Religiously Traditional,Unusually Supportive? Examining Who Gives,Helps,and Advises in Americans’ Close Networks

Abstract

A large literature is currently contesting the impact of religion on prosocial behavior. As a window into this discussion, I examine the close social networks of American adults and consider whether religious traditionalists are more likely than other network members to supply several basic forms of social support. Analysis of the Portraits of American Life Survey reveals three main findings. First, a majority of Americans—religious or not—count at least one perceived religious traditionalist among their close network ties. Second, American adults are more likely to receive advice, practical help, and money from ties identified as religious traditionalists than from other types of ties, a pattern that held among both kin and nonkin network ties. Finally, although perceived traditionalist network members appear especially inclined to assist highly religious people, they nevertheless offer social support to Americans across a broad spectrum of religiosity. Beyond its relevance for debates on religion and community life, this study also proposes a novel strategy to assess prosocial behavior. Asking people to recount the deeds of their network members can reduce certain self-reporting biases common to survey research and helps locate prosocial activity in concrete and meaningful social relationships.

Keywords

religion social support prosocial behavior social networks close ties traditionalism

Introduction

Pundits and observers often observe that religious traditionalism corresponds with several divisive fault lines in American society. Given the rapid processes of globalization, the diversifying demographic profile of the country, and the rapid change in attitudes toward issues such as sexuality and family life, religious traditionalists often find themselves swimming against many currents in contemporary American life. Though worthwhile to think through potential implications for politics (Layman 2001), science (Coyne 2012), and other key public institutions, equally important questions pertain to the informal, everyday activity that comprises community life. Within our religiously pluralistic society, do the 28 percent of Americans considered “religious conservatives” make good friends, neighbors, or extended family members?¹ Are they generous with their time, money, and attention in the context of everyday relationships? The present article considers informal personal support as an important form of prosocial behavior and examines what role apparent religious traditionalists play in its provision.

Though in the global context religion is often “linked to intolerance, violence, and mayhem—not to civic good manners” (Putnam and Campbell 2010:443), evidence from the United States also suggests that observant believers are most likely to help others with housework, assist peers in finding a job, and to simply appear “nice” relative to other people (Putnam and Campbell 2010:44; see also Ellison 1992). As one might expect of such an apparent paradox, the prosociality of religious believers—and traditionalists, in particular—is a topic of intense and ongoing debate (see Borgonovi 2008; Galen 2012a, 2012b; Saroglou 2006, 2012; Stark 2012; Wright 2010).

A common way to gauge the prosociality of religious people is to examine patterns of charitable giving and volunteer behavior. Observational studies on these topics—based mostly on self-reports—have largely concluded that religious Americans are more generous with their time and their money than less religious people (Hill and Vaidyanathan 2011; Hodgkinson and Weitzman 1996; Monsma 2007; Regnerus, Smith, and Sikkink 1998). One major caveat from the giving and volunteering research, however, is that religious people often give to causes that strictly promote their own values, and many of their volunteering hours are devoted to activities organized within their own congregation or religious in-group (Galen 2012a). Furthermore, many studies fail to obtain external verification of the prosocial behavior: Were the reports truthful, and—in the case of volunteering—was the individual’s behavior actually effective or helpful?

Another way to assess prosociality is to adopt an experimental approach, measuring trust, generosity, and cooperation among religious people in behavioral economic games or using semantic priming of religious concepts and detecting changes in social attitudes or behavior. In the experimental paradigm, prosocial behavior is usually targeted at an unknown experimental subject—the hypothetical “everyman” with whom an individual must interact in a novel environment. Relying on such approaches, many researchers have again documented heightened prosociality among the religious (Pichon, Boccato, and Saroglou 2007; Randolph-Seng and Nielsen 2007; Shariff and Norenzayan 2007; Sosis and Ruffle 2003). Yet some studies suggest that devout people are only cooperative when their exchange partners appear to be religious (Ben-Ner et al. 2009; Fershtman et al. 2005) and that religious primes can elicit antisocial attitudes such as antagonism toward out-group members and racial prejudice (Johnson, Rowatt, and LaBouff 2010; LaBouff et al. 2012).

The current study takes up the issue of religion and prosocial behavior, but departs from both the giving/volunteering paradigm and the experimental approach. Using a nationally representative study, I examine informal social support in the context of Americans’ core social networks. Specifically, I assess (1) who adults identify as their close network ties, (2) who among these ties provide various forms of informal assistance, and (3) whether the perceived religious orientation of a given tie has anything to do with that tie’s provision of support. This strategy offers the advantage of examining prosocial actions from the vantage point of a recipient—not a self-described benefactor (a limitation of the giving/volunteering paradigm)—and in the context of actual, ongoing relationships—not from an experimentally contrived scenario (a limitation of the experimental paradigm). My social network/social support approach has some limitations of its own, and these will be mentioned in the “Discussion” section below. For now, however, I turn to a brief elaboration of core theoretical constructs before presenting the empirical results.

I first provide a succinct overview of religious traditionalism. I then summarize several key currents in the social support literature and note how informal assistance captures an important—if underemphasized—aspect of prosocial behavior. Finally, I note the value of looking beyond congregations to identify religiously traditionalist network ties to and to assess their patterns of social support provision.

Theory and Background

Religious Traditionalism

Drawing from Hunter (1991) and from Davis and Robinson (1996a, 1996b), my analysis draws a basic distinction between several broad groups in American society: orthodox believers, more moderate or progressive religious adherents, and those without any particular religious affiliation. Orthodox believers—termed religious traditionalists in this study—suppose that God is the ultimate arbiter of good and evil, consider Scripture as authoritative and inerrant sources of Divine revelation, and believe that God is ever-present and actively engaged in human activity. Accordingly, traditionalists maintain a rather distinctive “moral cosmology”: They adhere to historic orthodoxy and dispute the modernist assumption that moral precepts are historically contingent and that sacred texts are bound by mere cultural processes. More moderate or progressive (modernist) believers, on the other hand, tend to align more closely with nonreligious Americans on a number of cultural indicators. Mainline Protestants, representing one dominant strand of religious progressivism, for instance, “have typically emphasized an accommodating stance toward modernity” (Steensland et al. 2000:293). The moral cosmology perspective emphasizes that the orthodox/modernist distinction is more influential than denominational affiliation in shaping people’s social attitudes and behaviors and that the actual frontline for the “culture wars” is not between faiths but amid the modernist and orthodox constituents within American society (Davis and Robinson 1996a, 1996b; Hunter 1991; Starks and Robinson 2007).

Though my analysis is not an explicit test of moral cosmology theory, religious conservatives represent an important ideal type for the current study. In the ongoing debate as to whether religion, on balance, enhances the common good (e.g., Hitchens 2007; Stark 2012; Wright 2010), orthodox believers represent a crucial data point because they hold most staunchly to received traditions and appear most out-of-step with a progressive milieu. Traditionalists of various sorts are accused of perpetuating in-group tribalism (Galen 2012a), fostering discrimination and intolerance (Terrizzi, Shook, and Ventis 2010), and retarding intellectual progress (Coyne 2012). When particular religions are disparaged for their social or moral deficiencies, fundamentalist adherents bear the brunt of the derision (Hitchens 2007). For better or for worse, religious traditionalists draw attention to the charged nature of religious disagreement in the contemporary United States. Yet despite the allegations and the purported acrimony, data used in this study suggest that a nontrivial share of the American population (religious and nonreligious, traditionalist and progressive alike) count religious traditionalist among their close social ties. I wish to examine whether people understood as traditionalists are distinctive in their social support provision.

Social Support and Prosocial Behavior

Social support refers to the informal exchange of resources with the intent to promote the recipient’s well-being (Shumaker and Brownell 1984). Hundreds of studies have documented how such exchanges are linked to many forms of individual and social well-being.

One common classification scheme distinguishes three forms of such help that gets transmitted within personal networks: emotional support, including a listening ear or a sense of camaraderie; information provision, which involves helpful insight about the external environment or about the individual herself; and instrumental aid, the sharing of practical resources, including money or assistance with a task (House 1981). Existing scholarship also draws attention to the social processes that elicit support. In an influential study of adults living in East Toronto, Wellman and Wortley (1990) investigated multiple factors that could shape the provision of support within personal networks and concluded that people’s varied needs are met by a mixture of role relationships. While kin-based ties are especially likely to offer financial assistance, neighbors, friends, and other nonkin close ties also provide a wide range of support, including companionship, emotional aid, and help with practical household tasks (Wellman and Wortley 1990). These findings point to the importance of considering various sources and forms of social support within people’s close networks.

Finally, it is important to note the conceptual significance of social support within a broader set of theoretical perspectives. Given the high degree of reciprocity within social support networks—for both nonkin (Phan et al. 2009) and within-family relations (Silverstein et al. 2002)—it is reasonable to expect that many ties are governed by the rules of social exchange. Nevertheless, social support arguably transcends an instrumentally motivated, exchange-based phenomenon, and in several ways, it denotes a distinct form of prosocial behavior (Shumaker and Brownell 1984).² For one, providers of social support must at some point decide to provide emotional, informational, or instrumental help; this often involves a person’s willingness to read social cues about the necessity or desirability of assistance. Second, social support can be susceptible to the diffusion of responsibility problem which hinders other forms of prosocial behavior; when there are multiple others who could conceivably offer support (e.g., a large and dense social network), each network tie could suppose that another person should assume responsibility. Together, these insights underline that support provision does not necessarily flow from the mere presence of a network tie—the tie must at some point decide to act in the interest of another.

