The Bridging Activity of Multiracial Congregations

The growing diversity of U.S. communities has led scholars to explore how racial/ethnic diversity effects social capital, civic engagement, and social trust. Less is known about the relationship between diversity and the work of community-based organizations (CBOs). In this study, we examine how the racial/ethnic composition of one ubiquitous type of CBO, religious congregations, is related to measures of organizational bridging social capital. Analyzing data collected through a census of congregations in one Midwestern county, we explore the relationship between racial/ethnic diversity and the bridging activity of religious congregations. We find that multiracial congregations are more likely to be involved with externally focused service programs, tend to support a larger number of programs, and report more interorganizational collaborators than other congregations. Our findings suggest that multiracial congregations can provide a valuable resource for increasingly diverse communities and civil society.


Introduction
Despite progress made over the last half-century, race relations continue to be a source of concern in U.S. communities.Incidents of violence against minorities and episodes of civil unrest in cities such as Ferguson, MO and Charlotte, NC make frequent headlines.Disagreements over federal and local policies that affect immigrants divide many communities.Further, according to at least one recent national poll, a majority of Americans (74%) feels race relations are bad (Dann 2017).Such realities have led observers to ask what impact racial/ethnic diversity may have on the social cohesion of communities and on the strength of civil society.Scholars have begun to explore the implications of diversity for civic behavior (Alesina and Ferrara 2002;Putnam 2007;Uslaner 2012).However, the relationship between racial/ethnic diversity and the work of community-based organizations (CBOs) remains relatively unexamined.
Observers going back to Alexis de Tocqueville (de Tocqueville 1969) have noted the importance of community groups for civil society; they play a role in drawing individuals into civic life, strengthening community bonds, and providing valuable resources (Ruef and Kwon 2016;Schneider 2007;Small et al. 2008).Yet, despite growing awareness of the challenges diversity poses for civic engagement, we know little about how demographic shifts relate to the work of CBOs.Does increasing diversity pose a challenge to the organizational life that has been so important for civil society in the past?Might it provide new opportunities for CBOs to realize bridge-building potential in communities?In this article, we extend research on the intersection of diversity and civic life by bringing the experience of CBOs into the conversation.Specifically, we explore the relationship between racial/ethnic diversity and the bridging activities of one ubiquitous type of CBO-religious congregations.
There are over 350,000 congregations in the U.S. today (Brauer 2017).The most recent General Social Survey (GSS) estimates that 45 percent of U.S. adults attend religious services at least once a month (Smith et al. 2016), making congregations one of the most commonly reported forms of associational involvement in the U.S. (Putnam and Campbell 2010).Despite the fact that most congregations continue to be segregated by race, the number of multiracial congregations-those in which the majority group makes up less than 80 percent-is increasing (Chaves and Anderson 2014).These congregations bring people together across social divisions (DeYoung et al. 2003;Emerson and Woo 2006).As such, we contend that these organizations provide a unique opportunity for examining the relationship between racial/ethnic diversity and the bridging activity of CBOs.After a review of research on the relationship between diversity and civic engagement and a review of what we know about multiracial congregations and bridging social capital, we analyze data drawn from a census of congregations in one U.S. county to explore how diversity is related to the bridging activities of these organizations.Specifically, we explore how diversity is related to the provision of community services and to the extent of congregations' interorganizational collaboration.

