Introduction

Individuals interface with a variety of social structures such as the family, educational systems, or religious institutions that provide a context to address social problems (Berger and Neuhaus 1977). Termed “mediating structures,” they provide unique contexts that may influence, or “mediate,” how individuals engage with society and social problems, and often serve as the practical bridges between individuals and society. Examining the role and function of these mediating structures is important in understanding how particular contexts may influence individual engagement with addressing social problems, including, for example, how religious congregations influence how and why individuals address social problems. Moreover, individuals may use the mediating structure itself (e.g., the religious congregation) as a vehicle for social justice engagement (e.g., volunteering in a congregational social justice activity). This study examines how one set of mediating structures, religious congregations, influences how individuals address social problems, with particular attention to social justice engagement, or one’s prioritization of and participation in social justice focused congregational activities.

The connection between religious congregations and efforts to address social problems is not a new revelation. Scholars within the field of community psychology have noted this connection, and called for further research to better understand how religious congregations may provide a context that promotes social justice engagement (e.g., Kloos and Moore 2000; Pargament 2008; Pargament et al. 1983). Moreover, previous studies have examined the different types of services and resources congregations offer to address social problems such as hunger and homelessness (Chaves 2001; Chaves and Tsitsos 2001). An examination of history also shows how religious individuals and groups have worked for the common good to address social problems regarding poverty (Marsden 1991) and civil rights (Sernett 1999), connecting a spiritual motivation to charitable, philanthropic, and political activity (Sider and Knippers 2005; Skillen 2004). This study extends research by focusing on social justice engagement, defined here as how much individuals believe their congregation should prioritize social justice as a part of congregational activity (i.e., prioritization), and how they personally participate in congregational social justice activities (i.e., participation). Thus, this study explores attitudes about the role of the congregation as a mediator for social justice (i.e., prioritization), and how the congregation itself serves to mediate participation in social justice activities (i.e., participation). Social justice in this study is defined as efforts to relieve poverty and hunger, to promote world peace and justice, or as addressing local issues; however, the exact form and content of these activities are not explicitly defined. Also, individual and congregational characteristics that may be associated with social justice engagement are examined to elucidate how aspects of the congregational setting may influence individual social justice engagement. Specifically, the extent to which theological orientation, frequency of participation in the congregation, bonding social capital, and bridging social capital are associated with social justice engagement are examined at individual and congregational levels of analysis. Examining these individual and congregational predictors of social justice engagement is an important step in understanding how congregations function as mediating structures to promote social justice.

Individual and Congregational Characteristics

Theological Orientation

Theological orientation refers to theological conservativism/liberalism, often defined by denominational affiliation, specific theological beliefs (e.g., otherworldliness, literalism of scriptural interpretation, “orthodox” Christian beliefs), or self-identification (i.e., fundamentalist, Evangelical, mainline-liberal; Marsden 1991; Smith 1987; Spilka et al. 2003; Woodberry and Smith 1998). Typically, theological conservatives promote literalist biblical interpretation, evangelism as outreach, and individual separation from society whereas theological liberals adhere to a non-literalist biblical interpretation, social works as outreach, and remain active in society (Hood et al. 2005). Historically in the last century, individuals and congregations have been defined along this continuum with direct implications for how these congregations engage with social problems (Chaves 2001; Wuthnow 1996). For example, research has shown liberal congregations provide more social services and collaborate more with secular and governmental organizations than conservative congregations who focus more on meeting the needs of congregational members through the congregational community (Chaves 2001; Hoge et al. 1998). Moreover, Kanagy (1992) found that congregational theological orientation, as defined by the congregation’s aggregated view of the Bible, related to the prioritization of social action with more liberal congregations placing a higher priority on social action. In addition, theological orientation was important in predicting where and how individuals volunteer, with theological conservatives volunteering within their congregation and liberals volunteering more in secular settings (Hoge et al. 1998; Park and Smith 2000). Therefore, theological orientation may be an important individual and congregational variable that is associated with social justice prioritization and participation, with theological liberalism predicting higher levels of both.

Frequency of Religious Participation

Frequency of religious participation, or how often individuals participate in their religious congregation, is a second variable that may be associated with social justice engagement. As a basic tenet of ecological theory, the influence of a setting or mediating structure depends on individual presence and engagement in the setting, with the influence of the mediating structure increasing with greater participation (Bronfenbrenner 1979). Wilson and Janoski (1995) specify that increased participation leads to the increased internalization of group norms. Thus, the question is not only about frequency of participation, but of participation in what type of setting, as different settings may socialize members according to different group norms. For example, Deckman (2002) found that associations between theological conservativism and particular social attitudes were moderated by frequency of attendance, with stronger associations present for those with higher attendance. This indicates that there may be some type of interaction between theological orientation and frequency of participation in predicting social attitudes. The present study extends the literature by examining the theological orientation of the congregation as a moderator of the association between individual level frequency of religious participation and social justice engagement. Given that liberal congregations may have greater prioritization of social justice (Kanagy 1992), it is expected that participation in liberal congregations will be associated more strongly with social justice prioritization and participation than participation in conservative congregations.

