Understanding the Religiosity-Political Participation Linkage among Muslim Women: Culturalism or Social Capital?

• 1 Introduction

  Political participation is deemed necessary for the functioning of its liberal democracies and groups being represented. But in this debate, Muslim women are regularly portrayed as submissive and oppressed (Azabar & Thijssen, 2022), and Muslim people in general are stereotyped as the dangerous ‘Other’ who opposes democracy (Cesari, 2014). These notions are in line with culturalist studies, where Islam and religiosity are depicted as a barrier for gender equality attitudes and women’s empowerment, including political participation.
  At the same time, the general political science literature and research in migration studies argue that (Islamic) religiosity is instead a mobilising force for political participation by yielding social capital and resources as such. This stems from a large literature showing that social capital, conceptualised either as attitudes or organisational integration, provides an important driver of social and political engagement (Putnam, 1993; Uslaner, 2003; Van Heelsum, 2005; Vermeulen & Berger, 2008). Given the contrary reasoning of the culturalist logic on the one hand and the social capital logic on the other, it might not be that surprising that studies on the role of Islamic religiosity in shaping women’s political participation in Western Europe show a mixed bag of findings (Azabar, 2023; Joly, 2017; Oskooii & Dana, 2017; Kollar et al., 2023), and in particular how Muslim women’s marginalised social positioning relates to the various forms of their political participation remains unclear (Azabar & Thijssen, 2022; Finlay & Hopkins, 2020; Joly, 2017). However, whether these proposed mechanisms are at play and through which factors they manifest themselves has hardly ever been theorised and tested systematically. This is where this study contributes, doing so for the Dutch case and thus particularly for Muslim women. Such an intersectional focus allows us to shed particularly light on the role of Islamic religiosity (which is gendered) as well as other factors in shaping Muslim women’s political participation. The Dutch case is particularly useful to lay bare which mechanisms are at play as the preconditions for the different mechanisms -theorized below in more detail- to manifest themselves are present.The culturalist approach builds a mechanism for an attitudes-behaviour linkage, and the social capital logic does so foremost via civic integration and skills (Putnam, 1993; Spierings & Vermeulen, 2023; Uslaner, 2003). Bringing these literatures together, we lay bare this theoretical tension and zoom in on gender attitudes, political interest, institutional and social trust, civic integration and collective action attitudes as potentially important pieces of the puzzle in understanding why results on Islamic religiosity and women’s political empowerment are not as clear-cut as ‘common knowledge’ dictates. By combining insights from these literatures, we aim to provide a more fine-grained picture of this issue and empirically assess the often-assumed underlying mechanisms at play.
  Painting this more fine-grained picture is moreover done by disentangling various dimensions of both Islamic religiosity and political participation. Building on the recent acknowledgement of Islamic religiosity being a multidimensional concept, we underline that religiosity has multiple facets (e.g. communal participation, individual affect and individual beliefs) (Ben-Nun Bloom & Arikan, 2012; Glas, 2020), of which some might be tied more clearly among women to one of the mechanisms than the other. Similarly, we theorise these mechanisms in terms of political participation as a concept with different manifestations: electoral (i.e. voting), other institutionalised and non-institutionalised (Van Deth, 2014; Stockemer, 2014). So while we will formulate general hypotheses in line with the dominant literature in the theory section, we will also provide theorisation and empirical testing of whether and how mechanisms play a part across these different dimensions, generating and building a more detailed theory based on the insights from this study.
  All in all, we theoretically contribute to a more fine-grained discussion of the underlying mechanisms that are supposedly at play in the link between Islamic religiosity to (women’s) political participation and empirically address a range of dimensions and factors to paint a more complete picture of the issue at stake. In sum, our central research question reads:

  To what extent are several dimensions of Islamic religiosity related to different forms of Muslim women’s political participation, and is this explained by gender role attitudes, institutional and social trust, civic integration, collective action attitude and/or political interest?

  Empirically, we take a novel multi-study approach, analysing this matter with two different survey datasets in the Netherlands, a case that allows strong comparison to the existing literature and in which gender and Islam are salient in the political context. This allows us to grasp a wider range of indicators to test proposed mechanisms as well as explore the research question across several ethnic groups in the Netherlands and assessing the replicability of the results on independent data within our study. The first, the SIM dataset (Survey Integratie Minderheden, collected in 2020), is somewhat less rich in items included (still covering all core concepts), but it does include a probability sample of a broader group of Muslim women (Iranian-, Moroccan-, Somalian- and Turkish-Dutch) and, thus, allowed migrant women to join the survey in their mother tongue. Second, the MI-ID dataset (Migrant Identity among Young Muslim Adults in the Netherlands, collected in 2019) allows us to include Muslim women’s religiosity and political participation in a uniquely fine-grained manner, and, in particular, it asks about equality attitudes in politics, as well as the other explanations.
  To embed these analyses, we start from an intersectional assessment of Muslim women’s political participation vis-à-vis ethno-religious majority and Muslim men and majority women, based on the SIM data. After this, we move to our main focus and assess the overall relationship between Islamic religiosity and political participation among Muslim women. Lastly, by adding the potentially mediating variables, we assess to what extent found effects between religiosity and political participation (and the lack thereof) can be ascribed to the mediating factors (and their countering each other). We test this simultaneously but separately for the two datasets to provide more robust assessments.

• 2 Contextual and Theoretical Background

  As mentioned earlier, we argue and emphasise that religiosity consists of various dimensions that – especially when specifying underlying mechanisms for its relationship with political participation – are of importance to distinguish. Following the distinctions made in previous research, we are able to study the role of mosque attendance (an indicator of communal practice or behaviour), affective belonging (an indicator of affective belief) and literalism (an indicator of doctrinal beliefs) (Ben-Nun Bloom & Arikan, 2012; Glas et al., 2018; Kollar et al., 2023).
  In addition, we explore these mechanisms’ relationship empirically with different dimensions of political participation: electoral, other institutionalised and non-institutionalised indicators (Azabar, 2023, Jeroense & Spierings, 2023; Van Deth, 2014). While we will not formulate specific hypotheses on these differences as such, we acknowledge these differences and empirically study these separately to provide a more fine-grained picture of the religiosity-political participation linkage among Muslim women. In this way, we explore to what extent certain dimensions of religiosity are more or less important for specific indicators of participation and whether the distilled theoretical mechanisms presented in the following section may be bound to certain indicators of both core concepts only. After having formulated general hypotheses based on previous literature, we therefore discuss the nuances with respect to how specific dimensions of both religiosity and participation may affect these overall expectations. We come back to this in Section 4.2.4 in terms of generating more specific theories based on our results. However, first we will provide a brief sketch of the context in which we theorise the relationships to start with.

  2.1 Muslim Women in the Dutch Context

  Specific systematic quantitative studies on the political participation of Muslim women (or intersectional general studies) are rare (but see Joly [2017], particularly in the context of being a gendered minority). Most studies addressing the religiosity-participation linkage among Muslims do not address the intersection with gender (e.g. Moutselos, 2020; Oskooii & Dana, 2018). This also holds for the Dutch case, even though it is arguably one of the most studied contexts when it comes to political behaviour and attitudes of ethno-religious minorities (e.g. Fennema & Tillie, 1999; Kranendonk & Vermeulen, 2019; Spierings & Vermeulen, 2023; Van Heelsum, 2005). As such, zooming in on the dynamics among Muslim women is in itself valuable, as we know little about specific dynamics for the Dutch case and doing so for the Dutch case implies having access to unique data and existing insight on the political participation of ethno-religious minorities as a reference point.
  Moreover, it is widely known that religious practices and manifestations are gendered (see Glas et al., 2018, for a summarising discussion thereof) and that this has ramifications of political outcomes (e.g. Kollar et al., 2023). For instance, mosque attendance is gendered (Glas et al., 2018) and can thus have differential effects for Muslim men’s and women’s political actions. Similarly, the role of gender equality attitudes might shape women’s actions differently. On the one hand, this further underlines the importance of intersectional analyses of political participation and, on the other, this draws us to the specific setting of the Dutch case, as the setting in a strong and gendered anti-Muslim context contributes to it being a high-profile case in which the mechanisms and complexity discussed in the following can be expected to be relatively visible.
  In the Netherlands, about 5% of the people considered themselves Muslim, which can be traced back mostly to migration from former colonised countries (former Dutch Caribbean, Indonesia), labour migration (e.g. Morocco, Turkey) and political and war refugees (e.g. Iran, Syria), as well as the descendants of migrants and some converts. The different communities have different degrees of organisation with particularly the Turkish community being said to be organised well politically (Fennema & Tillie, 1999; Van Heelsum, 2005; Vermeulen & Berger, 2008) as well as the Islamic schools and broadcasting organisations which are part of the public sphere. One of the manifestations of the political organisation of these groups and the relatively open politics system is that since 2017, the Islamic-inspired political party (DENK) has held seats in parliament, but so far all MPs of the party were men, which does not hold for the local politicians in different cities.
  More generally, the descriptive representation of women, migrant background and Muslim citizens in national parliament has been relatively high in Dutch politics, with particularly minority women being seemingly and relatively better advantaged (Mügge et al., 2024). However, only in 2021, the first Muslim MP wearing a headscarf obtained a seat in the parliament, and Muslim and ethnic-minority women politicians face much more transgressive behaviour, which is a part of the anti-Islam discourse (Mügge et al., 2024). Particularly since 9/11, the politics discourse has focussed on Islam as a marker of otherness, and, ever since Fortuyn’s rise, the anti-Islam discourse in Dutch politics has become strong, and it has been gendered all along, with Muslim women being depicted as victims, oppressed and unemancipated (Roggeband & Verloo, 2007).
  Given the existing organisation structures, the clear presence of a culturalist discourse and the political saliency of the anti-Islamic discourse, as well as the visibility of Muslim(-interest) politics and politicians, the Dutch case is particularly useful to lay bare which mechanisms are at play.