Diversity and Civic Life
Civil society has long been a hallmark of American democracy and most often refers to the aggregate of CBOs and associations that operate outside the state and market to support the welfare of citizens (Anheier 2014;Edwards 2014).Throughout U.S. history, this aggregate has comprised mutual aid groups, civic clubs, religious congregations, and nonprofit organizations, among others.Such groups mobilize citizens to address collective problems and bring people together in ways that contribute to the development of valuable bridging social capital-social ties that link individuals to one another and the wider community and that foster trust and reciprocity (Anheier 2014;Schneider 2009;Handy et al. 2014).An important element of bridging social capital is that it tends to link citizens and groups across social cleavages or divisions in a community (Lichterman 2005;Putnam 2000).Bridging ties also facilitates the flow of information and resources between groups (Coleman 1988;Putnam 2000).Scholars of civil society have also noted that many social reforms in the U.S. have emerged from the work of CBOs and associations; community development organizations, hospitals, educational institutions, arts and cultural organizations, and myriad social movements trace their origins to the work of such groups (Hall 2006;Skocpol et al. 2000).
Despite a robust tradition of civil society, however, some have raised questions about its long-term durability in the U.S. (Putnam 2000;Skocpol 2002).In the late 20th century, scholars found that components of civic life (i.e., civic engagement, social capital, and social trust) seemed to be declining among U.S. adults (Paxton 1999;Putnam 2000).One of the most well-known proponents of this thesis is Robert Putnam, whose best-selling book, Bowling Alone, argued that at the end of the 20th century, Americans were less likely to participate in community groups and associations or to interact with neighbors than they had been in the 1950s (Putnam 2000).Putnam lays much of the blame for declines in civic life on cultural and structural shifts occurring since the 1950s: a rise in the number of two-career families, suburbanization, the growth of electronic media, and generational change (Putnam 2000).More recently, he has argued that growing racial and ethnic diversity may have a negative effect on civic participation as well (Putnam 2007).He argues that, in the short run, rapid diversification poses challenges for communities, as it tends to have a dampening effect on factors important for civic life.Drawing on national survey data from the U.S., Putnam (2007) found that individuals living in neighborhoods that are more diverse tend to be less trusting of others and less engaged in their communities.Individuals respond to growing diversity by "hunkering down" and disengaging from civic life (Putnam 2007, pp. 144-49).
Putnam is not alone in raising concerns; other studies have also found a relationship between diversity and civic life (Letki 2008;Leigh 2006).Drawing on national data sets, Costa and Kahn (2003) found that individuals living in communities that are more diverse were less likely to trust others, volunteer, or join associations.Likewise, Alesina and Ferrara (2002) analyzed 20 years of GSS data and found that living in a racially mixed neighborhood was related to lower levels of trust.However, there has been criticism of this work in recent years (Ariely 2014;Sturgis et al. 2010), and researchers have sought to specify the conditions under which diversity has a negative effect on civic life (Portes and Vickstrom 2011;Van der Meer and Tolsma 2014).Some contend it is segregation rather than diversity leading to lower levels of cohesion and trust (Rothwell 2012;Sturgis et al. 2014;Uslaner 2012).Others argue there may be something unique to the U.S. context that makes diversity challenging (Van der Meer and Tolsma 2014).As the U.S. continues to diversify, additional research is needed to clarify the relationship between diversity and civic engagement.Further, social scientists interested in civil society should be asking what challenges and opportunities diversity poses for the many CBOs active in U.S. communities.

Congregations and Civic Life
Scholarly interest in one type of CBO, religious congregations, has increased since the adoption of Charitable Choice provisions in the 1990s, which made it easier for faith-based organizations to compete for federal funding.Specifically, researchers began to examine what these organizations were doing in communities and with whom they were doing it (Ammerman 2005;Chaves and Tsitsos 2001;Cnaan et al. 2002).Temples, churches, and mosques provide a range of services to community residents; in the process, many collaborate with other organizations (Chaves and Eagle 2016;Fulton 2016;Tesdahl 2015).In fact, research reveals that collaborating with community organizations is one of the primary ways congregations engage their communities (Ammerman 2005;Chaves 2001).In addition to providing services to individuals and families, connections made through such collaborations have the potential to strengthen local networks and represent a form of bridging capital-facilitating the flow of information and resources and promoting bonds of trust among citizens and organizations (Foster 2014;Schneider 2009).Not all congregations invest equally in the provision of services or the development of bridging relationships however (Chaves and Wineburg 2010).
Previous research demonstrates that congregational resources such as income and human capital can have a significant impact on a congregation's involvement in service provision, as can environmental factors such as the social composition of a congregation's neighborhood (Tsitsos 2003).Characteristics such as a congregation's size, demographic composition, and location tend to be related to engagement (Chaves 2004;Cnaan et al. 2002;Tsitsos 2003;Polson 2016).For instance, one of the most significant factors determining a congregation's capacity for providing services is its size (Chaves and Tsitsos 2001).Larger congregations have more economic and human resources to invest.Likewise, congregations in poorer neighborhoods and those comprised of a larger percentage of college-educated members provide a larger number of services (Chaves 2004).Previous research demonstrates that theologically conservative congregations tend to be less involved in service to the wider community (Chaves and Higgins 1992).Particularly relevant for the current study, some have found that the racial/ethnic composition of a congregation is related to community engagement (Brown 2008;Polson 2015).

