You are on page 1of 20

Rev Relig Res

DOI 10.1007/s13644-016-0280-3

Religious Conservatives and Outsiders: Determinants


of Cross-Racial Ties Among White Christians

Joseph Yi1 • Christopher Graziul2

Received: 19 April 2016 / Accepted: 8 December 2016


Ó Religious Research Association, Inc. 2017

Abstract This article analyzes how a congregation’s theology and denominational


affiliation influence the racial ties of its white members. We posit two distinct
pathways. In the first, theologically conservative congregations generate more
embedded social ties (measured by number of friendships) than do non-conservative
congregations, and more congregation friendships increase the likelihood of cross-
racial ties. In the second pathway, congregations not affiliated with historically
major denominational families report higher levels of racial diversity, and high
levels of congregation racial diversity increase the likelihood of cross-racial ties.
Our key methodological innovation is to divide Evangelical congregations into two
categories: those affiliated with the historically major families (Baptist, Lutheran,
Methodist, Presbyterian) and those not (e.g., Pentecostal, Seventh-Day Adventist).
Christian groups that join theological conservatism and outsider (non-major) status
generate high levels of friendships and racial diversity in their congregations, both
of which contribute to cross-racial ties among white members. Analysis of survey
data from a national probability sample of white Christians (2006 Faith Matters
Survey) mostly supports our hypotheses.

Keywords Christian  Evangelical  Race  Social ties (friendship)  Bridging

Joseph Yi and Christopher Graziul have contributed equally to the article.

& Joseph Yi
joyichicago@yahoo.com
Christopher Graziul
graziul@uchicago.edu
1
Political Science, Hanyang University, 222 Wangsimni-ro, Seongdong-gu, Seoul 04763, South
Korea
2
Postdoctoral Research Associate, Brown University, Spatial Structures in the Social Sciences
(S4), Box 1836, Providence, RI 02912, USA

123
Rev Relig Res

Introduction

What is the relationship, if any, between religious organizations and the cross-racial
ties of their members? One perspective claims that organizations with high levels of
embedded trust and ties among members, or bonding social capital, tend to be
socially homogeneous. Therefore, to the extent that congregations exude bonding
capital—which is more likely for theologically conservative than liberal congre-
gations—they should lack bridging, cross-racial ties. An alternative perspective
argues that some congregations promote a pan-racial, religiously based identity and
proselytize to members of different races: they potentially join high levels of
bonding and (racially) bridging social capital.
We hypothesize that a congregation’s general theology and denominational
affiliation influence the social ties of its members. We posit two distinct pathways.
In the first, theologically conservative congregations generate more embedded
social ties (measured by number of friendships) than do non-conservative
congregations; more congregation friends increase one’s general social network;
and a large social network increases the likelihood of cross-racial ties. In the second
pathway, congregations not associated with historically major denominational
families report higher levels of racial diversity (percent nonwhite), and high levels
of congregation racial diversity increase the likelihood of cross-racial ties among
members. Congregations that join theological conservatism and outsider (non-
major) status generate high levels of friendships and racial diversity and therefore
doubly contribute to cross-racial ties among their members.1
Our key innovation is to divide Evangelical denominations and groups
(Steensland et al. 2000) into two subcategories: those affiliated with historically
major denominational families (Baptist, Lutheran, Methodist, Presbyterian) and
those not (e.g., Pentecostal, Seventh-Day Adventist, non-denominational). We
hypothesize that members of the former (‘‘Major-Evangelical’’) shall be less likely
to attend racially diverse congregations than do members of the latter (‘‘Other-
Evangelical’’).
Analysis of a national probability sample of white Christians in the 2006 Faith
Matters Survey generally supports our hypotheses. We limit the analysis to white
Christians, largely because our religious categories (e.g., Mainline, Evangelical)
were originally constructed for whites and exclude blacks. Black Protestants
historically belonged to different denominations than white Protestants and are
coded into a separate religious category of ‘‘black Protestants.’’

Religious Bonding and Racial Bridging

A key question is how Christian congregations influence the social ties of its
members. One account holds that organizations with high levels of embedded trust
and ties among members, or bonding social capital, tend to be socially

1
In this article, ‘‘outsider’’ or ‘‘non-major’’ simply refers to being outside historically major
denominational families (Mainline, Major-Evangelical). Outsider does not mean sectarian or socially
marginal, in the sense of Jehovah’s Witnesses or early Mormons and Pentecostals.

123
Rev Relig Res

homogeneous and to lack close ties with outsiders (McPherson and Smith-Loving
1987; Putnam 2000). Therefore, to the extent that congregations exude bonding
social capital—which is more likely among theologically conservative than liberal
congregations—they should lack personal ties across societal cleavages, or bridging
social capital. This implies that tight-knit mostly conservative congregations would
be less racially diverse than loose, liberal congregations. A related argument focuses
on the particular ideology and life experiences of white Evangelicals. Their
religiously and socially conservative beliefs (e.g., individual responsibility and
opportunity), and their inclusion in the privileged, white majority, blind white
Evangelicals to structural discrimination against minorities and undermine authen-
tic, interracial ties (Jackman and Crane 1986; Emerson and Smith 2000).
An early study (analyzing 1978 and 1980 General Social Surveys) found that
white members of (mostly) conservative denominations (Baptists) were less likely
to attend service with blacks than were white members of mostly liberal
denominations (Presbyterian, Episcopalian) (Hadaway et al. 1984: 208). Recent
studies, however, show steady growth in interracial worship (Yi 2009; Chaves and
Anderson 2014), and this trend is more pronounced among conservative than liberal
denominations. Evangelicals are more likely to attend multiracial congregations
(Putnam and Campbell 2010) and to report African-American and Hispanic friends
(Wuthnow 2003) than are Mainline Protestants.
To explain, an alternative account argues that the characteristics of conser-
vative congregations, including bonding social capital, can be conducive to cross-
racial ties. The logic behind racial homogeneity is that it is more efficient for a
voluntary organization to focus its efforts on one distinct population than to
dilute them among multiple populations (Scheitle and Dougherty 2010). To
counter such homogeneity, some congregations, especially Evangelical ones,
promote a pan-racial, religiously based identity or ‘‘master status’’ (Marti 2009)
and actively recruit members of different races (Christerson et al. 2005; Emerson
and Woo 2006). Once recruited, members of different ethno-racial groups
develop close, interracial friendships in Evangelical congregations, due to their
relational and communal emphasis (Garces-Foley 2008; Putnam and Campbell
2010: 306); in contrast, Catholic and Mainline Protestant congregations empha-
size cultural pluralism and their members ‘‘balkanize around ethno-linguistic
groupings’’ (Perry 2013: 263).
The dual accounts suggest that for most congregations, measures of social
bonding and racial bridging are inversely correlated, but that some congregations
join high levels of bonding and bridging. Our article theorizes two ways that
congregations influence cross-racial ties. Firstly, theologically conservative
congregations generate more embedded social ties (measured by friendships)
than do liberal congregations, and more congregation friends increase the
likelihood of cross-racial ties. Secondly, congregations not associated with
historically major denominational families report higher levels of racial diversity,
and high levels of congregation racial diversity increase the likelihood of cross-
racial ties.