Religion, Social Networks, and Social Support: Looking beyond Congregational Boundaries

In the lives of many Americans, churches and other religious organizations represent a key conduit for interpersonal support. In a recent study of anticipated support, Edgell, Tranby, and Mather (2013) document that a majority of Americans feel that they could turn to those in their church—either friends or leaders—for emotional help during a troubling time. Congregations provide a forum for like-minded individuals to meet regularly, become aware of various needs, and mobilize assistance. Not surprisingly, then, numerous studies document the presence of supportive relationships among co-congregates and suggest that frequent church attendance augments people’s networks and increases the likelihood that they will receive help when it is needed (Ellison and George 1994; Taylor and Chatters 1988). The supportive role of co-congregates has been posited as one explanation for the health and longevity advantages enjoyed by religious people (Ellison and Levin 1998).

Though illuminating, the existing literature on religion and support is limited because it compresses social ties and assistance within the scope of congregational membership. The full connection between religion and support provision, however, is not confined within churches or other formal institutions. Indeed, contemporary American society is characterized by varied, diffuse, and differentiated personal networks that span organizations and nonoverlapping social groups, an emerging phenomena identified by several observers as networked individualism (Rainie and Wellman 2012). This has at least two key implications. First, religious people—including religious traditionalists—can maintain social ties with similarly devout people outside of their particular congregation. Perhaps more importantly, those unaffiliated (or only loosely connected) with a religious organization can also have close social ties and receive social support from religious traditionalists. Thoroughly secular adults with religiously disinterested spouses and children, for instance, may work alongside devout Christians or have regular contact with religiously conservative parents. Likewise, religious traditionalists who regularly associate with like-minded believers nevertheless intermingle in various settings among those with less religious views. Despite, then, the tendency toward similarity in close social ties (i.e., homophily), we should not be surprised at the existence of at least some cross-religious relationships among American adults.³ One recent study documents that nearly 70 percent of religiously affiliated people cite nonreligious people as their friends (Vargas and Loveland 2011). Unfortunately, it remains unclear to what extent religious traditionalists can be found in the close networks of the nonreligious, and I am unaware of any existing national estimates concerning this type of network heterogeneity. Furthermore, I know of no existing studies that document the acts of social support that people receive from kin and nonkin religious traditionalists outside the specific context of their religious organization.

The Current Study: Traditionalist Ties and Social Support

By one line of reasoning, religious traditionalists may be the individuals most inclined to provide various forms of social support to the important people in their lives. Religious texts are replete with commands to “love thy neighbor”: Followers are told to give without expectation of return, to put others’ needs in front of their own, and to protect the weak and the vulnerable. Presumably, traditionalists believe these to be literal directives for everyday life.

Some empirical evidence, moreover, links scriptural injunction to personal belief and suggests that religious motivations stimulate a broad range of prosocial activity. Einolf (2011), for instance, reports that born-again Christians view Christ’s sacrifice as pivotal inspiration to selflessly serve others. In his landmark study on American volunteerism, Wuthnow (1991) likewise notes that conservative Christian churchgoers are strongly motivated by biblical models of sacrifice.

Psychological evidence also suggests that the sense of God’s awareness and supervision produces an implicit motivation to act generously and justly (Baumeister, Bauer, and Lloyd 2010; McCullough and Willoughby 2009), an effect that is most pronounced among those who “fully internalize” the beliefs of their religion (Koole et al. 2010). The sense of omnipresent Divine observation typical of orthodox believers may therefore provide additional motivation for helping others. Finally, congregational dynamics may reinforce a connection between traditionalism and social support. The high levels of commitment and social embeddedness characteristic of traditionalist congregations likely produce “stronger norms concerning various helping activities, including social support, framing them not as a prosocial behavior or personal favor but as a divine call” (McClure 2013:701). Interpersonal influence from co-congregants may therefore stimulate and sustain social support behavior. In light of these various perspectives, it is plausible to hypothesize that people are most likely to receive social support from those in their close networks that are religiously traditional.

It is important to recognize, however, that this hypothesis may apply only to a select subset of possible dyadic relations. Numerous studies suggest that prosocial activity performed by highly religious people is disproportionately targeted to others that are similarly religious. This has been demonstrated most clearly in experiments that assess the religiosity of both a participant and an assistance recipient during controlled social interaction (Galen 2012a:882).⁴ Extending this idea to social support provision, it is plausible that religious traditionalists are most likely to provide multiple forms of social support only to those who share their religious orientation or perhaps only to those who are especially religious. An extreme interpretation of the in-group/out-group dynamics involving religious traditionalists, in fact, may lead us to anticipate that very few religiously progressive or nonreligious Americans would even have traditionalists in their close networks. Depending on the extent of these sorting and selection processes, then, it could be difficult to assess whether religious traditionalists are most likely to provide social support to a close tie irrespective of the recipient’s religious persuasion. Taken together, these considerations suggest that it is important to first assess how prevalent the presence of religious traditionalists is in the personal networks of Americans across a broad spectrum of religiosity. It is also critical to account for the religiosity of the support beneficiary—both as a statistical covariate and as a variable that may interact with the religious traditionalism of the network tie—when assessing the primary hypothesis.

Another possibility is that people are most likely to receive social support from religious traditionalist ties, but only in the context of a kinship relation. Religious traditionalists in the United States have long emphasized the centrality of the family among other social institutions and stressed the moral obligations embedded within family life (Edgell 2006; Pearce and Axinn 1998). Kin selection theory, in general, predicts that family members could be more likely than nonkin to provide help and assistance. It is conceivable that religious traditionalism magnifies this prevailing inclination and expands the range of existing social support tendencies within families, yet has little impact on the behavior of nonkin ties. Given that forms of social support such as monetary assistance are especially more common among kin ties (Wellman and Wortley 1990), it is also important to assess the potential joint impact of religious traditionalism and kinship across a range of helping behaviors.

Method

Sample

Data for this study come from the 2006 Portraits of American Life Study (PALS), a nationally representative study of 2,610 American adults aged 18 and above. I examine the social ties among 2,435 people who identified at least one close network tie outside of their home and who were not missing on any of the study variables.⁵ All variables used had missing <5 percent, and the multivariate regressions feature a total n of 2,308 individuals (and 8,544 dyadic relationships involving the PALS respondents).

The primary purpose of the PALS project was to understand the role of religion in everyday life and its influence on social behaviors and attitudes (Emerson, Sikkink, and James 2010). As such, the PALS is an ideal data source for the current study; I am unaware of any other large-scale social surveys containing questions about the religious traditionalism of close social network ties and about the social support provided by these ties.

PALS respondents were selected through a multistage process. Zip code areas were first randomly selected with probability proportionate to size, then addresses were randomly selected from each zip code area, and finally one randomly selected adult was selected for a full interview from each selected household (after being screened for eligibility in an initial interview). The survey yielded an 83 percent contact rate, an 86 percent screening rate, and an 82 percent cooperation rate. Thus, the overall response rate was 58 percent (.83 × .86 × .82). Further details on the sampling design are available at http://www.palsresearch.org/pals/researchers. Interviews were conducted in English and Spanish through face-to-face in-home interviews, and selected sensitive questions used computer-assisted self-interviewing.

Social Network Measures

This study’s hypothesis is assessed with respondents’ reports about their close network ties. To initially ascertain the respondent networks, interviewers asked participants to think of “persons outside your home that you feel closest too” including “friends, co-workers, neighbors, relatives, or anyone else who does not live here.” Respondents were then asked to provide the name or initials of the person with whom they felt closest. All but 163 individuals named at least one close tie. This was followed by a request for the names or initials of a second, third, and fourth person to whom the respondent felt close.⁶

After filling in the network roster based on this name generator technique, interviewers asked respondents a series of follow-up questions about their ties. Respondents were asked about informal assistance: (1) “Which [of your ties] have volunteered their time to help you in times of need in the past three years?”; (2) “Which have provided advice that helped you make an important decision in the past three years?”; and (3) “Which have ever helped you by giving you money or a loan?” This set of actions roughly encompasses the emotional, informational, and instrumental aspects of social support. I created a binary variable for each form of support denoting whether the network tie provided it or not (1 = provided support; 0 = otherwise).

To measure religious traditionalism, I use the following question: “Which [ties] have views on spiritual or religious matters that are conservative or traditional?” Respondents were shown their list of network ties and asked to indicate all that apply. I denote religious ties as religious but not traditional (i.e., religious moderates or progressives) if the respondent indicated that they were religious, but did not identify their views as conservative or traditional. Nonreligious ties were those who fell into neither category. It should be noted that this measurement of ties’ religious orientation has several limitations. For one, it relies on a person’s judgment about the religious beliefs of their close ties—an evaluation that may be based on incomplete information (though DiPrete et al. 2011 argue, regarding this point, that “one cannot readily hide behaviors and values in close networks”; p. 1272). Second, the religion measures do not establish the particular branch of religion to which their close tie belongs. This feature, however, is actually somewhat advantageous if respondents can identify their tie as a religious conservative but understandably lack the theological astuteness to differentiate between myriad denominational varieties (e.g., fundamentalist Pentecostals vs. reformed Presbyterians vs. traditionalist Serbian orthodox). Since World War II, moreover, a growing number of religious conservatives identify as nondenominational believers (Wuthnow 1989). Nevertheless, I acknowledge that additional interviews with each of the ties identified could have yielded valuable information about their religious identity. Finally, the network tie religious orientation indicators used do not disentangle traditionalism from religious intensity, another key measurement limitation.

