Religious Diversity and Change in American Social Networks: How Our Social Connections Shape Religious Beliefs and Behavior
Findings from the American National Social Network Survey
December 15, 2020 |
A new national survey released this past fall explored the character and composition of our immediate social networks. The American National Social Network Survey introduced an egocentric network design that describes the racial, religious, and political dimensions of American social relationships. This extensive battery of questions required respondents to identify members of their core social network—people with whom they “discussed important personal matters and concerns”—and provide information about their personal characteristics and background, such as their education level, gender, age, race and ethnicity, political orientation, and religion.
Based on this unique data, we released a report that found widespread racial segregation among Americans’ personal networks, despite growing racial and ethnic diversity in the US. Now, a new analysis explores the degree to which Americans’ religious networks are composed largely of those with similar beliefs and affiliations or those that are more diverse. It also explores how religious diversity among our close personal relationships serves to structure religious behavior and belief.
Religious Diversity and Personal Connections
The US is undergoing a profound demographic change that is reshaping the composition and character of American religious identity. A large part of this change is due to the rise of religiously unaffiliated Americans who now make up more than one-quarter (26 percent) of the adult population. But non-Christian communities, which include Buddhists, Hindus, Jews, and Muslims, are growing as well. White Christians have experienced pronounced declines over the past several decades, while American Catholics have experienced major shifts in ethnic composition because of immigration from Latin America.
The growing religious diversity in the US is significantly affecting American culture and religious life. Recent research has shown that rates of religious intermarriage are increasing among newly married couples. A report by the Pew Research Center finds that among couples married between 2010 and 2014, nearly four in 10 have a spouse with a different religious background. The religious composition and social context varies considerably by generation. Among young adults (age 18 to 29), the most common religious identity today is none; more than one in three (34 percent) young adults are religiously unaffiliated.
The degree of religious diversity in the US is also reflected in Americans’ core social networks. Few religious Americans count only members of their religious tradition as part of their immediate social circle. But there are important differences across traditions. White evangelical Protestants’ religious networks are 80 percent Protestant, while black Protestants’ networks are similarly composed primarily of other Protestants (82 percent). Sixty-nine percent of white mainline Protestants’ networks are Protestant. Religious homogeneity is weaker among Catholics. Fifty-eight percent of Catholic social networks are Catholic, while half of non-Christian networks are composed of members of non-Christian religions.
Despite evidence of self-sorting by religious affiliation, there is significant diversity among Americans’ religious networks, but this varies considerably across traditions. Two-thirds (67 percent) of black Protestants and more than six in 10 (62 percent) white evangelical Protestants report having social networks that include only other Protestants. Members of other religious traditions report greater levels of diversity in their immediate social circles. Only 46 percent of white mainline Protestants say their social networks include only other Protestants. Among Catholics, 39 percent say their core social network is composed exclusively of other Catholics. Thirty-six percent of Americans who belong to non-Christian religious traditions report that their network includes only people who are similarly members of non-Christian traditions, while 29 percent of unaffiliated Americans have social networks made up entirely of people who are also unaffiliated.
More Americans Have Close Connections with Someone Who Is Religiously Unaffiliated
As more Americans report having no religious affiliation, the percentage of the public who report having a close contact who is unaffiliated has increased. Overall, 42 percent of Americans have a close social connection with someone who is religiously unaffiliated. There are important differences across religious traditions and denominations, but Americans have more close relationships with those who are unaffiliated than ever before. More than one in four (27 percent) white Protestants—including 37 percent of white mainline Protestants and 23 percent of white evangelical Protestants—and about one-quarter (28 percent) of Catholics report having someone who is unaffiliated in their immediate social circle. Only 16 percent of black Protestants have a social network that includes someone who is unaffiliated. Notably, almost eight in 10 (79 percent) of Americans who are unaffiliated have someone who shares their lack of religious identity in their social network.
Roughly a decade and a half earlier, many fewer Americans reported having a religiously unaffiliated member of their core social network. In 2004, only 18 percent of the public reported having a close social connection to someone who is religiously unaffiliated. Again, there were significant differences by religious affiliation. About one in 10 white Protestants (12 percent) and Catholics (13 percent) had an unaffiliated social connection. Notably, over half (56 percent) of unaffiliated Americans in 2004 said they had at least one member of their social network who is also unaffiliated.