123
Rev Relig Res

Historically Major Protestants and Congregational Diversity

Racially diverse congregations tend to come from religious traditions outside of the
Evangelical and Mainline Protestant traditions, in particular the Catholic, Other-
Christian (e.g., Jehovah’s Witnesses), and non-Christian traditions (Dougherty and
Huyser 2008; Putnam and Campbell 2010: 292). To explain the racial heterogeneity
of Catholic parishes, one account claims that an ecclesiastical hierarchy organizes
all participating Catholics within a particular geographic area (Putnam and
Campbell 2010: 297; Dougherty 2003: 69). As historically white ethnic neighbor-
hoods see an influx of Catholic immigrants from Africa, Asia, and Latin America,
and with the conversion of native-born blacks and whites, the territorially defined
parish hosts a diverse set of believers. (Conversely, special dispensations allow
some Catholics to form ethnic or ‘‘personal’’ parishes based on cultural or linguistic
preferences, and this may reduce the likelihood of racial heterogeneity.)
In another account, Emerson (Emerson and Woo 2006: 39) argues that smaller,
religious traditions offer less opportunity for members to select homogeneous
congregations. Because of the limited number of congregations, members of diverse
races are often forced to worship together in the same congregation. Finally, small,
sectarian groups such as early Mormons and Pentecostals (Moore 1986) and
contemporary Jehovah’s Witnesses (Stark and Iannaccone 1997) exude strong moral
boundaries and social capital that attract socially marginal persons (c.f., Niebuhr
1929; Iannaccone 1994). If marginality is more evident among racial minorities, one
would expect that conservative churches attract more racial minorities than liberal
churches.
The congregation diversity literature rarely examines denominational differences
within the Evangelical and Mainline Protestant traditions. Yi (2009) posits that
congregations affiliated with the historically major, denominational families would
be more racially homogeneous than those not affiliated. In the USA, six Protestant
families (Episcopal, Presbyterian, Methodist, Lutheran, Baptist, United Church of
Christ/congregational) largely defined religious practice in the nineteenth and early
twentieth centuries (Roof and McKinney 1987: 236; Smith 1990), including the
pattern of racially separate worship. In the nineteenth century, free black members
of mostly Methodist and Baptist denominations separated from their white brethren
and formed autonomous black congregations and denominations (e.g., African
Methodist Episcopal, National Baptist Convention) (Lincoln and Mamiya 1990;
Emerson Emerson and Woo 2006: 12–17). Racially homogeneous congregations
were civic pillars of separate, white and black communities.
In contrast, relatively marginal, religious organizations were less likely to be
standard bearers for the larger white or black community and less constrained from
experimenting with interracial worship. Many Pentecostal denominations, for
instance, were interracial before the 1920s (MacRobert MacRobert 1997; Emerson
and Woo 2006: 17). As the Pentecostal movement grew, white leaders sought
societal respectability by creating white-only denominations (e.g., Assemblies of
God). Still, interracial worship continued in smaller denominations, black and white
Pentecostals retained a shared, historical memory (e.g., Azusa Street), and
interracial worship grew after the mid-twentieth century. In 1993, congregations

123
Rev Relig Res

from the Assembly of God reported more than double the racial–ethnic diversity of
congregations from Presbyterian, Lutheran, and Southern Baptist denominations
(Dougherty 2003: 76). In a pooled sample from the General Social Survey, 60% of
self-identified ‘‘Pentecostals’’ (excluding Assembly of God) claimed interracial
worship, which was much higher than Presbyterian, Lutheran, and Southern Baptist
respondents (Yi 2009: 201). Smaller, religious groups reported even higher rates of
interracial worship: 69% of Seventh-Day Adventists (est. 1863) and an astounding
86% of Jehovah’s Witnesses (est. 1870s) (c.f., Stark and Iannaccone 1997).
Experiments in interracial worship are also common among recently formed
(post-1950) denominations. Dougherty and Huyser (2008: 36) report that ‘‘congre-
gations with shorter institutional histories do not have the same path dependencies
that impede organizational change as do long-established congregations.’’ Actively
proselytizing congregations today either downplay their affiliations with the
historically major denominations or break with them altogether. For instance,
Mosaic (Marti 2009) and Saddleback (Putnam and Campbell 2010: 54–69) de-
emphasize their Southern Baptist affiliations and frame themselves as generically
Evangelical, biblically grounded churches. John Osteen, founder of Houston-based
Lakewood Church (and father of current pastor, Joel Osteen), was also affiliated
with the Southern Baptist Convention before becoming non-denominational.
Whether old (Adventists) or new (Lakewood), Christian organizations that distance
themselves from major, denominational ‘‘brand-names’’ are among the most racially
diverse.
To review, the literature finds that Christian groups outside of the Evangelical
and Mainline Protestant traditions are more racially diverse: this would include
Catholic, Mormon, and Other-Christian groups. We add that congregations inside
the Evangelical tradition, but outside the historically major, denominational
families, would also be more diverse. These ‘‘Other-Evangelical’’ groups include
both old, non-major denominations (e.g., Nazarenes, Seventh-Day Adventist) and
new, post-1950 churches (e.g., Lakewood). Our hypothesis helps explain why
previous studies (e.g., Hadaway et al. 1984; Emerson and Smith 2000) found
Evangelical congregations to be racially homogeneous; they largely drew from the
major denominational families (e.g., Methodist, Baptist). With the continuing
growth of Evangelical congregations that define themselves as outside the major
families, the Evangelical category as a whole should become more multiracial.
Attending a racially diverse congregation, in turn, is strongly associated with
personal, interracial ties, according to both qualitative (e.g., Marti 2009; Emerson
and Woo 2006) and quantitative studies (Yi 2009: 162; Putnam and Campbell 2010:
309). A racially diverse congregation does not directly indicate bridging social
capital between members of different races. Members of different races may
worship inside the same church building, but lack close, personal ties. Still,
compared to a mono-racial church, a multiracial church provides more opportunities
to interact with those of another race. We conceptualize congregation racial
diversity as a weak or potential form of bridging social capital that, depending on
other factors, can be converted into close, interracial ties.