Bonding and Bridging Social Capital

Putnam (2000) differentiates bonding and bridging as two types of social capital. Bonding social capital refers to connections among individuals within a particular group (i.e., one’s congregation), and norms of reciprocity and trust that are produced from such relationships. Putnam theorizes that such connections maintain strong in-group loyalty and mobilize solidarity around shared values. Studies examining bonding capital in religious communities have found greater bonding related to increased connections across class status (Wuthnow 2002), the facilitation of moral development in adolescence (King and Furrow 2004), and participation in civil rights activism in Black congregations (Brown and Brown 2003; Chaves and Higgins 1992). In the current study bonding is defined by the number of close friends within one’s congregation, with the expectation that increased individual and congregational bonding will facilitate greater social justice engagement, as connected members may be more aware of congregational needs, or of opportunities within their congregation to address wider issues of social justice. Bridging social capital refers to ties between heterogeneous group members (i.e., other congregations; Putnam 2000). These bridging connections may help counter homophilly, or the tendency for “birds of a feather to flock together” (McPherson et al. 2001). Furthermore, Granovetter (1973) notes the importance of bridging ties between diverse groups, even if they are weak, as they provide linkages between multiple groups and open up the possibility for collaboration, cooperation, and the sharing of resources to address larger scale social problems. Ammerman (2002) notes that congregational partnerships are one type of bridging tie and are often in service of addressing larger community needs or social justice issues, showing that bridging for congregations may be in direct service of social justice. Therefore, higher bridging may indicate greater prioritization of social justice, and may provide opportunities for members to participate in justice activities.

Demographic Controls

Previous research has shown that demographic variables are associated with the study variables. Wuthnow (1996) notes how the demographic variables of education and age are related to theological orientation, with those who are less educated and older more likely to be theologically conservative. Woodberry and Smith (1998) further note that education is associated with attitudes toward biblical literalism, and other authors (e.g., Felson and Kindell 2007) control for education when measures of biblical literalism are used for theological orientation. Although gender and race are not consistently associated with theological orientation (Wuthnow 1996), it is possible that these variables are associated with social justice engagement, with members of non-dominant groups (i.e., women, Blacks) more likely to prioritize and participate in congregational based social justice activities. Moreover, previous research has determined that the size and income level of the community where a church resides is associated with social justice activity, with more inner-city and poor members and congregations more involved in justice activities (Chaves 2001). Denominational affiliation also relates to theological orientation and involvement in social justice (Chaves 2001). Thus, the demographics of race, gender, education, income, community size (i.e., urbanicity), and denomination should be included as controls.

Multilevel Modeling

Examining the setting level characteristics of congregations as distinct from individual effects is an important step in understanding the degree to which, and how congregations may function as mediating structures to promote social justice. Multilevel modeling allows for this separation and is an appropriate analytic strategy to examine multiple levels of analysis while accounting for the dependence in the data (Raudenbush and Bryk 2002). Moreover, characteristics of the individual may also be important in understanding the impact of mediating structures, as individuals may respond differently to the same context depending on individual characteristics. Although previous scholarship has examined individual (Hoge et al. 1998) or congregational influences on social justice engagement (Chaves 2001; Kanagy 1992), few studies have examined both levels of analysis simultaneously. Moreover, the moderating impact of congregational setting can be tested by examining cross-level interactions, where particular individual-level associations (e.g., frequency of religious involvement and engagement with social justice) may be moderated by particular congregational settings (e.g., liberal vs. conservative congregations). Finally, multilevel modeling is able to accommodate groups of varying sizes. Therefore, there is a need for a multilevel approach that simultaneously considers individual, congregational, and cross-level interactive effects to extend the existing literature examining congregations as mediating structures for social justice.

Present Study

The current study provides an integrated multilevel examination of the individual and congregational associations of theological orientation, frequency of religious participation, bonding social capital, and bridging social capital with social justice prioritization and participation. Furthermore, this examination is at multiple levels of analysis, allowing for the specification and testing of individual and congregational effects. To date, little is known empirically about how aspects of congregational context are associated with social justice engagement. Thus, this study contributes to the growing body of literature regarding religious congregations as mediating structures for social justice engagement by examining how aspects of the congregational context uniquely predict individual social justice prioritization and participation, while controlling for other relevant variables.