The Disappearing Social Benefit of Religious Participation?
Past research has shown that regular participation in worship services is strongly associated with higher rates of sociability, civic involvement, and political engagement. But the social benefits of regular religious participation may be disappearing or more limited than previously thought. Americans who are involved in religious communities do not report having more extensive social networks than those who are not. Americans who attend religious services at least once a week have similarly sized social networks as those who seldom or never attend. Seventeen percent of Americans who attend services once a week or more often report having no close social ties, while an identical number (18 percent) of those who seldom or never attend also report having no immediate social connection. However, Americans who regularly attend religious services are slightly more likely to have social networks with at least six members (29 percent vs. 25 percent).
There are also few differences between Americans of different religious traditions and those who do not claim any religious affiliation. Religiously unaffiliated Americans (15 percent) are not any more likely to report having no close social contacts than are white evangelical Protestants (16 percent), white mainline Protestants (16 percent), Catholics (17 percent), or members of non-Christian religious traditions (17 percent). And Americans who identify with a particular religious tradition do not have larger social networks than those who are unaffiliated do. Interestingly, the social network size of religiously unaffiliated Americans and white evangelical Protestants is nearly indistinguishable.
Social Influence and Religious Beliefs and Behavior
Belief in God
Belief in God has been among the most stable measures of religious belief, behavior, or identity in the US public. Gallup has tracked belief in God for more than 70 years and found significant stability even as the US has experienced social upheaval and demographic shifts and abandoned many traditional cultural values. Nearly nine in 10 (87 percent) Americans report they believe in God when offered a single binary question. Despite the apparent consistency in belief over time, Americans express considerable uncertainty about this fundamental religious belief. A new version of this measure, requiring respondents to express their relative belief or disbelief in God along a 10-point scale, uncovers significant feelings of doubt about God’s existence among those who believe and those who do not.
Overall, more than half (53 percent) of Americans report they believe in God without any doubts at all. Conversely, 6 percent of Americans say they do not believe in God and express no uncertainty in their belief. More than four in 10 (41 percent) Americans express at least some uncertainty about their belief in God. About one in 10 (11 percent) Americans express some degree of disbelief in God but remain at least somewhat unsure. Nineteen percent of Americans are inclined to believe in God but are somewhat less than completely certain in their belief. Eleven percent of Americans report being completely uncertain in their views on God—locating themselves in the exact middle between certain belief and disbelief.
Certainty of Religious Belief Varies Considerably Across Religious Traditions
More than eight in 10 white evangelical Protestants (87 percent) and black Protestants (83 percent) say they are absolutely certain God exists (Figure 3). A majority (59 percent) of Catholics report they have no doubts about their belief in God. Only half (50 percent) of white mainline Protestants express complete certainty in God’s existence. Religiously unaffiliated Americans differ starkly from most major religious traditions in the lack of religious certainty they express. Only 13 percent say they are certain in their belief in God. A significantly larger number (22 percent) say they are completely confident in their disbelief. The majority of religiously unaffiliated Americans express some amount of uncertainty with their belief in God.
However, there are important differences among religiously unaffiliated Americans between those who identify as “atheist,” “agnostic,” and “nothing in particular.” More than six in 10 (61 percent) atheists report being absolutely certain they do not believe in God, while only 10 percent of agnostics and those who are nothing in particular say the same.
Belief in God and Personal Morality
Today, most Americans say it is not necessary for a person to believe in God to be moral and have good values, a remarkable shift in recent years. Close to six in 10 (59 percent) Americans say a belief in God is not a precondition to being moral and having good values. Forty-one percent of the public say a belief in God is essential.
Views about the relationship between being moral and belief in God vary considerably across religious traditions. A majority of black Protestants (65 percent), white evangelical Protestants (61 percent), and Hispanic Catholics (55 percent) say a belief in God is an important part of being moral. Less than half of white Catholics (43 percent), white mainline Protestants (38 percent), members of non-Christian traditions (34 percent), and unaffiliated Americans (13 percent) say it is necessary to believe in God to be moral. Eighty-seven percent of unaffiliated Americans disagree that belief in God is a requirement for moral behavior, including roughly three-quarters (74 percent) who strongly reject this idea.