Multiracial Congregations and Bridging Social Capital
As the demographic makeup of the U.S. has shifted in recent decades, so too has the profile of American congregations.In the most recent iteration of the National Congregations Study, Chaves and Anderson (2014) found that the number of multiracial congregations has grown significantly since the 1990s, and even among relatively homogenous congregations, there is more diversity than there was several decades ago.As these shifts have occurred, interest in multiracial congregations and the people who attend them has increased (DeYoung et al. 2003;Dougherty and Huyser 2008;Edwards et al. 2013;Emerson and Woo 2006).
Scholars of race and religion have found that multiracial congregations face a number of demographic and organizational challenges; compositional diversity is not easy for congregations to maintain (Christerson and Emerson 2003).Yet, these congregations also represent unique social contexts, and the individuals who attend them tend to differ from other attenders in significant ways (Emerson and Woo 2006;Johnson and Jacobson 2005;Tavares 2011;Yancey 2001).Emerson and Woo (2006, p. 99) use the term "Sixth Americans" to refer to them because their social networks seem to cut across traditional racial/ethnic categories.Indeed, they appear to move more comfortably among different groups.Researchers have also found that worshiping in a multiracial congregation is linked to reduced prejudice and social distance between whites and non-whites (Emerson and Woo 2006;Polson and Dougherty 2019;Yancey 2001).These findings suggest that important bridging capital exists within such congregations-individuals and families develop friendships and interact across racial/ethnic categories.How does the existence of bridging capital within affect the bridging activity of the congregations themselves?Does a more diverse congregation seek to also build more bridges to other groups in the community?
Scholars of civic life argue that bridging occurs at both the individual and organizational levels in communities (Briggs 1998;Lichterman 2005;Putnam 2000;Schneider 2008;Wuthnow 2004).Actions and patterns of interaction that connect individuals and groups across social divisions, whether through friendship ties or through interorganizational collaboration, have the potential to contribute to increased levels of trust and reciprocity.Such connections can also be important conduits for the flow of information and needed resources in a community (Coleman 1988;Putnam 2000;Small et al. 2008).Because the development of social ties and interaction between different groups is such a central element of the concept of bridging (Lichterman 2005;Putnam 2000;Wuthnow 2002), and because multiracial congregations seem to be particularly adept at sustaining bridging ties at the individual level, we theorize they may also demonstrate a propensity for increased bridging activity in their communities.We propose this link between internal and external bridging in multiracial congregations for several reasons.
First, we anticipate that the presence of a racially and ethnically diverse set of attenders with crosscutting social ties is likely to represent a more diffuse set of community connections than is present in more homogenous congregations.While most voluntary associations remain relatively homogenous and exist within racially and ethnically homogenous social networks (McPherson et al. 1992(McPherson et al. , 2001)), we anticipate that multiracial congregations' membership may connect them to a more diverse group of community groups and organizations.We theorize that diffuse connections create increased potential for bridging activity.Second, we theorize that the unique cultural and social interaction patterns that are valued and promoted within multiracial congregations likely contribute to a culture or ethos that values inclusion and bridge-building.Previous research demonstrates that members of multiracial congregations tend to hold more inclusive views on a variety of issues and are often looking for or desiring a unique form of religious experience, one that is more inclusive and welcoming (Emerson and Woo 2006;Perry 2011Perry , 2013)).Extrapolating from these findings, we theorize that the multiracial context is likely one that values bridging more generally, in terms of both interpersonal relationships and organizational activities.Lastly, because previous research has demonstrated that bridging social ties can be an important source of both information and resources (Coleman 1988;Small et al. 2008;Wuthnow 2002Wuthnow , 2004)), we theorize that the presence of a diverse membership may make multiracial congregations especially aware of the diverse resources, needs, and community organizations that exist in the larger community.The flow of information into the congregation through more diverse social networks may help to draw multiracial congregations into a wider variety of bridging activities, both service provision and interorganizational collaboration.Indeed, one recent study found that diverse congregations tend to provide a wider variety of community services than other congregations do (Polson 2015).For all of these reasons, we anticipate that multiracial congregations may have the potential to foster valuable bridging capital in communities.
Drawing on what we know about the short-term challenges diversity poses for social cohesion and civic engagement as well as what we know about the potential multiracial congregations possess for bringing individuals together across divisions, we seek to answer two research questions.First, do multiracial congregations represent stores of bridging social capital in the form of externally focused programs and services?Second, do multiracial congregations report more interorganizational collaborations than other congregations do as they engage in service provision?Both are important questions that may help us better understand the impact of diversity on civil society and the role diverse CBOs might play in strengthening communities.