123
Rev Relig Res

Congregational Friendships

Congregational diversity receives much academic attention, but it may not be the
central pathway to cross-racial friendships. We also consider how social embed-
dedness (i.e., the number of close ties) within a congregation cultivates cross-cutting
ties. A large network of friends encompasses a greater diversity of backgrounds and
lifestyles than a small network (Putnam 2007: 254). Wuthnow (2003) finds that
respondents with many friends (10 or more) are more likely to report friendships
with manual workers, welfare recipients, African Americans, and Hispanics than
those with few friends. We hypothesize that respondents with more congregation
friends shall have more friends overall, and that a large social network increases the
likelihood of ties with cross-racial ties. Congregation members potentially develop
other-race friends directly through their intra-congregational social network, or
indirectly though the extra-congregational networks of church friends (‘‘friends-of-
friends’’). Since the respondent’s number of close congregation friends contributes
to her total number of close friends, controlling for total friends should statistically
reduce the association between congregation friends and other-race friends.
In summary, we theorize two distinct pathways from Christian congregations to
cross-racial ties (Fig. 1). In the first, congregations associated with theologically
conservative categories generate more embedded social ties (measured by number
of friendships) than do congregations not affiliated; more congregation friends
increase one’s general social network; and a large social network increases the
likelihood of cross-racial ties.
In the second pathway, congregations not associated with historically major
denominational families report higher levels of racial diversity, and high levels of
congregation racial diversity increase the likelihood of cross-racial ties. Congre-
gations that belong to both theologically conservative and outsider (non-major)
categories generate high levels of friendships and racial diversity in their
congregations and therefore doubly contribute to cross-racial ties among their
members.
We posit four hypotheses:
H1 Members of congregations affiliated with historically major denominational
families shall report less racial diversity in their congregations than do members of
non-major (outsider) congregations.
H2 Members of conservative, outsider congregations shall report high levels of
close friends and racial diversity in their congregations.
H3 A high number of congregation friends shall increase the likelihood of close
ties (friendships) with members of racial out-groups; this association shall weaken
when controlling for total friends.
H4 Attending a congregation that is racially diverse shall increase the likelihood
of close ties with members of racial out-groups.

123
Organizational Conservative Outsider
Type
Rev Relig Res

Other-Evangelical
Religious Major-Evangelical Mormon Catholic
Categories Other-Christian

Congregation Congregation Friends Cong. Racial Diversity


Total Friends (Potential Bridging)

Cross-Group
Friendship
Black Friend
Hispanic Friend

Fig. 1 Path diagram

123
Rev Relig Res

Data and Method

We analyze a nationally representative survey [2006 Faith Matters Survey (FMS)],


directed by Robert Putnam and David Campbell. The 2006 sample has a total of
3108 adult respondents, age eighteen or older. Interviews were conducted by phone
in both English and Spanish, and the margin of error for total respondents is ±1.73
at the 95% confidence level (Roper Center, n.d.). FMS provides exceptionally
detailed information on respondents’ religious beliefs, belonging, and behavior;
social networks and out-group ties (e.g., black friends, gay friends); and political
views.
This article analyzes a subsample of respondents who report being non-Hispanic
whites and who attend specific Christian congregations (N = 1727). Following
Steensland et al. (2000), FMS categorizes Christians along religious traditions,
including Evangelical, Mainline, Mormon, and Catholic. We create a residual
category, Other-Christian, for respondents who claimed membership in a group
(e.g., Jehovah’s Witnesses) that did not fit the above religious traditions.
Our key innovation is to subdivide Evangelicals into two categories, depending
on their affiliations with the major denominational families of the nineteenth and
early twentieth centuries. We operationally define historically major families as
denominations coded as ‘‘Major Protestant Families’’ by the General Social Survey
(Smith 1990), which include Episcopalian, Presbyterian, Methodist, Lutheran, and
Baptist. The GSS codes all other Protestant denominations and groups as ‘‘Other
Protestant.’’ Similarly, we code Evangelical congregations affiliated with ‘‘Major
Protestant Families’’ as ‘‘Major-Evangelical,’’ and all Other-Evangelical congrega-
tions as ‘‘Other-Evangelical.’’2
Major-Evangelicals include theologically conservative (Evangelical) strains of
Baptist, Lutheran, Methodist, and Presbyterian families. Other-Evangelicals are
outside the major families and include historically non-major denominations (e.g.,
Pentecostal, Seventh-Day Adventist) and more recent groups (e.g., non-denomina-
tional Lakewood church). By analyzing Major- and Other-Evangelicals, we control
for the influence of a general religious tradition, and its body of beliefs and practices
(Steensland et al. 2000). Since the Major-Evangelical and Other-Evangelical
categories each claim about half of Evangelicals, we also control for the size of the
religious category (Emerson and Woo 2006: 39).
All respondents in Major-Evangelical, Other-Evangelical, and Mormon cate-
gories, and most respondents in the Other-Christian category, belong to denomi-
nations coded as theologically conservative (‘‘Fundamentalist’’) by the General
Social Survey (Smith 1990). Moreover, respondents in the Evangelical, Mormon,
and Other-Christian categories are more likely than those in Mainline to agree that
the Bible is the word of God and should be taken literally (descriptive statistics, not
shown).
2
The GSS scheme (Major Protestant Families/Other Protestant) is a parsimonious, established method to
code historically major, denominational families. The groups coded as ‘‘Other Protestant’’ were mostly
either nonexistent or considered less establishment than the ‘‘Major Protestant Families’’ denominations
before 1950, and we expect this institutional history to influence current race relations. The validity of the
categorization emerges in our analysis.