The Social Dimension of Religious Belief
Religious Americans whose immediate social circle includes members of other faiths, or none at all, express less certainty about their own religious beliefs. Conversely, Americans who have less religiously diverse social networks report having fewer doubts about their belief in God. This pattern is remarkably consistent across religious traditions.
Two-thirds (67 percent) of Catholics whose social network includes only other Catholics say they believe in God without any doubts. Catholics whose social network includes members of other faiths or none at all are less certain of their belief; 55 percent say they are completely certain. Similarly, while a majority (56 percent) of white mainline Protestants with social networks that include only other Protestants say they believe in God without any doubts, less than half (43 percent) of those with diverse networks express this level of confidence in their belief. The gap among white evangelical Protestants is considerably smaller. Eighty-nine percent of white evangelical Protestants whose social network is exclusive to Protestants say they believe in God without any doubts, compared to 81 percent of those whose networks include members of other religious traditions or those who do not belong to any tradition.
Religious Americans who have a close contact who is unaffiliated express much less certainty in their belief in God. Nearly two-thirds (65 percent) of Catholics whose immediate social circle does not include someone who is unaffiliated say they are completely certain in their belief in God. In contrast, 36 percent of Catholics who have at least one member of their inner circle who is unaffiliated express this level of confidence in their belief.
A similar pattern is evident among white mainline Protestants. Those with an unaffiliated member of their network are less likely to say they are certain in their belief in God (55 percent vs. 36 percent). Religious doubts are also more common among white evangelical Protestants with close personal contacts with people who are unaffiliated. Eighty-nine percent of white evangelicals whose core social network includes no unaffiliated members say they are certain in their belief about God, compared to 76 percent who have a close relationship with someone who is unaffiliated.
Perspectives about the relationship between morality and belief in God vary between those with close ties with someone who is religiously unaffiliated, a pattern largely consistent across religious traditions. A majority (63 percent) of white evangelical Protestants whose immediate social circle does not include someone who is unaffiliated say believing in God is a necessary precondition to being moral. In contrast, this view is held by less than half (48 percent) of those who have at least one unaffiliated social connection. Half (50 percent) of Catholics whose core social network does not include someone who is unaffiliated say belief in God is essential for being moral, while only 33 percent of those who have such a connection express this view. Finally, white mainline Protestants who lack a social connection with someone who is unaffiliated are more likely to say believing in God is required to be moral and have good values than those who do (42 percent vs. 27 percent).
Traditionalist vs. Modern Approach to Religious Worship and Belief
Religious Americans are divided over whether their religious tradition should preserve traditional beliefs and practices or adapt to changing societal norms and beliefs. Close to half (47 percent) of Americans who belong to a specific religious tradition say their religion should work to preserve traditional religious practices and beliefs. Thirty-eight percent say it should adjust in certain aspects due to changing circumstances in society, while 14 percent say their religion should fully adopt modern beliefs and practices. Notably, interest in altering religious beliefs and approaches to worship has not changed much over the past decade. In 2008, 44 percent of religious Americans said they would prefer their denomination to preserve traditional practices and beliefs, while 35 percent said it should make some changes. Only 12 percent supported adopting modern beliefs and practices.
There are modest differences between generations. Seniors (age 65 or older) are more likely to say their denomination or tradition should preserve traditional beliefs and practices than are young adults (48 percent vs. 38 percent, respectively). Young adults are about twice as likely as seniors are to say their religion should adopt modern beliefs and practices (17 percent vs. 9 percent).
Religious Americans whose close social contacts reflect their own religious affiliation express much more interest in retaining their traditional beliefs and practices than those with more diverse religious networks do. More than four in 10 (42 percent) Catholics whose core social networks include only other Catholics say the church should preserve traditional beliefs and practices. Less than one-third (31 percent) of Catholics with a more diverse religious social circle express this view. There is a comparable gap among white mainline Protestants. Forty-four percent of white mainline Protestants whose close social ties include only other Protestants say their church or denomination should continue to embrace traditional teachings and practices, a view shared by 32 percent of those with more diverse social connections. White evangelical Protestants with homogeneous religious networks are more inclined to support maintaining traditional approaches to their religion than are those with more diverse social ties (76 percent vs. 69 percent).