Data and Methods
To address our research questions, we utilize secondary data drawn from the Kent County Congregations Study (KCCS), a survey of congregational leaders in Kent County, MI. In 2006-2007, researchers from the Calvin College Center for Social Research, Grand Valley State University Community Research Institute, and Douglas and Maria DeVos Foundation collaborated to conduct a census of congregations in Kent County.Researchers conducted a street-by-street driving-and-walking census.Further, congregational lists derived from telephone directories, denominational websites, and InfoUSA's American Church Lists service informed the census.This effort resulted in a list of 720 total congregations, 409 of which were located in a primary study area (PSA) defined by the researchers as U.S. Census block groups containing a public school with high percentages of students receiving free or reduced lunch (Hernandez et al. 2008).The remaining 311 congregations in the Kent County census lay outside the PSA.KCCS researchers invited congregational leaders to participate in a face-to-face or telephone survey.One of the primary goals of the survey was to document the community engagement of congregations.Leaders from 583 congregations agreed to participate, representing a response rate of approximately 81 percent (Hernandez et al. 2008).
Preliminary analyses of available data on nonparticipating congregations revealed several slight differences between these congregations and those that chose to participate.Nonparticipating congregations were less likely to be located in a PSA or one of the census block groups with higher-need schools.There were also slight differences in the religious tradition of nonparticipating congregations.Analyses reveal that Black Protestant and non-Christian congregations were less likely to participate than congregations of other religious traditions.These findings suggest that white Protestant and Catholic congregations located within one of the study's PSAs are slightly overrepresented among participating congregations.Non-Christian, Black Protestant and congregations located in more affluent areas may be underrepresented.To ensure that differences did not affect findings, we constructed and applied sample weights to account for non-response as described in our methodology.For the current study, we restrict analyses to data collected from 500 congregations responding to the survey items under examination.
While our findings are not generalizable to the population of U.S. congregations, KCCS data are ideal for addressing our questions.Derived from a census of congregations in one county, KCCS data make it possible to compare the collaboration and service provision of congregations sharing geographic, social, and political space.Further, the details gathered about congregational programs allow us to examine the extent of collaborations in ways previous surveys have not.Finally, utilizing data from the 2000 U.S. Census, KCCS researchers matched participating congregations with demographic data for the census block in which they were located, making it possible to estimate the racial/ethnic makeup of each congregation's neighborhood.

Kent County, MI
The presence of a diverse and growing county population and a robust congregational marketplace make Kent County, MI a valuable context for examining the intersection of diversity and the bridging activities of congregations.Located in Western Michigan, Kent County had a population of approximately 599,524 in 2006 (U.S. Census Bureau 2018).The county comprises approximately 35 villages, townships, and cities with Grand Rapids, the county seat, being the largest with a population of 163,736 in 2006 (Kent County 2018; U.S. Census Bureau 2018).At the time of the KCCS survey, Kent's population was slightly less diverse than the U.S. The county had a higher percentage of white residents (77.7%) than the U.S. (66.2%); and Hispanic/Latino and African American individuals each made up approximately 9 percent of the county-lower than U.S. estimates for each group (15% and 12%).Other racial and ethnic groups made up an estimated 4 percent of the population compared to 6.7 percent in the U.S. (U.S. Census Bureau 2018).Data from the KCCS census also indicate that Kent County has a robust religious sector.With 720 identified congregations, we estimate that the county was home to approximately 1.20 congregations per 1000 residents in 2006 (Hernandez et al. 2008).This is slightly higher than the 1.11 congregations per 1000 residents estimated for the U.S. as a whole in 2010 (Grammich et al. 2012).

