INTRODUCTION
While Omar Mateen was murdering 49 people at the Pulse nightclub in Orlando, FL, he called 9-1-1 and claimed to be an “Islamic soldier” seeking vengeance for the killing of Abu Wahib, an ISIS commander. Subsequent media coverage of the call highlighted Wahib's relative obscurity, and suggested that Mateen's familiarity with a low-profile commander indicated the extent of his radicalization and disillusionment (Kirby Reference Kirby2016). The question then became, how and where did U.S.-born Mateen become radicalized? Attention focused on local mosques, and it was discovered that Mateen had prayed at a mosque in Kissimmee, FL a few days before the attack. Though the imam disclaimed any association with Mateen, an arsonist later set fire to the mosque in an apparent attempt to defend the United States from the threat of Islamic terrorism that he believed was inculcated in American mosques (Freytas-Tamura Reference de Freytas-tamura2017). Then-presidential candidate Donald Trump made the same association immediately after the attack when he called for surveillance of mosques as a means for preventing terrorist attacks (Diamond Reference Diamond2016).
These responses beg the question of whether and how the mosque reinforces disillusionment or disincentivizes political and social incorporation among Muslim Americans. This paper contributes to a growing body of literature investigating the relationship between mosque attendance and political engagement. In contrast to the assumptions made by Donald Trump and many others before him, this paper finds the mosque to be a site of civic inclusion rather than alienation. However, the extent of the inclusive effect depends on participation in social events at the mosque. Statistical models reveal that the frequency with which someone attends a mosque to pray does not significantly predict political engagement across any political measures considered here. However, participation in social or religious activities other than prayer services at a mosque or Islamic center is associated with an increased probability of political engagement across nearly all of the political indicators. Taken together these results make an important adjustment to existing research on the relationship between mosque attendance and political behaviors. They suggest that not all mosque involvement is equally likely to promote political involvement, and that religious participation must do more than enrich the spiritual life or fulfill religious obligation to have a substantive political effect—religious leaders and congregations must influence the social lives of their congregants. Furthermore, this research suggests that prayer services at the mosque are not directly comparable with Christian church services, requiring scholars to nuance research connecting religious associational life with political participation in order to consider divergent effects associated with a diversity of activities within western and non-western religious traditions.
This argument is developed after reviewing the theorized links between religious associational life and political participation and discussing the dynamic and diverse character of the American mosque. I posit two hypotheses distinguishing between the effects of attending mosque for prayer versus attending for other social or religious programs. The hypotheses are tested using statistical analysis of survey data collected from a probabilistic sample of Muslim Americans collected by the Pew Research Center. The results section explains that the latter form of mosque participation is found to have a consistent relationship with political engagement. The final section concludes with a discussion of the implications of the research and future opportunities for researching congregational effects on political engagement.
THE POLITICAL INFLUENCE OF RELIGIOUS INSTITUTIONS
As the most widespread form of voluntary organizations association in the United States, American churches are particularly effective at bringing people together and transmitting and maintaining group norms, both of which could influence political outlooks or behaviors (Wald Owen and Hill Reference Wald, Owen and Hill1988, Reference Wald, Owen and Hill1990; Djupe and Gilbert Reference Djupe and Gilbert2009). A well-developed body of research has explored the way religious institutions create political communities through helping the congregants develop basic civic skills (Peterson Reference Peterson1992; Verba, Schlozman and Brady Reference Verba, Schlozman and Brady1995; Smidt Reference Smidt1999; Jones-Correa and Leal Reference Jones-Correa and Leal2001; Schwadel Reference Schwadel2002; Brown and Brown Reference Brown and Brown2003) and expand their social networks (Putnam Reference Putnam1993; Putnam Reference Putnam2001; Schwadel Reference Schwadel2005), and they provide a venue through which congregants can get information about political issues and discuss public affairs (Brown and Brown Reference Brown and Brown2003; Wald and Calhoun-Brown Reference Wald and Calhoun-Brown2014). All of these mechanisms link religious attitudes to political behaviors through increasing the likelihood of political participation or through influencing individual political ideologies by promoting ideological convergence within the group.
While churches may serve as a venue for civic skill development or behavioral contagion, social network formation, or information transmission, not all religious institutions or behaviors activate the mechanisms that would link religious institutions to political behavior. Wald, Owen and Hill (Reference Wald, Owen and Hill1988) argue that a political contagion effect requires more than mere attendance at a religious service, and that the communication of political information and the opportunity to observe the reactions of co-religionists to political messages is necessary in order to link church attendance to political outcomes. They find that congregational theology is strongly associated with personal political conservativism (conservative attitudes with regard to abortion, gender roles, censorship, sexuality, drug use, and birth control) even more than personal moral commitments, providing evidence that these mechanisms linking group attributes to personal political attitudes are at work in some Protestant congregations, though political coherence across congregations varies widely (Wald, Owen and Hill Reference Wald, Owen and Hill1988; Reference Wald, Owen and Hill1990).