Discussing Religion with Family Members and Friends
Americans are divided over whether it is better to discuss religious beliefs and ideas with family members and friends who do not share the same perspective. Close to half (46 percent) of Americans say it is better to talk about religious differences to find common ground, while more than half (52 percent) say it is better to avoid talking about these differences because doing so usually makes things worse.
Having a diverse religious network does not increase the appetite for having a religiously focused discussion. Americans with a uniform religious network are more likely than are those with a diverse network to say it is better to discuss religious differences with friends and family (52 percent vs. 41 percent). However, views vary dramatically by religious tradition. White mainline Protestants with diverse social networks are much less likely to endorse interreligious discussion than are those with uniform social circles (42 percent vs. 62 percent). Notably, white evangelical Protestants and Catholics express nearly identical opinions regardless of the composition of their social network.
Invitations to Attend Worship Service
Being asked to attend a worship service is not a common experience for most Americans. Less than half (47 percent) of Americans report being invited to attend a religious service in the past 12 months, and only 23 percent say they received an invitation in the past month. A majority (54 percent) of Americans say they have not been asked to participate in a religious service in the past 12 months or have never been asked.
There is a strong relationship between being asked to participate in religious services and worship attendance. Two-thirds (67 percent) of Americans who have been asked to attend services in the past week report attending at least weekly or more often. In contrast, only 14 percent who have never been asked to attend report they go at least weekly. Nearly half (47 percent) of Americans who have never been asked say they never go.
Americans with religiously diverse social networks report they are less likely to receive invitations to attend worship services. Only 18 percent of Americans who have a diverse religious social circle say they were encouraged to attend a religious service in the past month, while roughly one-third (32 percent) of those whose religious networks are uniform report receiving such an invitation. However, there is one exception to this general pattern. Religiously unaffiliated Americans with diverse networks are more likely to be asked to participate in a worship service. Less than half (46 percent) of unaffiliated Americans with social networks that include only other unaffiliated people say they have ever received an invitation to attend, while 71 percent of those with diverse networks report having this experience.
The invitation discrepancy between those with diverse and homogeneous religious networks may explain the sharp division in patterns of religious attendance. Americans with homogeneous social networks are much more likely to attend religious services regularly. Only about one in five (21 percent) Americans with diverse religious networks report attending services weekly or more often, compared to 37 percent of those with uniform social network. A majority (58 percent) of those with diverse social networks say they seldom or never attend services.
Religious Discrimination and Being a Religious Minority in the US
Christians are much more likely to say it has become more difficult to live in the US as a Christian than to say it has become easier, but most say things have not really changed much. White evangelical Protestants are unique among Christians in believing it is more difficult to be a Christian in the US today. A majority (54 percent) of white evangelical Protestants say it has become more difficult, while only 6 percent say it has gotten easier. A majority of white mainline Protestants (57 percent), white Catholics (57 percent), black Protestants (61 percent), and Hispanic Catholics (68 percent) say it has not gotten easier or has gotten more difficult to live in the US as a Christian.
Perceptions about life in the US as a Christians are filtered through ideology. Conservative Christians are far more likely to say it has become more difficult to be Christian in recent years, while most liberal Christians disagree. A majority (55 percent) of Christians who identify as politically conservative say it has become more difficult, while only 30 percent of Christians identifying as liberal agree.
Religious Groups Experiencing Discrimination
American perceptions about the nature of discrimination experienced by religious communities vary widely. More than seven in 10 (71 percent) Americans say Muslims in the US experience a lot of discrimination. About half (49 percent) the public say Jews face a lot of discrimination, while fewer than one in three Americans say Christians (29 percent), evangelical Christians (28 percent), Mormons (27 percent), and atheists (24 percent) experience a considerable amount of discrimination in the US.
Across the religious spectrum, there is substantial agreement about the challenges Muslims face in American society. A majority of every major religious group say Muslims experience a lot of discrimination in the US, including white evangelical Protestants (55 percent), white mainline Protestants (65 percent), white Catholics (66 percent), Hispanic Catholics (75 percent), black Protestants (78 percent), and the religiously unaffiliated (81 percent).