Dependent Variables
The dependent variables under analysis represent three measures of congregations' civic activity.First, leaders were asked whether their congregation had participated in any social service, community development, or neighborhood organizing projects in the past year.Leaders who responded "no" were asked if their congregation had engaged in any human service projects, outreach ministries, or other activities intended to help people who are not members.Combining responses to these items, we constructed one dichotomous measure, external engagement, which indicates whether a congregation had been involved in programs intended to engage groups outside of the congregation during the previous year (1 = yes, 0 = no).
If a leader indicated that their congregation had participated in such activities, they were asked to list each of the programs.Each congregation's list was tallied to create a count variable, service programs, representing the number of externally focused programs supported by the congregation.The number of programs listed ranged from 0 to 16. Leaders were then asked to provide additional information about each program, including the names of the two most important organizational collaborators.KCCS researchers tallied the number of unique organizational collaborators listed for each congregation, excluding congregations, to construct a count variable, unique collaborations.The number of collaborators listed ranged from 0 to 15.These three variables allow us to examine the extent of congregations' service provision and interorganizational linkages.Descriptive information for variables is provided in Table 1.

Independent Variable
Our independent variable reflects the racial composition of each congregation.Leaders were asked to estimate the percent of regular participants belonging to each of four racial/ethnic groups (i.e., white/non-Hispanic, African-American, Hispanic or Latino, Asian or Pacific Islander).Drawing on the widely used criterion that a multiracial congregation is one in which no more than 80% of participants belong to one racial/ethnic group (Emerson and Woo 2006), we constructed a dichotomous variable indicating multiracial congregations (1 = multiracial).

Control Variables
Controls included in multivariate analyses are congregation age (in years), congregation size measured as four dichotomous variables indicating the number of regular participants (i.e., 0-99, 100-349, 350-999, 1000 and greater), the number of full-time paid staff, an item indicating the financial stability of the congregation (1 = not at all stable, 2 = not very stable, 3 = somewhat stable, 4 = very stable), whether the congregation is located in an urban area (1 = urban, 0 = non-urban), and the pastor's highest level of education (1 = less than high school, 2 = high school, 3 = some college, 4 = college, 5 = some graduate education).We included a dichotomous measure indicating whether the congregation is located in a PSA (1 = in PSA).Defined by the researchers as a U.S. Census block group containing a public school with high percentages of students receiving free or reduced lunch, we propose using this measure as a proxy for communities experiencing higher economic need (Hernandez et al. 2008).We also included a dichotomous measure indicating whether the pastor would characterize the congregation as theologically conservative (1 = theologically conservative) and controls for religious tradition measured as five dichotomous variables (i.e., evangelical Protestant, mainline Protestant, Black Protestant, Roman Catholic, other).The "other" category includes all non-Christian congregations that participated in the study.We also utilize estimates provided by leaders to control for aspects of congregations' demographic composition.We include estimates for the percent of attenders with household incomes less than $25K, the percent with household incomes greater than $100K, and the percent of attenders having earned a college degree.Finally, drawing on population data from the 2000 U.S. census, we include a dichotomous variable indicating whether the census block in which each congregation is located would be considered multiracial (1 = multiracial).

Modeling Strategy
Because our first dependent variable, external engagement, is dichotomous, we utilize logistic regression to examine whether multiracial congregations are more or less likely than other congregations to be engaged in externally focused service programs.The two remaining dependent variables represent a count of programs and a count of unique collaborators.Because count variables are not truly linear and not characterized by normal distributions, they violate assumptions of ordinary least squares estimation.We utilize Poisson regression for analyses of these variables.Based on the results of initial diagnostics to determine goodness of fit and test for overdispersion, we also ran a series of negative binomial models for analysis of unique collaborations.Results did not differ significantly; for ease of interpretation, we report results from Poisson regression models (see Berk and MacDonald 2008).We also constructed and applied non-response weights for all analyses to assess whether non-response bias had any effect on findings.To construct sample-based, non-response weights, we utilized two variables available for all nonparticipating congregations, religious tradition and a dichotomous measure indicating whether a congregation was located in a municipality associated with the study's PSA (Lynn 1996).Separate analyses with both weighted and unweighted data revealed no significant or substantive effects on our findings.As a result, we report results from analyses of unweighted data below.