I employ several additional pieces of information about respondents’ ties. American women tend to be more religious than men; they also are more likely to provide various forms of aid to friends and neighbors (Campbell and Lee 1990), so I include the gender of each network tie as a covariate in the analysis. Types of assistance may also differ according to kinship status, and so I differentiate between network ties that are family members from those who are not. I also denote whether the tie is from the same faith tradition as the respondent (Christian [non-Catholic], Catholic, Muslim, Jewish, Mormon, Buddhist, Hindu, or no religious affiliation).

Respondent Measures

Several variables corresponding to the individual respondent are used in the analysis. Religious salience is measured by the question “How important is religion or religious faith to you personally?” Response categories include (1) “not at all important,” (2) “somewhat important,” (3) “very important,” (4) “extremely important,” and (5) “by far the most important part of your life.” Religious participation was assessed with a question frequency of “worship service” attendance, excluding weddings and funerals. Response options ranged from 1 = never to 8 = three times a week or more. Religious affiliation was categorized using the Steensland et al. (2000) classification scheme. Religious orientation was measured by asking about “views on religious or spiritual matters,” ranging on a 7-point scale from 1 = very conservative to 7 = very liberal.

Basic demographic variables include gender (coded as 1 = female, 0 = male), age (ranging from 18 to 80+), education (coded as approximate years of formal schooling based on highest degree completed), and race/ethnicity (denoted with dummy variables for black, Hispanic, or other race, with white non-Hispanic as the excluded reference group). To adjust for regional variation in exposure to religious traditionalists, I include dummy variables indicating residence in the American South, Midwest, or West (Northeast is the excluded reference group). Finally, I include two indicators of recent hardship that could signify an acute need for social support. Dummy variables denote whether the respondent “suffered a serious illness, injury, or an assault” and “a major financial crisis” in the past three years.

Analysis

Assessing the study’s research question requires a multilevel analysis: Social network ties and their characteristics are nested within the PALS participants. I use a random-intercept logistic regression model of social support provision/receipt to account for clustering of multiple observations within each respondent and to assess predictor variables corresponding to multiple levels of analysis (respondents’ network ties and the respondents themselves). The model takes the form,

L o g [p_{i j} / (1 - p_{i j})] = α + x_{i j} β + ω_{0} γ + ζ_{0 j} + ϵ_{i j} .

Here, the outcome variable is the log odds of a respondent (j) receiving social support from a network tie (i); network ties, the level 1 units, are indexed I = 1, . . ., I; respondents, the level 2 units, are indexed j = 1, . . ., J; α is a fixed mean intercept; $x_{i j}$ are variables corresponding to network ties; $β$ represents a vector of level 1 coefficients; $ω_{0 j}$ are variables corresponding to respondents; $γ$ represents level 2 coefficients; $ζ_{0 j}$ denotes level 2 residuals; and $ϵ_{i j}$ denotes level 1 residuals. The residuals at the respondent level are assumed to follow a Gaussian distribution, while those at level 1 are assumed to follow the standard logistic variance structure (π²/3).

Evaluation of the general model proceeds in three stages. First, I estimate a simple model for each social support outcome without any explanatory variables (model 0). The random-intercept modeling framework enables us to partition the variance of the outcome variable at level 1 and at level 2, and so this initial model establishes how much variance in support is attributable to differences between ties and how much is attributable to differences between respondents (Rabe-Hesketh and Skrondal 2008). This is completed by calculating the intraclass correlation (ICC or ρ) with the following formula for logistic regression:

ρ = \frac{σ_{i n t}^{2}}{σ_{i n t}^{2} + (\frac{π^{2}}{3})} .

In this formula, π²/3 represents the level 1 variance, while $σ_{i n t}^{2}$ represents variance at the level of the respondent (level 2).

The next model includes level 1 and level 2 predictors (model 1). This provides a test of whether religious traditionalism of the tie is associated with the likelihood of support received by the respondent from the tie, while holding constant the tie’s gender, kinship status to the respondent, and multiple characteristics of the respondent. This model is estimated for each of the three social support outcomes.

The final model includes cross-level interaction terms between tie-level religious traditionalism and respondent-level religious importance (model 2a) and religious orientation (model 2b). To aid in interpretation, the religious importance variable was centered so that “0” is equivalent to “somewhat important,” negative values indicating “not at all important,” and values ≥ 1 indicating heightened levels of religious importance.⁷ This centering strategy means that in model 2, the main effect coefficient for tie’s religious traditionalism is interpreted as the effect among those who see religion as only “somewhat important”; the interaction term indicates how the effect of ties’ religious traditionalism increases or decreases as the respondent’s religious importance score surpasses the neutral point. Religious orientation was centered at “moderate,” so that positive values indicate increasing conservativeness, while negative values indicate increasing liberality.

Finally, to evaluate kin/nonkin distinctions, I differentiate religious traditionalist ties by kinship status and replicate models 1 and 2 for each social support outcomes. All multilevel models use the Stata GLLAMM program, which accommodates the level 1 weighting and clustering scheme necessary because of the PALS complex survey design. Models are estimated with maximum likelihood.

Results

Descriptive and Bivariate Results

Table 1 presents descriptive information for the sample using the PALS survey weights. In addition to providing proportions and mean values for each respondent-level variable, I examine how key aspects of people’s social networks differ according to these factors. Several findings are notable.

Table 1.

Weighted Descriptive Statistics, the Portraits of American Life Survey.

	Weighted mean/proportion	Social network data
		n of ties (range = 1–4)	Have religious traditionalist tie (proportion)			Proportion of all ties providing social support
		n of ties (range = 1–4)	Any	Nonkin	Kin	Help	Advice	Money
Overall		3.73	0.70	0.44	0.44	0.79	0.59	0.26
Religious importance
Not at all	0.11	3.76	0.57	0.38	0.32	0.78	0.62	0.29
Somewhat	0.27	3.71	0.64	0.38	0.38	0.79	0.59	0.28
Very	0.29	3.69	0.71	0.44	0.43	0.76	0.55	0.24
Extremely	0.18	3.83	0.79	0.50	0.52	0.84	0.63	0.25
Most important	0.16	3.72	0.80	0.55	0.39	0.80	0.59	0.24
Religious orientation
Very liberal	0.11	3.80	0.59	0.36	0.35	0.82	0.65	0.28
Somewhat liberal	0.17	3.75	0.67	0.37	0.44	0.79	0.57	0.28
Moderate/liberal	0.14	3.73	0.66	0.42	0.36	0.83	0.64	0.27
Moderate	0.02	3.73	0.32	0.18	0.18	0.78	0.57	0.29
Moderate/conservative	0.22	3.74	0.75	0.49	0.43	0.78	0.56	0.22
Somewhat conservative	0.22	3.73	0.81	0.52	0.53	0.79	0.61	0.27
Very conservative	0.12	3.73	0.84	0.54	0.56	0.77	0.54	0.24
Religious affiliation
Black Protestant	0.05	3.56	0.67	0.43	0.42	0.78	0.60	0.44
Evangelical Protestant	0.27	3.77	0.79	0.57	0.48	0.80	0.56	0.24
Mainline Protestant	0.14	3.79	0.70	0.42	0.43	0.82	0.62	0.18
Catholic	0.26	3.65	0.70	0.39	0.45	0.75	0.59	0.25
Jewish	0.02	3.92	0.65	0.44	0.40	0.74	0.56	0.15
Other	0.06	3.78	0.58	0.35	0.36	0.76	0.54	0.21
Unaffiliated	0.16	3.79	0.62	0.38	0.36	0.80	0.60	0.33
Protestant, other	0.04	3.76	0.75	0.40	0.52	0.83	0.58	0.29
Church attendance (past year)
Never	0.26	3.69	0.63	0.38	0.39	0.77	0.54	0.28
Once or twice	0.16	3.78	0.72	0.43	0.42	0.82	0.65	0.30
Several times	0.13	3.79	0.70	0.43	0.44	0.79	0.59	0.30
Once a month	0.06	3.58	0.69	0.37	0.47	0.76	0.63	0.28
2–3x a month	0.09	3.72	0.74	0.47	0.49	0.79	0.62	0.24
Once a week	0.19	3.75	0.76	0.48	0.46	0.80	0.57	0.21
Twice a week	0.06	3.75	0.72	0.57	0.39	0.78	0.58	0.21
Three or more times a week	0.05	3.83	0.84	0.62	0.52	0.83	0.50	0.15
Female	0.52	3.79	0.71	0.45	0.47	0.83	0.63	0.25
Male	0.48	3.72	0.70	0.44	0.40	0.75	0.54	0.26
White	0.72	3.79	0.72	0.46	0.46	0.82	0.60	0.22
Black	0.11	3.55	0.66	0.39	0.41	0.77	0.61	0.44
Hispanic	0.12	3.55	0.65	0.39	0.37	0.70	0.53	0.32
Other race	0.06	3.73	0.60	0.38	0.29	0.68	0.48	0.19
Age (range = 18–80)	45.12	r = .06	r = .09	r = .03	r = .07	r = −.14	r = −.28	r = −.30
Education (range = 10–23)	14.46	r = .13	r = .03	r = .06	r = −.01	r = .10	r = .10	r = −.17
Northeast	0.17	3.74	0.65	0.39	0.39	0.78	0.57	0.23
South	0.34	3.72	0.71	0.46	0.44	0.80	0.57	0.26
Midwest	0.23	3.73	0.75	0.47	0.47	0.82	0.59	0.26
West	0.26	3.75	0.69	0.43	0.42	0.76	0.60	0.26
Recent health problem	0.22	3.71	0.74	0.46	0.45	0.78	0.53	0.28
Recent financial problem	0.23	3.66	0.73	0.48	0.45	0.81	0.63	0.38

Note. r = Pearson correlation coefficient (shown for continuous variables in Column 1).