  2.2 Political Participation

  We build on previous definitions and distinctions made with respect to political participation, where often institutional (including both electoral and non-electoral) and non-institutional forms of political participation are distinguished (Van Deth, 2014; Jeroense & Spierings, 2023), also for Muslims in particular (Azabar, 2023). In this article, we will study electoral participation in terms of voting or the intention thereto as well as engagement with (local) politics as indicators of institutional political participation and include a range of indicators for non-institutional political participation, including attending a protest and signing a petition. Importantly, previous research has shown that specific mechanisms such as institutional trust may work differently for institutional political participation than for non-institutional participation. For instance, a lack of institutional trust may hamper institutional participation yet stimulate non-institutional participation (Azabar, 2023). In what follows, we will first derive the main hypotheses on explaining the association between Islamic religiosity and political participation based on the culturalist and social capital logic. After that, we will provide some nuances that can be applied to these general hypotheses.

  2.3 The Link between Religiosity and Political Participation: The Culturalist Logic

  Muslim women are often depicted as submissive and oppressed (Azabar & Thijssen, 2022), in which the culturalist logic assumes that women who are more religious will participate less in politics because of conservative attitudes towards gender roles (Dancygier, 2017), which may suggest women have no role to play in the political arena. Related to this logic, Muslim women have been found to be less interested in politics and therefore arguably participate less in politics. In the following, we will specify the underlying mechanism that assumes this attitudes-behaviour link situating Muslim women’s political participation. We elaborate on this logic reflecting on several dimensions of Islamic religiosity and how this can be tied to political participation.
  When looking at mosque attendance in particular, the context of mosques in Western Europe is often assessed as a place where attendees are exposed to conservative thoughts with respect to gender equality. In addition, as mosques in general remain characterised by patriarchal gender regimes in which power lies with the men in the mosque board to decide on women’s ability to participate (Azabar & Thijssen, 2022), it is argued that women who attend these mosques hold (and are subsequently socialised with) more conservative gender attitudes, also with respect to political participation. As put by Van Klingeren and Spierings (2020), such communal religiosity hampers the adoption of progressive gender equality attitudes among Muslim women in Europe. This is seemingly in line with Read’s (2007) study illustrating that Muslim women who visit mosques often participate in politics to a lesser degree. As such, it can be expected that those Muslim women who visit mosques more often hold more conservative attitudes on gender equality and less (personal) political interest causing them to participate less in politics.
  With respect to affective belonging – the degree to which one identifies with and feels attached to their religion – it can be argued in a similar way that those who more strongly identify as Muslim hold more specific attitudes on gender roles. As most dominant interpretations of (monotheistic) religions, including Islam, contain strong patriarchal norms, we expect that Muslim women who identify more strongly as such are likely to be socialised within these gender norms and values (Burdette et al., 2005; Van Klingeren & Spierings, 2020), in which, for example, young men are encouraged to be interested in politics and young women are not (Azabar & Thijssen, 2022). This link between affective belonging and more conservative gender attitudes is also found in other contexts such as the Middle East (Glas et al., 2018). Subsequently, it can be derived that women who identify more strongly as Muslim hold more conservative gender attitudes and show less (personal) political interest which function as obstacles to political participation.
  Finally, with respect to how literal one’s interpretations of religious texts are, it is supposed that supporting a literalist interpretation is more likely to go hand in hand with rejecting (more progressive) interpretations of Islam beyond the common superficial ones (see Glas, 2020; Koopmans, 2019).. Despite research illustrating that some Muslim women in and beyond Western societies actively dispute an interpretation of Islam that prescribes patriarchal values and practices (Glas, 2020), it is assumed that those who believe that the Quran, Hadith and Sunnah have one meaning tend to follow more conservative and direct interpretations, also with respect to gender roles and participation in democracies. As such, having more literalist interpretations will bring about less interest in politics and more conservative attitudes with respect to gender roles and, therefore, result in less political participation. The hypothesis following the culturalist logic is therefore as follows:

  The more religious Muslim women are, the less likely they are to have more progressive gender equality attitudes (H1a) and political interest (H1b), which decreases political participation in the Netherlands.

  2.4 The Link between Religiosity and Political Participation: The Social Capital Logic

  Seemingly contrarily to the culturalist logic discussed above, previous research offers plenty of examples of Muslim women being politically active in one way or the other (Azabar & Thijssen, 2022; Finlay & Hopkins, 2020; Joly, 2017). These results can be understood from the so-called social capital logic, in which (Islamic) religiosity can be a powerful and mobilising source that promotes political participation, through a range of interlocked elements, most notably summarised in terms of gaining abilities, information (i.e. learning), being more mentally involved with a group and society, and the active recruiting via religious networks (e.g. Kranendonk & Vermeulen, 2019; Moutselos, 2020; Spierings & Vermeulen, 2023; Verba et al., 1995), where the focus in this study¹x For reasons of empirical availability. is on civic integration and mentality (e.g. interest, trust, collective action aptitude).
  Being more integrated within a (religious) community that has specific interests is likely to bring about political interest and collective action attitudes (Geurts & Van Klingeren, 2023), whereby political participation is seen as a way to voice and act on these to improve the position of the group in society, that is, Muslims. Similarly, institutional trust and social trust are viewed as a result of embeddedness within a civic organisation, such as a religious network, that brings about civic skills and political information and mentality (Oskooii & Dana, 2018), whereby the integration into social networks and organisation beyond the religious network can be a result of the latter (e.g. Putnam, 1993) and thus also further foster these elements of social capital. Integration in social organisations, political interest, collective action attitudes, and institutionalised and social trust are therefore often used as indicators of socio-political attitudes that increase political participation.
  Are these mechanisms reflected in or connectable to the literature on Muslim women’s political participation and the different religious dimensions particularly? Mosque attendance has been shown to feed into broader integration on civic organisations, to be a place for recruitment and to bring about knowledge and duties with respect to the electoral system among its attendees as well as be a source of institutional trust and feelings of efficacy that stimulate political participation (Azabar & Thijssen, 2022; Calhoun-Brown, 1996; Jones-Correa & Leal, 2001; Sobolewska et al., 2015). These expectations apply and have been tested on both more institutionalised forms of participation, like voting (Moutselos, 2020), and non-institutionalised forms of participation, like protesting and signing petitions (Dana et al., 2017; Jamal, 2005). Moreover, previous research has shown that mosque attendance can amplify group consciousness and consciousness of the position of Muslim people in society and, accordingly, the wish for a change for their group, which is likely to stimulate their political participation (Oskooii, 2016; Azabar, 2023).
  Looking at affective belonging, previous research suggests that identifying (more) as Muslim can also relate positively to political participation as Muslim faith is perceived to be an important driver for active citizenship and, as both a manifestation and consequence thereof, and, therefore, political participation (Harris & Roose, 2014; Just et al., 2014; Peucker, 2018). Peucker (2018) moreover concluded that among those who more strongly identify as Muslim, such political engagement is both a religious duty and an act of serving humanity, going hand in hand with political interest to do so. This also applies to one’s attitude towards the position of Muslims in society and wanting to create a positive image of the group, which is more likely to be present among those who identify more strongly as Muslim, and subsequently triggers political participation to bring about such change (Geurts & Van Klingeren, 2023, but see Kollar et al., 2023). A similar expectation can be formulated with respect to trust, where those identifying more strongly to a societal group will bring about more social and political trust (Bègue, 2002; Wisneski et al., 2009). Love for and acceptance of others are pro-social principles that are at the core of Islam, suggesting that identifying as Muslim more will relate to having a trusting disposition. We therefore expect that affective belonging relates positively to political participation because of the experienced trust, collective action attitude and political interest to participate in such a way.
  Finally, with respect to literalism, as an indicator of religious beliefs, an association is less likely to be expected. For the Middle East, trust has been linked to literal readings of the Quran via arguments to do good (see Spierings, 2019), and some have argued political engagement is a religious duty (e.g. with respect to environmentalism, as studied by Hancock, 2020). However, these interpretations strongly depend on the way texts are interpreted. Existing theories and empirical studies provide too little information, particularly so in the contexts of migration and of being a minority religion, to argue for clear unidirectional effects on political participation of literalism via social capital.
  All in all, it thus seems reasonable from a social capital perspective to connect religiosity, social capital indicators and political participation positively, across each of the dimensions:

  The more religious Muslim women are, the more likely they are to experience trust (H2a), political interest (H2b), civic integration (H2c) and collective action attitude (H2d) which increases political participation in the Netherlands.