Method

Data

The data used in this research, individual responses to the Church and Community Planning Inventory (CCPI), are from a large multi-method project conducted by the Center for Church and Community Ministries, funded by the Lilly Endowment, and collected by Carl Dudley. The purpose of the project was, “to study congregational transformation toward social ministry” (Dudley 1991, p. 197) and encompassed multi-year involvement with Christian congregations in Illinois and Indiana to facilitate the development of social justice initiatives. The CCPI was administered at the beginning of the project in 1987 to collect baseline data on participating congregations and is publicly available (http:\\www.thearda.com). As noted by Dudley, many of the CCPI items were drawn from previous scholarly work though items were also circulated and discussed by congregational leaders prior to dissemination. The final data set used in this study consists of 5,123 members of 62 congregations representing 11 different denominations. Participants per congregation ranged from 14 to 222.

A multi-faceted sampling method was used to select demographically “typical” mainline and evangelical churches within Illinois and Indiana (see Dudley 1991 for full description of sampling methods). The demographics of interest were geographic population (one-third rural, one-third small cities and suburbs, one-third metropolitan areas), denominational affiliation, congregation size, race/ethnicity, and liberal/conservative identification. To facilitate this representation, four geographic regions were selected (Chicago, Central Illinois, Northern Indiana, and Indianapolis). Within each region the main investigator and several denominational leaders from those areas selected prospective congregations that were “typical” and also were identified by these leaders as not heavily engaged in social ministry. Congregations not heavily engaged in social ministry were selected as the larger goal of the project was to increase sustainable social ministry. Prospective congregations were approached for participation in the larger multi-year project, offered seed money and support for the development of social ministries, were asked to plan and document their progress, and were asked to have lay leaders facilitate the endeavor. Dudley (1991) notes that about one-third of the congregations declined, with a lack of interest by lay leaders cited as the main reason. Out of the remaining congregations, survey packets were mailed to all members, though “sampling procedures” were used in congregations larger than 250. The average return rate was 55%, a rather high rate attributed to buy-in and support from church leaders.

Measures

Social Justice Prioritization

A four item scale was used to assess social justice prioritization. Each item was in response to the statement, “There are many tasks that a church can do. Of those listed below, what priority would you give to each for your church?” Sample items include “Encourage individual members to support local social reforms to relieve poverty and hunger” and “Develop church programs which would help people understand local programs and issues.” Items are rated on a Likert-type scale ranging from 1 (none) to 5 (highest). For the present investigation, the internal consistency estimate of the prioritization scale was .82. See Table 1 for individual and congregational level descriptive statistics.

Table 1 Summary of intercorrelations, means, and standard deviations for study variables at individual and congregational levels
Full size table

Social Justice Participation

One item was used to assess social justice participation. First, participants were asked, “Does your congregation participate directly in any programs to provide a community social service or promote peace and justice?” Seventy-five percent of participants answered yes to this question, and were then asked “How much time have you been able to give to these programs?” to assess their level of participation. Response options ranged from 1 (I am not able to give any time) to 5 (More than 5 h a week). The distribution of responses was positively skewed. Therefore, participation was dichotomized into those who gave no time (69.4%) and those who gave any amount of time (30.6%), coded 0 and 1 respectively. Participation thus represents participation or not for people who were aware of congregational social justice programs.

Theological Orientation

Belief about literal interpretation of the Bible was used to indicate theological liberalism/conservativism. Participants were asked “Which of the following statements best expresses your view of the Bible?” with four response options ranging from 1 (The Bible is the actual Word of God and is to be taken literally) to 4 (The Bible is a valuable book because it was written by wise and good people, but I do not believe it is really God’s word). Higher numbers represent theological liberalism, and smaller numbers theological conservativism. At the individual level this variable represents personal theological orientation, and at the level of the congregation the average theological orientation in the congregation. Although one item cannot capture the complexity of theological orientation (Woodberry and Smith 1998), the use of a view of Bible question to indicate theological liberalism/conservativism is relatively frequent and well established in the sociological and psychological literature (Kellstedt and Smidt 1993). Such an item predicts self-identification as theologically liberal or conservative (Wuthnow 1996), other theological beliefs that are considered conservative such as “orthodox” or “fundamentalist” beliefs (Altemeyer and Hunsberger 2005), political identification as liberal or conservative and associated presidential voting behavior (Kellstedt and Smidt 1993), other political attitudes associated with conservativism such as attitudes toward gay marriage (Olson et al. 2006), abortion (Kellstedt and Smidt 1993), and social conservativism (Felson and Kindell 2007).

Frequency of Religious Participation

Frequency of religious participation was assessed by a three-item scale. Each item was in response to the question, “How often do you personally do the following” with the response options ranging from 0 (never) to 7 (every day). The three items were “Attend worship service,” “Participate in religious courses or Sunday school,” and “Participate in other church activities other than worship.” Frequency of religious participation thus assessed both participation in worship services and extra congregational activities such as religious education. This variable represents personal religious participation at the individual level, and at the level of the congregation the average amount of participation in the congregation. For the present investigation, the internal consistency estimate of the frequency of religious participation scale was .64.