First, core network size differed according to some demographic traits (e.g., white respondents had the largest networks; education was positively correlated with network size), but it does not appear to vary systematically by participants’ religiosity. Analysis of variance (ANOVA) tests comparing mean number of ties across levels of personal religiosity fail to detect statistically significant differences, for religious importance, F(4, 2430) = 2.33, p = .053; for church attendance, F(7, 2430), = 1.61, p =.128.

Second, across the board, there is a notable presence of perceived religious traditionalists in respondents’ core networks. Unsurpri-singly, a high proportion of those reporting very high levels of personal religiosity and frequent church attendance indicated they have at least one traditionalist on their network roster (e.g., 79 percent and 80 percent of those saying religion is “extremely” important and the “most import” thing in their life, respectively; 76 percent of those reporting weekly church attendance and 84 percent among those who go 3+ times a week to church). Perhaps more remarkable, however, is that well over half of those who say that religion is “not at all” or that it is only “somewhat” important had at least one such tie in their close network (57 percent and 64 percent, respectively).⁸ Similarly high proportions also emerge for those at the liberal end of the religious orientation distribution, the religiously unaffiliated, and those at the low end of the service attendance distribution. Respondents across the religiosity distributions were about as likely to report a traditionalist kin tie as they were a traditionalist nonkin tie. Other factors associated with reporting a religious traditionalist include being white, relative to other racial/ethnic groups, and living in the South and Midwestern United States as compared with the Northeast and the West.

Finally, Table 1 indicates that the receipt of social support, as indicated by proportion of close ties providing respondents with each form of assistance, does not differ systematically according to personal religiosity. Respondents reporting high levels of religious importance received help from a slightly higher proportion of their close ties than do people with lower levels of religious importance, but those reporting the highest levels of religious importance and (especially) the most frequent church attendance received monetary assistance from a lower proportion of their ties. Religious liberals tended to receive more help and advice than did the most religiously conservative respondents. In general, this set of findings mitigates the concern that people of a particular religious profile are simply more likely to receive all types of support from each of their close ties and also reinforces the importance of looking across different forms of social support behavior because prevalence of assistance clearly varies across subsets of the population.

Overall, monetary assistance was the rarest form of support (26 percent of all ties across all respondents provided such support), while practical help was quite common (79 percent of all ties provided such support). Although not the focus of this study, Table 1 reveals differences in support receipt by gender, race, age, and education.

Multivariate Results

The results of random-intercept logistic regression models for three social support outcomes are presented in Table 2. The first set of models (model 0 for each outcome) is an intercept-only model used to show the variation in social support receipt attributable to PALS respondents. For each of the social support variables, between 52 percent and 63 percent of the variance is between respondents. This indicates that while characteristics of a person’s social ties are important for understanding whether or not they offer help, advice, or money, much of the explanation rests on characteristics of the focal individual him or herself.

Table 2.

Estimates from Random-Intercept Logistic Regression Models Predicting Social Support Provision.

	Model 0			Model 1			Model 2a			Model 2b
	Help	Advice	Money	Help	Advice	Money	Help	Advice	Money	Help	Advice	Money
Tie level (level 1)
Religious traditionalist^a				0.75*** (0.18)	0.60*** (0.11)	0.49*** (0.10)	0.43* (0.20)	0.38* (0.15)	0.47*** (0.12)	0.72*** (0.17)	0.56*** (0.11)	0.49*** (0.10)
Nonreligious^a				−0.29 (0.40)	−0.26 (0.26)	0.11 (0.27)	0.10 (0.67)	−0.15 (0.46)	−0.08 (0.40)	−0.29 (0.40)	−0.23 (0.28)	0.12 (0.28)
Same religious faith as R				0.30* (0.14)	0.20* (0.09)	0.07 (0.13)	0.28 (0.15)	0.18* (0.09)	0.07 (0.12)	0.30* (0.15)	0.18 (0.09)	0.07 (0.12)
Kin				0.53*** (0.12)	0.26* (0.10)	1.50*** (0.13)	0.55*** (0.12)	0.27** (0.10)	1.50*** (0.13)	0.54*** (0.12)	0.27** (0.10)	1.51*** (0.13)
Female				0.10 (0.10)	0.14 (0.10)	−0.28** (0.09)	0.11 (0.10)	0.16 (0.10)	−0.28** (0.09)	0.10 (0.10)	0.15 (0.10)	−0.28** (0.09)
Respondent level (level 2)
Religious importance (RI)				0.05 (0.12)	0.05 (0.07)	0.04 (0.06)	−0.07 (0.12)	−0.05 (0.08)	0.03 (0.07)	0.05 (0.12)	0.04 (0.07)	0.04 (0.06)
Religious orientation (RO)				−0.08 (0.04)	−0.03 (0.04)	−0.07* (0.03)	−0.09* (0.04)	−0.03 (0.04)	−0.07* (0.03)	−0.11* (0.05)	−0.09* (0.04)	−0.07 (0.04)
Church attendance				0.02 (0.05)	0.002 (0.04)	−0.06 (0.03)	0.01 (0.05)	−0.01 (0.04)	−0.06 (0.03)	0.02 (0.05)	−0.002 (0.04)	−0.06 (0.03)
Black Protestant^b				0.09 (0.42)	−0.59 (0.31)	0.10 (0.26)	0.09 (0.42)	−0.59 (0.31)	0.10 (0.26)	0.13 (0.42)	−0.55 (0.30)	0.10 (0.26)
Evangelical Protestant^b				0.03 (0.31)	−0.57** (0.19)	−0.09 (0.18)	0.03 (0.31)	−0.58** (0.19)	−0.09 (0.17)	0.03 (0.31)	−0.59** (0.19)	−0.09 (0.18)
Mainline Protestant^b				0.06 (0.29)	−0.08 (0.20)	−0.30 (0.25)	0.07 (0.28)	−0.07 (0.20)	−0.30 (0.25)	0.07 (0.29)	−0.07 (0.20)	−0.30 (0.25)
Jewish^b				−0.98 (0.57)	−0.78 (0.45)	−0.19 (0.66)	−0.98 (0.52)	−0.81 (0.46)	−0.19 (0.66)	−0.93 (0.56)	−0.76 (0.47)	−0.18 (0.66)
Other religion^b				0.004 (0.39)	−0.36 (0.48)	−0.31 (0.27)	−0.002 (0.38)	−0.37 (0.49)	−0.31 (0.27)	−0.01 (0.39)	−0.39 (0.49)	−0.31 (0.27)
Unaffiliated^b				0.22 (0.46)	−0.30 (0.30)	0.26 (0.22)	0.15 (0.46)	−0.33 (0.30)	0.25 (0.22)	0.19 (0.45)	−0.32 (0.29)	0.26 (0.22)
Other Protestant^b				−0.07 (0.45)	−0.53 (0.31)	−0.35 (0.31)	0.02 (0.55)	−0.53 (0.30)	−0.35 (0.31)	−0.04 (0.46)	−0.53 (0.31)	−0.35 (0.31)
Gender (male)				−0.89*** (0.22)	−0.64*** (0.15)	−0.08 (0.12)	−0.89*** (0.23)	−0.64*** (0.14)	−0.08 (0.12)	−0.88*** (0.22)	−0.63*** (0.14)	−0.08 (0.12)
Age				−0.04*** (0.01)	−0.05*** (0.01)	−0.05*** (0.00)	−0.04*** (0.01)	−0.05*** (0.01)	−0.05*** (0.00)	−0.04*** (0.01)	−0.05*** (0.01)	−0.05*** (0.00)
Education				0.13*** (0.04)	0.11*** (0.02)	−0.10*** (0.03)	0.13*** (0.04)	0.10*** (0.02)	−0.10*** (0.03)	0.13*** (0.03)	0.10*** (0.02)	−0.10*** (0.03)
Black^c				−0.74* (0.32)	0.02 (0.18)	1.28*** (0.28)	−0.74* (0.34)	0.04 (0.20)	1.29*** (0.28)	−0.75* (0.32)	0.03 (0.18)	1.29*** (0.28)
Hispanic^c				−1.05*** (0.29)	−0.94*** (0.17)	0.38* (0.19)	−0.99*** (0.28)	−0.91*** (0.17)	0.38* (0.19)	−1.04*** (0.29)	−0.93*** (0.17)	0.38* (0.19)
Other race^c				−1.53*** (0.32)	−1.05** (0.34)	−0.05 (0.29)	−1.56*** (0.34)	−1.05** (0.34)	−0.05 (0.28)	−1.53*** (0.32)	−1.05*** (0.35)	−0.04 (0.28)
Midwest^d				0.40 (0.33)	0.03 (0.24)	0.20 (0.27)	0.39 (0.31)	0.03 (0.24)	0.20 (0.27)	0.42 (0.33)	0.07 (0.24)	0.20 (0.27)
South^d				0.14 (0.22)	−0.08 (0.22)	−0.05 (0.25)	0.12 (0.21)	−0.09 (0.21)	−0.05 (0.24)	0.16 (0.22)	−0.07 (0.22)	−0.05 (0.25)
West^d				0.04 (0.19)	0.31 (0.23)	0.25 (0.27)	0.02 (0.18)	0.29 (0.22)	0.25 (0.27)	0.05 (0.20)	0.32 (0.23)	0.25 (0.27)
Recent health problem				0.04 (0.20)	−0.33 (0.18)	0.28* (0.14)	0.03 (0.21)	−0.33 (0.18)	0.28* (0.14)	0.05 (0.21)	−0.31 (0.18)	0.28* (0.13)
Recent financial problem				−0.003 (0.21)	0.19 (0.15)	0.81*** (0.13)	0.01 (0.22)	0.20 (0.15)	0.81*** (0.13)	0.001 (0.21)	0.20 (0.15)	0.81*** (0.13)
Cross-level interactions
Religious traditionalist × RI							0.31** (0.12)	0.22** (0.08)	0.02 (0.08)
Nonreligious × RI							−0.22 (0.38)	−0.05 (0.29)	0.15 (0.29)
Religious traditionalist × RO										0.08 (0.08)	0.14*** (0.04)	−0.002 (0.06)
Nonreligious × RO										0.03 (0.22)	−0.17 (0.15)	−0.09 (0.16)
Constant	2.43*** (0.12)	0.78*** (0.09)	−1.66*** (0.11)	1.87*** (0.63)	1.45*** (0.45)	0.62 (0.46)	2.07*** (0.62)	1.62*** (0.45)	0.71 (0.44)	1.56* (0.75)	1.57*** (0.45)	0.71 (0.43)
Respondent-level variance	4.96 (0.33)	5.48 (0.75)	3.59 (0.59)	4.97 (0.47)	3.35 (0.35)	2.46 (0.35)	5.00 (0.47)	3.33 (0.36)	2.46 (0.35)	4.99 (0.48)	3.31 (0.36)	2.46 (0.35)
ICC	0.60	0.63	0.52	0.60	0.50	0.43	0.60	0.50	0.43	0.60	0.50	0.43
−2LL	3,760.8	5,174.11	4,481.31	3,585.64	4,958.71	4,021.85	3,577.95	4,952.18	4,021.59	3,584.33	4,949.46	4,021.56
AIC	7,525.6	10,352.22	8,966.62	7,227.28	9,973.43	8,099.71	7,215.91	9,964.36	8,103.18	7,228.66	9,958.92	8,103.13
BIC	7,939.7	10,366.33	8,980.73	7,424.76	10,170.91	8,297.19	7,427.50	10,175.95	8,314.77	7,440.25	10,170.51	8,314.72
Observations
Level 1	8,544	8,544	8,544	8,544	8,544	8,544	8,544	8,544	8,544	8,544	8,544	8,544
Level 2	2,308	2,308	2,308	2,308	2,308	2,308	2,308	2,308	2,308	2,308	2,308	2,308