Results
We first examined whether multiracial congregations were more likely than others to have been involved with externally focused service programs.While most congregations (86%) reported some type of external engagement, logistic regression results presented in Table 2 reveal that multiracial congregations were more likely to have participated in such programs (Odds Ratio = 3.585, p < 0.05).In fact, they were over 200% more likely to report this type of engagement.Congruent with previous research, we found that larger congregations and those with a higher percentage of college graduates were also more likely to participate in externally focused programs, while congregations with wealthier attenders and poorer attenders were less likely to do so.We also found that Roman Catholic congregations and other congregations were less likely to participate.The diversity of the neighborhood in which a congregation was located had no effect.These findings support the notion that multiracial congregations represent stores of bridging social capital; they appear to be particularly active in the provision of externally focused services.Note: a The omitted reference category for all analyses is 0-99 attenders.b The omitted reference category for all analyses is Mainline Protestant.* p < 0.05.** p < 0.01.*** p < 0.001.
Table 3 presents coefficients and incident rate ratios (IRR) from a series of Poisson regression models predicting the number of externally focused service programs (Models 1 and 2) and unique collaborations (Models 3-5) reported by congregations.Model 1 includes only control variables and indicates that larger congregations located in a PSA tend to support more service programs than other congregations do.In other words, larger congregations located in communities where higher levels of economic need are likely to exist support more service programs.Table 3 reveals that religious tradition is also related to the number of programs reported.The comparison group, mainline Protestant congregations, supports more programs than Roman Catholic congregations.Roman Catholic congregations report being involved in 44% fewer programs (IRR = 0.563, p < 0.01).There is no significant difference between the number of programs supported by mainline Protestant and evangelical, Black Protestant, or other congregations.In Model 2, we introduce our variable for multiracial congregations.Because the racial/ethnic composition of the community within which a congregation exists may have an impact on a congregation's demographic makeup, we include a control for multiracial neighborhoods.Findings reveal that, in addition to being more likely to participate in externally focused programs, multiracial congregations supported a larger number of programs than their less-diverse counterparts.Multiracial congregations in the study reported 22% more external service programs than other congregations (IRR = 1.221, p < 0.05).This effect does not appear to be influenced by neighborhood composition.Congregations in multiracial neighborhoods supported no more or no fewer programs than those in more homogenous neighborhoods.
Model 3 includes only control variables and predicts unique collaborations.Similar to the number of programs offered, we found that larger congregations located in a PSA reported a higher number of collaborations than other congregations.The largest congregations, those with 1000 or greater participants, reported 64% more collaborations than those with fewer than 100 participants (IRR = 1.642, p < 0.05).This is not surprising as larger congregations are likely to possess larger stores of human and social capital, which has been linked with service provision (Chaves 2004).We also found that congregations with more full-time paid staff reported more collaborations (IRR = 1.016, p < 0.01).Each additional full-time staff member increased the collaborations reported by one percent.Religious tradition also influences the number of collaborations.Roman Catholic and other congregations reported fewer collaborations than did mainline Protestant congregations.Interestingly, evangelical Protestants reported approximately 27 percent fewer collaborations than mainline Protestants (IRR = 0.731, p < 0.05).Despite no significant difference in the number of programs supported, evangelical Protestant congregations reported fewer collaborations.
In Model 4, we introduce our variables for multiracial congregations and neighborhoods.Control variables continue to operate largely as they did in the base model.Model 4 also indicates that multiracial congregations reported significantly more collaborations than did less-diverse congregations (IRR = 1.542, p < 0.001).On average, they reported 54% more collaborations.The diversity of a congregation's neighborhood had no effect.Because the number of collaborations reported by a congregation is likely to be influenced by the total number of externally focused programs, and because multiracial congregations tend to support more programs, we developed an additional model controlling for the number of service programs reported.Poisson results in Model 5 reveal that even when controlling for the number of programs supported, multiracial congregations reported more collaborations (IRR = 1.354, p < 0.01).Holding the number of service programs constant, multiracial congregations reported 35% more collaborations than other congregations.Multiracial congregations in Kent County are not only engaged with their communities through a larger number of programs; they also collaborate with a larger number of other organizations.

