burqinis difficult to tolerate. Although meant to hide the female body from the looks of non-related men, burqinis de facto lead to “a paradox hypervisibility” in German swimming pools (c.f. Karstein and Burchardt 2017, 213). Covering most of the body, burqinis may be (mis-)understood by other pool users as contradicting current trends of defining one’s identity through investments in body capital, which promises success and can be reached through personal effort (Bourdieu 1982 cited in Gugutzer 2012, 289). Gugutzer (2012, 289) underlines the strength of such ideas by classifying body-optimising practices as this-worldly oriented, functional equivalents for religion. To the extent that burqini-wearing women are perceived as rejecting this trend, social boundaries are further enhanced.

The case

In Germany, swimming pools are public institutions. Municipalities are legally obliged to maintain public pools to allow school swimming lessons. Most public pools are governed by municipalities (81 per cent in our data set), tax-financed and open to the general public. There are different types of pools which to some extent attract different kinds of users. While some users seek a physical work-out, others – especially during summer – want to cool down and relax. Except for the recreation-oriented covered pools, open-air pools usually attract a much broader clientele than covered pools. Swimming pools are public places of socialization where different social groups meet. They are and most likely have always been places where individuals present their bodies and flirt (Kaschuba 2018). Open-air pools in Germany – this is substantiated by our analysis – are more permissive spaces than covered pools because people sunbathe on the lawns surrounding the wet area, some women are topless, and people sometimes change on the lawn where others can see them.

Swimming pools usually have a clearly-defined and well-limited space which everyone can access by paying an entrance fee. Although fees are on the rise, especially summer pools remain a comparatively cheap activity for those who cannot afford a vacation at the seaside (in the past an upper-class summer activity). In the twentieth century, swimming pools were also a venue of racist exclusion. In 1933, many German towns prohibited Jews access to public pools, and by 1937 such prohibitions were in place throughout Germany (c.f. Andryszak and Bramkamp 2016). In the US, Afro-Americans were excluded from public pools during times of segregation and de-facto excluded from swimming after desegregation (Zaubler 2015, 82) because public pools were then less well maintained and many better-off whites built private pools in their backyards (Smith 2012, 46). Smith (2012, 43) also underlined that “due to the intimate nature of swimming, public pools were enforced with the most stringent of segregation codes to prevent this race-mixing” and de-segregation was particularly slow.

In the clearly delimited area of the swimming pool, a different dress code applies than in most other public spaces. Beyond the changing rooms towards the wet area, showing a lot of naked skin is normal. Also, inside the pool and in particular inside the showers (note that in Germany, showers are gender-separated and many people shower nude) wearing covering clothes is easily perceived as a breach of norms. Most swimwear is more revealing than usual clothing. This first and foremost has a functional role since uncovering and tight-fitting swimwear facilitates swimming. The scarce dress in swimming pools, however, may also be an item of fashion and it can gain a sexual connotation when the body is exposed or observed. It is difficult to hide one’s body in a swimming pool as even a T-shirt or a towel worn around the waist usually needs to be taken off when entering the water. Another particularity of German pools is that many pool visitors change in common changing rooms.

The pool regulates nudity. It formulates organizational rules that define what swimwear may be worn inside the pool. Furthermore, its architectural arrangements have an influence on how much nudity one encounters in a public pool. For example, gender-separated showers that are directly connected with gender-separated common changing rooms foster nudity more than gender-separated showers that are connected with gender-mixed single changing rooms. Swimming pools can also regulate whether naked swimming hours or areas for naked sunbathing exist and whether they offer separate swimming hours for women. The common prohibition to film and take pictures also falls under organizational regulations of nudity. Furthermore, the swimming pool must regulate not only consensual intimacy but also cases of sexual harassment.