  2.5 Nuances of the Main Expectations

  The foregoing paragraphs already hint that a certain level of diversity in the main associations is to be expected. A core one here is the variation in political activities. For instance, when discussing the social capital logic, the institutional trust mechanism may more strongly apply to institutional forms of participation, where a certain level of institutional trust is deemed necessary for such activities (Spierings & Vermeulen, 2023; Jeroense & Spierings, 2023). For non-institutional political participation, especially for activities that are against the (political) system, that is, protesting, one may argue that this is less strongly affected by institutional trust and political interest. At the same time, a lack of institutional trust can also stimulate this participation (Azabar, 2023), whereas social trust is arguably vital in participating in all these activities.
  With respect to the culturalist logic, we can expect the mechanisms to work less strongly for non-institutional forms of participation for which certain gender role attitudes may be more blurry and where political interest is less of a necessity in order to participate. Additionally, it has been shown that relatively few factors drive electoral participation in first-order elections as there is still a norm of participation and the costs are relatively low (see Jeroense & Spierings, 2023; Lefevere & Van Aelst, 2014).
  These considerations and nuances on the different facets of political participation and how they might link differently to religiosity and provided mechanisms will be explored empirically by explicitly testing the hypotheses separately for electoral, other institutional and non-institutional forms of political participation.

• 3. Data and Methods

  3.1 Multi-study Approach: SIM 2020 and MI-ID 2019

  Instead of arguing why one dataset is superior to the other, we decided to take a multi-study approach as on some fronts one dataset provides better data and on others the other does so. By using two datasets, we reap the benefits from both while at the same time offering a more rigorous test of the associations at hand. What the datasets used here share is that they include data on Muslim women in relatively large numbers from one and the same context (the Netherlands) and that they include data that allow us to distinguish between individual and collective religiosity dimensions and rather uniquely include religious literalism and other items to distinguish between voting, other institutional activities and non-institutional political activities. Moreover, they need to include information on at least several of the potentially mediating factors of interest. Both the most recent (2020) SIM dataset (SCP-CBS, 2022) and the 2019 MI-ID dataset(Spierings & Van Klingeren, 2020; see also Peelen et al., 2022) meet these criteria. Other datasets such as the Dutch Ethnic Minority Elections Study or the Dutch Eurislam data include no literalism and, respectively, too few Muslim women, and only two mediators or no political participation items beyond voting.
  The SIM (Survey Integration Minorities) 2020 oversampled several ethnic minorities in the Netherlands, including several in which Muslim people are the predominant religious group, and did so based on random sampling from the population registry. Questions could be answered in both Dutch or the language native to or dominant in those groups. We selected all women who self-identify as Muslim, leading to starting subsamples of 96 Iranian, 270 Moroccan, 285 Somalian and 326 Turkish women (total: 977). Particularly the presence of Iranian and Somalian women is noteworthy as most Dutch ethnic-minority data focus on Moroccan- and Turkish-background respondents when it comes to Islamic religiosity. While SIM provides randomly sampled and dual-language data, it does not include information on interpersonal trust and collective action attitudes, and the measurement of political activities is more restricted than in MI-ID.
  The MI-ID (Migrant Identity among Young Muslim Adults in the Netherlands) data have been collected in 2019 and include adult self-identified Muslim respondents between 18 and 45 years of age (~Generation Y and Z), who were younger than 13 upon arrival or who are (grand)children of migrant people, with Moroccan and/or Turkish roots. The dataset is richer than any other in terms of the items included with regard to political participation (various items per category), religiosity (including literalism) and the theorised mediators, which include multiple gender attitudes, trust and strategies on making change. As it is not a random sample, but a (diverse) convenience sample, we should not use the data to provide descriptive point estimates, but it is suitable for assessing relationships, particularly given the richness of the data. The number of Muslim women in MI-ID is 232.
  Both datasets are thus collected in the same context, include the same control variables and allow us to test the same religious dimensions and multiple mediating variables. At the same time, the sample and interview methods differ as do some variables. If we find the same results, this can be considered strong proof for a relationship to exist. As results differ, we will reflect on the specific differences and discuss whether these differences might be attributed to difference between the models and datasets. We propose that it is better to have these differences within an article, including the discussion there of, presenting two articles with different results which are not related to each other. Regarding the differences in samples, we also reran the analyses on the SIM data, only including Moroccan- and Turkish-Dutch Muslim women who were either born in the Netherlands or who were born abroad but migrated together with their parents, which by and large reproduces the MI-ID population.

  3.2 Dependent Variables: Political Participation

  Based on the literature and factor analyses of the political participation items, we distinguish between three types of participation: voting, (other) institutional and non-institutional. The descriptions of all variables are presented in Table 1.
  In the SIM data, two variables on voting, or the electoral part of institutional participation, were present: whether people voted during the most recent parliamentary elections and whether they would vote if elections were to be held today. Both included an answering option that the respondent was not eligible to vote, such that these respondents were excluded from the voting models. The other two variables are analysed separately, as often the intention to vote exists across groups relatively more strongly while the rate of actual act of voting varies more. In MI-ID, it was asked whether respondents would vote if elections were held today, including the explicit ‘would not vote’ option. This variable is used to create a dichotomous outcomes variable indicating the intention to vote (or not). All three outcome variables discussed here are dichotomous. In order to make model comparison easier, we estimated linear probability models (LPM), but the main models have been rerun as logistic regression models to check whether the relations hold in those too. These would not lead to other conclusions.
  For institutional participation beyond voting, SIM includes two items that were combined in one scale (0-2): whether someone did or did not contact a politician or political party and whether or not they partook in a political hearing or discussion meeting, both in the last two years. In MI-ID, a broad list of activities was included of which it was asked whether people had done so at least once in the last five years. Five of these are combined as an institutional participation index (0-5) here: contacting a politician or civil servant; visiting a municipal participation evening; member of a political participation; attended a meeting of a political party; and participation in a neighbourhood citizens’ initiative. Based on the discrete nature and distribution of both the variables (many respondents scoring 0), negative binomial models are most fitting. Accordingly the core models have been run with such specification; however, for the purpose of model comparison and to improve the interpretability of results, we present standard OLS models as they show the same relations and lead to the same conclusions.
  Last, non-institutional participation could only be measured with one item in SIM: whether one participated in a demonstration or protest march in the last two years (no or yes). In MI-ID, however, four different items fitted this concept and loaded on one factor in an exploratory factor analysis: signed a petition; posted or shared political messages on social media; discussed politics on social media; active in an action group. Based on this an index running from 0 to 4 was created. Regarding estimating LPM instead of logistic and OLS instead of negative binomial models, the same arguments and observations apply as given earlier.
  For both institutional and non-institutional participation, only those questions in SIM were included in split-half A, which reduces the sample size and makes it impossible to link them to literalism with SIM.

  Table 1 Descriptions of Variables Included

  SIM data			MI-ID data
  Variable	Mean	Std. Dev.	Variable	Mean	Std. Dev.
  Political participation
  Electoral participation: voting (0-1)	0.66	0.48
  Electoral participation: vote intention (0-1)	0.84	0.30	Electoral participation: vote intention (0-1)	0.87	0.34
  Institutional participation (0-2)	0.05	0.17	Institutional participation (0-5)	0.54	0.75
  Non-institutional participation (0-1)	0.05	0.22	Non-institutional participation (0-4)	0.81	0.96
  Islamic religiosity
  Communal attendance (0-4)	1.0	1.1	Communal attendance (0-6)	3.4	1.6
  Affective belonging (0-4)	3.2	0.8	Affective belonging (0-9)	6.9	1.4
  Literalism (0-4)	2.6	1.2	Literalism (0-4)	2.8	0.9
  Mechanisms
  Gender equality: financial roles in household (0-10)	7.5	2.2	Gender equality: political leadership (0-4)	1.2	0.9
  			Gender equality: perceived gender discrimination (0-9)	4.7	2.0
  Political interest: interest domestic/political news (0-3)	2.2	0.7	Political interest: domestic/political news consumption (0-5)	4.1	1.3
  Trust: Institutional (0-10)	5.8	1.9	Trust: Institutional (0-10)	5.4	1.9
  			Trust: Interpersonal (0-10)	5.5	2.0
  Active in civic organisations (0-2)	0.59	0.74
  			Collective action strategy (0-10)	7.8	2.1
  Control variables
  Education			Education
  Low	0.15		Low	0.10
  Middle	0.30		Middle	0.67
  High	0.31		High	0.23
  Unknown	0.23
  Highest education in country of birth (0-1)	0.04	0.19
  Paid employment (0-1)	0.45	0.50	Paid employment (0-1)	0.33	0.47
  Born in Netherlands (0-1)	0.34	0.47	Born in Netherlands (0-1)	0.93	0.26
  Age: groups (0-6)	1.6	1.6	Age: in years	25	5.4
  Country of origin (of parents)			Country of origin (of parents)
  Morocco	0.28		Morocco	0.56
  Turkey	0.33		Turkey	0.42
  Somalia	0.29		Both	0.02
  Iran	0.10
  			Treatment effect of MI-ID design (for collective action)
  			Positive news	0.31
  			Negative news	0.42
  			Unrelated news	0.17