123
Rev Relig Res

We construct a dummy variable for each religious category (Major-Evangelical,


Other-Evangelical, Mormon, Other-Christian, Catholic), with Mainline Protestants
as the reference category in regression analysis. We also construct dummy variables
for high number of congregation friends (more than 10) and medium number of
congregation friends (6 to 10); high level of congregational racial diversity (white
respondent reports about half or less of the congregation are white) and medium
level of congregational racial diversity (about three-quarters white). Previous
studies define a multiracial congregation as one in which any racial group must be
less than 80 or 75% of the congregation (Emerson and Woo 2006; Putnam and
Campbell 2010). Therefore, if a respondent reports that her congregation is about
three-quarters of the same race, the congregation is coded as homogeneous by
Putnam and Campbell (2010: 291). We test the assumption that only a high level of
diversity influences cross-racial ties, by creating separate variables for high and
medium levels of congregational racial diversity. A third variable (‘‘multiracial
congregation’’) includes both high and medium levels of diversity (three-quarters or
less white).3
Figure 1 (Path diagram) shows the expected relationships among the variables.
White respondents of congregations from predominantly, theologically conservative
categories (Major-Evangelical, Other-Evangelical, Mormon, Other-Christian) are
more likely to claim high numbers of friends in their congregations than respondents
from non-conservative categories (Mainline, Catholic). Respondents from Outsider
categories (Other-Evangelical, Mormon, Other-Christian, Catholic) attend more
multiracial congregations than respondents from historically major Protestant
families (Mainline, Major-Evangelical). The respondent’s number of close congre-
gation friends contributes to her total number of close friends. More friends or racial
diversity in one’s congregation predicts higher odds of other-race friends (e.g.,
black, Hispanic).
We limit the analysis to non-Hispanic whites, largely because our key religious
categories (e.g., Mainline, Evangelical) were historically constructed for whites.
Mainline and Major-Evangelical encompass historically major denominations for
white Protestants, but not for black Protestants. Most black Protestants historically
belonged to different denominations than white Protestants (Lincoln and Mamiya
1990). Following Steensland et al. (2000), FMS codes all respondents who identity
as racially black and religiously Protestant into a separate category of ‘‘black
Protestants.’’ Future research can unpack and recode denominational families for
black Protestants, as we do for white Protestants; it could divide black Protestant
denominations into historically major (e.g., African Methodist Episcopal, National
3
The variables for black friend and Hispanic friend are dichotomous; therefore, we utilize logistic
regression. Congregation racial diversity is originally an ordinal variable; for purpose of analysis, it is
appropriate to construct dichotomous variables for high and medium levels of diversity. To facilitate
comparisons among adjusted odds-ratios, we also construct dichotomous variables for high and medium
levels of congregational friends and total friends. Wuthnow (2003) also creates dichotomous variables for
friends (10 or more friends; less than 3 friends).
Alternately, one can analyze congregational friends and total friends as simple count variables. In
separate analysis (not shown), we conducted ordinary least squares regression on congregation friends
and used logged counts of congregation friends and total friends to predict black and Hispanic friends.
The results are substantively similar.

123
Rev Relig Res

Baptist Convention) and non-major categories, and analyze whether the latter are
more racially diverse.
We also exclude Asian and Hispanic Protestants, since adding them would
increase the number of out-group variables (i.e., white friend) and complicate the
theoretical analysis. Future research can consider whether to pool Asian and
Hispanic Protestants with white Protestants or to analyze them separately (like black
Protestants), with different schema of historically major and non-major denomina-
tions. By restricting to non-Hispanic whites, this article focuses on how
congregations in historically white Protestant categories (Mainline, Evangelical)
influence the cross-racial ties of their white members.
Our analysis includes typical predictors of attending a multiracial congregation:
sex, age, education, income, urban residence, geographic region, political/social
ideology (Dougherty and Huyser 2008; Emerson and Woo 2006: 49, 91), and size of
one’s congregation.4 We also include church attendance and biblical literalism as
measures of personal religious commitment and religious conservatism, respec-
tively, and as predictors of close friends or racial diversity in one’s congregation.
An interesting anomaly is the relationship between close congregation friends
and total number of close friends. Logically, the respondent’s number of close
congregation friends should be smaller than, or at most equal to, the total number of
close friends. However, 20% of respondents report higher numbers of close
congregation friends than total close friends. This seems illogical, unless
congregation friendship and generic friendship represent distinct or partly overlap-
ping concepts. To explore this pattern, we conduct separate analyses of congre-
gation friends and total friends.

Analysis

To assess our hypotheses, we present descriptive statistics (Table 1) and analyze


multiple regressions (Table 2, 3). Table 1 shows the raw proportions of non-
Hispanic whites from different religious categories. Respondents from predomi-
nantly theologically conservative categories (Major-Evangelical, Other-Evangeli-
cal, Mormon, Other-Christian) are uniformly more likely to claim high number of
friends (more than ten) in their congregations than respondents from non-
conservative categories (Mainline, Catholic).
Consistent with the first hypothesis, respondents from historically major
denominational families (Mainline, Major-Evangelical) are the least likely to report
attending multiracial congregations (about three-quarters or less white): 32% of
4
Putnam and Campbell (2010 : 296) find that the most important predictor of congregation diversity is
congregation size (c.f., Chaves 2011), followed by Catholic, county racial diversity, Latino (race of
respondent), West region, Evangelical, age, and female. The publically available dataset does not include
county diversity, so we use urban residence as a proxy. Instead of West, we use the South region.
Contrary to Emerson (2006: 49), participation in small church groups is not a significant predictor of
congregation racial diversity (analysis not shown). The Faith Matters dataset does not include a variable
for charismatic worship (cf., Emerson 2006; Dougherty and Huyser 2008).
An appendix of our variables and measurements and a table of summary statistics (not shown) are
available upon request from the authors.