As mentioned above, most of the German pools are run by municipalities which are free to decide on the exact formulation of these rules. Rules can be changed through simple administrative decision as well as through political decision in the municipal council as was the case for some burqini rules. Such adaptations of pool rules can be influenced by the lobbying of Muslim and non-Muslim civil society organizations as well as by the umbrella organization of all swimming pool operators, the German Bathing Society (Deutsche Gesellschaft für das Badewesen). Sociologists of organization assume that organizational change can flow from the organization itself, e.g. through the action of internal lobbyists (Dobbin, Kim, and Kalev 2011) and that it can also be initiated by the organization’s environment (e.g. through legal changes, public debates). Such organizational change may find its expression in the change of both formal and informal organizational rules (Kühl 2015; Luhmann [1964] 1995). Unlike for example the military, swimming pools are intimately linked to their environment through the public and their municipal governance.

Hypotheses

Following Christensen et al. (2018) we assume that the users partially determine how the framework for accommodation provided by the organization is lived in practice. It is thus not only important to study the pool rules on swimwear but also to take into account the permissiveness of the users’ swimwear. In fact, next to the formal organizational regulations, there is a lot of social regulation among users for example through the looks they carry on each other’s swimwear and the complaints they make to the pool personnel about other users. We expect differences across pools in terms of the composition of their users – some pools having a more homogenous and others a more heterogeneous population of users in terms of body practices. We expect enhanced conflict in pools with users with more heterogeneous body practices (H1). In a similar vein, we expect that a higher share of immigrants (Muslim and non-Muslim) in the area around the swimming pool and thus by extension among the pool users increases feelings of alienation among certain visitors and thus leads to more complaints about Muslim body practices in swimming pools (H2). Referring to the literature on cross-religious differences in body practices, we also expect differences in complaints about the burqini according to whether a pool is located in a dominantly Catholic or Protestant region in Germany (H3). Relying on the literature about claims-making presented above we expect that burqinis are more likely to be accommodated than separate swimming hours for Muslim women (H4) since they can be framed as a parity of treatment while the latter involve special treatment for a small group and cut into access for the majority. We also assume that organizational rules influence the pool visitors’ perception of social norms. This means that we expect pools which have explicitly allowed the wearing of burqinis for more than five years to receive fewer complaints about this swimwear (H5). With regard to the organizational environment we expect that an increased share of right-wing populist votes in the area surrounding the swimming pool leads to more complaints about Muslim body practices in swimming pools (H6a). Koopmans, Michalowski, and Waibel (2012) have shown that higher vote shares for extreme right-wing populist parties are positively correlated with a restriction of religious and cultural citizenship rights for immigrants. We thus expect that an increased share of right-wing populist votes in the area surrounding the swimming pool leads to less accommodation of Muslim body practices both in terms of swimwear and swimming hours (H6b).

Methods of data collection and analysis

This analysis is based on a survey carried out in cooperation with the German Bathing Society (DGfdB) among the roughly 5100 German swimming pools that are registered in a DGfdB dataset (www.baederatlas.de). 339 swimming pools participated in the survey (Michalowski and Behrendt 2020). Taking the 5100 swimming pools as the universe of cases, our sample is roughly representative of the swimming pools in Germany with pools in rural areas being slightly under- and larger pools that combine an indoor and an outdoor area and are located in densely populated areas being slightly overrepresented. The survey asked a question about the pool’s zip code. This allowed us to connect the dataset with standard geodata, e.g. on the population share with immigrant background and on voting behaviour. For most of our analyses, we excluded the questionnaires without zip code which reduced the sample from 339 to 323 cases. The questionnaire was filled in either by personnel working in the pool or by the pool operator. The questionnaire included questions about the permissiveness of the pool visitors especially regarding different body practices including swimwear, about pool rules and their implementation, about sexual harassment, separate swimming hours, and about the organizational structure.