The development of civic skills is also not an automatic feature of church attendance but is primarily observed in those who are involved in leadership and decision-making within the church (Peterson Reference Peterson1992; Verba, Schlozman and Brady Reference Verba, Schlozman and Brady1995). Those who take these leadership roles are more likely to practice skills like speech giving, letter writing, group discussion, or chairing a meeting, all of which can be translated into political activity (Verba et al. Reference Verba, Schlozman, Brady and Nie1993). However, Djupe and Grant (Reference Djupe and Grant2001) question whether the relationship operates in reverse, where the church merely provides an opportunity for congregants to practice civic skills gained in secular contexts, supplying encouragement to participate in politics in other more critical ways. Either way, civic skills appear to be transferrable between religious and secular contexts if they are actively practiced and developed.
Similarly, merely attending religious services is not expected to automatically activate political activity like organizational membership or voting. Schwadel (Reference Schwadel2005) finds that participation in church activities other than religious services is positively associated with the likelihood of participating in non-church organizations, while a conservative congregational context that promotes conservative theology and within-church social ties has a suppressive effect on non-church organizational membership. This finding demonstrates that political behavior is contingent on the nature of the congregational context and the kinds of interactions fostered within the church. Djupe and Grant (Reference Djupe and Grant2001) also make a contextual argument when they find that an explicit link between religious and political activity must be made intentionally within the church in order for church participation to have an effect on political behavior. On occasion, this might be done by co-religionists, who will use the church to recruit congregants into the political process (Djupe and Grant Reference Djupe and Grant2001), though in many cases this connection might be made by religious leaders who can play an important role by communicating religious interpretations that would encourage political participation, through associating particular issues with a religious worldview, or through holding political meetings (Harris Reference Harris1994; Calhoun-Brown Reference Calhoun-Brown1996; Smidt Reference Smidt2004; McDaniel Reference McDaniel2009).
The role of the church as a source of political engagement and mobilization is especially pronounced in minority communities. African-American churches are very effective at political mobilization, activism, community organizing, and community building, often because they are the only or most prominent non-governmental organization in black communities and the largest organized expression of black community group interests (Calhoun-Brown Reference Calhoun-Brown1996; McDaniel Reference McDaniel2009; Smith Reference Smith2013). As in white churches, while simply attending church does not seem to be enough to propel African Americans into political activities, activist black churches with a strong civic culture or politicized church networks are significantly associated with individual political activism (Harris Reference Harris1994; Calhoun-Brown Reference Calhoun-Brown1996; Brown and Brown Reference Brown and Brown2003). The role of clergy is particularly important for promoting church life as a resource for voter mobilization among African-American churchgoers because black pastors tend to explicitly connect the religious and personal for African-American believers (Lincoln and Mamiya Reference Lincoln and Mamiya1990). For example, in 2002, 66% of pastors in African-American churches reported that their sermons “always” delivered advice for daily living by addressing the current needs of black communities (Crabtree Reference Crabtree2002). Three-fourths of pastors also reported taking part in voter registration of voter information drives.
Harris (Reference Harris1999) highlights the importance of considering African-American political activism as multi-dimensional, rooted in the socialization experience that encouraged political tactics both inside and outside mainstream political processes. In particular, the exclusion of African Americans from civic and social life with Jim Crow institutionalized a racialized public sphere, which led to the development of a “black counterpublic” (Dawson Reference Dawson1994). The counterpublic includes religious institutions, communication networks, and a number of other groups and organizations that encouraged oppositional political tactics that challenged the white supremacist status quo. The strong mobilizing role of black churches was especially critical in the North, West, and South during the civil rights movement (Smith Reference Smith2013).
Just as churches have played a distinctive role in mobilizing African Americans, they have been important sites for promoting socialization and integration into the wider American society for immigrant groups (Gordon Reference Gordon1964). Religious organizations provide a ready-made community with opportunities for friendship and fellowship, often in a culturally familiar environment. The congregations comprised of fellow immigrants or co-ethnics can help new arrivals cope with the difficulties of adjusting to a new status as a minority and can serve as a refuge from social discrimination (Hirschman Reference Hirschman2004). Religious institutions also provide many services to their congregants through providing information about jobs, housing, classes and seminars, and business opportunities, and like churches among U.S.-born Americans, they provide opportunities for developing and practicing civic skills (Foley and Hoge Reference Foley and Hoge2007). Because religious identities are tolerated in the United States, immigrants in the United States will use religion as a basis on which to construct their culture and identity within their new country and a mechanism through which they can become Americans.
The discussion above samples the extensive research on the relationship between attending religious institutions and politics. The general consensus is that attendance has at least an indirect effect on civic engagement and political participation, and that the relationship between attending church and political engagement is attributable to group dynamics more than to any particular belief system. The effect should theoretically transfer to any religious tradition that involves similar congregational dynamics. Nevertheless, the general acceptance of the relationship between religious associations and political engagement does not automatically transfer to non-western or non-Christian religions (Foner and Alba Reference Foner and Alba2008), and the existence of a relationship between Islam and political engagement or incorporation is the subject of intense debate in the popular press and academic literature. Some politicians, journalists, and academics suggest that Islam is incompatible with democratic ideals and institutions (ex. Huntington Reference Huntington1993, Reference Huntington2004; Lewis Reference Lewis2002), and they associate Muslim religious institutions and high levels of religious commitment among Muslims with acts of political violence around the world (Powell Reference Powell2011).Footnote 1 In response, researchers investigation the link between mosque attendance and political behavior have explicitly addressed this implicit assumption, like Oskooii and Dana (Reference Oskooii and Dana2017), who frame their research question as investigating whether mosque attendance is associated with civic and political disengagement and the endorsement of violence. The next section considers the way the mosque may differ from the Christian church in its relationship with and promotion of political participation.