Bonding and Bridging Social Capital

Bonding social capital was assessed using one item regarding friendships within the congregation. Participants were asked, “Of your five closest friends, how many are members of this congregation?” with response options ranging from 0 (0 friends) to 5 (5 friends). At the individual level, more friends indicate more personal bonding social capital, whereas at the congregational level higher numbers indicate a more densely bonded congregation. Bridging social capital was assessed with a single item asking about partnerships with churches of other denominations. Participants were asked, “To what extent do you agree that [this] statement describes your congregation? Cooperative projects and joint workshops with churches of other denominations are highly valued” with response options ranging from 1 (don’t know) to 5 (agree strongly). Due to the ambiguity of a “don’t know” response, participants (6.6%) were excluded if they endorsed such a response. At the individual level, bridging represents an awareness of congregational bridging activities whereas at the congregational level bridging is an aggregate perception that the congregation values cooperative projects with other denominations.

Denomination

The sixty-two congregations in this study were nested within the 11 denominations of American Baptist, AME Zion, Brethren, Disciples of Christ, ELCA, Evangelical Covenant, Presbyterian USA, Southern Baptist, United Church of Christ, United Methodist, and Catholic. These denominations represent the full spectrum of liberal to conservative denominations, though scholars have noted within denominational variability on theological orientation (Woodberry and Smith 1998). Denomination was included as a control at the congregational level, given that some of the congregational variability may have been explained by denominational affiliation. Ten dummy-coded variables were used to represent denomination, with Catholic as the reference group.

Demographic Control Variables

Six demographic controls were assessed. Age was recorded in years, with an average age of 51.46 (SD = 18.18). Education was assessed on a one (did not graduate high school) to six (post graduate degrees) scale, with an average education of 3.38 (SD = 1.53) which is between having some college and not finishing a college degree. Income was assessed on a one (1,000–1,999) to seven (60,000 or more) scale, in increments of 10,000. The average income was 3.74 (SD = 1.81) which is closest to the 30,000–39,000 income bracket. Urbanicity was assessed on a one (midtown) to nine (rural settlement) scale, with smaller numbers representing a more urban location. The average urbanicity was 4.76 (SD = 2.10), closest to the “metro suburb” location. Gender was coded zero for male (35.5%) or one for female (62.1%). Race was coded zero for White (86%) or one for Black (10%).

Multilevel Modeling Data Analysis Strategy

Multilevel modeling was used to examine the individual and congregational influence of theological orientation, frequency of religious participation, bonding social capital, and bridging social capital on the two social justice engagement outcomes of prioritization and participation, while controlling for individual demographics and congregational denomination. Congregation level variables were created by computing the mean for each congregation. Multilevel modeling allows for the separation and simultaneous testing of level I (i.e., individual) and level 2 (i.e., congregational) effects for nested data structures (i.e., individuals nested within congregations) while accounting for dependence in the data (Raudenbush and Bryk 2002; Snijders and Bosker 1999). Moreover, interactions between level I and level II (i.e., cross-level interactions) are tested to assess if a level II variable moderates associations at level I. These cross-level interactions test if there are different patterns of associations at the individual level in different types of congregations (e.g., liberal or conservative, Shinn and Rapkin 2000).

To examine both the individual and congregational effects of the study variables, five models were examined that sequentially added demographics, level I variables, denomination, level II variables, and the cross-level interaction between individual frequency of religious participation and congregational theological orientation to predict each outcome. Parameters for fixed effects were examined in each model to determine the presence of individual, congregational, and cross-level interactive effects, as displayed in Tables 2 and 3. All independent variables were standardized. Grand-mean centered individual level variables were used so that the congregational level fixed effects would be a pure estimate of the “compositional” congregational impact, representing the congregational impact over and above individual effects (see Raudenbush and Bryk 2002 for a discussion of how level I centering impacts level II parameter estimates).

Table 2 Multilevel modeling of social justice prioritization
Full size table
Table 3 Multilevel logistic regression modeling of social justice participation
Full size table

The two outcomes of interest, prioritization and participation, were two different types of variables which influenced the multilevel modeling approach. Prioritization was a continuous outcome, and PROC MIXED in SAS 9.2 was used to model the data. Participation was a binary outcome, thus PROC GLIMMIX in SAS 9.2 was used to conduct multilevel logistic regression. This procedure allows for multilevel modeling of binary outcomes, which was specified with a logit link and binomial distribution. For participation, the Laplace method of estimation was used, as this numeric maximum likelihood method of estimation allows for the computation of likelihood estimates (Snijders and Bosker 1999). Finally, parameter estimates for the social justice participation fixed effects are reported in Table 3 as standardized coefficients (\( \hat{\beta } \)), though they are converted to odds ratios (i.e., \( {\text{e}}^{\hat{\beta }} \)) and discussed as such in the text. For example, \( \hat{\beta } \) = .21 for bonding indicates that for a one unit increase on the bonding scale, the odds of social justice participation are 1.23 times larger (e.21 = 1.23). For both outcomes, the same five models were compared.