We use complaints as indicators of conflicts about social norms. To better understand which factors increase the likelihood of complaints in particular about burqinis, on the one hand, and about particularly revealing swimwear on the other, we developed explorative models on the basis of our survey data. Our central goal was to identify general characteristics of the public that correlate with the likelihood of complaints. We differentiate between characteristics of the public that relate to a certain degree of permissiveness in showing one’s body in public (swimwear, use of communal dress rooms, acceptance of children of the opposite sex in the showers) as well as characteristics of the public that relate to political preferences (voting behaviour) and immigrant background (regional share of individuals with immigrant background). Our dependent variable measuring conflict among users is an estimate of the number of complaints by pool visitors about other visitors provided by the person who filled in the questionnaire. For the model estimating the likelihood of complaints about the burqini, our dependent variable is “How often does your pool receive complaints by pool visitors about other visitors wearing a burqini?” (coded 0= never/rarely; 1= sometimes/often). For the model estimating the likelihood of complaints about pool visitors with particularly permissive body practices, our dependent variable is an index composed of the items:

How often does your pool receive complaints about: (a) visitors wearing a string bikini?, (b) topless women?, (c) visitors getting changed in areas visible to the opposite sex (e.g. in front of the lockers), (d) visitors having physical contact in front of others, (e) children of the opposite sex in the showers?

The independent variables are identical for the two models and always scaled in four categories (never = 1; rarely = 2; sometimes = 3; often = 4). They first comprise the variable “Are there women wearing the burqini in your pool?” as well as the individual items measuring the permissiveness of the pool users (Frequency of … men in particularly revealing swimwear; women in particularly revealing swimwear; use of communal dress rooms; users getting dressed in areas visible to the opposite sex, e.g. in front of the lockers; users getting dressed on the lawn and women sun tanning topless on the lawn; note that the last two items were only asked in outdoor pools). For each pool, we calculated an average that was standardized from 0 to 1.

Both models control for the share of people with immigrant background among the inhabitants of the area in which the pool was located and for the voting behaviour in the area. The regional share of individuals with immigrant background was taken from the micro census 2017. It should be noted though that here data are provided for the level of districts comprising between one and five million inhabitants (former Regierungsbezirke), not municipalities. Voting behaviour was measured through party votes (Zweitstimmen) in the federal elections of 2017 on the level of Kreise and Landkreise, usually comprising between 50,000 and 500,000 inhabitants. We also use these variables on voting behaviour and share of immigrant population in further models to estimate the likelihood of burqinis being allowed in a pool.

Results: user behaviour and organizational rules

We present our results in two steps. First, we discuss results relating to the organization’s public, i.e. the (potential) pool users. In a second step, we focus on the pool rules, i.e. their adaptation as well as the potential consequences of such adaptations.

The users: complaints about burqinis and particularly revealing clothing

The survey’s first section dealt with the permissiveness of the users. We chose this topic to study how swimming pools adapt to increasing religious heterogeneity in the form of body practices marked by religion. Users may seek to impose their own social norms, for example by complaining to the pool attendant about the behaviour of other users.

Descriptively, we first of all see that only 7 per cent of the 339 pools frequently and 25 per cent sometimes count women in burqinis among their guests. The other pools rarely (45 per cent) or never (21 per cent) have women in burqinis among their guests. Almost all of the latter pools (86 per cent in case of “rarely women in burqinis” and 91 per cent in case of never women in burqinis) say that they rarely or never receive complaints about the burqini. Among the pools, however, where women in burqinis are a more frequent sight, the share that rarely or never receives complaints about the burkini is considerably lower (55 per cent in the case of “sometimes women in burqinis” and 46 per cent in the case of “often women in burqinis”). In total only 32 per cent of the pools that participated in our survey declared that they sometimes or frequently encounter the body practice of wearing a burqini and 19 per cent declared that they sometimes or often receive complaints about the burqini. Remarkably, among the pools which sometimes or often count women in burqini among their guests, only about half state that they sometimes or often receive complaints while another half notice less resistance against the renegotiation of secularity, i.e. the admission of a religiously motivated piece of clothing in a public space.