THE MOSQUE AND POLITICS
The Muslim population is one of the fastest-growing religious minority groups in the United States, with the total population hovering at around 3 million people.Footnote 2 The number of American mosques has grown with the population: in 2011, the U.S. American Mosque Report identified 2,106 mosques, representing a 74% increase from the 1,209 mosques identified in 2000 and a 119% increase from the 962 identified in 1994 (Bagby Reference Bagby2012). Mosque attendance for weekly and Eid prayers is also increasing: in 2011, 2.6 million people attended Eid prayers, up from 2 million in 2000. These trends make questions about the political influence of mosque and Muslim communities more relevant and interesting.
Research on Muslim communities in the American context demonstrates that Muslims are rapidly acculturating, and that their religious communities drive political inclusion and engagement with mainstream society (Jamal Reference Jamal2005a, Reference Jamalb; Ayers and Hofstetter Reference Ayers and Hofstetter2008; Bagby Reference Bagby2009; Dana, Barreto and Oskooii Reference Dana, Barreto and Oskooii2011). Many scholars use survey data to identify links between mosque involvement and political participation. Mosque involvement among Arab Americans mobilizes political activities like making contributions to political candidates, attending political rallies, engaging in political agenda-setting through writing petitions, and considering oneself an active party member (Jamal Reference Jamal2005a). The Muslim American Public Opinion Surveys (MAPOS) from 2007 and 2008 revealed that respondents involved in mosque activities were more likely to write letters to officials, to protest, and to participate in community meetings (Barreto and Dana Reference Barreto and Dana2009).Footnote 3 Dana, Barreto and Oskoii (Reference Dana, Barreto and Oskooii2011) find that religiously devout Muslims are more likely to support political participation, while Ayers and Hofstetter (Reference Ayers and Hofstetter2008) find that religious beliefs were negatively associated with political participation and measures of religious resources (frequency of mosque attendance, prayer, and volunteerism) and were positively related to political participation. There is also survey evidence suggesting that Muslim religious leaders and institutions actively promote political involvement and engagement with mainstream society. The data from the 2011 U.S. Mosque survey suggests that Muslim religious leaders explicitly encourage political engagement, demonstrating almost unanimous support for Muslim involvement in American political life: 98% of Muslim religious leaders condoned institutional involvement and 91% encouraged the political involvement of Muslims (Bagby Reference Bagby2012).
These findings imply that there is nothing fundamentally different about the role of the mosque in promoting civic and political engagement compared with other religious institutions and that the same mechanisms connecting church attendance to political engagement should work for mosque attendance. However, while scholars have demonstrated that church attendance is not associated with political behavior unless there is access to political information or socialization through small groups and the development of politically rich relationships (Djupe and Grant Reference Djupe and Grant2001; Schwadel Reference Schwadel2005; Djupe and Gilbert Reference Djupe and Gilbert2009), no one has investigated these divergent effects of attendance versus group activity within mosques. This paper theorizes that while the direction and mechanisms of the relationships between church versus mosque attendance and political behavior may be similar, the magnitude of the relationships might be stronger and the gap between the political effects of attendance and small group activity could be wider in mosques than they are in churches. There are three areas of substantial difference between mosques and churches which may influence the ways Muslim religious associations activate information transfer or small group formation, all three of which influence political behavior: (1) the nature of prayer services at mosque, (2) the ethnic composition of the mosque communities, and (3) the position of Muslims within the current socio-political landscape. These differences and the way they are expected to alter the mechanisms connecting the mosque to political participation are discussed below.
First, the mosque plays a slightly different role in the daily lives of Muslims compared with the role that other dominant religious institutions play in the United States. The mosque can be visited several times a day for congregational prayer lead by an imam, and the prayers are typically accompanied by Qur'anic reading (Salah prayer). Both prayer and reading are typically done in Arabic and are highly ritualized. On Friday at midday, Muslims gather for congregational prayer (Jumah prayer) and the imam will deliver the Islamic equivalent of a sermon in Christian services (Khutbah). Mosque activities, therefore, differ from American church activities in the daily, ritualized prayer services. In addition, many mosques will host other kinds of community events, both for their congregants and the community at-large. Many events are targeted at Muslims, attempting to bring the community together for prayer, classes, or support groups. Other activities provide community services, neighborhood service projects, or visits from political leaders.