  3.3 Independent Variables: Religiosities

  As mentioned in the theoretical section, religiosity is a multidimensional concept, and so we should be careful not to conflate different dimensions. The focus in this article is on three dimensions: the affective dimension of individual belonging; the collective behavioural dimension of service or mosque attendance; and the individual beliefs dimension, in particular whether people hold a more literalist view on how the Quran and other religious sources should be interpreted. Based on this conceptual-theoretical focus and a series of exploratory factor analysis and scale reliability analyses, we create three different variables in both datasets.
  First, affective belonging to one’s religion in SIM is measured by an index of three items that are conceptually related to this and have a Cronbach’s alpha of 0.78 (which would drop if any of the items were left out): ‘My belief is an important part of me’; ‘It hurts if someone says something bad about my religion’; ‘That I am Muslim is something I often think about’ (answering options: fully agree, agree, neither agree nor disagree, disagree, fully disagree). The simple average of the items was taken (as long as two or three were available)²x The metrics of the scale remains the same (e.g. the mean), but more cases could be included. and linearly rescaled to run from 0 to 4, a higher score indicating stronger affective belonging. In MI-ID, the measurement was based on two items tapping the same concept: ‘How religious would you say you are?’ (0-9) and a question selecting one of four statements about how strong one feels connected to their Muslim identity (response options: not, somewhat, strongly, very strongly). Both items are rescaled to run from 0 to 9 and the average is taken.
  Second, religious communal attendance in SIM was asked by the question: ‘How often do you go to a religious meeting?’ Daily, weekly, monthly, a few times a year, never, or less frequent than yearly. This scale was inverted to run from 0 to 4. In MI-ID, it was asked, ‘How often do you go to a place of worship for a religious service?’ The response options included daily, weekly, several times per month, monthly, several times per year, only on special occasions, never. This was also included as a scale running from 0 to 6. In both cases, a higher score indicates more frequent attendance.
  Third, literalism is included in both surveys with one item too. In SIM, however, the items were part of the split-sample design, so only half of the respondents were (randomly) asked to answer this question. The item reads, ‘Every Muslim is allowed to interpret the Quran as they prefer’ (response options: fully agree, agree, neither agree nor disagree, disagree, fully disagree). A lower score indicates more freedom of interpretation, and, although disagreement does not per se indicate a specific interpretation, this item has been used to tap literalism as particular clerics and people arguing to follow the sources literally are those who do not believe in divergent interpretations. The variable runs from 0 to 4. In MI-ID, the literalism item – which was the basis of the SIM item – reads as follows: ‘Quran verses and the Hadith texts may be interpreted by anyone in their own way’ (1 = completely disagree to 5 = completely agree), which was recoded from 0 to 4, with a higher score being more literalist.
  Sometimes, there is a worry about overlap and multicollinearity between these dimensions, but, as seen in previous research, in our sample, too, there is a clear variation between the dimensions. The correlations between the three items in both datasets vary from 0.09 to 0.32, with only ‘belonging’ and ‘attendance’ in MI-ID correlating higher (0.51), but additional multicollinearity analyses of the models presented in the following section do not show VIF scores that even come close to 2.5 or higher.

  3.4 Mediators: Gender attitudes, Social Activities, Political Interest, Trust and Collective Action Strategies

  SIM2020’s full sample includes four typical 5-point Likert items on gender role attitudes, all focusing on financial roles and power relations in the household. These four have a Cronbach’s alpha of 0.79, and that would drop if any of the items were left out. We combined them in one scale, if at least three items were answered (see Note 2), by taking the simple average and linearly rescaling that to run from 0 (least supportive of financial gender equality) to 10 (most supportive). The items are: ‘The husband can best be responsible for the money’; ‘For boys it is more important to earn their own money than for girls’; ‘Household decision on big expenditures can best be made by the man of the household’; ‘Women should stop working when they become a mother.’ All had the same answering options: fully agree, agree, neither agree nor disagree, disagree, fully disagree. While we will not claim that our measurement is representative of all types of gender role attitudes, the dimension we focus on does tap the issue at the core of gender equality and likely to be strongly correlated to other dimensions (not included in SIM).
  In the MI-ID data, Muslim women’s position on gender equality and women’s empowerment could be included by two different items, arguably more closely linked to the mechanism focussed on here compared to the SIM data. First, the standard item from comparative survey on political gender equality was used: ‘men are better political leaders than women’ (response options ranging from fully disagree to fully agree), which was coded to run from 0 to 4, with the higher score implying that men are not better political leaders. Second, we included Muslim women’s perception of gender discrimination: ‘How often would you say the following groups are discriminated against in the Netherlands: women?’ (0 = never to 9 = very often).
  Next, we included whether people were civically active in social or intermediary organisations as a classic indicator of social capital (Putnam, 1993). This was included in SIM only and could be measured combining two yes/no items: ‘Have you volunteered in the last 12 months?’ and ‘Did you visit a meeting of an association or organisation in the last 12 months?’ Adding them up (0 = no; 1 = yes) leads to a scale running from 0 to 2 (Cronbach’s alpha = 0.54).
  For political interest, both surveys include items on domestic and political news. In SIM, it was asked to what degree people are interested in domestic news/news on the Netherlands, after a linear recode running from not at all (0) to very interested (3). Also this item was part of the split-sample design. Unfortunately, in the other half on the literalism item, with SIM we cannot test the potential mediation of a literalism effect by political interest. In MI-ID, the focus was also on political and domestic news, more particularly news consumption: ‘How often do you use Dutch media sources (newspapers, online, television, radio) to get information about current affairs and politics?’ (response options: daily, weekly, several times per month, monthly, several times per year, never), which was recoded to run from 0 to 5, with 5 being the highest interest score. The MI-ID data overrepresented the politically interested; yet, 44% of the Muslim women did not consume some news daily.
  As the fourth form of mediation, we focus on trust. In SIM, a question on trust in institutions was asked referring to three institutions separately (none on interpersonal trust): government, police and judges (response options: completely no trust, no so much trust, much trust, very much trust). We combined the three items (Cronbach’s alpha = 0.82; this would not increase by dropping items) to create a scale running from 0 to 9, with a higher score indicating more trust. This variable was only part of the split-half sample as was the case with political interest. In MI-ID, both institutional trust and interpersonal trust were present. On institutional trust, we included three items directly focusing on political institutions: ‘How much trust do you have in each of the following institutions—the Dutch government, the House of Representatives, the city council of your municipality?’ (1 = no trust at all to 10 = complete trust). We took the average of these three items and rescaled it to run from 0 to 10. For interpersonal trust, the standard item – ‘Do you think that in general most people can be trusted (10) or that you cannot be too careful in dealing with people (1)?’ – was rescaled to run from 0 to 10 as well.
  Finally, in MI-ID also collective action attitudes were included in line with the measure proposed and used by Geurts and Van Klingeren (2023). Four items are combined into a scale running from 0 to 10, with a higher score indicating more preferences for a collective action strategy to deal with the position of Muslim people in Dutch society. The items: ‘I would like to create more chances with other Muslims for Muslims in society’; ‘I would like to actively contribute to giving Muslims a more positive image’; ‘I feel the need to explain people that veils are no sign of oppression’; and ‘I feel the need to explain people that Muslims are in general peaceful.’ Empirically the four fit one latent concept (Geurts & Van Klingeren, 2023) and theoretically all refer to positive collective actions to change the larger society’s view of Muslim people. As the items were part of the outcome variable of a vignette study, we added the treatment (different newspaper articles) as additional control variables.

  3.5 Control Variables

  The models are controlled for variables that are related to our core independent variables and the outcome variables, while being unlikely of mediating the effects between them. This prevents us from making either Type I or Type II errors. The controls included: highest education level in three categories, and for SIM including a category ‘don’t know’ (by using dummified variables, including a dummy for respondents with missing values, we prevent the loss of a selective part of our sample); location of highest education (SIM, Netherland or abroad); employment status (paid employee or not); age as an interval variable with eight categories, as provided in SIM (0-7) or in years (MI-ID); country of origin (Morocco, Turkey, Somalia [SIM only], Iran [SIM only] or Morocco and Turkey [MI-ID only]); and migration generation (descendant of migrant parent(s), that is, second generation; or migrant, that is, first generation).