Further statistical analyses show a robust, statistically significant correlation between the number of burqini-wearing women in a pool and complaints about women wearing a burqini (c.f. see the linear regression in model 1). The model also shows that, when controlling for the number of burqini-wearing women, the likelihood of complaints about the burqini increases the higher the share of AfD votes in the 2017 parliamentary elections in the region. This provides support for hypothesis H6a (Table 1).³

Table 1. Complaints about other pool visitors wearing a burqini.

Variables	Complaints about the burqini
Visitors wearing burqini (rarely)	0.0772* (0.0395)
Visitors wearing burqini (sometimes)	0.381*** (0.0676)
Visitors wearing burqini (often)	0.458*** (0.103)
Regional share of individuals with immigrant background	1.005*** (0.361)
Index permissiveness of visitors	0.267** (0.128)
AfD party vote 2017	1.636** (0.816)
CDU party vote 2017	−0.431 (0.434)
SPD party vote 2017	0.288 (0.502)
Constant	−0.370 (0.336)
Observations	313
R-squared	0.222

Note: Robust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1.

A region’s confessional orientation (Catholic/Protestant) does not impact the likelihood of complaints about the burqini (no support for H3), but pools with more permissive users and pools in a region with a higher share of immigrants are more likely to receive complaints about the burqini. In consequence, we find support for H2. We also see that pools which have more particularly permissive and more particularly prude users than the average are more likely to receive complaints about the burqini. Apparently, the encounter of the extremes does not foster a live-and-let-live-attitude (Table 2). In parallel, we calculated another model using the same explanatory variables to estimate their effect on complaints about particularly permissive behaviour, measured through an index. This model shows that complaints about particularly permissive behaviour increase when the number of women wearing a burqini increases. This provides support for H1 that the encounter of more heterogenous body practices (particularly prude and particular permissive) among pool users enhances complaints. At the same time – and different from complaints about the burqini – we find no significant effect of the share of individuals of immigrant origin in the region, nor of the right-wing populist vote on complaints about particularly permissive behaviour. It is important to note though that pools in which particularly permissive users are more numerous are also more likely to receive complaints about this particularly permissive behaviour. These findings suggest that there is limited habituation towards and acceptance of body practices falling outside the norm and that conflict increases if two body practices that fall outside the norm become more frequent in a pool.

Table 2. Complaints about the permissiveness of other pool visitors.

Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Visitors wearing burqini (rarely)	0.168*** (0.0485)	0.169*** (0.0489)	0.117*** (0.0454)	0.114** (0.0461)	0.114** (0.0464)	0.111** (0.0469)
Visitors wearing burqini (sometimes)	0.359*** (0.0557)	0.360*** (0.0572)	0.283*** (0.0539)	0.277*** (0.0545)	0.279*** (0.0562)	0.272*** (0.0567)
Visitors wearing burqini (often)	0.422*** (0.105)	0.423*** (0.106)	0.382*** (0.0953)	0.376*** (0.0952)	0.378*** (0.0957)	0.376*** (0.0970)
Regional share of individuals with immigrant background		−0.0358 (0.260)	0.384 (0.257)	0.319 (0.297)	0.320 (0.297)	0.276 (0.297)
Index permissiveness of visitors			0.621*** (0.0934)	0.620*** (0.0935)	0.621*** (0.0935)	0.598*** (0.0957)
AfD party vote 2017				−0.257 (0.535)	−0.246 (0.541)	−0.730 (0.663)
CDU party vote 2017					0.0330 (0.305)	−0.207 (0.379)
SPD party vote 2017						−0.540 (0.417)
Constant	1.239*** (0.0396)	1.246*** (0.0687)	0.995*** (0.0724)	1.044*** (0.127)	1.030*** (0.177)	1.304*** (0.282)
Observations	315	315	315	315	315	315
R-squared	0.135	0.135	0.241	0.241	0.241	0.245

Note: Robust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1.