The second way the American mosque differs from churches is in the way they cater to an especially socio-culturally diverse and dynamic Muslim population.Footnote 4 The Muslim population is the most racially diverse major religious group. Thirty-eight percent of Muslim in the United States identify as white, 28% are black, 28% are Asian, 4% are Latino, and 3% identify as another race (Lipka Reference Lipka2015). The diversity is especially striking when compared with the United Methodist Church (the largest mainline Protestant Church) or the Evangelical Lutheran Church in America (another mainline denomination), both of which are more than 90% white. Catholics and Pentecostal denominations are between 60% and 70% white, while Hispanics constitute the largest minority group making up between 25% and 35% of adherents.Footnote 5 American mosques are also especially diverse compared with mosques outside the United States, which will often have congregations primarily comprised of a particular ethnic or denominational group. Meanwhile, only 3% of American mosques are found to have a congregation of a single ethnicity. While a majority of American mosques are found in urban areas where they might draw a large number of worshipers from the same ethnic community, the percentage of mosques in urban locations is decreasing, and mosques can be found throughout the American landscape (Bagby Reference Bagby2012). In more rural communities there will likely be a single mosque serving a great variety of believers from the surrounding area. This diversity provides an opportunity for the development of rich social capital, where mosque attendees develop relationships with others from different racial, ethnic, class, and denominational backgrounds. At the same time, these diverse identities intersect with a shared religious identity. Jamal (Reference Jamal2005a) and Simmons (Reference Simmons2008) assert that the diversity of the American-Muslim community creates an opportunity for the development of a pan-ethnic Muslim identity, which could help the Muslim community achieve group consciousness within the United States.Footnote 6
The third way the mosque may function differently from the church is through promoting political engagement in a hostile socio-political context, where Muslim minority status and related experiences with discrimination influence the degree and nature of Muslim political incorporation. Like with African Americans and immigrants, the minority status of Muslims has served to develop political solidarity and a civic identity as religious Americans (Welborne et al. Reference Welborne, Westfall, Russell and Tobin2018). As discussed above, houses of worship play a critical role in political mobilization for all of these groups. While the diversity of the Muslim-American community provides a rich source of social capital, it is possible that congregational diversity compromises efforts to develop a coherent set of Muslim-American interests, making the mosque a weaker incubator for group identity than the African-American church, for example. That said, the current political and social crisis has started to forge common interest through necessity, and mosque leaders have taken a leadership role in promoting outward-facing activism. Mosque leaders have been particularly eager to engage in community outreach and encouraging integration with the larger community—many mosques have created full or part-time community organizer positions. This outreach is motivated by a desire to bridge the gap between Muslims and non-Muslims, to reduce the fear and suspicion of Islam, and also to provide counter-radicalization in the face of recruitment efforts by extremist groups. The political engagement of Muslims has been aided by interest groups such as the Islamic Society of North America, the American-Muslim Task Force, and the Council on American-Islamic Relations, all of which focus on Muslim rights but also make concerted efforts to make connections with the wider non-Muslim community. It is possible that compared with ethnic minority Christian churches, mosques are less focused on articulating an inward-facing common civic identity and are more committed to outreach, resulting in political activity that is more mainstream and less activist.
Despite these efforts to connect Muslims with non-Muslims, American society tends to view Muslims and Islamic associations with suspicion and reserve. The hostile environment may distort the mobilizing effect of the religious organization or lead to the disenfranchisement of Muslim Americans. Importantly, the precipitating factor in this scenario is the rejection of Muslims by society at large, not Muslim withdrawal. For example, in their investigation of the association between Muslim religiosity and party identification, Barreto and Bozonelos (Reference Barreto and Bozonelos2009) find that contrary to what has been found for Christian and Jewish religiosity and party identification, high levels of Islamic religiosity, together with disenfranchisement as a result of discrimination against Muslim Americans, may lead Muslims to refrain from identifying with a main political party. Alternatively, the interests of the Muslim community in the face of discrimination may incentivize even greater political involvement compared with other non-Muslim Americans through a desire to contradict fearful public perception of Muslims and to demonstrate belonging and membership within the American citizenry. Djupe and Calfano (Reference Djupe and Calfano2012) provide evidence that personal experiences with discrimination constitute a form of civic education, and that these experiences are associated with higher levels of tolerance of disagreeable speech and actions in Muslim Americans.
There are many reasons to expect that the mosque functions differently in encouraging political engagement when compared with role of Christian churches. The ritualistic character of mosque services, the unique diversity of the Muslim-American population, and the discriminatory environment could all impact the nature and degree of mobilization. These distinctive attributes warrant a renewed investigation into how political mobilization functions within the mosque.
HYPOTHESES
Though the literature investigating political mobilization through the mosque is growing, most of the research using survey data to examine the role of the mosque in influencing political participation relies on a monolithic measure indicating a degree of general mosque involvement.Footnote 7 This measurement obscures the reality that like churches, mosques serve distinct but overlapping needs of their congregants at different times and through different services. While all mosque activities have a social component that increases the possible transfer of political information or civic skills, this paper tests the supposition that not all participation in mosque activities is equally associated with political engagement.
Though Muslims can pray anywhere, congregational prayer through daily prayer services at the mosque (salat) is encouraged in order to foster community development. This specific social objective could link more frequent mosque attendance to political engagement, though much depends upon the behavior of the individual congregant at the prayer service. Attendance does not require the congregant to engage in the wide variety of activities that connect attending religious services to political engagement, especially where services are short, ritualistic, and conducted in Arabic, which the congregant may not speak or understand. Someone who habitually attends prayer service under these conditions could derive weaker benefits of civic skill development from the associational membership, especially if the person does not socialize around prayer time.
The weekly service of Jumah prayer held just after noon on a Friday resembles the daily ritual prayers except that the imam or prayer leader will also deliver a two-part sermon or khutbah. The Jumah prayer will typically draw more congregants, as attendance at the mosque for weekly prayer is considered to be an obligatory act. Given the higher rates of attendance and the inclusion of a spoken message (though the khutbah is often delivered in Arabic), the Jumah prayer service may provide more opportunity to acquire political information, expand one's social networks, or practice civic skills. However, like with the daily prayer services, much depends on the individual congregant, whether they socialize around the prayer service, and how they engage with the message.