123
Rev Relig Res

Table 1 Descriptive statistics by religious categories


Mainline Major- Other- Mormon Other- Catholic Total
Protestant Evan Evan Christian

Cong. racial High 9% 9% 23%a 10% 29%a 27%a 17%


diversity Medium 23% 26% 34% a
34% 25% 31% 28%
Total 32% 35% 58%a 44% 55%a 58%a 45%
diversity
N 380 302 280 41 55 379 1437
Cong. friends High 17% 25% 23% 22% 22% 19% 20%
Medium 9% 15% 12% 14% 11% 11% 11%
N 381 304 284 41 55 387 1452
Total friends High 15% 18% 14% 26% 17% 15% 16%
Medium 25% 24% 24% 28% 25% 25% 25%
N 468 346 326 46 69 470 1725
Black friend Yes 31% 43%a 45%a 35% 42% 35% 38%
N 467 346 326 46 69 468 1722
Hispanic Yes 19% 26% 33%a 47%a 48%a 33%a 29%
friend N 468 346 326 45 69 469 1723
a
Significantly different from Mainline Protestant (Tukey–Kramer pairwise comparisons)

Mainline respondents and 35% Major-Evangelicals. This compares to 58% of


Other-Evangelicals, 44% Mormons, 55% Other-Christians, and 58% Catholics. The
difference is starker for high levels of racial diversity (about half or less white): 9%
of Mainline and Major-Evangelical respondents report highly diverse congrega-
tions, compared to 23% of Other-Evangelicals and 29% Other-Christians.5
Consistent with the second hypothesis, respondents from conservative, outsider
categories (Other-Evangelical, Mormon, Other-Christian) report high levels of close
friends and racial diversity in their congregation. 23% of Other-Evangelicals claim
ten or more congregation friends (compared to 25% of Major-Evangelicals) and
58% report multiracial congregations (compared to 35% of Major-Evangelicals).6
Binary logistic regression models in Table 2 affirm that religious categories are
significant predictors of congregation friends and diversity.7 In Model 1,

5
In Table 1, 17% of white Christians attend highly diverse congregations (less than three-quarters
white). The 17% figure matches the results of the 2006–07 National Congregations Survey: 17% of all
congregations report no race making up 80% or more of the majority (Chaves and Anderson 2014). The
NCS reports a steady, linear growth in such congregations, from 15.3% (1998) to 19.7 (2012).
6
One might infer that the differences between Major- and Other-Evangelicals are explained by
geography. Major-Evangelicals, mostly Baptists, are concentrated in the historically segregated South. In
further analysis (not shown), we find the differences are not geographic in nature. Even among Southern
whites, Other-Evangelicals are 11% more likely to frequently attend service, nearly twice as likely to
report multiracial congregations, and more likely to claim black and especially Hispanic friends than are
Major-Evangelicals.
7
Ideally we would have utilized structural equation modeling or some sort of path analysis, as our
theoretical model implies the existence of endogeneous variables, but our variables of interest are binary
(e.g., close black friends) and this is a non-trivial case for such approaches. Thus, we opted to use logistic
regression using listwise complete cases to model outcomes.

123
Rev Relig Res

Table 2 Odds ratios for many congregational friends, multiracial congregation, and many total friends
Covariates High congregational friends Multiracial congregation High total friends

Model 1 Model 2 Model 3 Model 4 Model 5

Major-Evangelicala 1.745* 1.428 1.019 1.051 1.324


(0.404) (0.343) (0.188) (0.197) (0.337)
a
Other-Evangelical 2.115** 1.593? 2.454*** 2.566*** 1.068
(0.509) (0.398) (0.450) (0.483) (0.290)
Mormona 2.677* 1.803 1.516 1.617 3.014*
(1.248) (0.848) (0.537) (0.581) (1.389)
Other-Christiana 1.521 1.098 2.807** 2.946** 0.773
(0.624) (0.462) (0.920) (0.979) (0.394)
Catholica 1.486? 1.367 1.740** 1.753** 1.122
(0.357) (0.335) (0.302) (0.305) (0.285)
Multiracial congregationb 0.686* 0.707* 1.254
(0.112) (0.117) (0.220)
High cong. friendsc 0.685* 0.705* 6.078***
(0.111) (0.115) (1.148)
High attendanced 2.330*** 0.887 0.662*
(0.419) (0.119) (0.127)
High biblical literalisme 1.356? 0.934 0.961
(0.239) (0.134) (0.189)
Congregation size (lg10) 1.375* 1.293? 2.068*** 2.090*** 0.972
(0.212) (0.201) (0.250) (0.254) (0.166)
Female 1.238 1.145 1.006 1.017 0.762
(0.193) (0.181) (0.121) (0.123) (0.129)
Education 0.885** 0.882** 0.995 0.995 0.977
(0.0377) (0.0388) (0.0345) (0.0349) (0.0469)
Income 1.119** 1.143** 0.971 0.967 1.057
(0.0488) (0.0512) (0.0328) (0.0330) (0.0509)
South 1.525** 1.422* 1.055 1.067 1.154
(0.247) (0.234) (0.137) (0.139) (0.210)
Social conservative 1.466* 1.247 0.788? 0.813 1.011
(0.240) (0.211) (0.0979) (0.104) (0.182)
Age (centered) 1.047*** 1.042*** 0.989** 0.990** 1.018**
(0.00569) (0.00575) (0.00393) (0.00399) (0.00587)
Urban 0.600** 0.640** 1.319* 1.304* 0.852
(0.096) (0.104) (0.161) (0.160) (0.149)
(Constant) 0.067*** 0.051*** 0.129*** 0.132*** 0.117***
(0.028) (0.022) (0.042) (0.044) (0.053)
v2 147.4 177.7 131.6 132.8 140.2

123
Rev Relig Res

Table 2 continued

Covariates High congregational friends Multiracial congregation High total friends

Model 1 Model 2 Model 3 Model 4 Model 5

Degrees of freedom 14 16 14 16 17
N 1275 1275 1275 1275 1274

* p \ 0.05;** p \ 0.01; *** p \ 0.001; ? p \ 0.1


a
Ref = Mainline protestant
b
Ref = Almost all or all white congregations
c
Ref = Less than 10 friends in congregation
d
Ref = Attend less frequently than once a week
e
Ref = Scripture is inspired word of God or ancient book recorded by men