Organizational rules: burqini allowances and separate swimming hours for women

Given the potential conflicts among users, one may ask how pool rules can alleviate or increase such conflicts. We first look at burqini admissions and prohibitions, then at separate swimming hours for women. Pools with naked swimming hours (FKK) only represent 7 per cent of our sample (25 pools) and do not differ systematically from the rest of the sample, e.g. in terms of burqini allowance. 75 per cent of the 339 German pools (255 pools) which participated in the survey allow the burqini. Only 32 pools, i.e. 10 per cent explicitly prohibit it while 52 pools (15 per cent) left this question unanswered – maybe because the situation is unclear (these pools did also not answer the question on the frequency of burqini-wearing women or stated that there are never or rarely women in burqini among the pool visitors). Among the pools that allowed the burqini and that answered an additional question about when the burqini was first allowed (n = 195), only 33 per cent had already allowed it in 2013 while in 2016, the year after the arrival of large numbers of refugees in Germany, 83 per cent allowed the burqini. By the end of 2016, the burqini allowance was also endorsed by the German Bathing Society, the umbrella organization of swimming-pool operators. This shows a clear change of organizational rules towards the accommodation of conservative Muslim users. In line with our hypothesis H6b, burqini prohibitions are more likely where the share of right-wing populist votes and – to a smaller extent – the share of centre-right Christian conservative votes is higher (c.f. see the linear regression in Table 3).

Table 3. Influence of regional factors on burqini allowances.

Variables
Regional share of individuals with immigrant background	−0.179 (0.222)
AfD party vote 2017	−1.827*** (0.582)
CDU party vote 2017	−0.833** (0.328)
SPD party vote 2017	−0.371 (0.341)
Constant	1.505*** (0.189)
Observations	273
R-squared	0.042

Note: Robust standard errors in parentheses ***p < 0.01, **p < 0.05, *p < 0.1.

Pools with and without burqini admission do not differ significantly in terms of the share of the population of immigrant origin in their region, but pools with burqini bans obviously have fewer guests (wanting to) wear a burqini. These pools are also less likely to receive complaints about the burqini than pools with burqini admission. Interestingly, however burqini bans do not fully prevent complaints: 25 per cent of the pools allowing the burqini and still 17 per cent of the pools prohibiting it sometimes or often receive complaints about the burqini. If we now compare the pools that permitted the burqini before and after 2013 we find, other than expected in hypothesis H5, no adaptation effect in the sense of a decrease of complaints about burqinis over time: even in the pools that were the forerunners of the currently wide-spread burqini admission, a higher number of women in burqinis goes along with more complaints about this piece of clothing.

Pools in our sample were more reticent to provide separate swimming hours for Muslim women. In general, public pools may either rent out opening hours to private associations, which may then offer separate swimming hours for Muslim women only, or they may offer separate swimming hours for women with or without adaptations for Muslim women. Our survey shows that only 32 of the 323 pools (10 per cent) offer separate swimming hours for women and among them only 7 state that these swimming hours are de facto intended for Muslim women. These 7 and further 9 pools have exclusively female pool attendants during these separate swimming hours. Given that the other half of the pools also employs male personnel, their offer is open only to a limited extent to conservative Muslim women. Thus, separate swimming hours for women do not necessarily seek to accommodate Muslim women. They are also not a new phenomenon: approximately one third of these pools already had separate swimming hours for women in 1999 and none of the pools that introduced such hours in the 1960s, 1970s, 1980s or 1990s has special hours for Muslim women. This is a newer phenomenon with the first separate swimming hours for Muslim women in our sample starting in 2003. Interestingly, six of the 32 pools with separate swimming hours for women declared that they introduced them in response to political pressure. Overall, this finding lends support to H4 predicting that swimming pools accommodate burqinis more frequently than separate swimming hours for Muslim women. This is in line with political claims theory predicting that claims for religious accommodation are less likely to be successful if they also require the general population to change its behaviour. Other than burqini admissions which only require a toleration of other practices, separate swimming hours for (Muslim) women are a claim for equal participation (c.f. Koenig 2005) which is highly obtrusive (c.f. Carol and Koopmans 2013) because it will require all men and potentially even non-Muslim women to renounce on swimming during those hours. The main argument used by the proponents of separate swimming hours for women is that this may well be the only way of letting conservative Muslim women participate in swimming and teaching those who have not been to school in Germany how to swim. The main argument used by opponents of separate swimming hours is that they separate citizens according to gender or even faith, thereby jeopardizing a common public good. Opponents of separate swimming hours for women also refer to a German federal administrative court ruling deciding that students must participate in mixed-gender swimming classes at school under the condition that the burqini is allowed.