Services at the mosque differ from many Christian church services in the ritualized nature of the service and the use of Arabic. The service on a Friday may also constrain social interactions around the service for those who work a conventional work week. The mechanisms that have linked church attendance and political engagement may not function as effectively with regard to attendance at prayer services in the mosque, though admittedly attending the prayer services provides more opportunities for social engagement than not attending at all.
As discussed above, many mosques also provide other types of social or religious programs address a wider range of religious, social, political, and economic needs. Because these activities are not centered on religious ritual, the language of instruction is typically English or the dominant language used in the community served. The nature of the programming also requires more interaction with others in order to be used effectively. These considerations along with the findings of previous research lead to the following hypotheses:
H1:
Increased frequency of attending mosque for prayer services will be weakly though positively associated with political engagement.
H2:
Attending mosque for religious or social events other than prayer will be positively associated with political engagement.
Admittedly, many people will attend social and religious programs in addition to prayer services. However, a congregant may prioritize one type of activity or program over the others, and the nature of the prioritization could impact the political benefits derived from mosque attendance. Those who attend mosque for reasons beyond prayer will be more likely to be engaged in the mosque as a religious and social institution, and will, therefore, be most likely to reap the informational, associational, and skills benefits from the mosque.
Figure 1 displays data on Muslim-American mosque attendance from a 2011 survey of Muslim Americans by the Pew Research Center, and demonstrates that most of the people who attend mosque to pray frequently also attend mosque for other reasons, but there are those who attend mosque only to pray, and others who only attend for other social or religious reasons (e.g. religious instruction, holiday celebrations, charity events). Thirty-four percent of those who attend prayer services daily do not attend other kinds of events at the mosque, and of those who never attend prayer services, 5% attend the mosque for other reasons. These data suggest that Muslim Americans differentiate between the different services provided by the mosque, and these services could, therefore, play different roles in the lives of congregants.
Mosque attendance.
DATA AND METHODOLOGY
To investigate the link between Muslim-American mosque attendance and political behavior, this paper uses data from the second national probability survey of Muslim Americans conducted by the Pew Research Center in 2011.Footnote 8 The survey includes data from phone interviews of 1,033 adult Muslims living in the United States. Definitions and descriptive statistics of all of the variables derived from the dataset are provided in Table A1.
The concept of political engagement is captured by a range of indicators that will serve as dependent variables in separate models. The first variable captures whether the respondents say they follow what is going on in government or political affairs, captured with an indicator ranging from 0 to 3 where 0 indicates that they “hardly” follow politics, and 3 indicates that they follow politics “most of the time.” The Muslim Americans in the Pew sample are generally interested in politics, with an average score of 2.12 indicating that most of them follow politics at least some of the time. The second variable measures whether the survey respondent considers himself or herself to be a member of a major political party (Republican or Democrat). About 55% of the sample identifies as members of either the Republican or Democrat parties, though 89% of these party members identify as Democrats. Other research has explored the determinants of self-identifying as either a Republican or Democrat, and there have been substantial shifts in party identification from supporting Republican to supporting Democratic candidates after the 2000 presidential elections (Cho, Gimpel, and Wu Reference Cho, Gimpel and Wu2006; Ayers Reference Ayers2007; Djupe and Green Reference Djupe, Green and Wilson2007; Barreto and Bozonelos Reference Barreto and Bozonelos2009). In the 2000 presidential elections, 40% of Muslim Americans supported George W. Bush for president-elect. By 2004, that number had shifted to only 7% of Muslims, while 86% supported presidential candidate, John Kerry (Ayers Reference Ayers2007, 191; Read Reference Read2007, 1075). Subsequent presidential elections have consistently shown higher levels of support for the Democratic Party. Within the 2011 Pew data, only 6% of respondents report identifying with the Republican Party. These trends demonstrate a transition from Muslims’ historical preference for more socially conservative politics (Cho, Gimpel, and Wu Reference Cho, Gimpel and Wu2006; Djupe and Green Reference Djupe, Green and Wilson2007; Pew 2007; Barreto and Bozonelos Reference Barreto and Bozonelos2009). The limited variance in party identification has reoriented research on Muslim-American party identification to focus more on partisan identification more generally, and to consider the disenfranchisement of Muslim Americans from the political party system (see Barreto and Bozonelos Reference Barreto and Bozonelos2009; Westfall et al. Reference Westfall, Russell, Welborne and Tobin2017). This paper likewise focuses on whether a survey respondent identifies with a primary political party, versus identifying as an independent, some other party, or indicating no preference for a political party.
Two additional variables capture political engagement through voting behavior. The first measures whether or not the respondent voted in the 2008 presidential election and the second measures whether respondents are certain that they are registered to vote, excluding those who were not citizens at the time of the election. Seventy-four percent of the Pew sample reports voting in the 2008 presidential election is, compared with 64% turnout reported in the general population. Seventy-six percent of the eligible participants in the Pew sample reports being registered to vote compared with the 71% registration rate in the general population (File and Crissey Reference File and Crissey2012). These rates are slightly higher than those found in other research finding that Muslim Americans tend to register and vote at slightly lower rates than non-Muslims in the United States. For example, in a 2007 survey by Pew Research Center, 63% of Muslim-American respondents reported that they were registered to vote, compared with 76% of non-Muslim respondents, and 58% of eligible Muslim Americans reported voting in the 2004 elections compared with 74% of non-Muslim respondents (Kohut, Lugo, and Keeter Reference Kohut, Lugo and Keeter2007). The different results in the 2011 Pew data could be a product of over-reporting, or it could reflect the enhanced mobilization of Muslims during and after the 2008 elections.