• 4 Results

  4.1 Descriptive Results

  Based on the SIM data, we first chart the degrees of political participation of Muslim women, and we do so in a comparative intersectional perspective, also presenting figures for the ethno-religious majority members (for SIM) and men (Figures 1 and 2).
  The major differences in political participation between the minority and majority citizens are found for electoral participation (particularly the actual vote), with minority citizens participating less in the electoral process. This is not found consistently for the other political participation variables. Moreover, for electoral participation, neither of the datasets shows any gender gap.
  For other institutional and for non-institutional participation, an inter-categorical intersectional approach proved important. For instance, among men, participation is higher in the minority sample compared to the majority sample, but not among women. Also, in the SIM minority sample, women’s other institutional and non-institutional participation is lower than that of men, which we also see for other institutional participation in the MI-ID data. In those data, however, Muslim women are shown to be more active than Muslim men. A closer look suggests that social media activities and petitions are mainly responsible, not action groups, which are closer to the SIM item on protest activities.
  Altogether, ethno-religious minority women with a Muslim background do seem to participate less in institutional politics than other groups and ethno-religious minority citizens with a Muslim background participate less in elections than majority members, but in some non-institutional activities a gender gap or ethno-religious gap is less evident. In what follows, we zoom in on which factors drive Muslim women’s political participation, focusing on religiosity and cultural attitudes linked to gender and social capital.

  

  

  4.2 Explanatory Results

  In the following sections, per outcome (and per dataset), we first estimate the overall role of the different dimensions of religiosity, after which we included the culturalist and social capital indicators. Where relevant, the models on the SIM data also present the total models rerun on only the Moroccan- and Turkish-background women raised in the Netherlands, which increases the comparability with the MI-ID data, although this does not always explain the found differences.

  4.2.1 Voting

  For Muslim women’s electoral participation, Table 2 shows mixed results regarding the relationships with religiosity (Models A1-C1, A8-B8). Of the thirteen coefficients, seven are positive and five are negative, and only one is statistically significant (p < 0.05): affective belonging for intention to vote in MI-ID.
  The absence of an overall relationship (also per religiosity dimension) might be due to counteracting mechanisms, nullifying each other in the overall models. If so, the models including the additional variables would show this. However, regardless of the mediating variables included, the coefficients for the religiosity variables hardly change. Social capital and culturalist mechanism neither suppress nor explain (the minimal) effects of religiosity on electoral participation among Muslim women in the Netherlands. This holds true for both the datasets and their respective models.
  In themselves, some of the factors we identified as part of the social capital and culturalist logic do influence Muslim women’s electoral participation though. Most particularly, we find clear results for political interest (manifested in news consumption) being positively related to electoral participation (p < 0.05 or p < 0.10), but only in the SIM data and less clearly so when we restrict the SIM data to the sample similar to the MI-ID data. News consumption seems to feed into more electoral participation among Muslim women, but not so in the case of Turkish and Moroccan women who grew up in the Netherlands, although news consumption does increase their vote intention.

  4.2.2 Other Institutional Participation

  Muslim women’s other institutional participation – the least frequent form of participation – is more clearly linked to Islamic religiosity (Table 3): both datasets show a clear positive relationship for attendance; no effect is found for literalism; and, in MI-ID, a clear and consistent positive effect is found for affective belonging (Models A1, A4, B1). Overall, we find no clear negative effects, and a stronger (organisational or affective) connection with one’s religions is found to mobilise women’s participation in institutional politics.
  The models in Table 3 show that even though there is some indication that attitudes towards women’s position have a link with participation (B8), this does not explain the religiosity-participation linkage (cf. Models B2-B3 with B1). The culturalist logic can thus be refuted for other institutional participation.
  More support is found for the social capital mechanism. In both datasets, political interest is positively related to Muslim women’s institutional participation (although its p-value turns 0.13 in Model A7, in an alternativ model without civic activities it is p < 0.05); in the MI-ID data also, a positive relationship for interpersonal trust was found; and, in the SIM data, it can be seen that Muslim women’s participation in civic activities relates to higher institutional participation. Initial positive effects were also found for institutional trust and collective action strategies. These, however, disappear when the other social capital variables are added. Moreover, and importantly, particularly adding civic activities, interest and trust substantially decreases the coefficients for attendance and/or belonging, even to the degree that they are not statistically significant anymore in the final model (and Models B5 and B6). If anything, we find clear indications that Islamic religiosity fosters Muslim women’s participation in institutional politics via increase in social capital, civic, activities, interpersonal trust and political interest most clearly).

  4.2.3 Non-institutional Participation

  Last, Table 4 on Muslim women’s non-institutional participation shows partly different dynamics, whereby it should also be noted that the results for the two datasets are different to the degree that they lead to different conclusion. Given the more consistent results for the other forms of political participation, at last in main terms, this is mostly likely related to the participation items included varying considerably.
  In both datasets, clear negative relationships between religiosity and participation are found. In the MI-ID data, a strong negative linkage between attendance and participation was registered (Model B1) and in the SIM data one for affective belonging (Model A1, A3). No effect for literalism is found (B1), and the linkage with attendance in SIM (A1) is actually positive, and the difference for attendance is even stronger when looking at Model A7 which replicated the MI-ID sample in the SIM data. Linking this to the difference in the outcome variable, we surmise that attendance might be a mobilising agent for protest (i.e. people are recruited at religious meetings), but not for the more individual and online acts central to the MI-ID data. Indeed, if we rerun the MI-ID models with only social media activities/petition signing and joining in an action group separately, it shows the former fully drives the negative relationship with attendance.
  Across the models, few indications are found that either the culturalist indicator or the social capital indicator explains the linkage with religiosity. Only in Models B6 and B8 we see indications that interpersonal trust partly links mosque attendance and non-institutional participation. More particularly, it seems that mosque attendance links positively to interpersonal trust and more trusting people seem to avoid these individual non-institutional activities (while trust fosters institutional participation, see Table 3). It seems religiosity fosters social capital and this steers Muslim women towards engaging in politics via institutional activities like party membership and meetings and contacting politicians, while (thus) steering them away from individualist and often anonymous non-institutional participation.

  4.2.4 Islamic Religiosity and Political Participation among Muslim Women

  The sceptical interpretation of the divergent results discussed earlier could be that the relationship expected should simply be refuted as mere chance might lead to sometimes finding a (statistically) significant relationship. However, the number of relationships found and the patterns in them do align with some insight from the literature, like electoral participation in first-order elections leading to relatively few individual-level results because so many people have the intention to vote (Lefevere & Van Aelst, 2014), and that non-institutional activities are partly fuelled by distrust (Jeroense & Spierings, 2023). What does this all imply for our hypotheses?
  Hypotheses 1a and 1b can be squarely refuted. Even though we found some linkages between some gender attitudes and political interest and some dimensions of political participation, there was no single instance in which this (partly) explained the negative effect of religiosity indicators.³x Or increased the positive effect, signalling suppression effects which would also confirm a mechanism at work as hypothesized by culturalists. This conclusion holds across all dimensions of political participation and all dimensions of religiosity; even in literalism, which is often linked to traditionalism.
  For Hypotheses 2a to 2d, the results were somewhat more complex. For electoral participation and non-institutional participation, Hypotheses 2a to 2d are refuted. Regarding the former, the relative importance and low costs of participating in parliamentary elections seem to suppress the expected mechanisms. Regarding the latter, we did find a linkage between religious attendance and non-institutional participation via interest, but in the opposite direction as expected in H2b. Religiosity does not decrease trust, it only increases it, but interpersonal trust steers Muslim women away from individualist non-institutional politics (as discussed in our theoretical section on potential nuances; see also Azabar, 2023). Lastly, our results suggest that for institutional participation, interpersonal trust (H2a), political interest (H2b) and civic activities (H2d) explain why more frequently attending and more affectively religious Muslim women are more likely to participate politically. No such effect was found for doctrinal literature, which is in line with the our observation in the theory section that the link between literalism and social capital is less evident. For institutional participation, we largely corroborate our second hypothesis.
  Taking one step back, we can conclude that these results have several implications for understanding the religiosity-political participation linkage: (1) Islamic religiosity only matters for the more selective forms of political participation, and not the cheap and salient activity of voting in national elections; (2) as for religiosity, only social integration and affective integration matter, and the beliefs dimension in the form of literalism do not; (3) participation is not linked to religiosity via gender equality attitudes, even though behavioural motivations of Muslim women are often culturalised; the connections run via aspects of social capital; and (4) social capital is fostered by religiosity dimensions of an integrative nature and this mobilises Muslim women in institutional politics, but withholds them (particularly via higher trust) from engaging in non-institutional activities, particularly if these are online, individual and anonymous, suggesting the latter are a fall-back option when institutional routes are not considered viable.