Results

Social Justice Prioritization

To determine the amount of variance in prioritization that could be explained at the congregational level, the intra-class correlation was computed from the variance components of the random intercept null model (i.e., no independent variables). The result indicated that 9.52% of the variance in scores could be explained at the congregational level. Although most of the variability is at the individual level (i.e., 89.48%), almost 10% is a sizeable amount that should be further explored to understand what aspects of the congregation are associated with individual social justice prioritization. Further examination of descriptive graphs confirmed the need to model a random intercept for prioritization, which is included in all further modeling.

Table 2 shows the results from the five models examining individual and congregational predictors of social justice prioritization. The first model shows that the demographic variables of age, gender, race, urbanicity, and income were significantly associated with prioritization (see Table 2). Overall, these results indicate that demographic variables are associated with prioritization, with those from underprivileged groups (i.e., women, Blacks, urban, and poor) exhibiting higher levels of social justice prioritization.

The second model added the individual level religious variables, while controlling for all demographic variables. All of the religious variables had a significant positive association with prioritization. This indicates that those who are more theologically liberal, participate more in their congregation, have higher numbers of close friends in the congregation, and who view their congregation as a bridging congregation had higher levels of social justice prioritization. Moreover, bridging had the strongest effect of all the variables in the model. Furthermore, the same demographic variables from the first model remained significant after adding the individual level religious variables.

The third model added the denominational affiliation of the congregation to the prediction of social justice prioritization as a level two control. Six of the ten denominations had significantly different levels of social justice prioritization than Catholics (see Table 2). Furthermore, none of the estimates for the level 1 variables changed appreciably after adding in denomination. Finally, the addition of denomination improved the model fit indices.

The fourth model added in the congregational level variables to assess the influence of congregational context on individual social justice prioritization, after controlling for denomination and other level 1 variables. Congregational bridging was the only significant congregational effect, indicating that congregational theological orientation, frequency of participation, and bonding did not have an influence on individual prioritization. Overall, individual level bridging remained the strongest predictor, as well as congregational bridging emerged as a congregational level predictor of social justice prioritization.

The fifth model added the cross-level interaction between individual-level frequency of religious participation and congregational theological orientation, which was significant in this model, indicating that the association between frequency of religious participation and prioritization was not the same in liberal versus conservative congregations. Follow up regressions were conducted for liberal and conservative congregations (e.g., as defined by a median split using congregational theological orientation) to examine this significant cross-level interaction. There was no association (β = .02, SE = .01, p = .27, 95% CI [−0.02, 0.06], n = 2,127) between frequency of participation and prioritization for people in conservative congregations, whereas there was a significant and positive association (β = .09, SE = .01, p < .05, 95% CI [0.06, 0.11], n = 2,761) for people in liberal congregations, as displayed in Fig. 1. Moreover, congregational bridging remained a significant congregational level predictor. The pattern for denominations was similar to model four. All individual level religious predictors remained significant, with bridging still evidencing the strongest association with prioritization. Overall, demographic (i.e., age, gender, race, urbanicity, and income), individual (i.e., theological orientation, frequency of religious participation, bonding, and bridging) congregational (congregational bridging), and cross-level interactive effects (frequency of religious participation * congregational theological orientation) were present in predicting individual social justice prioritization. A summary model of the findings for social justice prioritization is presented in Fig. 2.

Fig. 1
figure 1

Congregational theological orientation moderating religious participation and social justice prioritization

Full size image
Fig. 2
figure 2

Summary model of results for social justice prioritization. Note Solid arrows indicate significant associations, dashed arrows indicate non-significant associations. Level-two variables are in shaded boxes. Perpendicular lines indicate cross-level moderation

Full size image

Social Justice Participation

An intra-class correlation was not computed for social justice participation as this is not recommended for multilevel logistic regression (Raudenbush and Bryk 2002). Instead, a conditional likelihood ratio test was used to compare a null model with and without a random intercept, assessing how the inclusion of the random intercept may improve model fit. The model with the random intercept improved the model fit (p < .05); thus, a random intercept was included in all subsequent modeling. The same five models used for prioritization were examined for participation, with statistical information presented in Table 3.

The first model shows that the demographic variables of race, education, and income were significantly associated with social justice participation. This indicates that those who are Black, more educated, and who had less income had greater odds of participation in congregational social justice activities (ORs = 1.17, 1.26, and 1.08 respectively) than those who are White, have less education, or who have more income.

The second model added the level 1 variables. Frequency of participation and bonding social capital were both significant, with frequency of participation emerging as the strongest predictor. Thus, those who participate more in their congregation, or who have more close friendships within the congregation, have greater odds of participating in congregational social justice activities (ORs = 2.20, 1.23 respectively) than those who participate less or who are less bonded. Theological orientation and bridging social capital were not significant. Finally, the same demographic predictors were significant.