The final dependent variable captures a more informal element of political engagement through the following question: “In the past 12 months, have you worked with other people from your neighborhood to fix a problem or improve a condition in your community or elsewhere?” This question captures community involvement and the exercise of political social capital on a local level outside formal political institutions and about 36% of the sample reports engaging in community-based problem-solving.
Because most of the dependent variables are dichotomous, it is appropriate to fit generalized linear models, where various measures of political engagement are the dependent variables, the primary independent variables of interest measure mosque involvement, and including an array of control variables:
logit[political engagement] = α + + β1(mosque pray) + β2(mosque other) + β3(frequency of prayer) + β4(religion important) + β5(follow politics) + β6(political party member) + β7(satisfaction) + β8(employed) + β9(Muslim friends) + β10(female) + β11(age) + β12(education) + β13(US born) + β14(second generation) + β15(black) + β16(Asian) + + β17(other race)
The primary independent variables of interest measure mosque involvement. Pew asks two questions about mosque attendance that allow for the exploration of the potentially different effects of mosque participation on political participation, depending on the sociability of the mosque experience. The first is an ordinal variable asking how often participants attend mosque for prayer, and the second is a dichotomous variable asking whether the respondent attends other religious or social activities at the mosque outside of prayer.
In order to distinguish the effects of mosque attendance from other forms of religiosity, controls for religiosity beyond mosque attendance must be included in the model. This allows for the differentiation between political engagement that might be a product of personal religious beliefs, habits, or convictions and engagement encouraged by the sociopolitical environment of the mosque. Pew asks two additional questions about religiosity in addition to the questions about mosque attendance: one measures the self-reported frequency of prayer ranging from never to five times daily, and the other asks how important religion is in the respondent's life, ranging from not at all important to very important. These indicators are moderately correlated with each other and with the indicator capturing the frequency of mosque attendance for the purpose of prayer, but not strongly correlated with attending mosque for other social or religious reasons (see Table A2), again suggesting that attending mosque for reasons other than prayer could be motivated by concerns beyond personal religiosity.
The model includes a number of control variables representing competing political, religious, and social explanations for various forms of political engagement. The two indicators of following politics and party membership are converted to independent variables for the equations predicting formal political participation and community involvement. Those who regularly follow political developments should be more likely to participate in the political system in a variety of ways, and those who belong to a political party are more likely to register and vote (Erikson Reference Erikson1981). Whether respondents say they are satisfied with the way things are going in the country is measured with a dichotomous variable. Dissatisfaction might suggest general disengagement with the political system, which should disincentivize political participation. Alternatively, someone who is dissatisfied might feel motivated to participate in politics as a way to bring about change. In the models predicting political engagement in ways other than party membership, the variable measuring whether the respondent belongs to a main political party is included in the analysis.
An indicator of the composition of friend networks is included in the model in order to capture the religious homogeneity of the respondents’ friend networks. Extensive research has shown how the density and specificity of social networks affects political participation (Lake and Huckfeldt Reference Due Lake and Huckfeldt1998; Leighley Reference Leighley1990; Mutz Reference Mutz2002; McClurg Reference McClurg2003; Kwak, Shah, and Holbert Reference Kwak, Shah and Holbert2004; Scheufele et al. Reference Scheufele, Nisbet, Brossard and Nisbet2004; Scheufele et al. Reference Scheufele, Hardy, Brossard, Waismel-Manor and Nisbet2006). Possessing a diverse social network, where members’ views are not consistent with one's political beliefs, can increase political participation through exposing individuals to a variety of opinions, forcing group members to engage with and compromise across conflicting ideas, or motivating deep individual inquiry into the source of friends’ differences (Leighley Reference Leighley1990; Scheufele et al. Reference Scheufele, Nisbet, Brossard and Nisbet2004, Reference Scheufele, Hardy, Brossard, Waismel-Manor and Nisbet2006). If friend groups are more homogenous, they may disincentivize political participation through group insularity. Wald Owen and Hill (Reference Wald, Owen and Hill1990) find support for this relationship when they find that strong friendship ties among members leads to political conformity, but not necessarily political action. While diversity is multi-dimensional, the indicator in the model focuses on religious homogeneity through a question asking how many of the respondents’ close friends are Muslim. This variable is measured on a scale from 0 to 4, ranging from having “hardly any” Muslim friends to having an exclusively Muslim friend network.
Hypothesis 2 predicts that attending mosque for reasons other than prayer should more effectively activate mechanisms linking mosque attendance to political engagement, compared with attending for prayer. This should be the case even for someone who attends both types of activities at the mosque. However, it is possible that different types of people have different incentives to attend mosque or different access to mosque programming. Table A3 provides a rough demographic breakdown of who is attending mosque for prayer daily, who reports attending weekly for Jumah services, and who is attending other social or religious events at the mosque. Women make up a much larger proportion of those who attend mosque for social and religious events than they do of those who attend prayer, and a slightly larger proportion of the attendees at the social and religious events are college educated, employed, and born in the United States.