Similarly, the public generally agrees on views about atheists. Few Americans say there is a lot of discrimination against atheists in the country today. Less than one in four white evangelical Protestants (14 percent), white Catholics (17 percent), white mainline Protestants (22 percent), and Hispanic Catholics (22 percent) say atheists face a lot of discrimination. Twenty-seven percent of black Protestants say the same. Notably, less than one-third of religiously unaffiliated Americans, including 40 percent of self-identified atheists, say there is a lot of discrimination against atheists in the US.
Regarding discrimination Christians experience, there is considerable disagreement. More than half (51 percent) of white evangelical Protestants believe Christians experience a lot of discrimination in the US, a view shared by 38 percent of black Protestants and about one-third of white Catholics (33 percent) and white mainline Protestants (29 percent). Only about one in eight Hispanic Catholics (15 percent), members of non-Christian religions (13 percent), and unaffiliated Americans (13 percent) share this view.
Who Is a Religious Minority?
A sizable number of Americans say they think of themselves as a member of a minority because of their religious beliefs. Twenty-seven percent of Americans, including many Christians who are a numerical majority in the US, say they feel like a minority because of their religious beliefs.
Americans who belong to non-Christian religious traditions, such as Buddhism, Hinduism, Islam, and Judaism, are most likely to see themselves as a member of a minority community. Seventy-one percent of Americans who belong to these traditions say they consider themselves a minority. In contrast, no more than one in four black Protestants (23 percent), Catholics (19 percent), and white mainline Protestants (16 percent) say they are a religious minority. More than one in three white evangelical Protestants (37 percent) and atheists (35 percent) think of themselves as a minority community.
However, across denominations, religious Americans who attend services regularly are more likely to feel like a minority. For instance, close to half (43 percent) of white evangelical Protestants who attend services at least once a week say they feel like a member of a religious minority group, compared to 21 percent of white evangelical Protestants who attend services a few times a year or less often.
American Religious Identity and Culture
Most Americans believe either the US was a Christian nation in the past or the US still is a Christian nation today, even as a growing number believe the country’s religious identity has changed. Thirty-seven percent of Americans believe the US has always and continues to be today a Christian nation. More than four in 10 (42 percent) say the US was a Christian nation but is not any longer. About one in five (19 percent) say the US has never been a Christian nation. A decade earlier slightly more—42 percent—Americans said the US was still currently a Christian nation.
Perceptions vary among religious groups about the country’s religious identity. However, there is general agreement that the US is not a Christian nation. Less than half of white Catholics (46 percent), white evangelical Protestants (40 percent), white mainline Protestants (39 percent), Hispanic Catholics (38 percent), and black Protestants (36 percent) say the US is a Christian nation. Less than one in three (30 percent) religiously unaffiliated Americans agree. Notably, about half (49 percent) of Americans who belong to non-Christian religious traditions say the US is still a Christian nation. White evangelical Protestants are unique to the extent that they believe the country’s Christian identity is a thing of the past. A majority (52 percent) of white evangelical Protestants say America was a Christian nation in the past but not today. Conversely, atheists stand out for their belief that the US has never been a Christian country; 41 percent express this view.
Americans are divided over the type of country the US should be: a country with an essential culture and values system or one that evolves as new people arrive. Most (53 percent) Americans say the US culture and values should adapt as immigrants come, while close to half (45 percent) say the US should hold on to its culture and identity that immigrants should adopt.
The question of whether the US should have an essential culture divides Americans by race and religious affiliation. White Christians are most likely to believe the US culture and values should be immutable that immigrants should take on. Nearly two-thirds (64 percent) of white evangelical Protestants and almost six in 10 white mainline Protestants (58 percent) and white Catholics (57 percent) say the US ought to be a country with an identifiable culture and have a set of values that are adopted by immigrants coming over. Less than half of Hispanic Catholics (40 percent), black Protestants (35 percent), members of non-Christian religious traditions (33 percent), and unaffiliated Americans (32 percent) agree.
Americans with greater religious diversity in their immediate social circle are much more likely to believe the US should be a country with a dynamic and evolving culture. A majority (57 percent) of Americans with religiously diverse social networks say the US should be a country made up of many cultures and values that respond to people coming here. Less than half (47 percent) of Americans with uniform religious networks express this view.