The introduction of denomination in model three had little effect on the individual level demographics or religious predictors. The model fit indices only had a slight decrease when denomination was added, as noted in Table 3. Nevertheless, denomination is included in further modeling to serve as a level 2 control for congregations.

The fourth model added the congregational level variables to assess the influence of congregational setting on individual social justice participation, after controlling for denomination. None of the congregational level predictors were significant, indicating that congregational theological orientation, frequency of religious participation, and congregational bonding and bridging social capital did not have direct effects on individual participation in social justice activities. Finally, the model fit statistics remained relatively constant, indicating that these congregational level variables did not appreciably improve the model fit.

The fifth model added the cross-level interaction between individual frequency of religious participation and congregational theological orientation to examine if there were different associations between frequency of participation and social justice participation within liberal and conservative congregations. The cross-level interaction was significant. Follow up logistic regressions showed a significant association between frequency of religious participation and social justice participation for people in conservative congregations (β = .53, SE = .05, p < .05, 95% CI [0.43, 0.63], OR = 1.70, n = 1,608). There was also a significant and slightly stronger association for people in liberal churches (β = .70, SE = .04, p < .05, 95% CI [0.62, 0.78] OR = 2.02, n = 2,227). This shows that congregational theological orientation moderated the association between frequency of religious participation such that there were greater odds of social justice participation with increased frequency of participation, with a more pronounced effect in liberal congregations. Finally, the inclusion of the cross-level interaction did not alter effects for other study variables. The final model showed significant effects for race, education, income, individual level frequency of religious participation and bonding, and a more pronounced effect for frequency of religious participation in liberal congregations. A summary model of the findings for social justice participation is presented in Fig. 3.

Fig. 3
figure 3

Summary model of results for social justice participation. Note Solid arrows indicate significant associations, dashed arrows indicate non-significant associations. Level-two variables are in shaded boxes. Perpendicular lines indicate cross-level moderation

Full size image

Discussion

The purpose of this study was to examine religious congregations as mediating structures for social justice. Using multilevel modeling, the study found that at the individual level, bonding and bridging social capital and theological orientation were associated with prioritization, whereas bonding social capital predicted social justice participation. Individual demographics also predicted social justice prioritization and participation. Yet, results indicated that congregational bridging social capital predicted only social justice prioritization. Furthermore, congregational theological orientation served to moderate the associations between frequency of religious participation and both prioritization and social justice participation. This demonstrates that particular aspects of congregational context, such as congregational bridging and theological orientation, were associated with social justice prioritization and participation over and above individual effects and demographic controls. These findings suggest that religious congregations serve as mediating structures of social justice, with particular congregational and individual characteristics influencing how individuals prioritize and participate in congregational social justice activities.

Theological Orientation and Frequency of Religious Participation

The findings from this study suggest that the theological orientation of individuals and congregations are related to social justice engagement. Such findings affirm observations from history, (Marsden 1991; Woodberry and Smith 1998) sociology (Chaves 2001), and psychology (Hood et al. 2005) that theological dimensions influence social justice attitudes (i.e., prioritization) and behavior (i.e., participation). Furthermore, these findings extend these literatures by simultaneously examining theological orientation at individual and congregational levels of analysis, allowing for implications to be drawn about both personal and congregational theological orientation. At the individual level, the findings indicated that theological liberalism was positively associated with the prioritization of, but not participation in congregational justice activities. It appears, at the individual level, that liberalism facilitated a priority on justice activities within a religious congregation. This interpretation is consistent with historic accounts of the theological conservative movement pushing away social justice as a role of the church (being apart from the world); whereas a defining feature of theological liberalism was embracing social justice (being a part of the world; Marsden 1991). To be clear, these findings do not indicate that conservative individuals or congregations are anti-justice or are not engaged in activities they see as making the world a better place, but the findings do reflect differences in how social justice is articulated and enacted between liberal and conservative individuals and settings. As discussed in the limitations section, the current study does not allow for conclusions or comparisons about the types or even amounts of social justice activities, but rather shows that theological liberals prioritize social justice as part of the mission and function of the congregation. This limitation points to the need of future research to examine how theologically liberal and conservative people articulate an understanding of social justice, and how both groups may link their spirituality to doing good in the world, however defined.