respondents from three, predominantly conservative categories (Major-Evangeli-


cals, Other-Evangelicals, Mormons) have higher odds of claiming a high number of
close congregation friends (more than ten) compared to our reference category
(Mainline Protestants). In Model 2, adding high church attendance and biblical
literalism reduce the significance of all religious categories. High attendance (every
week or more) is significantly associated with reporting ten or more congregation
friends. This suggests that attendance is a critical, intervening variable between
religious categories and congregation friendships.
Partially supporting our first hypothesis, respondents from three, outsider
categories (Other-Evangelical, Other-Christian, Catholic) have higher odds of
attending multiracial congregations than Mainline Protestants (Models 3 and 4).
When controlling for high levels of congregation friends and high attendance in
Model 4, Other-Evangelicals, Other-Christians, and Catholics show even higher
odds of attending multiracial churches. In fact, high congregation friends and high
attendance are negatively associated with attending a multiracial congregation (but
only congregation friends are statistically significant). This is consistent with the
claim that tight-knit mostly conservative congregations would be less racially
diverse than loose, liberal congregations (McPherson and Smith-Loving 1987;
Putnam 2000). It also suggests that despite their high levels of religious
commitment and social capital (congregation friends), Other-Evangelicals and
Other-Christians, and to a lesser extent Catholics, report more multiracial
congregations than Mainline Protestants.
Reporting ten or more congregation friends is associated with lower odds of
reporting a multiracial congregation (Models 3 and 4). However, supporting the
second hypothesis, the largest category of conservative outsiders (Other-Evangel-
icals) is more likely to report high levels of friends (Models 1 and 2) and racial
diversity (Models 3 and 4) in their congregations than Mainline Protestants. Smaller
categories of conservative outsiders (Mormons, Other-Christians) are also more
likely to report high congregation friends and congregation diversity, but the
associations are not always statistically significant, likely because of small sample
sizes.

123
Rev Relig Res

Respondents with ten or more congregation friends have six times higher odds of
reporting ten or more total friends, compared to those with fewer than ten
congregation friends (Model 5).8 Interestingly, high church attenders have lower
odds of reporting many total friends (Model 5), even though they report many
congregation friends (Model 2). This suggests that for many (especially highly
involved) Christians, congregation friends and generic friends represent overlap-
ping, but distinct social networks.
Logistic regression models in Table 3 assess whether the social ties and racial
makeup of one’s congregation are significant predictors of her personal ties with
racial out-groups.9 Consistent with the third hypothesis, white respondents with high
numbers of congregation friends (more than ten) show much higher odds of
reporting close black friends (Model 1) and Hispanic friends (Model 3) than the
reference category (respondents with fewer than six congregation friends); these
associations weaken after controlling for total friends (Models 2, 4), but black
friends remain significantly associated with high congregation friends (Model 2).
Consistent with the fourth hypothesis, whites attending high diversity congre-
gations (half or less whites) have much higher odds of reporting close black friends
(Model 2) and Hispanic friends (Model 4) than the reference category (respondents
attending all or nearly all white congregations). Those attending medium diversity
congregations (three-quarters whites) also have higher odds of reporting black and
Hispanic friends than the reference category.
Among religious categories, Other-Evangelicals have the highest odds of
claiming black friends (Model 2). Major-Evangelicals, Mormons, and Other-
Christians have significantly higher odds of claiming Hispanic friends than Mainline
Protestants (Model 4).

Discussion

Analysis of white Christians from the 2006 Faith Matters Survey mostly supports
our hypotheses. Respondents from the largest category of outsider-conservative
congregations (Other-Evangelical) have significantly higher odds of reporting high
levels of close friends and racial diversity in their congregations than do Mainline
Protestants. Smaller, outsider-conservative categories (Mormon, Other-Christian)
also have higher odds, but the associations are not always significant. In contrast,
the racial diversity of conservative, major denominational congregations (Major-
Evangelical) is indistinguishable from that of Mainline Protestants. Finally, whites
who report high levels of close friends or racial diversity in their congregations have
higher odds of close Hispanic and black friends, but when controlling for total

8
In additional analysis, the absolute number of friends in one’s congregation is highly correlated with
her total number of friends (r = 0.367, p \ .001).
9
Due to the way church diversity and congregational friendships are operationalized, we are not able to
investigate the intuitive possibility that having many congregational friendships within a diverse church is
more likely to lead to cross-racial friendships than either having many congregational friendships or
attending a diverse church alone. We hope to test this interaction in the future using either different data
or a different analytic approach.

123
Rev Relig Res

Table 3 Odds ratios for close black friend and Hispanic friend
Covariates Close black friend Close Hispanic friend

Model 1 Model 2 Model 3 Model 4

Major-Evangelicala 1.364? 1.224 1.656* 1.562*


(0.251) (0.239) (0.340) (0.342)
a
Other-Evangelical 1.590* 1.571* 1.434? 1.352
(0.294) (0.308) (0.291) (0.292)
Mormona 1.220 0.818 2.864** 2.732*
(0.463) (0.342) (1.063) (1.101)
Other-Christiana 1.098 0.943 2.353* 2.674**
(0.373) (0.343) (0.785) (0.946)
Catholica 0.951 0.886 1.317 1.386
(0.165) (0.165) (0.246) (0.276)
High cong. friendsb 2.592*** 1.618** 1.927*** 1.005
(0.416) (0.292) (0.330) (0.198)
Medium cong. friendsb 1.255 0.973 1.134 0.831
(0.241) (0.203) (0.236) (0.190)
High diversity cong.c 2.838*** 2.218*** 3.290*** 2.706***
(0.479) (0.406) (0.565) (0.497)
Medium Diversity Cong.c 1.961*** 1.762*** 1.820*** 1.546**
(0.276) (0.265) (0.273) (0.248)
High total friendsd 3.347*** 3.313***
(0.649) (0.666)
Medium total friendsd 1.916*** 1.676**
(0.289) (0.272)
Close Hispanic friend 3.421***
(0.494)
Close black friend 3.476***
(0.501)
Female 0.776* 0.825 0.851 0.952
(0.0940) (0.107) (0.110) (0.132)
Education 0.988 0.977 1.050 1.062
(0.034) (0.036) (0.039) (0.042)
Income 1.030 1.017 1.012 0.998
(0.035) (0.037) (0.0366) (0.039)
South 1.761*** 2.059*** 0.728* 0.595***
(0.227) (0.285) (0.103) (0.091)
Social conservative 0.823 0.838 0.878 0.935
(0.104) (0.112) (0.118) (0.133)
Age (centered) 1.004 1.004 0.996 0.993
(0.004) (0.004) (0.004) (0.005)
Urban 1.094 0.985 1.604*** 1.674***
(0.133) (0.129) (0.207) (0.231)

123
Rev Relig Res

Table 3 continued

Covariates Close black friend Close Hispanic friend

Model 1 Model 2 Model 3 Model 4

(Constant) 0.254*** 0.166*** 0.122*** 0.061***


(0.068) (0.048) (0.036) (0.020)
v2 129.8 281.2 114.1 257.7
Degrees of freedom 16 19 16 19
N 1332 1331 1332 1331