Source. Portraits of American Life Survey.

Note. Table presents unstandardized regression estimates. Robust standard errors are shown in parentheses.

Religious nontraditionalist is reference group.

Catholic is reference group.

White is reference group.

Northeast is reference group.

p < .05. **p < .01. ***p < .001.

Model 2 adds respondent- and tie-level explanatory variables. For each of the three outcome variables, respondents were more likely to receive social support from religiously traditional ties than from nontraditionalists (the reference group).⁹ This association, moreover, is net of the ties’ gender and kinship status, as well as the respondent’s own religiosity, education level, gender, and other demographic characteristics. For instance, perceived religious traditionalism doubled the odds that a given tie would offer help (0.75 exponentiated = 2.12) and increased the odds of advice giving and monetary assistance by 82 percent (e⁶⁰ = 1.82) and 63 percent (e⁴⁹ =1.63), respectively. Interestingly, nonreligious ties did not differ statistically from the nontraditionalist reference group for any of the three social support outcomes. Ties were also more likely to provide each form of support if they were a kin member. This is especially pronounced in the case of money, where kinship status predicts a more-than-fourfold increase in the odds of support (e^1.51 = 4.53). Monetary assistance also stands out in that it is the only support behavior for which gender of the tie is a statistically significant predictor; female ties are less likely than males to give or to loan money.

Women, however, are more likely than men to receive help and advice from a given tie in their network. Other relevant covariates at level 2 include age, which decreased the odds of the respondent receiving each form of support; education, which generally manifested the opposite association (except in the case of monetary support); and race/ethnicity, which also manifested a pattern in which racial minorities are less likely than whites to obtain help and advice from close ties, yet more likely than whites to get monetary help.

Adding explanatory variables did not change the ICC in model 1 for help, perhaps an unsurprising finding because this form of assistance is so ubiquitous within close networks (i.e., 78 percent of close ties provided it, and 94 percent of the sample had at least one tie that gave help). The ICC for models of advice and money, however, did decrease somewhat (from .63 to .50 for advice, .52 to .43 for money), indicating that about 10 percent of the between-person variability associated with the odds of receiving social support from a given tie can be accounted for by the respondent-level variables.

To assess whether the social support boost associated with perceived religious traditionalism is concentrated mainly among people who themselves are highly religious or are particularly conservative, I turn to model 2, which includes a cross-level interaction term between ties’ religious traditionalism (level 1) and respondents’ religious importance (2a) or orientation (2b; level 2). For each of the three outcomes, religious traditionalism in the uppermost row remains a positive and statistically significant predictor of social support. In model 2a, the main effect term represents the association between tie-level traditionalism and social support for those respondents who see religion as only “somewhat important.” Notably, two of the three coefficients for the interaction term are also positive and significant—those in model 2 predicting help and in model 2 predicting advice. This indicates that people reporting relatively low levels of religious importance are more likely to receive social support from their religiously traditionalist ties, yet that the odds of receiving help or advice from a traditionalist tie increase substantially as a person’s own religiosity increases. Interestingly, however, such a pattern does not manifest for monetary assistance. Here, the interaction term coefficient is small and far from significant (b = .02, p = .80), and the main effect coefficient is largely unchanged from model 1. This indicates that respondents are most likely to receive monetary assistance from religious traditionalists, and the importance of religiosity for the recipient has little or nothing to do with the monetary exchange.¹⁰

Model 3b focuses on the respondent’s own religious orientation. Here, there is somewhat less evidence for homophily-driven social support, as the only statistically significant interaction term is for the advice outcome (b = .14, p < .001). The main effect terms for perceived religious traditionalist shift very little from their counterparts in model 2. Overall, the results suggest that a respondent’s own connection to traditional beliefs does little to enhance the general likelihood that he or she will receive help or money from a perceived traditionalist, but receipt of advice from traditionalists is most pronounced among religiously conservative respondents. Nevertheless, even those at the moderate to liberal end of the distribution report being more likely to receive advice from those they perceive to be traditionalists than from other types of ties.¹¹

Kin, Nonkin Distinctions

Table 3 probes deeper into the association between ties’ religious traditionalism and social support behavior by differentiating between kin and nonkin close ties. The results have already demonstrated the importance of kinship for predicting social support—especially in the case of monetary assistance. Religious traditionalism may operate differently within families than it does among neighbors, co-workers, or more general friendships. To explore this possibility, I created kinship-specific traditionalism/nontraditionalism/nonreligious variables and repeated the analyses from Table 2 (excluding the intercept-only model). Nonkin, nontraditionalist ties are the reference category in Table 3, and the models adjust for all other covariates included in Table 2. For the sake of space, Table 3 focuses on model 1 (full tables with models 2a and 2b and all covariates are available upon request).

Table 3.

Estimates from Random-intercept Logistic Regression Models Predicting Social Support Provision and Differentiating Ties by Kin/Nonkin Status.

	Help	Contrast to tradition in relevant kin/nonkin comparison	Advice	Contrast to tradition in relevant kin/nonkin comparison	Money	Contrast to tradition in relevant kin/nonkin comparison
Nontraditionalist, nonkin	ref		ref		ref
Nonreligious, nonkin	−0.07 (0.44)	p = .080^a	−0.35 (0.36)	p = .008^a	−0.014 (0.33)	p = .188^a
Religious traditionalist, nonkin	0.69*** (0.19)		0.62*** (0.11)		0.45** (0.17)
Nontraditionalist, kin	0.49*** (0.11)	p = .004^b	0.28* (0.13)	p = .026^b	1.48*** (0.20)	p = .010^b
Nonreligious, kin	−0.85 (0.72)	p = .0002^b	−0.03 (0.36)	p = .001^b	0.27 (0.64)	p = .0001^b
Religious traditionalist, kin	1.31*** (0.21)		0.85*** (0.15)		1.99*** (0.17)
Constant	1.59* (0.76)		2.21*** (0.59)		1.33* (0.56)
Respondent-level variance	4.95 (0.46)		3.35 (0.35)		2.45 (0.35)
ICC	0.60		0.50		0.43
−2LL	3,583.65		4,956.13		4,019.58
AIC	7,229.30		9,974.26		8,101.17
BIC	7,447.94		10,192.90		8,319.81
Observations
Level 1	8,544		8,544		8,544
Level 2	2,308		2,308		2,308

Source. Portraits of American Life Survey.