At the level of the congregation, theological orientation played a more subtle role in the prediction of social justice activities by acting as a moderator of individual level relationships. Specifically, the theological orientation of the congregation did not demonstrate a main effect (i.e., the average congregational theological orientation did not predict individual social justice prioritization), but interacted with the frequency of individual religious participation, such that the more involved people were in liberal congregations, the more they prioritized and participated in social justice activities whereas no association (e.g., for prioritization) or a weak association (e.g., for participation) was present in conservative congregations. Thus, it was not only being more involved that was associated with social justice prioritization and participation, but involvement in a particular type of setting, showing how a characteristic of the setting (i.e., theological orientation) accounted for the extent to which the setting does indeed mediate individual social justice engagement. Given that previous scholarship has found that liberal congregations provide more social services (Chaves 2001), it is possible that there are more social justice options in a liberal congregation, leading to a higher likelihood of involvement in social justice activities for those who are more involved. More social justice opportunities may also perpetuate a self-selection process, where individuals interested in social justice action select into liberal justice focused congregations as this is a good person-environment fit with the congregation providing a backdrop of support and opportunities for these individuals (Moos 2002). A community culture promoting social justice, possibly created through sermons, teachings, music, and programs aimed at issues of justice may also influence how the individuals value justice (i.e., prioritization), and how their time is spent in congregational activities (i.e., participation). Though speculative, these findings point to the need of understanding more concretely how and why liberal congregations promote this association between religious participation and the prioritization and participation in congregational justice activities. Finally, these findings show how religious congregations serve as mediating structures for social justice, influencing both the prioritization of and participation in social justice, and by providing a concrete location for people to be involved in justice activities, mediated through the congregation.

Bridging and Bonding Social Capital

Bridging

The results suggest that bridging, or congregational interdenominational collaboration, also influenced how individuals prioritized, but not participated in, congregational social justice activities. Specifically, individual awareness of congregational bridging was associated with individual prioritization of social justice. Given that congregations often bridge with other congregations for social justice of activities (Ammerman 2002), it is possible that individuals see and internalize a commitment to justice by observing the justice actions of the congregation. Of course the directionality of this association cannot be determined, and those who are more aware of bridging may simply be those who already have a higher priority for social justice. Nevertheless, it is this acknowledgement of an outward, collaborative focus that is associated with how individuals prioritize justice as a congregational activity. Moreover, bridging at the level of the congregation also predicted individual prioritization, but not participation in social justice activities. Higher congregational bridging may indicate that the congregation in fact has more collaborations, and that the presence of these collaborations communicates to members the importance of social justice, possibly because they see their congregation and congregational leaders involved in such activities.

Furthermore, higher congregational bridging may promote a climate of openness and collaboration, having an outward focus on meeting larger community and societal needs. Although this study shows how congregational bridging influences individual prioritization, future research could explore how this setting level bridging translates into individual prioritization. Finally, at individual and congregational levels, bridging related to prioritization of, and not participation in, congregational social justice activities. Apparently, bridging influenced the value placed on justice, but not actual involvement behavior. It is possible that observing others involved in social justice may increase individual prioritization, but may not provide the structure or support necessary for personal involvement. In addition, the two outcomes of prioritization and participation may tap into different aspects of social justice (e.g., an abstract ideal versus concrete actions) and may hold different meaning to members, thus being predicted by different individual and congregational level variables. Future research may help to elucidate these differences.

Bonding

In contrast to bridging, bonding was related to both social justice prioritization and participation at the level of the individual. This indicates that the degree of connection within a congregation provided a link to both prioritize justice and to be involved in congregational justice activities. According to social capital theory (Putnam 2000), bonding ties can create obligations, systems for reciprocity, or trust. If bonding creates trust and solidarity within a group, greater connection and involvement may engender a sense of trust and/or obligation to being more involved in the tasks of the setting (i.e., social justice participation). Furthermore, information about personal need or congregational activities may be more accessible to higher bonded individuals as they are more densely connected, thus they are positioned to be more aware of and to participate in congregational activities. In regards to prioritization, higher bonded individuals may be more aware of the needs of others in the congregation, and may feel that it is the congregation’s responsibility to meet these needs. Previous scholarship suggests that people give more time and energy to higher bonded religious congregations (i.e., volunteering within the congregation, Hoge et al. 1998), thus higher bonded individuals may prioritize many congregational activities, including social justice, as part of the congregations mission. These findings also add to the bonding social capital literature by showing that another outcome of bonding ties, in addition to social support or status connections (Wuthnow 2002), is the prioritization of and participation in justice activities. Although bonding may not always have prosocial outcomes (i.e., see Schwadel 2005 for an example of higher bonding relating to less civic involvement), these findings show that individual bonding is associated with the prioritization and participation in congregational social justice activities.

Demographics and Social Position

Demographic variables were associated with both social justice prioritization and participation. One common thread between these demographics and the patterns of association, was that individuals from underprivileged groups (i.e., Black, poor, women, urban) were more likely to prioritize congregational social justice activities. This finding is not surprising given that previous research has found that congregations do more social justice work in economically disadvantaged and urban areas (Chaves 2001). However, these results indicate that congregations may be valued more highly as mediating structures for social justice by those who are disadvantaged. Given the lower prioritization of justice, it is possible that privileged group members may not have the same awareness of social inequality and may thus not see the relevance of social justice activities to congregational life. Scholarship affirms this link, by showing that justice work is indeed more important and relevant for people from underprivileged groups (Adams et al. 2007). In addition, race, income, and education predicted the likelihood of participating in congregational justice activities. Although previous research has shown levels of income to relate to the amount of congregational social justice activities (Chaves 2001), the finding that higher education predicts more social justice participation is intriguing, and shows that income and education may function differently in predicting participation in social justice.