Since these demographic groups have slightly different patterns of mosque attendance, it is necessary to include them in the model. Employment in full or part-time work, respondent race, gender, whether the respondent is foreign-born, and whether he or she is a second-generation immigrant are all coded as dichotomous variables. Education is measured on a five-point scale. Employment and education should be positively associated with political participation because employment or attending educational institutions generally exposes individuals to political information and helps with the development of civic skills, much like religious associational membership is purported to do. Gender is controlled for because it has been shown to differently predict political participation: Burns, Schlozman and Verba (Reference Burns, Schlozman and Verba2001) find that among the general population, religious institutions motivate political activity more for men than they do for women. Furthermore, gender roles and consciousness may channel and enhance political participation for women in certain ways. For example, Verba, Burns and Schlozman (Reference Verba, Burns and Schlozman1997) find a gender gap in a political engagement where women are less interested in and informed about politics, and they argue that this gap has implications for political participation. American-Muslim women may present a special case due to the active renegotiation of gender identities within American mosques.Footnote 9
The race of our respondents is included in the model because of the very different experiences with Islam within specific racial groups. For example, the African-American community participated in the formation of a racialized political Islam in the United States through the teachings of Elijah Mohammad and Wallace Fard Mohammad (Simmons Reference Simmons2008). Race consciousness may also channel political participation in particular ways. For example, the African-American Muslim community aligned its interests with the civil rights movement during the 1950s and 1960s, and more recently Jamal (Reference Jamal2005b) argues that the mosque is more strongly associated with civic involvement among Arab Americans due to their desire to improve their popular image which often portrays them as terrorists, and to influence foreign policy in the Middle East. In the models considered here, the race is controlled for with a variety of categorical variables excluding the racial category “white,” which likely includes Arab respondents, as the baseline for comparison.Footnote 10 Whether a respondent is foreign-born could matter for linguistic or cultural access to avenues for political socialization. Age, measured with a categorical variable, is also included in the model since the elderly are more likely to participate in formal politics in the United States (Campbell Reference Campbell2003).
RESULTS
The results of the analysis are presented in Table 1. The most notable findings are in the first two rows, which show that attending mosque to pray is not a significant predictor of political engagement across any of the models. The insignificant relationship between attending mosque to pray and all the measures of political engagement contrasts sharply with the many significant relationships between social reasons for attending mosque and the likelihood that a person follows politics, voter registration, voting, and community involvement. Further, compared with all the other independent variables, attending mosque for social reasons is the most consistent predictor of political engagement.
Logistic models estimating odds of political engagement
Whether someone attends mosque for religious or social events increases the probability of following politics by 5%, holding all other variables at their mean. It also increases the probability of voting and registering to vote by 9% and the probability of collaborating with neighbors to solve a communal problem by 15%. Together these results confirm hypothesis 2, which predicts a positive relationship between political engagement and mosque attendance for social or religious reasons, but they lead to a reconsideration of hypothesis 1, which predicted attending mosque for prayer services would be weakly and positively associated with political engagement. There does not even appear to be a weak statistical relationship between the frequency of attending prayer services at the mosque and political engagement, unless the frequency of prayer is removed from the model, which allows attending mosque for prayer to increase the likelihood of collaborating with neighbors.Footnote 11 This contrasts with the literature linking the frequency of attending church services to engagement and as theorized above, could be the result of the timing and the ritualistic nature of mosque services, which may prevent the transmission of information or fail to engage civic skills in the same way as regular attendance at church services. I hypothesized that gathering in the community creates opportunities for engaging for motivated or interested people, but if so, the effects were not strong enough to be captured by the model.
The consistently positive relationships between attending mosque for social and religious activities and multiple indicators of political engagement send a strong signal that these activities enforce political engagement. The only indicator of political participation that does not share a relationship with social mosque attendance is party membership. This affirms the findings of Barreto and Bozonelos (Reference Barreto and Bozonelos2009) who found that the frequency of mosque attendance did not predict partisanship in their original survey data. They do find that individuals’ degree of faith and the degree of their Islamic practice determine partisanship, with more religious individuals being less likely to belong to either primary political party. While the model considered here cannot replicate these exact indicators, they are approximated by the religiosity variables, which do not share a significant relationship with party membership or any other indicator of political participation. This different result could be attributable to the ambiguous character of the questions measuring religiosity (frequency of prayer, self-reported importance of religion), compared with Barreto and Bozonelo's use of more targeted measures of Islamic knowledge and Qur'anic influence. Comparable questions were not available in the Pew data.
When party membership is converted to an independent variable, it significantly increases the probability that the survey respondent follows politics by 3%, registering to vote by 12%, and voting by 10%, but does not significantly change the probability of community involvement. Party membership serves primarily as a political information and voter mobilization tool around elections, while informal participation through community-based problem solving can happen at any time, and does not require a particular affiliation. Regularly following politics, also a measure of political information is a significant predictor of all forms of political participation except voter registration. These different results for the information-producing mechanisms may reflect the one-time and administrative nature of voter registration—it does not require any political information to register, and often registration corresponds with other administrative tasks like driver's licensure. By contrast, the extent to which a survey respondent follows politics requires regular effort that may speak to a greater propensity to engage with all types of political activity. When the extent to which a survey respondent follows politics is converted to an independent variable, comparing someone who hardly follows politics at all to someone who follows it most of the time, following politics significantly increases the probability of belonging to a political party by 13%, increases the probability of voting by 20%, and increases the likelihood that someone will work with their neighbors to solve a problem by 23%.