Note. Table presents unstandardized regression estimates. Robust standard errors are shown in parentheses. Analyses adjust for all additional covariates shown in Table 2.

Comparison is with religious traditionalist, nonkin.

Comparison is with religious traditionalist, kin.

p < .05. **p < .01. ***p < .001.

Results from Table 3 indicate that there is some utility to disaggregating kin from nonkin ties with respect to religious traditionalism. First, the results indicate that ties’ religious traditionalism predicts a higher likelihood of social support that extends beyond kinship; indeed, as indicated by the statistically significant and positive coefficients across the third row of Table 3, nonkin traditionalists are more likely than nontraditionalist, nonkin ties to provide generic help, advice, or money. Nonreligious, nonkin ties, on the other hand, did not differ statistically from their nontraditionalist counterparts. Wald χ² tests are used to compare the coefficients for nonreligious ties with traditionalist ties among nonkin network members, and the results of these comparisons generally support the conclusion that traditionalist ties are distinctive in support from both alternative groups in the analysis (e.g., the contrast for help approaches conventional significance levels, and the contrast for advice has a p value of .008).¹² Interestingly, nonkin traditionalist ties are even more likely than kin nontraditionalists to provide advice that respondents found helpful, as indicated by post hoc Wald χ² tests of coefficient differences (p = .035).

Second, the results from Table 3 demonstrate that religious traditionalism is associated with a pronounced increase in the odds of social support among kin members themselves. Bearing in mind that the reference group is nontraditionalist, nonkin, the coefficients for each kin dummy variable must be compared with that designation. Therefore, the positive coefficients for nontraditionalist kin ties indicate that such network members are more likely than their nonkin counterparts to provide each form of social support. For our purposes, however, the key comparisons are between (1) nontraditionalist kin and traditionalist kin and (2) nonreligious kin and traditionalist kin (e.g., within kinship comparisons). For each social support behavior, χ² tests reveal that coefficients are statistically larger for kin traditionalists relative to kin nontraditionalists; indeed, for help and for advice, the unstandardized coefficients are more than twice as large (1.31 vs. 0.49; 0.85 vs. 0.28, respectively). The magnitude of difference is somewhat lower for monetary support, but kin traditionalists are still 34 percent more likely than kin nontraditionalists to have given such assistance ([(((1.99 − 1.48)/1.48) × 100) = 34.46 percent]; p = .01) Similar findings emerge when comparing the traditionalists with the nonreligious. Here, the coefficient differences are even more pronounced, and the p values for χ² comparisons are quite low (all ≤.001).

To again assess whether the effects attributed to ties’ traditionalism is concentrated solely among certain segments of the population, I also estimated the kin-specified model with a series of interaction terms (akin to models 2a and 2b in Table 2; full results available upon request). In brief, the results from the religious salience cross-level interactions indicated that the respondents reporting the lowest levels of religious importance were the most likely to receive helpful advice from nonreligious kin. No other interactions were statistically significant, and conclusions about traditionalism were unchanged. For the religious orientation cross-level interactions, respondents with the most liberal views were the most likely to receive help from kin nontraditionalists. The advice outcome yielded several significant interactions: Religious conservatives were the most likely to receive helpful advice from kin traditionalists (though the main effect term for kin traditionalism remained positive and statistically significant [b = .84, p < .001]), the same pattern was observed for nonkin traditionalists (main effect coefficient for traditionalism was b = .61, p < .001), and religious liberals were the respondents most likely to report getting helpful advice from their nonreligious kin. Taken together, these additional findings suggest that some kin-specific religious homophily matters for social support exchange—particularly when it comes to receiving helpful advice—but provide very weak support for the idea that an association between perceived traditionalism and social support is concentrated solely among a select group of highly religious or religious conservative recipients.¹³

Discussion

Much has been made of the growing religious divide in the United States, and the consequences of religious traditionalism, in particular, have been the subject of intense scholarly criticism and public debate (e.g., Hitchens 2007; Stark 2012). Prosociality and social support represents one important offshoot of the broader discussion. Entering this dialogue, I investigated the straightforward issue of social support within people’s personal networks. Acts such as sharing money, providing advice, and offering practical assistance are among the many informal prosocial behaviors that sustain community life (Wellman and Wortley 1990), yet few studies have addressed their connection to religious traditionalism. I hypothesized that ties understood to be religious traditionalists would be most likely to offer each type of assistance, but I also offered several points of specificity that could complicate the general association between religious traditionalism and support receipt/provision. The analyses reveal several noteworthy findings.

First, people perceived as religious traditionalists are found within the close networks of a surprising number of American adults. It is to be expected, of course, that those who attend church regularly, who espouse theological conservatism, and who endorse high personal importance of religiosity would be most likely to identify religious traditionalists within their core network. What is notable, however, is the fact that more than half of those who indicate that religion is “not at all” important and nearly two-thirds of those who say it is only “somewhat” important nevertheless count at least one traditionalist among their close ties. Similarly, perceived traditionalists are found in the networks of 59 percent of the Americans who say their religious views are very liberal and in 60 percent among those who have not attended a religious service at all within the past year. This underscores the importance of examining the role of religion in social networks beyond formal institutional bases of affiliation (e.g., churches). In the current context of highly differentiated personal networks that cross-cut diverse social groups and institutions (Rainie and Wellman 2012), it is essential to recognize that religious traditionalists can find their way into the close networks of even the most secular American men and women. Fischer and Mattson (2009) have recently argued that strong accounts of “culture war” attitude polarization have been overblown; while the extent to which this is true may be subject to debate, the findings of the current study likewise suggest that sharp network polarization on the basis of religious orientation is not an accurate portrayal of the contemporary United States. Previous research reports that regardless of their own religious preference, about 7 in 10 Americans have at least one nonreligious friend (Vargas and Loveland 2011), but I am unaware of any previous work that documents the presence of traditionalists in Americans’ close networks.

Second, I find that even after adjusting for core aspects of their own religiosity, American adults are more likely to receive various forms of social support from those they identify as religious traditionalists than from other people in their close networks. This pattern was consistent for both kin and nonkin relations. Kin ties—who, overall, were more likely than nonkin ties to provide help, advice, or money—demonstrated greater levels of support provision if they were understood to be traditionalists than if they were not. For help and for advice, perceived traditionalism more than doubled the odds of support among kin members. Similarly, the likelihood of nonkin ties offering support was boosted by perceived religious traditionalism. In the case of help and advice, nonkin traditionalists were actually somewhat more likely than nontraditionalist kin members to offer support (though the difference was only statistically significant for the advice outcome). While the data do not enable us to identify the particular reasons why traditionalists appear to offer more social support than others, several explanations are plausible. Social causation would be operant if traditionalist beliefs and church-based teachings motivate prosocial behavior; those who most deeply internalize the demands of their faith and see them as literal directives for day-to-day life could well be those who act most altruistically in the lives of their close associates. On the other hand, social selection could imply that people who are inclined to do the most good toward their friends and family feel most drawn to traditionalist belief and practice (or perhaps are perceived to identify with traditionalism—more on that point below).

Third, the findings offer only limited support for the “minimal prosociality” thesis (Galen 2012a; Saroglou 2006)—the idea that traditionalists provide support only for those who share their faith commitments. Results do suggest that the odds of a perceived religious traditionalist offering social support—particularly useful advice—increase at higher levels of a respondent’s own religiosity. This shows that while perceived traditionalists are generally more likely than others to offer various forms of assistance, they are indeed especially supportive among peers that strongly endorse religious belief. Yet even when including interaction terms between perceived tie traditionalism and several relevant dimensions of personal religiosity, a main effect for tie traditionalism remained positive and statistically significant. The findings were further supported when close network ties were differentiated by their kin/nonkin status.

In drawing attention to how the perceived religious orientation of people’s general network ties—not only those constrained by organizational co-membership—is associated with social support provision, I do not intend to downplay the centrality of congregational contexts for facilitating bonding social contact among members (McClure 2013) or for organizing civic engagement that bridges to those beyond the congregation (Cnaan 2002; Lewis, MacGregor, and Putnam 2013). Such processes indeed demonstrate how religious organizations are “a rich source of social capital in the United States” (Beyerlein and Hipp 2006:97). Recognizing the importance of these dynamics, I see the current approach as complementary with the congregational focus of prior research (e.g., Edgell et al. 2013; McClure 2013). The frequent and consistent gathering of like-minded traditionalists would be expected to efficiently facilitate social support transactions. Such congregations, however, can also matter beyond their immediate context, by inspiring or commanding “love of neighbor” or by providing a pool of additional co-social-supporters that could be enlisted to aid friends or family in need (e.g., congregant friends, ministers, or other staff). An extra-congregational focus also emphasizes, however, that modern life—including that inhabited by religious traditionalists—entails hopping between differentiated institutional worlds as well as maintaining complex kinship relations that are also becoming increasingly religiously differentiated (see, for example, Bengtson 2013).