* p \ 0.05; ** p \ 0.01; *** p \ 0.001; ? p \ 0.1


a
Ref = Mainline Protestant
b
Ref = Less than 6 friends in congregation
c
Ref = Almost all or all white congregations
d
Ref = Less than 6 friends total

number of friends, the association between high congregation friends and Hispanic
friend becomes insignificant.
We highlight three contributions of this article. Firstly, both high and medium
levels of congregational diversity are significantly associated with reporting other-
race friends (Table 3). This shows the validity of analyzing different levels of racial
diversity in congregations, instead of just one level (high diversity) (cf., Emerson
and Woo 2006; Putnam and Campbell 2010).
Second, we explain some of the contradictions between congregational social
bonding and racial bridging. Consistent with the literature that social bonding and
racial bridging are inversely related (McPherson and Smith-Loving 1987; Putnam
2000), we find that white respondents with high numbers of congregational friends
are less likely to attend multiracial congregations. On the other hand, these socially
embedded respondents are also more likely to report high numbers of total friends
and other-race friends. (Hispanic friend, but not black friend, loses statistical
significance when controlling for total friends.) This implies that congregations with
high bonding social capital—which is more likely among theologically conservative
than liberal congregations—are more likely to be racially homogeneous and to
include members with large, racially mixed social networks.
The third, and most novel, finding is that congregations inside the historically
major denominational families are less racially diverse than congregations outside.
This includes a sharp divide between Major- and Other-Evangelical congregations.
Both are part of the Evangelical tradition, but the percentage of Major-Evangelicals
attending multiracial congregations (35%) look like Mainline Protestants (32%),
while Other-Evangelicals (58%) are akin to Other-Christians (55%) and Catholics
(58%). Mormons (44%) show lower diversity, but this may be largely attributed to
their concentration in the more homogeneous, Mountain West region. Whether
theologically liberal (Mainline) or conservative (Evangelical), congregations
affiliated with the historically major denominational families are the least diverse.

123
Rev Relig Res

Previous literature separately analyzed the congregational diversity of Catholic,


sectarian, or marginal traditions. For the first time, this article analyzes the
congregational diversity of all white Christians outside the historically major
denominational families. We find that outsider Christian categories—including
Other-Evangelicals—all display higher levels of congregational diversity. Other-
Evangelicals, Other-Christians, and Catholics are three, very different religious
categories, whose only commonality is that they are outside of the historically major
Protestant families. Their nearly identical multiracial congregation percentages
(55–58%) suggest that this may be the ‘‘new normal’’ for the growing number of
white Christians outside the major denominational families. (Combined with
Mormons, the outsider categories make up 53% of white Christian respondents, as
of 2006.)
Marginality or sectarian separation from society (Niebuhr 1929; Iannaccone
1994) does not seem to be a significant determinant of congregational diversity. If it
were, Other-Christians would report higher diversity than Other-Evangelicals, but
that is not the case. Institutional similarity also does not seem to be a key
determinant: many non-denominational congregations (e.g., Lakewood) are close
offshoots of Major-Evangelical denominations (e.g., Baptist), with similar beliefs
and practices, but the racial diversity of the non-denominational category is
substantially higher than that of Major-Evangelicals (additional analysis, not
shown).
More influential determinants may be how congregations are framed (Goffman
1974) and their institutional ties. Congregations inside the historically major
families are popularly associated with an institutional legacy of (white) racial
homogeneity. Switching affiliation from Baptist to non-denominational, as did
Lakewood Church, symbolizes a break with a historically segregated, denomina-
tional family and the start of a ‘‘new’’ church open to new practices and persons.
Moreover, formal affiliation with historically major families may be tied to
institutional practices that (inadvertently) reinforce homogeneity; by cutting the
affiliation, the non-denominational church reduces such practices. Future research
can examine the higher diversity of Other-Evangelical denominations and
congregations, including those that are close offshoots of Major-Evangelical
denominations.

Self-selection

We theorize that the general theology and denominational affiliation of a


congregation influence the social networks of attendees. An alternative theory
would stress reverse causality: respondents with prior cosmopolitan or tolerant
attitudes are more likely to form diverse social networks and to join racially
heterogeneous congregations. In fact, persons who attend multiracial congregations
are more likely to have grown up in integrated schools and neighborhoods
(Emerson, Kimbro, and Yancey 2002), and interracial couples search for diverse
congregations (Christerson, Edwards, and Emerson 2005; Emerson and Woo 2006).

123
Rev Relig Res

To address the alternative claim, our control variables include educational status
and sociopolitical ideology. College-educated, socially liberal individuals are more
likely to valorize diversity, and better-educated persons enjoy greater opportunities to
interact with different social groups at work and in the civic arena (Nie et al. 1996).
Even controlling for education and ideology, however, our key variables remain
significant. In fact, religious categories (e.g., Other-Evangelical, Mormon) show
stronger associations with interracial friendships than do education or ideology.
The contact literature shows that social networks more often influence political
attitudes than vice versa (Pettigrew and Linda 2006). Congregation-based social
networks transmit political information and shape respondent attitudes (Djupe and
Gilbert 2009). In multiracial congregations, attendees report that their social ties
became more racially diverse and that they changed their racial and social views
because of their involvement (Emerson and Woo 2006).
Conservative denominations generally impose more requirements on members,
in terms of beliefs and behavior, than do liberal denominations; it seems unlikely
that many people would switch from liberal (Mainline) to conservative denomi-
nations, primarily to interact with other races. It also seems unlikely that people
would switch religious traditions or categories, such as from Evangelical to
Mormon, Other-Christian, or Catholic, primarily for racial interaction. On the other
hand, it seems plausible for racially inclusive Christians to switch from relatively
homogeneous to diverse denominations within a particular religious tradition; this
may account for the higher, cross-racial ties of members in the Other-Evangelical
category, as racially inclusive members switch (self-select) from Major-Evangelical
to Other-Evangelical denominations and groups, for instance, from Baptists to
Nazarenes or non-denominational.