A few variables that can be interpreted as indicators of societal integration share relationships with some measures of political engagement. Employment positively increases the probability of registering to vote and voting behavior and education improves the probability of following politics regularly. An immigrant experience also appears to make a difference for political engagement with those born outside the United States being 11% less likely to vote, and 14% less likely to work with their neighbors to solve a community problem. Immigrants may have fewer civic skills than native-born Americans by virtue of being socialized outside the United States. The skills barrier should not affect second-generation immigrants in the same way, and the model demonstrates no significant difference between second-generation immigrants and those with no immigrant background when it comes to political engagement.
Some demographic controls shared significant relationships with a few indicators of political participation. Like in the general population, older Muslim Americans are more likely to register and vote. Asian Americans have a lower probability than white Americans of following politics. Muslim-American women are more likely to belong to a political party, but gender is not a consistently significant predictor of the other measures of political participation. This non-finding may speak to the relatively gender inclusive nature of the American Muslim where Muslim women are actively re-negotiating their roles within religious institutions and are more active in the mosque than Muslim women elsewhere.Footnote 12 Perhaps this involvement transfers to the political sphere, where Muslim women are as likely as Muslim men to participate in voter registration, voting, and community involvement.
DISCUSSION
The results of the statistical models suggest that the mosque fulfills diverse functions and that the relationship between mosque attendance and political participation depends on what a person does at the mosque. The frequency with which one attends the mosque to pray is not significantly associated with the political variables considered here, meaning that it is neither associated with political activity nor with political inactivity or withdrawal/disenfranchisement from party membership, voter registration or voting. Attending a mosque to pray does not influence formal political behavior one way or the other. This is an important point in light of popular discourse accusing the mosque of disenfranchising and radicalizing congregants. In contrast, attending mosque for other social or religious events is predictive of voter registration, voting, and community involvement.
Taken together, these results make an important adjustment to existing research on the relationship between mosque attendance and political behaviors—they suggest that not all mosque involvement is equally beneficial for promoting political involvement, and that religious participation must do more than enrich spiritual life or fulfill religious obligation to have a broader social or political effect. Such a finding can help reconcile dissonant experiences with the mosque as in the case of Omar Mateen, who attended mosque to pray, but may not have used the mosque to connect with other community members within or outside of the mosque setting, and went on to commit violence against other Americans rather than engaging with the civic institutions to register his discontent.
These findings encourage those who research the link between the religious association and political participation to be more expansive in their measurement of political participation and religious involvement in order to better understand the mechanisms theorized to be at work here. Surveys including more indicators of various forms of political engagement and particularly better measures of volunteerism could nuance findings with more important implications for the politics of everyday life. Similarly, including indicators capturing involvement in the variety of services offered by the mosque or other religious institutions or better measurement of congregational characteristics like racial, ethnic, or class diversity could provide new opportunities for more explicitly exploring the mechanisms linking different activities performed within religious associational life to political engagement and participation.
Many of the arguments developed here are built off of an assumed comparison of Christian churches and mosques because the literature on the political effects of churches is relatively well developed. Future work should empirically engage with the ways the mosque is different from the Christian church and make the comparison direct and explicit, like Oskooii and Dana (Reference Oskooii and Dana2017) do use data from the UK. I argue that the differences are rooted in the nature of the religious services, the diversity of the Muslim-American population, and in the discriminatory social environment surrounding mosques, but I could not test the comparison because the data in the survey used here is restricted to Muslim Americans. A dataset including substantial representative samples of adherents of many different religions could help test the assertion of a comparative difference, and with more precise measurements of religious associational life and political participation, could make a major contribution in identifying how the links between attending religious services or religiosity and political engagement.
The data considered in this paper were collected in 2011. They cannot, therefore, speak to developments around the 2016 presidential election, which have challenged and changed the Muslim-American community in profound ways. Anecdotal evidence suggests that mosques and religious institutions were critical agents of mobilization during and after the election. For example, U.S. Council of Muslim Organizations, the Islamic Circle of North America, the Council on American-Islamic Relations, and the Arab-American Institute collaborated to organize a national bus tour and voter registration drives in order to register a million new voters. As explained by Robert McCaw, the government affairs director at CAIR, “When you look at the toxic political climate in the United States, minority community members take the brunt of many political attacks now more than ever. We have to join together so we're heard. The best way is to go to the polls” (Dizard Reference Dizard2015). Trump's victory dramatically altered the social and political environment for Muslim Americans as a spike in Islamophobic attacks and the presumed tolerance of intolerance put many Muslim Americans on the defensive. Some Muslim religious institutions proactively responded by seeking out opportunities to engage with the wider community in order to prevent marginalization. Many have also used the social discrimination as an opportunity to make connections with other marginalized groups, especially Jewish communities.Footnote 13 It is possible that stronger relationships between mosque attendance may appear in data gathered during this critical time or that the mechanisms may change as the external environment stimulates greater consideration of collective Muslim identities and interests. It is vital that researchers remain vigilant in their examination of how relationships between religious associations and political outcomes change over time and across different religious communities.
Appendix
Variable descriptive statistics
Demographic breakdown of attendees at mosque events by percentage of total
Correlations between religiosity indicators