While the results of this analysis lend support to my central hypothesis, it is important to avoid overgeneralization when interpreting the current findings in light of the broader literature on religion and prosociality. Batson and Powell (2003:465) note that “variables accounting for variance in one form of prosocial behavior in one setting are not likely to account for the same amount of variance (if any) in other forms of behavior or in other settings.” This implies that religious traditionalists could indeed exhibit “minimal prosociality” when dealing with strangers or acquaintances, even if they are most likely to provide help to those with whom they share a close relationship. Charitable giving—one customary benchmark of prosocial behavior—was evaluated in the analysis as financial assistance to a specific close relation. But again, we can say nothing about charitable giving to the benefit of anonymous recipients. There is some empirical evidence to suggest that traditionalists give away relatively high amounts of money to various causes, religious and secular alike (see Saroglou 2012), yet some scholars debate the utility of donation measures commonly found in the literature (see Galen 2012a).

The current results also pertain only to personal networks and informal support behavior, not to civic prosocial engagement and helping behavior in formal organizations. One influential sector of religious traditionalism—Evangelical Protestantism—is relevant in this regard, as involvement in such churches appears to suppress participation and leadership rates in civic organizations such as labor unions, neighborhood associations, or arts and culture interest groups (Beyerlein and Hipp 2006; Driskell, Lyon, and Embry 2008). According to some observers, this phenomena may reduce the supply of volunteer labor and curb opportunities for bridging social capital in the broader community (Beyerlein and Hipp 2006; Schwadel 2005). It may be that many traditionalists intentionally withdraw (or feel closed-off from) the broad-based civic institutions that do not fully share their particular values, yet compensate for their ostensibly lower rates of civic prosociality with high levels of informal supportive behavior. Good neighboring and supportive extended family relationships are certainly critical beams of community infrastructure and crucial for individual well-being (Fischer 1982; Wellman and Wortley 1990), yet prosocial behavior unmoored from civic organizations could well be less stable and have less capacity to meet large-scale community needs. Careful, nuanced research is needed to uncover how traditionalists of different religious backgrounds navigate the institutional field of civic life, how they allocate their prosocial energies, to what extent such groups see their behavior as complementing and/or replacing formal civic institutions, and how they construct the sense of “a common good.” From a pragmatist position, it would be helpful to “accept that people have different versions of what is ‘good,’ to respect these differences, and then to interrogate how the differences matter for the ways people evaluate society and choose to engage” (Bennett et al. 2013:524).

Limitations and Conclusion

The present study adopted a novel approach for empirically assessing the prosociality of religious traditionalists—obtaining reports from a random sample of American adults about the social support behavior or those in their close social network. This strategy avoids the trap of having to believe respondents’ reports about their own prosocial activity; it also locates prosocial activity in a real-world relational context rather than in the contrived setting of an experimenter’s lab. Still, my approach deserves critical scrutiny on the basis of several important limitations.

First, this study is not immune to the possibility that people interpret the religiosity of their close tie as a result of their social support provision. This scenario is plausible in light of widespread religion-morality stereotyping in which “most individuals have a strong tendency to assume that there is an association, and even a causal connection, between religiosity and morality” (Galen 2012a:878). If operant, this endogeneity could confound my basic conclusions. One advantage of the traditionalism question used in the PALS survey, however, is that it refers to a particular strand of religious life and a pertinent point of cultural division regarding attitudes and worldview rather than invoking a simple dividing line between the religious and nonreligious. Undoubtedly, for many people, the very words “conservative” or “traditional” prime mental images of narrow-mindedness, intolerance, or bigotry (cf. Brandt et al. 2014; Yancey 2010). These implicit associations would certainly counteract the general tendency to unconsciously associate religion with “niceness” or prosociality. What is more, the findings suggest that even those with liberal theological views report being relatively likely to receive various forms of social support from perceived religious traditionalists.

Second, these data provide no way of knowing whether respondents are more likely to ask the traditionalists in their network for support. American adults may simply assume—correctly or otherwise—that traditionalists are most likely to lend money, advice, or assistance, and so they may be more frequently sought out for help. The PALS survey measure is not equipped to assess potential issues such as network members’ sense of initiative, their delivery of reluctant or unenthusiastic support, and a range of other subtle processes that certainly arise in social support exchanges.

A third critical limitation is the fact that the conclusions are drawn chiefly from the network members of a representative sample of Americans—these network ties do not themselves represent a sample of American adults. The population distribution of prosociality among traditionally religious people could be bimodal, with equal numbers very supportive and extremely unsupportive. To the extent that people tend to eschew the most misanthropic religious people in the population, my estimates could be biased by selection effects (i.e., only the remarkably supportive religious individuals show up in people’s network rosters). In any event, our assessment of ties’ supportiveness—regardless of their religious orientation—is contingent on them being nominated as close ties in the first place.

Despite these limitations, the current study has contributed to the growing body of research on religion and prosociality by addressing social support, an important, but somewhat understudied topic in the sociology of prosocial behavior. Future research could probe further into how religious traditionalism—as well as other aspects of American religiosity—shapes the everyday, give-and-take behavior of ordinary social life. By design, my analysis focused on close relations, but other studies should examine processes that occur within weak-tie networks, such as religious people’s willingness to share a job lead or to make a medical care referral. Assessing if, how, and when religious believers act in prosocial ways will certainly require far more analysis and methodological innovation, but I hope the present study raises new issues for reflection and offers some helpful leads for future research directions.

Footnotes

Acknowledgements

I thank Editors Calasanti and Roscigno, as well as the Social Currents reviewers, for their constructive feedback. An earlier version of this research was presented at the 2014 meeting of the American Sociological Association in San Francisco.

Author’s Note

Data for this study come from the Portraits of American Life Study, a publicly available dataset available at

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Batson

C. Daniel

Powell

Adam A.

2003. “Altruism and Prosocial Behavior.” in Handbook of Psychology, Vol. 5, pp. 463-484: Personality and Social Psychology, edited by Millon

Theodore

Lerner

Melvin J.

Hoboken, NJ: John Wiley.

Baumeister

Roy F.

Bauer

Isabelle M.

Lloyd

Stuart A.

2010. “Choice, Free Will, and Religion.” Psychology of Religion and Spirituality 2(2):67–82.

Bengtson

Vern L.

2013. Families and Faith: How Religion Is Passed Down across Generations. New York: Oxford University Press.

Ben-Ner

Avner

McCall

Brian P.

Stephane

Massoud

Wang

Hua

. 2009. “Identity and In-group/Out-group Differentiation in Work and Giving Behaviors: Experimental Evidence.” Journal of Economic Behavior & Organization 72(1):153–70.

Bennett

Elizabeth A.

Cordner

Alissa

Klein

Peter T.

Savell

Stephanie

Baiocchi

Gianpaolo

. 2013. “Disavowing Politics: Civic Engagement in an Era of Political Skepticism.” American Journal of Sociology 119(2):518–48.

Beyerlein

Kraig

Hipp

John R.

2006. “From Pews to Participation: The Effect of Congregation Activity and Context on Bridging Civic Engagement.” Social Problems 53(1):97–117.

Borgonovi

Francesca

. 2008. “Divided We Stand, United We Fall: Religious Pluralism, Giving, and Volunteering.” American Sociological Review 73(1):105–28.

Brandt

Mark J.

Reyna

Christine

Chambers

John R.

Crawford

Jarret T.

Wetherell

Geoffrey

. 2014. “The Ideological-conflict Hypothesis: Intolerance among Both Liberals and Conservatives.” Current Directions in Psychological Science 23(1):27–34.

Campbell

Karen E.

Lee

Barrett A.

1990. “Gender Differences in Urban Neighboring.” Sociological Quarterly 31(4):495–512.

10.

Cnaan

Ram A.

2002. The Invisible Caring Hand: American Congregations and the Provision of Welfare. New York: New York University Press.

11.

Coyne

Jerry A.

2012. “Science, Religion, and Society: The Problem of Evolution in America.” Evolution 66(8):2654–63.

12.

Davis

Nancy J.

Robinson

Robert V.

1996a. “Are the Rumors of War Exaggerated? Religious Orthodoxy and Moral Progressivism in America.” American Journal of Sociology 102(3):756–87.

13.

Davis

Nancy J.

Robinson

Robert V.

1996b. “Religious Orthodoxy in American Society: The Myth of a Monolithic Camp.” Journal for the Scientific Study of Religion 35(3):229–45.

14.

DiPrete

Thomas A.

Gelman

Andrew

McCormick

Tyler

Teitler

Julien

Zheng

Tian

. 2011. “Segregation in Social Networks Based on Acquaintanceship and Trust.” American Journal of Sociology 116(4):1234–83.

15.

Driskell

Robyn L.

Lyon

Larry

Embry

Elizabeth

. 2008. “Civic Engagement and Religious Activities: Examining the Influence of Religious Tradition and Participation.” Sociological Spectrum 28(5):578–601.

16.

Edgell

Penny

. 2006. Religion and Family in a Changing Society. Princeton, NJ: Princeton University Press.

17.

Edgell

Penny

Tranby

Eric P.

Mather

Darin M.

2013. “Profiles of Anticipated Support: Religion’s Place in the Composition of Americans’ Emotional Support Networks.” Journal for the Scientific Study of Religion 52(2):293–308.

18.

Einolf

Christopher J.

2011. “The Link between Religion and Helping Others: The Role of Values, Ideas, and Language.” Sociology of Religion 72(4):435–55.

19.

Ellison

Christopher G.

1992. “Are Religious People Nice People? Evidence from the National Survey of Black Americans.” Social Forces 71(2):411–30.