Discussion and conclusions

The current study seeks to contribute to the existing literature on organization-specific approaches to religious accommodation by focusing on differences across organizations of the same type that are located within the same country. The article explored how and why swimming pools differ in their accommodation of conservative Muslim body practices and what factors might explain conflict between different types of users.

In terms of outcome-differences we looked at organizational rules relevant to accommodate conservative Muslims (i.e. burqini regulations and separate swimming hours for women) as well as at conflict among users (in particular at complaints about other users’ body practices). In terms of potential explanatory factors, we looked at variations between users in terms of body practices (from prude to permissive), at organizational regulations, at the share of immigrants in the area surrounding the swimming pool, at Christian confessional distribution in the larger region and at the local share of right-wing populist AfD votes in the last federal elections 2017. We surveyed 339 geo-located German swimming pools about their users and their pool rules. Our survey is roughly representative of swimming pools in Germany, even though pools in larger cities are slightly over- and pools in rural areas slightly under-represented.

With regard to conflict between users we find that complaints about women in burqinis increase with the number of women wearing it. We also find, when controlling for the number of women wearing a burqini, that complaints are more frequent in pools located in regions that had higher shares of right-wing populist AfD votes in the last federal elections (support for H6a). We also find support for H2, to the extent that a higher share of individuals of immigrant origin in the region surrounding the pool is positively correlated with complaints about the burqini, still controlling for the actual number of burqini-wearing women. Complaints about burqinis do not correlate with a region’s Catholic or Protestant orientation (no support for H3) but they are more likely in pools where more users show a particularly permissive behaviour measured mainly in terms of swimwear and public nudity. This finding supports H1, predicting that greater variance in body practices increases conflict (rather than fostering a live-and-let-live attitude). Finally, with regard to users we find that pool rules only seem to have a limited effect on complaining behaviour because swimming pools that have allowed the burqini for five and more years do not experience a significant decrease in complaints about this piece of clothing. We thus find no support for H5. In terms of organizational rules, we find that burqini prohibitions are more likely in municipalities where the share of right-wing populist votes and – to a smaller extent – the share of centre-right Christian conservative votes is higher, providing support for H6b. The sample size was too small to observe similar effects on separate swimming hours for women. This relates to the fact that we found support for H4, predicting that swimming pools would be more likely to accommodate the less obtrusive burqini compared to the more obtrusive separate swimming hours for (Muslim) women that potentially require the majority to change its habits.