The Survey Center on American Life of the American Enterprise Institute is grateful to the John S. and James L. Knight Foundation for its generous support of the American National Social Network Survey.
 Pew Research Center, “In U.S., Decline of Christianity Continues at Rapid Pace,” October 17, 2020, https://www.pewforum.org/2019/10/17/in-u-s-decline-of-christianity-continues-at-rapid-pace/.
 Caryle Murphy, “Interfaith Marriage Is Common in U.S., Particularly Among the Recently Wed,” Pew Research Center, June 2, 2015, https://www.pewresearch.org/fact-tank/2015/06/02/interfaith-marriage/.
 Due to limitations of sample size, Buddhist, Hindu, Jewish, and Muslim respondents were collapsed into a non-Christian category. Among any one of these groups, rate of homogeneity could be higher than 50 percent. Differential rates of intermarriage among these groups supports this possibility.
 This analysis excludes Americans without anyone in their social network.
 This analysis relies on a similarly constructed social network battery embedded in the 2004 General Social Survey (GSS). A few important differences are worth noting. First, the GSS social network battery allowed respondents to identify only up to five social contacts, while the current social network battery allows respondents to identify up to seven. Despite this important distinction, there is no evidence that it alters the composition of religious networks. Restricting the social network battery to five in the current social network analysis does not change the results by a significant margin across any religious subgroups.
 The 2004 GSS data did not include a variable to identify evangelical Protestants.
 Daniel A. Cox et al., “A Loneliness Epidemic? How Marriage, Religion, and Mobility Explain the Generation Gap in Loneliness,” American Enterprise Institute, September 26, 2020, https://www.americansurveycenter.org/research/a-loneliness-epidemic-how-marriage-religion-and-mobility-explain-the-generation-gap-in-loneliness/.
 Gallup, “Religion,” https://news.gallup.com/poll/1690/Religion.aspx.
 Daniel A. Cox, “Doubting Disbelievers: A New Approach to Measuring Religious Uncertainty,” American Enterprise Institute, April 1, 2019, https://www.americansurveycenter.org/research/doubting-disbelievers-a-new-approach-to-measuring-religious-uncertainty/.
 This question was piloted in a 2019 survey. The results were included in Cox, “Doubting Disbelievers.”
 Pew Research Center, Religion and Public Life, “U.S. Religious Landscape Survey: Religious Beliefs and Practices,” June 1, 2008, https://www.pewforum.org/2008/06/01/u-s-religious-landscape-survey-religious-beliefs-and-practices/.
The survey was designed and conducted by the American Enterprise Institute. Interviews were conducted among a random sample of 4,067 adults (age 18 and older) living in the United States, including all 50 states and the District of Columbia. Interviews were conducted both online using a self-administered design and by telephone using live interviewers. All interviews were conducted among participants using a probability-based panel designed to be representative of the national US adult population run by NORC at the University of Chicago. Interviewing was conducted between July 31 and August 11, 2020.
Weighting was accomplished in two separate stages. First, panel base weights were calculated for every household based on the probability of selection from the NORC National Frame, the sampling frame that is used to sample housing units for AmeriSpeak. Household level weights were then assigned to each eligible adult in every recruited household. In the second stage, sample demographics were balanced to match target population parameters for gender, age, education, race and Hispanic ethnicity, division (US Census definitions), housing type, and telephone usage. The telephone usage parameter came from an analysis of the National Health Interview Survey. All other weighting parameters were derived from an analysis of the US Census Bureau’s Current Population Survey. The sample weighting was accomplished using an iterative proportional fitting (IFP) process that simultaneously balances the distributions of all variables. Weights were trimmed to prevent individual interviews from having too much influence on the results.
The use of survey weights in statistical analyses ensures that the demographic characteristics of the sample closely approximate the demographic characteristics of the target population. The margin of error for the survey is +/– 2.0 percentage points at the 95 percent level of confidence. The design effect for the survey is 1.83.
 The National Frame is representative of over 97 percent of US households and includes additional coverage of hard-to-survey population segments, such as rural and low-income households, that are underrepresented in other sample frames. AmeriSpeak uses US mail notifications, NORC telephone interviewers, and in-person field interviewers to recruit panel households.