20.

Ellison

Christopher G.

George

Linda K.

1994. “Religious Involvement, Social Ties, and Social Support in a Southeastern Community.” Journal for the Scientific Study of Religion 33(1):46–61.

21.

Ellison

Christopher G.

Levin

Jeffrey S.

1998. “The Religion-health Connection: Evidence, Theory, and Future Directions.” Health Education & Behavior 25(6):700–20.

22.

Emerson

Michael O.

Sikkink

David

James

Adele D.

2010. “The Panel Study on American Religion and Ethnicity: Background, Methods, and Selected Results.” Journal for the Scientific Study of Religion 49(1):162–71.

23.

Fershtman

Chaim

Gneezy

Uri

Verboven

Frank

. 2005. “Discrimination and Nepotism: The Efficiency of the Anonymity Rule.” The Journal of Legal Studies 34(2):371–96.

24.

Fischer

Claude S.

1982. To Dwell among Friends: Personal Networks in Town and City. Chicago, IL: University of Chicago Press.

25.

Fischer

Claude S.

Matson

Greggor

. 2009. “Is America Fragmenting?” Annual Review of Sociology 35:435–55.

26.

Galen

Luke W.

2012a. “Does Religious Belief Promote Prosociality? A Critical Examination.” Psychological Bulletin 138(5):876–906.

27.

Galen

Luke W.

2012b. “The Complex and Elusive Nature of Religious Prosociality: Reply to Myers (2012) and Saroglou (2012).” Psychological Bulletin 138(5):918–23.

28.

Hill

Jonathan P.

Vaidyanathan

Brandon

. 2011. “Substitution or Symbiosis? Assessing the Relationship between Religious and Secular Giving.” Social Forces 90(1):157–80.

29.

Hitchens

Christopher

. 2007. God is Not Great: How Religion Poisons Everything. New York, NY: Twelve Books.

30.

Hodgkinson

Virginia A.

Weitzman

Murray S.

1996. Giving and Volunteering in the United States: Findings from a National Survey. Washington, DC: Independent Sector.

31.

House

James S.

1981. Work Stress and Social Support. Reading, MA: Addison-Wesley Longman.

32.

Hunter

James Davison

. 1991. Culture Wars: The Struggle to Define America. New York: Basic Books.

33.

Johnson

Megan K.

Rowatt

Wade C.

LaBouff

Jordan

. 2010. “Priming Christian Religious Concepts Increases Racial Prejudice.” Social Psychological & Personality Science 1(2):119–26.

34.

Koole

Sander L.

McCullough

Michael E.

Kuhl

Julius

Roelofsma

Peter H. M. P.

2010. “Why Religion’s Burdens Are Light: From Religiosity to Implicit Self-regulation.” Personality and Social Psychology Review 14(1):95–107.

35.

LaBouff

Jordan P.

Rowatt

Wade C.

Johnson

Megan K.

Finkle

Callie

. 2012. “Differences in Attitudes toward Outgroups in Religious and Nonreligious Contexts in a Multinational Sample: A Situational Context Priming Study.” International Journal for the Psychology of Religion 22(1):1–9.

36.

Layman

Geoffrey

. 2001. The Great Divide: Religious and Cultural Conflict in American Party Politics. New York: Columbia University Press.

37.

Lewis

Valerie A.

MacGregor

Carol A.

Putnam

Robert D.

2013. “Religion, Networks, and Neighborliness: The Impact of Religious Social Networks on Civic Engagement.” Social Science Research 42(2):331–46.

38.

Marsden

Peter V.

1988. “Homogeneity in Confiding Relations.” Social Networks 10(1):57–76.

39.

McClure

Jennifer M.

2013. “Sources of Social Support: Examining Congregational Involvement, Private Devotional Activities, and Congregational Context.” Journal for the Scientific Study of Religion 52(4):698–712.

40.

McCullough

Michael E.

Willoughby

Brian L. B.

2009. “Religion, Self-regulation, and Self-control: Associations, Explanations, and Implications.” Psychological Bulletin 135(1):69–93.

41.

Monsma

Stephen V.

2007. “Religion and Philanthropic Giving and Volunteering: Building Blocks for Civic Responsibility.” Interdisciplinary Journal of Research on Religion 3(1):2–28.

42.

Pearce

Lisa D.

Axinn

William G.

1998. “The Impact of Family Religious Life on the Quality of Mother-child Relations.” American Sociological Review 63(6):810–28.

43.

Phan

Mai B.

Blumer

Nadine

Demaiter

Erin I.

2009. “Helping Hands: Neighborhood Diversity, Deprivation, and Reciprocity of Support in Non-kin Networks.” Journal of Social and Personal Relationships 26(6–7):899–918.

44.

Pichon

Isabelle

Boccato

Giulio

Saroglou

Vassilis

. 2007. “Nonconscious Influences of Religion on Prosociality: A Priming Study.” European Journal of Social Psychology 37(5):1032–45.

45.

Putnam

Robert D.

Campbell

David E.

2010. American Grace: How Religion Divides and Unites Us. New York: Simon & Schuster.

46.

Rabe-Hesketh

Skrondal

Anders

. 2008. Multilevel and Longitudinal Modeling Using Stata. College Station, TX: Stata Press.

47.

Rainie

Lee

Wellman

Barry

. 2012. Networked: The New Social Operating System. Boston: MIT Press.

48.

Randolph-Seng

Brandon

Nielsen

Michael E.

2007. “Honesty: One Effect of Primed Religious Representations.” International Journal for the Psychology of Religion 17(4):303–15.

49.

Regnerus

Mark D.

Smith

Christian

Sikkink

David

. 1998. “Who Gives to the Poor? The Influence of Religious Tradition and Political Location on the Personal Generosity of Americans toward the Poor.” Journal for the Scientific Study of Religion 37(3):481–93.

50.

Saroglou

Vassilis

. 2006. “Religion’s Role in Prosocial Behavior: Myth or Reality?” Psychology of Religion Newsletter 31(2):1–8.

51.

Saroglou

Vassilis

. 2012. “Is Religion Not Prosocial at All? Comment on Galen (2012).” Psychological Bulletin 138(5):907–12.

52.

Schwadel

Philip

. 2005. “Individual, Congregational, and Denominational Effects on Church Members’ Civic Participation.” Journal for the Scientific Study of Religion 44(2):159–71.

53.

Shariff

Azim F.

Norenzayan

Ara

. 2007. “God Is Watching You Priming God Concepts Increases Prosocial Behavior in an Anonymous Economic Game.” Psychological Science 18(9):803–809.

54.

Shumaker

Sally A.

Brownell

Arlene

. 1984. “Toward a Theory of Social Support: Closing Conceptual Gaps.” Journal of Social Issues 40(4):11–36.

55.

Silverstein

Merril

Conroy

Stephen J.

Wang

Haitao

Giarrusso

Roseann

Bengtson

Vern L.

2002. “Reciprocity in Parent-child Relations over the Adult Life Course.” The Journals of Gerontology Series B: Psychological Sciences and Social Sciences 57(1):S3–13.

56.

Sosis

Richard

Ruffle

Bradley J.

2003. “Religious Ritual and Cooperation: Testing for a Relationship on Israeli Religious and Secular Kibbutzim.” Current Anthropology 44(5):713–22.

57.

Stark

Rodney

. 2012. America’s Blessings: How Religion Benefits Everyone, Including Atheists. Conshohocken, PA: Templeton Press.

58.

Starks

Brian

Robinson

Robert V.

2007. “Moral Cosmology, Religion, and Adult Values for Children.” Journal for the Scientific Study of Religion 46(1):17–35.

59.

Steensland

Brian

Robinson

Lynn D.

Wilcox

W. Bradford

Park

Jerry Z.

Regnerus

Mark D.

Woodberry

Robert D.

2000. “The Measure of American Religion: Toward Improving the State of the Art.” Social Forces 79(1):291–318.

60.

Taylor

Robert Joseph

Chatters

Linda M.

1988. “Church Members as a Source of Informal Social Support.” Review of Religious Research 30(2):193–203.

61.

Terrizzi

John A.

Jr. Shook

Natalie J.

Ventis

W. Larry

. 2010. “Disgust: A Predictor of Social Conservatism and Prejudicial Attitudes toward Homosexuals.” Personality and Individual Differences 49(6):587–92.

62.

Vargas

Nicholas

Loveland

Matthew T.

2011. “Befriending the ‘Other’: Patterns of Social Ties between the Religious and Non-religious.” Sociological Perspectives 54(4):713–31.

63.

Wellman

Barry

Wortley

Scot

. 1990. “Different Strokes from Different Folks: Community Ties and Social Support.” American Journal of Sociology 96(3):558–88.

64.

Wright

Bradley R. E.

2010. Christians Are Hate-filled Hypocrites . . . and Other Lies You’ve Been Told. Minneapolis, MN: Bethany House.

65.

Wuthnow

Robert

. 1989. The Restructuring of American Religion: Society and Faith since World War II. Princeton, NJ: Princeton University Press.

66.

Wuthnow

Robert

. 1991. Acts of Compassion: Caring for Others and Helping Ourselves. Princeton, NJ: Princeton University Press.

67.

Yancey

George

. 2010. “Who Has Religious Prejudice? Differing Sources of Anti-religious Animosity in the United States.” Review of Religious Research 52(2):159–71.