Future Research

This article encourages new thinking on how organizations navigate the complex
dynamics of bonding and bridging social capital. The findings suggest that bridging
social capital (personal ties across societal cleavages) is more likely among
organizations that join strict doctrines with relatively new or outsider status.
Organizations with strict doctrines (e.g., Evangelical congregations) generate more
social capital (e.g., friendships) than less strict organizations (e.g., Mainline
congregations), but this social capital is more likely to be cross-cutting or bridging
among relatively outsider organizations (e.g., Other-Evangelical). By outsider, we
do not mean social marginality, but rather a degree of separation (real or perceived)
from institutions associated with particular cleavages. For white Christians, the
historically major denominational families are closely associated with the history of
racially segregated worship; non-major denominations and groups are generally less
associated and generate more racial bridging.
The next step is to replicate our analysis for nonwhites, starting with black
Christians. One can recode denominational families for black Protestants into
historically major and non-major categories and analyze whether the latter are more
racially diverse. Researchers can also examine organizations from other religious-

123
Rev Relig Res

cultural traditions. For instance, many Asian-style martial arts schools combine
hierarchical, Confucian doctrines with racially outsider status in the USA. Because
they are framed as neither white nor black, martial arts attract practitioners of all
races (Yi 2009; Yi and Silver 2015). How and why certain organizations bridge
societal cleavages remains a central question of plural societies.

Acknowledgements This article was supported by Hanyang University Research Fund.

References
Chaves, Mark, and Shawna Anderson. 2014. Changing American congregations: Findings from the third
wave of the National Congregations Study. Journal for the Scientific Study of Religion 53 (4):
676–686.
Christerson, Brad, Korie L. Edwards, and Michael O. Emerson. 2005. Against all odds: The struggle for
racial integration in religious organizations. New York: New York University Press.
Djupe, Paul A., and Christopher P. Gilbert. 2009. The political influence of churches. Cambridge:
University Press.
Dougherty, Kevin D. 2003. How monochromatic is church membership? Racial-ethnic diversity in
religious community. Sociology of Religion 64 (1): 65–85.
Dougherty, Kevin D., and Kimberly R. Huyser. 2008. Racially diverse congregations: Organizational
identity and the accommodation of differences. Journal for the Scientific Study of Religion 47 (1):
23–44.
Emerson, Michael O., and R.M. Woo. 2006. People of the dream: Multiracial congregations in the
United States. Princeton: Princeton University Press.
Emerson, Michael O., Rachel T. Kimbro, and George Yancey. 2002. Contact theory extended: The effects
of prior racial contact on current social ties. Social Science Quarterly 83: 745–761.
Emerson, Michael O., and Christian Smith. 2000. Divided by faith: Evangelical religion and the problem
of race in America. New York: Oxford University Press.
Garces-Foley, Kathleen. 2008. Comparing Catholic and Evangelical Integration efforts. Journal for the
Scientific Study of Religion 47: 17–22.
Goffman, E. 1974. Frame analysis: An essay on the organization of experience. Cambridge: Harvard
University Press.
Hadaway, C.Kirk, David G. Hackett, and James Fogle Miller. 1984. The most segregated institution:
correlates of interracial church participation. Review of Religious Research 25 (3): 204–219.
Iannaccone, Laurence. 1994. Why strict churches are strong. American Journal of Sociology 99 (5):
1180–1211.
Jackman, Marry R, and Marie Crane. 1986. Some of my best friends are black...: Interracial friendship
and whites’ racial attitudes. The Public Opinion Quarterly 50 (4): 459–486.
Lincoln, Eric C., and Lawrence H. Mamiya. 1990. The black church in the African–American experience.
Durham: Duke University Press.
MacRobert, Iain. 1997. The black roots of pentecostalism. In African American religion: Interpretive
essays in history and culture, ed. Timothy E. Fulop, and Albert J. Raboteau. New York: Routledge.
Marti, Geraldo. 2009. Affinity, identity, and transcendence: The experience of religious racial integration
in diverse congregations. Journal for the Scientific Study of Religion 48 (1): 53–68.
Moore, R.Laurence. 1986. Religious outsiders and the making of Americans. New York : Oxford
University Press.
Nie, Norman H., Jane Junn, and Kenneth Stehlik-Barry. 1996. Education and democratic citizenship in
America. Chicago: University of Chicago Press.
Niebuhr, H.R. 1929. The social sources of denominationalism. Gloucester: Henry Holt and Co.
Perry, Samuel L. 2013. Multiracial church attendance and support for same-sex romantic and family
relationships. Sociological Inquiry 83 (2): 259–285.
Pettigrew, Thomas F., and Linda R. Tropp. 2006. A meta-analytic test of intergroup contact theory.
Journal of Personality and Social Psychology 90 (5): 751–783.
Putnam, Robert D. 2000. Bowling alone: The collapse and revival of American community. New York:
Simon & Schuster.

123
Rev Relig Res

Putnam, R.D. 2007. E Pluribus Unum: Diversity and community in the twenty-first century, The 2006
Johan Skytte prize lecture. Scandinavian Political Studies 30 (2): 137–174.
Putnam, Robert D., and David E. Campbell. 2010. American grace: How religion divides and unites us.
New York: Simon & Schuster.
Roof, Wade Clark, and William McKinney. 1987. American mainline religion: Its changing shape and
future. New Brunswick: Rutgers University Press.
Scheitle, Christopher P., and Kevin D. Dougherty. 2010. Race, diversity, and membership duration in
religious congregations. Sociological Inquiry 80 (3): 405–423.
Stark, Rodney, and Laurence Iannaccone. 1997. Why Jehovah’s witnesses grow so rapidly: A theoretical
application. Journal of Contemporary Religion 12 (2): 133–157.
Smith, Tom W. 1990. Classifying protestant denominations. Review of Religious Research 31: 225–245.
Steensland, Brian, Jerry Z. Park, Mark D. Regnerus, Lynn D. Robinson, W. Bradford Wilcox, and Robert
D. Woodberry. 2000. The measure of American religion: Toward improving the state of the art.
Social Forces 79: 291–318.
Wuthnow, Robert. 2003. Overcoming status distinctions? Religious involvement, social class, race, and
ethnicity in friendship patterns. Sociology of Religion 64 (4): 423–442.
Yi, Joseph. 2009. God and Karate on the southside: Bridging differences, building American
communities. Lanham: Lexington Books.
Yi, Joseph, and Daniel Silver. 2015. God, yoga and karate. Journal for Scientific Study of Religion 54 (3):
596–615.

123

You might also like