  Table 2A Regression analysis results for the link between electoral participation and religiosity among Muslim women, Including Potential Mediators – Vote (SIM)

  	1	2	3	4	5	6	7	8	9	10	11
  	Religiosity: Total sample	Mediator 1	Mediator 2	Religiosity: Split-half sample A	Mediator 3	Mediator 4	Mediators 1-4	Religiosity: Split-half sample B	Mediator 1	Mediator 2	Model 6 on Gen1.5 up Turkish- and Moroccan-Dutch
  Attendance	0.03 (0.11)	0.03 (0.08)	0.02 (0.24)	0.03 (0.23)	0.04 (0.13)	0.03 (0.22)	0.04 (0.09)	0.02 (0.47)	0.02 (0.42)	0.02 (0.55)	0.00 (96)
  Affective belonging	0.01 (0.55)	0.01 (0.53)	0.01 (0.58)	0.00 (0.94)	−0.01 (0.79)	0.00 (0.92)	−0.01 (0.79)	0.01 (0.70)	0.01 (0.68)	0.01 (0.69)	−0.01 (0.88)
  Doctrinal literacy								−0.01 (0.60)	−0.01 (0.69)	−0.01 (0.62)
  Gender attitudes: financial		0.02 (0.09)					0.02 (0.74)		0.02 (0.14)		−0.01 (0.78)
  Active in civic organisations			0.06 (0.03)				0.05 (0.21)			0.02 (0.62)	0.05 (0.46)
  Political interest: domestic news					0.10 (0.02)		0.10 (0.04)				0.02 (0.77)
  Institutional trust						0.01 (0.77)	−0.00 (0.85)				−0.00 (0.97)
  Control variables	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.
  Adjusted R2	0.16	0.16	0.16	0.13	0.15	0.13	0.15	0.19	0.19	0.18	0.09
  N	657	657	657	257	257	257	257	296	296	296	117

  Table 2B Vote Intention (SIM)

  	1	2	3	4	5	6	7	8	9	10	11
  	Religiosity: Split-half sample A	Mediator 1	Mediator 2	Religiosity: Split-half sample A	Mediator 2	Mediator 3	Mediators 1-3	Religiosity: Split-half sample B	Mediator 1	Mediator 2	Model 6 on Gen1.5 up Turkish- and Moroccan-Dutch
  Attendance	0.01 (0.58)	0.01 (0.48)	0.00 (0.71)	0.01 (0.62)	0.01 (0.50)	0.01 (0.54)	0.01 (0.52)	0.00 (0.89)	0.00 (0.86)	−0.00 (0.96)	−0.02 (53)
  Affective belonging	−0.01 (0.44)	−0.01 (0.51)	−0.01 (0.43)	−0.03 (0.17)	−0.04 (0.11)	−0.03 (0.2)	−0.03 (0.14)	0.00 (0.97)	0.00 (0.91)	0.00 (0.95)	−0.01 (0.76)
  Doctrinal literacy								0.00 (0.99)	0.00 (0.97)	0.00 (0.99)
  Gender attitudes: financial		0.01 (0.02)					0.01 (0.53)		0.01 (0.25)		0.01 (0.68)
  Active in civic organisations			0.01 (0.37)				0.02 (0.33)			0.02 (0.40)	0.01 (0.84)
  Political interest: domestic news					0.06 (0.02)		0.05 (0.06)				0.08 (0.10)
  Institutional trust						0.01 (0.17)	0.01 (0.41)				0.00 (0.92)
  Control variables	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.
  Adjusted R2	0.04	0.05	0.04	0.03	0.05	0.03	0.04	0.01	0.01	0.01	0.07
  N	739	739	739	283	283	283	283	329	329	329	130

  Table 2C Vote Intention (MI-ID)

  	1	2	3	4	5	6	7	8
  	Religiosity	Mediator 1a	Mediator 1b	Mediator 2	Mediator 3a	Mediator 3b	Mediator 4	Total model
  Attendance	0.01 (0.40)	0.01 (0.43)	0.01 (0.43)	0.01 (0.41)	0.02 (0.34)	0.01 (0.49)	0.01 (0.40)	0.01 (0.41)
  Affective belonging	0.05 (0.01)	0.06 (0.01)	0.06 (0.01)	0.05 (0.01)	0.06 (0.01)	0.05 (0.01)	0.06 (0.00)	0.06 (0.01)
  Doctrinal literacy	−0.02 (0.49)	−0.02 (0.53)	−0.02 (0.51)	−0.02 (0.53)	−0.02 (0.45)	−0.02 (0.51)	−0.01 (0.68)	−0.01 (0.68)
  Gender attitudes: politics		0.01 (0.61)						0.01 (0.85)
  Gender attitudes: women are discriminated			−0.00 (0.71)					−0.00 (0.91)
  Political interest: domestic news				0.01 (0.44)				0.02 (0.36)
  Institutional trust					−0.01 (0.58)			−0.02 (0.23)
  Interpersonal trust						0.01 (0.59)		0.02 (0.20)
  Collective fight strategy							−0.02 (0.14)	−0.02 (0.10)
  Control variables	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.
  Adjusted R2	0.07	0.07	0.07	0.07	0.07	0.07	0.07	0.06
  N	195	195	195	195	195	195	195	195

  bold in p < 0.05; reported: B coefficient (p-value); estimated as linear probability models; controlled for level of education (SIM,MI-ID), having finished education in country of origin (SIM), migration generation (SIM, MI-ID), age (SIM, MI-ID), country of origin (SIM), ethnic group (MI-ID), employment (SIM, MI-ID), treatment group, in the model with the collective strategy (MI-ID).

  Table 3 Regression analysis results for the link between Institutional Political Participation electoral participation and religiosity among Muslim women, Including Potential Mediators

  	1	2	3	4	5	6	7	8
  A: Institutional Participation: Contact Party and Attend Meetings (SIM)
  	Religiosity: Total sample	Mediator 1	Mediator 2	Religiosity: Split-half sample A	Mediator 3	Mediator 4	Mediators 1-4	Model 6 on Gen1.5 up Turkish- and Moroccan-Dutch
  Attendance	0.03 (0.00)	0.03 (0.00)	0.02 (0.01)	0.04 (0.00)	0.04 (0.00)	0.04 (0.00)	0.03 (0.00)	0.04 (0.02)
  Affective belonging	−0.01 (0.59)	−0.01 (0.61)	−0.01 (0.51)	−0.01 (0.63)	−0.01 (0.49)	−0.01 (0.55)	−0.01 (0.37)	−0.02 (0.38)
  Doctrinal literacy
  Gender attitude: financial		0.00 (0.32)					0.00 (0.52)	0.00 (0.75)
  Active in civic organisations			0.06 (0.00)				0.07 (0.00)	0.07 (0.01)
  Political interest: domestic news					0.03 (0.08)		0.03 (0.13)	0.03 (0.31)
  Institutional trust						−0.01 (0.09)	−0.01 (0.13)	−0.01 (0.38)
  Control variables	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl
  Adjusted R2	0.03	0.03	0.09	0.04	0.05	0.05	0.11	0.07
  N	448	448	448	333	333	333	333	149
  B: Institutional Participation: Contact, Party Meetings, Party Member, Neighbourhood Initiative (0-100) (MI-ID)
  	Religiosity	Mediator 1a	Mediator 1b	Mediator 2	Mediator 3a	Mediator 3b	Mediator 4	Full model
  Attendance	0.08 (0.03)	0.09 (0.01)	0.09 (0.02)	0.08 (0.02)	0.06 (0.11)	0.05 (0.19)	0.07 (0.05)	0.05 (0.16)
  Affective belonging	0.13 (0.01)	0.12 (0.01)	0.12 (0.01)	0.10 (0.02)	0.12 (0.01)	0.10 (0.03)	0.09 (0.05)	0.05 (0.26)
  Doctrinal literacy	0.01 (0.87)	0.00 (0.99)	0.01 (0.94)	0.02 (0.71)	0.02 (0.67)	0.01 (0.82)	−0.02 (0.69)	−0.01 (0.89)
  Gender attitudes: politics		−0.07 (0.19)						0.00 (0.95)
  Gender attitudes: women are discriminated			0.03 (0.18)					0.05 (0.04)
  Political interest: domestic news				0.16 (0.00)				0.14 (0.00)
  Institutional trust					0.06 (0.04)			−0.01 (0.87)
  Interpersonal trust						0.11 (0.00)		0.10 (0.00)
  Collective action attitude							0.07 (0.00)	0.04 (0.18)
  Control variables	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.
  Adjusted R2	0.12	0.12	0.12	0.19	0.13	0.19	0.15	0.26
  N	211	211	211	211	211	211	211	211

  bold in p < 0.05; reported: B coefficient (p-value); estimated as OLS; controlled for level of education (SIM,MI-ID), having finished education in country of origin (SIM), migration generation (SIM, MI-ID), age (SIM, MI-ID), country of origin (SIM), ethnic group (MI-ID), employment (SIM, MI-ID), treatment group, in the model with the collective strategy (MI-ID).