Discussion
In this article, we have sought to extend research on the intersection of diversity and civic life by bringing the work of one common type of CBO, religious congregations, into the conversation.Previous research contends that diversity may deter some forms of civic participation (Putnam 2007).However, few studies have directly examined the relationship between racial/ethnic diversity and the work of CBOs and associations.Drawing on survey data from a census of congregations in Kent County, MI, our findings suggest that multiracial congregations may represent unique stores of bridging social capital in communities.These organizations, comprising 16 percent of the current sample, were more likely to support externally focused programs, tended to support a larger number of programs, and developed more interorganizational collaborations than other congregations developed.In other words, they appeared to play a more significant role in providing services and connecting citizens through service activities than other congregations did.
We were surprised to find that the diversity of a congregation's neighborhood had no effect on the number of service programs or the number of interorganizational collaborations reported.This suggests that it is not merely diverse contexts or the presence of diverse needs that contribute to congregations' bridging activity, but something unique about diverse organizations.While causality cannot be determined using current data, we suggest that the presence of diverse attenders within a congregation represents a unique resource and may encourage the development of organizational bridging ties.Congregations that bring members together across significant social cleavages may be more likely to engage in bridging activities in a community.
We also found that evangelical Protestant congregations and other (i.e., non-Christian) congregations reported fewer unique collaborators.While both of these groups reported supporting no fewer programs than other congregations did, they collaborated with fewer partners.These congregations may collaborate with a few organizations that support multiple programs, they may collaborate primarily with other congregations, or they may provide service activities without collaborating.Future research should examine more closely, how and with whom evangelical Protestant and other non-Christian congregations collaborate to serve the community.
The current study has several limitations.First, as noted earlier, findings are not generalizable to the population of U.S. congregations.Rather, they paint a more nuanced picture of congregational engagement and collaboration in one community than previous surveys have.Studies conducted in other counties and regions of the U.S. are needed to provide additional support or clarification for these findings.Additionally, we note that Black Protestant, non-Christian, and congregations outside of a PSA were less likely to participate in the congregational survey.Future scholarship should explore the ways in which diversity affects the community engagement of congregations in often-understudied groups as well as those in rural, suburban, and more affluent communities.Second, while ubiquitous, congregations are only one type of CBO and may be distinct due to their religious mission.Additional research is needed to examine how diversity is related to the work of other types of CBOs.Third, while survey questions posed by KCCS researchers allowed for analysis of unique collaborations, the dependent variable dealing with collaboration has limitations.Congregational leaders were asked to provide the names of the two most important collaborators for each service program with which their congregation was involved.Because leaders were limited to two collaborators for each program, the extent of some congregations' collaborations is likely truncated.In the future, researchers can improve this item by allowing congregations to list all collaborators for each service program.Still, the current item provides a more complete picture of collaboration than previous surveys have.We also acknowledge again that it is not possible to determine causality using the current data.Multiracial congregations represent unique stores of bridging capital.However, we cannot determine whether diversity itself generates interorganizational connections or whether congregations' connections contribute to organizational diversity.Additional research is needed to clarify this issue.We suggest that causality is likely multidirectional.Finally, while our research suggests that these congregations contribute to stronger communities through the bridging ties they foster, future research would do well to explore the practical and specific ways in which multiracial congregations and other diverse CBOs engage in bridging at the community level.This may be a particularly fruitful area of research at the current time, when race/ethnic relations remain a significant issue for many U.S. communities.
At a time when scholars are concerned about declining levels of civic participation and the health of civil society, the activity of multiracial congregations suggests that diverse CBOs and associations might play an important role in fostering and sustaining bridging social capital in communities.Further, the work of these organizations may be a particular boon to communities experiencing rapid demographic change.These organizations represent one context within increasingly diverse communities where people can come together, get to know neighbors, develop trust, and work across social divisions to address common issues and social problems.
Author Contributions: E.C.P. conceptualized the current manuscript, conducted quantitative analyses, drafted the literature review, data and methods, results, and discussion sections, revised and edited the entire paper.R.G. assisted with the literature review, revised and edited the entire paper.
Note: a N represents the number of cases for which data is available for each variable.b Percentages reported for all dichotomous variables.PSA: primary study area.

Table 2 .
Odds ratios for effects of diversity on external engagement.

Table 3 .
Poisson Coefficients and Rate Ratios for Effects of Diversity on Service Programs and Unique Collaborations.= incident rate ratios; AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion.a The omitted reference category for all analyses is 0-99 attenders.b The omitted reference category for all analyses is Mainline Protestant.* p < 0.05.** p < 0.01.*** p < 0.001.