What do our findings tell us about potential explanatory factors of variations across different organizations of the same type? The existing literature as well as additional data currently collected in our project suggest that the idiosyncracies of local actor constellations are important for the final accommodation outcomes. If more and more pools today (inter alia because of legal pressures) include burqinis in their list of authorized swimwear, it still very much depends on the way this rule is implemented (maybe even informally), whether burqini-wearing women actually use the pool and how other groups of users react to this. Nonetheless, the present analysis has also identified systematic factors that can explain local variation in terms of conflict and organizational rules, namely the composition of users and the political environment. These can only partially be influenced by the organization itself. So, what can the organization do to facilitate social change? Our data suggests that formal organizational change does not necessarily lead to social change. If this was possible, one would expect pool users in pools that have allowed the burqini for more than five years to stop complaining about this piece of clothing. If social change could be enforced by social realities one might expect pool users in pools that frequently count women in burqinis among their users to get used to this sight and stop complaining about burqinis. Yet, none of this is the case. Burqinis are – unlike bikinis used to be – not just another provocative piece of swimwear. They are (mis-) understood as a marker of a bright group boundary between observant Muslims and increasingly secular Christians. Ezli (2014) pointedly argued that burqinis should not be understood as a sign of separation from society because conservative Muslim women wearing burqinis neither evade swimming pools, nor restrict themselves to female-only swimming hours. Yet, right-wing populist campaign posters showing bikini-wearing women and slogans such as “Burqa? We like bikinis – our land, our rules” use swimwear to make the social boundary brighter. Also current social trends of enhanced bodily fitness and beauty taking on quasi-religious forms of a marked individualist character (Gugutzer 2012) might enhance this social boundary. From this perspective, putting on a burqini can be understood as an expression of non-participation in the body cult and of belonging to a group that distances itself from individualism (Mellor and Shilling 2010) and favours practices considered emblematic of the legitimation of differences between the sexes (Karstein and Burchardt 2017). Yet, not all pools that frequently or sometimes have women in burqinis among their guests also experience widespread conflict. In more than half of these pools burqinis are apparently tolerated. This finding suggests that at least in some cases a fragile equilibrium of accommodating change for conservative Muslims in German pools may have been reached. In a next step, the BODYRULES project qualitatively explores how swimming pool guests legitimize their own body practices as well as their tolerance or rejection of other practices. This will help to better understand the nuances of the concerns and complaints pool visitors have about each other’s body practices that are not captured in the current approach.

Finally, what can we say about the external validity of our findings on German swimming pools for other countries and organizations? Germany as a country-case is a liberal democracy, comparable for this purpose to other Western European countries that host large Muslim populations of immigrant origin. Although the public debate about Muslim accommodation in pools may have been more vivid in Germany than in some neighbouring countries, a quick research suggests that Austria, Belgium, France and the Netherlands also discuss burqinis and that there is variation in accommodation across pools within one country. Even in France where laïcité is frequently used as an argument to ban burqinis from public pools, some municipalities (e.g. Rennes) allow them. We therefore expect that our findings on the composition of users and on the political environment can be extended to other national contexts. Whether this is also true for other organizations will be studied in more detail in an upcoming comparison of swimming pools with two other organizations where the regulation of the human body stands central, namely hospitals and schools.

Acknowledgements

We would like to thank our colleagues at the WZB Berlin as well as our project partners Maja Apelt (University of Potsdam) and Liane Schenk (Charité Berlin) with their teams for comments on an earlier version of this paper. We also thank three anonymous reviewers for their helpful suggestions.

Disclosure statement

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

Additional information

Funding

The project BODYRULES was supported by the Federal Ministry of Education and Research (BMBF) [grant number 01UM1811BY].

Notes

1 For many see: www.independent.co.uk/travel/europe/german-spas-the-naked-truth-6269555.html; www.allemagne-au-max.com/forum/la-nudite-en-allemagne-vt15707.html; www.huffingtonpost.it/elena-cerizza/5-comportamenti-assurdi-nelle-piscine-tedesche-_b_7983090.html; www.vice.com/es_co/article/bm4gja/en-berln-puedes-pasear-desnudo-en-un-parque-y-las-autoridades-no-te-dirn-nada, last consulted in August 2019.

2 With differences across legal schools, the male awra is usually defined as reaching from navel to below the knee and as including the entire body except for hands, feet and face in the case of women.

3 To avoid multi-collinearity, we calculated alternative models in which the centre-right Christian Democrats (CDU/CSU) were replaced with the liberal democrats (FDP) and the centre-left Social Democrats (SPD) were replaced with the Greens.