  Table 4 Regression analysis results for the link between Non-institutional Political Participation electoral participation and religiosity among Muslim women, Including Potential Mediators

  	1	2	3	4	5	6	7	8
  A : Non-institutional Participation: Attending Protest (SIM)1
  	Religiosity: Total sample	Mediator 1	Mediator 2	Religiosity: Split-half sample A	Mediator 3	Mediator 4	Mediators 1-4	Model 6 on Gen1.5 up Turkish- and Moroccan-Dutch
  Attendance	0.02 (0.01)	0.03 (0.01)	0.02 (0.03)	0.03 (0.01)	0.03 (0.01)	0.03 (0.01)	0.03 (0.02)	0.05 (0.04)
  Affective belonging	−0.05 (0.00)	−0.04 (0.00)	−0.05 (0.00)	−0.06 (0.00)	−0.06 (0.00)	−0.06 (0.00)	−0.06 (0.00)	−0.08 (0.02)
  Doctrinal literacy
  Gender attitude: financial		0.01 (0.37)					0.01 (0.45)	0.00 (0.95)
  Active in civic organisations Active in civic organisations			0.03 (0.07)				0.03 (0.07)	0.07 (0.04)
  Political interest: domestic news					0.01 (0.65)		0.01 (0.71)	−0.02 (0.56)
  Institutional trust						−0.01 (0.22)	−0.01 (0.31)	−0.01 (0.52)
  Control variables	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.
  Adjusted R2	0.06	0.06	0.06	0.07	0.07	0.07	0.08	0.09
  N	448	472	472	333	333	333	333	149
  B: Non-institutional Participation: Petition, Social Media, Action Group (0-100)2 (MI-ID)
  	Religiosity	Mediator 1a	Mediator 1b	Mediator 2	Mediator 3a	Mediator 3b	Mediator 4	Full model
  Attendance	−0.21 (0.00)	−0.21 (0.00)	−0.21 (0.00)	−0.21 (0.00)	−0.20 (0.00)	−0.18 (0.00)	−0.19 (0.00)	−0.18 (0.00)
  Affective belonging	0.05 (0.41)	0.05 (0.38)	0.05 (0.41)	0.04 (0.45)	0.05 (0.40)	0.07 (0.24)	0.06 (0.34)	0.06 (0.29)
  Doctrinal literacy	0.05 (0.49)	0.06 (0.39)	0.05 (0.49)	0.05 (0.47)	0.05 (0.51)	0.05 (0.50)	0.07 (0.33)	0.09 (0.24)
  Gender attitudes: politics		0.10 (0.16)						0.07 (0.36)
  Gender attitudes: women are discriminated			0.00 (0.96)					0.01 (0.79)
  Political interest: domestic news				0.02 (0.64)				0.03 (0.55)
  Institutional trust					−0.01 (0.85)			0.06 (0.19)
  Interpersonal trust						−0.08 (0.03)		−0.11 (0.01)
  Collective action attitude							−0.03 (0.42)	−0.00 (0.95)
  Control variables	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.	Incl.
  Adjusted R2	0.12	0.13	0.12	0.19	0.13	0.14	0.15	0.16
  N	211	211	211	211	211	211	211	211

  bold in p < 0.05; reported: B coefficient (p-value); estimated as OLS; controlled for level of education (SIM,MI-ID), having finished education in country of origin (SIM), migration generation (SIM, MI-ID), age (SIM, MI-ID), country of origin (SIM), ethnic group (MI-ID), employment (SIM, MI-ID), treatment group, in the model with the collective strategy (MI-ID).

• 5 Conclusion and Discussion

  Using a multi-study design, this study provides new insights into the possible underlying mechanisms of the association between Islamic religiosity and political participation among Muslim women in the Netherlands. We provided answers to the question whether Islamic religiosity hampers political participation among Muslim women because of a lack of interest and conservative attitudes or stimulates such participation by yielding political knowledge and civic attitudes and skills.
  First, our descriptive results illustrate that (in the Netherlands) migrant women participate less in institutional politics than migrant men and that this gender gap is more pronounced among those with a Muslim migrant background. Such findings underline the contribution in studying Muslim women intersectionally and, in particular, to understand the role of Islamic religiosity in their political participation, although the (limited) size of these differences also underlines that Muslim women are not as distinctive as often framed in public debate.
  Second, our results as summarised above (in Section 4.2.4) provide clear support for rejecting the religion-attitudes and the attitudes-behaviour linkages that the culturalist logic assumes. We conclude that the social capital logic proves more worthwhile in understanding the association between Islamic religiosity and political participation among Muslim women than the culturalist logic. In line with previous research (Moutselos, 2020; Oskooii & Dana, 2018; but see Kollar et al., 2023), we find that mosque attendance is often positively associated with political participation, except for non-institutional participation. However, there is a negative linkage between attendance and non-institutional participation and to understand that (interpersonal) trust seems pivotal, the lack thereof driving Muslim women away from institutional action and towards individualist non-institutional acts (see also Azabar, 2023; Jeroense & Spierings, 2023).
  In addition and, more generally, our results show that different dimensions of Islamic religiosity, various measurements of political participation and focusing on different subgroups of Muslim women yield different results. We consider this vital for future research and conclude that mixed effects in previous research may be attributed to collapsing different religiosity dimensions and participation activities together as well as focusing much on specific – often the largest – ethnic groups. This study, by exploring and using various measurements and groups in two datasets, underlines that increasing our understanding of specific mechanisms at play requires specifying which forms of religiosity and participation are put central, systematically assessing differences, for instance, in infrastructure and degree of organisation which are often linked to the ethnic minority a person belongs to. A multi-study approach like the one used here enabled disentangling these mechanisms, while also showing that specific samples and items may affect conclusions drawn in previous research. For example, and although not of central attention in this study, one of the reasons that the two datasets showed – to a certain extent – different results may have to do with different ethnic groups included. Among Muslims with a Moroccan or Turkish migration background, we find that mosque attendance has a seemingly larger (positive) effect with respect to attending a protest, which may link to the large community present in the Netherlands (in comparison to those with a Somalian or Iranian background). Importantly, the number of respondents of each ethnic group did not allow us to study this in the present research, which leaves us with a recommendation for future research to explore how ethnic background intersects with the picture painted here.
  In addition, previous research has argued that political participation among Muslims is context-dependent (Moutselos, 2020; Kollar et al., 2023), for instance, due to different policies, attitudes among the majority population and international (political) events like the attempted coup in Turkey (Lubbers et al., 2024). The present study was not able to acknowledge such context dependency, in which the focus was on political participation in the Netherlands. Questions that remain and deserve further consideration revolve around acknowledging this context dependency and transnational political participation.
  We conclude that in general the effects of (the various dimensions of) Islamic religiosity are limited, while the Dutch case provides a context which would make effects likely. And if there are effects, they are more likely to stimulate than to hinder political participation. As such, these findings dismiss the notion that Muslim women participate less in politics because of Islamic religiosity. Actually, particularly, for the type of participation where Muslim women are most clearly behind the other intersectional groups, it is positively associated with more participation. Future research that compares Muslims to non-Muslims in their political participation should thus be careful in attributing potential differences to Islam.

References

• Azabar, S., & Thijssen, P. (2022). The Electoral Agency of Muslimahs: An Intersectional Perspective on Preferential Voting Behaviour. Journal of Ethnic and Migration Studies, 48(17), 4106-4128. https://doi.org/10.1080/1369183X.2021.1902793.

• Azabar, S. (2023). Halal Resistance: Unraveling Muslims’ Political Agency in Belgium [Doctoral dissertation], University of Antwerp.

• Bègue, L. (2002). Beliefs in Justice and Faith in People: Just World, Religiosity and Interpersonal Trust. Personality and Individual Differences, 32(3), 375-382. https://doi.org/10.1016/S0191-8869(00)00224-5.

• Ben-Nun Bloom, P., & G. Arikan (2012). A Two-edged Sword: The Differential Effect of Religious Belief and Religious Social Context on Attitudes towards Democracy. Political Behavior, 34(2), 249-276. https://doi.org/10.1007/s11109-011-9157-x.

• Burdette, A., Hill, T., & Moulton, B. (2005). Religion and Attitudes toward Physician-Assisted Suicide and Terminal Palliative care. Journal for the Scientific Study of Religion, 44(1), 79-93. https://doi.org/10.1111/j.1468-5906.2005.00266.x.

• Calhoun-Brown, A. (1996). African American Churches and Political Mobilization: The Psychological Impact of Organizational resources. The Journal of Politics, 58(4), 935-953. https://doi.org/10.2307/2960144.

• Cesari, J. (2014). Political Participation among Muslims in Europe and the United States. Engaging the Other: Public Policy and Western-Muslim Intersections (pp. 173-189).

• Dana, K., B. Wilcox-Archuleta, & M. Barreto. (2017). The Political Incorporation of Muslims in the United States: The Mobilizing Role of Religiosity in Islam. Journal of Race, Ethnicity, and Politics, 2(2),170-200. https://doi.org/10.1017/rep.2017.4.

• Dancygier, R. 2017. Dilemmas of Inclusion. Princeton University Press.

• Jamal, A. (2005). The Political Participation and Engagement of Muslim Americans: Mosque Involvement and Group Consciousness. American Politics Research, 33(4), 521-544. https://doi.org/10.1177/1532673X04271385.

• Fennema, M., & Tillie, J. (1999). Political Participation and Political Trust in Amsterdam: Civic Communities and Ethnic Networks. Journal of Ethnic and Migration Studies, 25(4), 703-726. https://doi.org/10.1080/1369183X.1999.9976711.