Limitations and Suggestions for Future Research

Although current study findings contribute a multilevel understanding of religious congregations as mediating structures for social justice, they are not without limitations. First, the definition of social justice engagement is narrowly defined (i.e., limited to congregations) and relativistic. Although this definition allows for the exploration of the congregation as a mediating structure, such a narrow definition does not allow for a determination of how religious settings may influence social justice activities outside of the congregation. In addition, the definition of social justice engagement is relativistic, with no way of knowing the content or purpose of these programs. Consequently, programs that individuals are prioritizing or participating in may or may not be social justice programs at all, or may even be diametrically opposed to definitions of social justice. In fact, Paul Speer (2009) proposed that congregations may be moderated mediating structures for social justice, with different types of congregations (e.g., moderator) mediating particular types of social justice activities (e.g., liberal congregations working for and conservative congregations working against particular social issues such as abortion or gay rights). Future research should explore this proposition, examining how the definitions, content, and purpose of congregational social justice activities may differ based on different types of congregations. Second, although the current study examines religious congregations, the only religious variable beyond frequency of religious participation was theological orientation. Future research should examine how other specifically spiritual (i.e., beliefs, practices, traditions) aspects of congregations may influence social justice engagement. Third, three of the study variables relied on single item measures and another had low reliability, possibly attenuating relationships that may have been observed if a stronger measurement strategy was used. Fourth, generalizability of these findings to other religious congregations should be made with caution, especially since these congregations represent a limited number of denominations, are exclusively Christian, and represent a small geographic area. Furthermore, these congregations were not randomly sampled, but were selected due to their lack of focus on social justice. Future research could randomly select congregations across denominations for wider generalizability. Sixth, the study design does not allow for a determination of causality, and it is possible that people self-select into congregations that match their values rather than congregations exerting a causal effect on social justice engagement.

A final limitation common to secondary analyses of archival data is the generalizability of findings from data collected in 1987 to the present. There have been important shifts in the religious landscape in the U.S. since 1987 with an increased fusion between religion and politics as evidenced by the rise of the religious right and moral majority in the 1990s (Woodberry and Smith 1998) and an increased connection between religiosity and Republican Party identification (Putnam and Campbell 2010). Putnam and Campbell (2010) also document a growing concern by the general public over the role of religious influence in politics, a growth of those who identify as having no religious affiliation and an increased political homogeneity within religious congregations as people sort themselves into homogeneous groups. Rather than nullifying our findings, we believe that these shifts, if anything, show an increased importance in understanding how religion and religious congregations shape the political and social justice attitudes and behaviors of members. In fact, our general finding regarding the importance of bonding social capital in religious congregations is also found in relation to political attitudes in more current research (Putnam and Campbell 2010). We are also confident that future research will show religious congregations to be mediating structures for social justice attitudes and behaviors; however, the nuances of how and to what end religious congregations mediate social justice may change over time. Future research is needed to explore these questions in detail, and to document the evolving influence of religious congregations on social justice engagement.

Conclusions

The findings in this study support the assertion that religious congregations serve as mediating structures for social justice in society. First, aspects of the congregation (i.e., congregational bridging) were predictive of how strongly individuals prioritized social justice activities as part of the mission of the congregation. Second, congregations provided a space for individuals to participate in congregationally sponsored social justice activities, linking individuals into larger community based social justice participation. Moreover, aspects of the congregation, such as congregational theological orientation, augmented if individuals participated in social justice activities. This shows that congregations provided not only the space for social justice participation, but that the theological orientation of the congregation facilitated this social justice participation. These findings showed that congregations served a dual mediating role, both by influencing the prioritization of social justice and by providing a space to participate in social justice activities. Future research should continue to unpack these and other ways in which congregations mediate if and how individuals are involved in social justice.

These findings have direct implications for the field of community psychology. First, there is promise for partnerships with religious organizations where community psychologists can bring their unique skills and social justice values to help catalyze social justice engagement within a congregation that holds similar social justice values. A congregation may be able to better realize its social justice agenda through such collaborative partnerships with community psychologists, and community psychologists may be able to further their own social justice agenda through such partnerships. Second, as noted by many community psychologists (e.g., Kloos and Moore 2000) religious organizations are an excellent place to locate theory, research, and social justice action. The current study provides an example of one way to examine community psychology questions within religious congregations, whereas future theory, research, and action could build upon such findings to systematically examine how such religious organizations can be utilized for positive social change. Indeed, there is much promise in the examination of religious organizations as mediating structures for social justice action.