• Finlay, R., & Hopkins, P. (2020). Resistance and Marginalisation: Islamophobia and the Political Participation of Young Muslims in Scotland. Social & Cultural Geography, 21(4), 546-568. https://doi.org/10.1080/14649365.2019.1573436.

• Geurts, N., & van Klingeren, M. (2023). Do Religious Identity and Media Messages Join Forces? Young Dutch Muslims’ Identification Management Strategies in the Netherlands. Ethnicities, 23(1), 26-45. https://doi.org/10.1177/14687968211060538.

• Glas, S. (2020). Where Are the Muslim Feminists? Religiosity and Support for Gender Equality in the Arab Region. Radboud Repository.

• Glas, S., Spierings, N., & Scheepers, P. (2018). Re-understanding Religion and Support for Gender Equality in Arab Countries. Gender & Society, 32(5), 686-712. https://doi.org/10.1177/0891243218783670.

• Hancock, R. (2020). Muslim Environmentalists, Activism, and Religious Duty. In M. Peucker & M. R. Kayikci (Eds.), Muslim Volunteering in the West (pp. 141-160). Palgrave Macmillan

• Harris, A., & Roose, J. (2014). DIY Citizenship amongst Young Muslims: Experiences of the ‘Ordinary’. Journal of Youth Studies, 17(6), 794-813. https://doi.org/10.1080/13676261.2013.844782.

• Jeroense, T., & Spierings, N. (2023). Political Participation Profiles. West European Politics, 46(1), 1-23. https://doi.org/10.1080/01402382.2021.2017612.

• Joly, D. (2017). Muslim Women’s Political Engagement in British and French Society. Revue europeenne des migrations internationales, 2(3), 157-181. https://doi.org/10.4000/remi.8643.

• Jones-Correa, M., & Leal, D. (2001). Political Participation: Does Religion Matter?. Political Research Quarterly, 54(4), 751-770. https://doi.org/10.2307/449233.

• Just, A., Sandovici, M., & Listhaug, O. (2014). Islam, Religiosity, and Immigrant Political Action in Western Europe. Social Science Research, 43, 127-144. . https://doi.org/10.1016/j.ssresearch.2013.09.005.

• Kollar, R. Geurts, N., & Spierings, N. (2023). Religiosity and Electoral Turnout among Muslims in Western Europe. European Political Science Review. Online First 1-20. https://doi.org/10.1017/S1755773923000334.

• Koopmans, R. (2019). Het vervallen huis van de islam: Over de crisis van de islamitische wereld. Prometheus.

• Kranendonk, M., & F. Vermeulen (2019). Group Identity, Group Networks, and Political Participation: Moroccan and Turkish Immigrants in the Netherlands. Acta Politica, 54(4), 625-666.

• Lefevere, J., & Van Aelst, P. (2014). First-Order, Second-Order or Third-Rate? A Comparison of Turnout in European, Local and National Elections in the Netherlands. Electoral Studies, 35, 159-170. https://doi.org/10.1016/j.electstud.2014.06.005.

• Lubbers, M., Fleischmann, F., Geurts, N., & Gijsberts, M. (2024). Political Turmoil and Attitude Change Among the Diaspora. The Impact of the 2016 Attempted Military Coup on Homeland Orientation Among Recent Turkish Immigrants in the Netherlands. International Journal of Public Opinion Research, 36(1), edad043.

• Moutselos, M. (2020). Praying on Friday, voting on Sunday? Mosque Attendance and Voter Turnout in Three West European Democracies. Journal of Ethnic and Migration Studies, 46(11), 2275-2292. https://doi.org/10.1080/1369183X.2019.1662283.

• Mügge, L. M., Runderkamp, Z., & Spierings, N. (2024). Descriptive representation of marginalized groups in Dutch Politics. In De Lange, S., Louwerse, T., ‘t Hart, P. and Van Ham, C. (Eds.), The Oxford Handbook of Dutch Politics (Ch.28). Oxford: Oxford University Press. (Forthcoming)

• Oskooii, K. (2016). How Discrimination Impacts Sociopolitical Behavior: A Multidimensional Perspective. Political Psychology, 37(5), 613-640. https://doi.org/10.1111/pops.12279.

• Oskooii, K., & Dana, K. (2018). Muslims in Great Britain: The Impact of Mosque Attendance on Political Behaviour and Civic Engagement. Journal of Ethnic and Migration Studies, 44(9), 1479-1505. https://doi.org/10.1080/1369183X.2017.1330652.

• Peelen, A., Geurts, N., Spierings, N., & Van Klingeren, M. (2022). Codebook for MI-ID: Migrant Identity among Young Muslim Adults in the Netherlands. Radboud Social Cultural Research. DANS. https://doi.org/10.17026/dans-xkj-7ek4.

• Peucker, M. (2018). On the (In) Compatibility of Islamic Religiosity and Citizenship in Western Democracies: The Role of Religion for Muslims’ Civic and Political Engagement. Politics and Religion, 11(3), 553-575. https://doi.org/10.1017/S1755048317000700.

• Putnam, R. (1993). Making Democracy Work: Civic Traditions in Modern Italy. Princeton University Press.

• Read, J. (2007). More of a Bridge than a Gap: Gender Differences in Arab-American Political Engagement. Social Science Quarterly, 88(5), 1072-1091. https://doi.org/10.1111/j.1540-6237.2007.00492.x.

• Roggeband, C., & Verloo, M. (2007). Dutch Women Are Liberated, Migrant Women Are a Problem: The Evolution of Policy Frames on Gender and Migration in the Netherlands, 1995-2005. Social Policy & Administration, 41(3), 271-288. https://doi.org/10.1111/j.1467-9515.2007.00552.x.

• SCP-CBS (Sociaal en Cultureel Planbureau; Centraal Bureau voor de Statistiek). (2022). Survey Integratie Minderheden 2020 – SIM2020 (V1). DANS Data Station Social Sciences and Humanities. https://doi.org/10.17026/dans-26h-xn4n.

• Sobolewska, M., Fisher, S., Heath, A., & Sanders, D. (2015). Understanding the Effects of Religious Attendance on Political Participation among Ethnic minorities of Different Religions. European Journal of Political Research, 54(2), 271-287. https://doi.org/10.1111/1475-6765.12081.

• Spierings, N. (2019). Social Trust in the Middle East and North Africa: The Context-Dependent Impact of Citizens’ Socio-Economic and Religious Characteristics. European Sociological Review, 35(6), 894-911. https://doi.org/10.1093/esr/jcz038.

• Spierings, N., & Van Klingeren, M. (2020) MI-ID: Migrant Identity among Young Muslim Adults in the Netherlands. https://doi.org/10.17026/dans-xkj-7ek4.

• Spierings, N., & Vermeulen, F. (2023). Minding the Political Engagement Gap: How Discrimination and Religion Impact Dutch Voters with an Immigrant Background. Acta Politica, 1-24. https://doi.org/10.1057/s41269-023-00286-7.

• Stockemer, D. (2014). What Drives Unconventional Political Participation? A Two-Level Study. The Social Science Journal, 51(2), 201-211. https://doi.org/10.1016/j.soscij.2013.10.012.

• Uslaner, E. (2003). Volunteering and Social Capital: How Trust and Religion Shape Civic Participation in the United States. In Dekker, P. & Uslaner, E. (Eds.) Social Capital and Participation in Everyday Life (pp. 104-117). Routledge.

• Van Deth, J. (2014). A Conceptual Map of Political Participation. Acta Politica, 49(3), 349-367. https://doi.org/10.1057/ap.2014.6.

• Van Heelsum, A. (2005). Political Participation and Civic Community of Ethnic Minorities in Four Cities in the Netherlands. Politics, 25(1), 19-30. https://doi.org/10.1111/j.1467-9256.2005.00225.x.

• Van Klingeren, M., & Spierings, N. (2020). Acculturation, Decoupling, or Both? Migration’s Impact on the Linkage between Religiosity and Gender Equality Attitudes. Journal of Ethnic and Migration Studies, 46(15), 3079-3100. https://doi.org/10.1080/1369183X.2020.1733947.

• Verba, S., Schlozman, K. L., & Brady, H. E. (1995). Voice and equality: Civic voluntarism in American politics. Harvard University Press.

• Vermeulen, F., & Berger, M. (2008). Civic Networks and Political Behavior: Turks in Amsterdam and Berlin. In S. Ramakrishnan & I. Bloemraad (Eds.), Civic Hopes and Political Realities: Immigrants, Community Organizations and Political Engagement (pp. 160-192). Russell Sage.

• Wisneski, D., Lytle, B., & Skitka, L. (2009). Gut Reactions: Moral Conviction, Religiosity, and Trust in Authority. Psychological Science, 20(9), 1059-1063. https://doi.org/10.1111/j.1467-9280.2009.02406.x.

Noten

• * This study was part of a research project funded by NWO (grant number: VI.Vidi.191.023).

• 1 For reasons of empirical availability.

• 2 The metrics of the scale remains the same (e.g. the mean), but more cases could be included.

• 3 Or increased the positive effect, signalling suppression effects which would also confirm a mechanism at work as hypothesized by culturalists.