Module 5 (Soc 4)

UNIVERSITY OF CALIFORNIA, BERKELEY

2420 BOWDITCH STREET BERKELEY, CA 94720-5670

Press summary March 2013

More Americans Have No Religious Preference: Key Finding from the 2012 General Social Survey

Michael Hout, University of California, Berkeley Claude S. Fischer, University of California, Berkeley

Mark A. Chaves, Duke University

7 March 2013

ABSTRACT. Twenty percent of American adults said in 2012 that they had no religious preference, according to the latest General Social Survey (GSS), a nationally represen- tative survey of American adults. This continues a trend of Americans disavowing a specific religious affiliation that began in the 1950s but has accelerated greatly since 1990. The GSS has asked adults the following question for forty years: “What is your religious preference? Is it Protestant, Catholic, Jewish, some other religion, or no re- ligion?” The percentage answering “no religion” was 18 percent two years earlier in 2010, 14 percent in 2000, and 8 percent in 1990. The upward trend in the “no reli- gion” choice is very broad. While some types of Americans identify with an organized religion less than others, Americans in almost every demographic group increasingly claim “no religion” since the trend began to accelerate in 1990. Preferring no religion is not atheism which is still very rare; in 2012, just 3 percent of Americans said they did not believe in God. Comparing religious origins with current religion we find that while 20 percent of adults currently have no religious preference, only 8 percent were raised without one. The GSS is an especially important source, because it conducts 83 percent of its interviews in person and has an uncommonly high response rate of over 70 percent.

Introduction

The American religious landscape is changing rapidly. Among the biggest changes is the retreat

from identification with organized religions. Once a central identity for adults, this kind of identifi-

cation is far less prevalent than it was twenty or twenty-five years ago. The General Social Survey

(GSS) has been tracking trends in religious preference since 1972. Everyone in a sample represen-

tative of the adult population of the United States is asked “What is your religious preference? Is it

Protestant, Catholic, Jewish, some other religion, or no religion?” In 1972, just 5 percent answered

“no religion”; by 1990, 8 percent did. The percentage preferring no religion has risen sharply since

1990. In the 2012 data (released March 7, 2013), 20 percent of Americans answered “no religion”

— that is an increase of 12 percentage points in 22 years. We find no evidence of a slowdown. The

change from 2010 to 2012 was 2 percentage points, within the margin of error of what we would

expect based on the overall rate of increase from 1990 to 2008.1

0

5

10

15

20

N o

re lig

io n

(% )

1972 1980 1988 1996 2004 2012 Year

Note: Line smoothes the data to highlight main trend; the circles are each survey's estimate. Source: General Social Surveys, 1972-2012.Figure 1. Percentage Preferring No Religion by Year, 1972-2012

Note: Data smoothed using locally estimated regression (loess); trend line shown in blue. Circles show raw data.

Source: General Social Surveys, 1972-2012.

1The GSS has collected data in even-numbered years since 1994, although it was originally an annual survey.

1

The percentage with no religious preference in each survey is shown in Figure 1 along with a

statistical trend line that eliminates fluctuations attributable to survey sampling. The preference for

no religion rose from 5 to 7 percent between 1972 and 1975, then rose barely perceptibly from 7 to

8 percent between 1975 and 1990. Since the early 1990s, the preference for no religion has risen

at a nearly constant rate of 0.6 percentage points per year.

The trend line in the figure is a very good summary of the changes that have occurred. The

biggest departures from the basic trend are in 1991 when the observed percentage fell almost two

percentage points below the trend line and 1998 when it was about a point-and-a-half above the

trend line. There is no evidence here or elsewhere that the trend has slowed; nonetheless we offer

no predictions about the likelihood of changes in the future.

Trends in subpopulations

Preferring no religion is a very widespread trend, evident in most major subpopulations. We show

this for gender, race, age, education, region, and political views in Table 1.

Women are generally more religious than men, and that generalization holds here. Women in

2012 were less likely than men to prefer no religion — 16 percent compared to 24 percent. Both

groups changed significantly compared to 1990. Women’s percentage preferring no religion was

10 percentage points higher than that of women in 1990; men’s was 14 percentage points higher

than in 1990.

Whites are generally less religious than either African Americans or Mexican Americans.2

That generalization holds with respect to having a religious preference. In 2012, 21 percent of

whites, 17 percent of African Americans, and 14 percent of Mexican Americans had no religious

preference. The rate of change in religious preference differed slightly across groups, but the

differences in rate of change are not statistically significant. That is, each group’s change is within

the margin of error of the overall change which was 12 percentage points.

Younger Americans are much less likely to state a religious preference than are their elders.

Among 18-to-24 year olds, 32 percent prefer no religion; among people 75 years old and over,

only 7 percent prefer no religion. The biggest gap in the age pattern is the ten percentage-point

2We would prefer to base our calculations on a broader aggregation of the hispanic population but it is not possible

at this time.

2

difference between 25-to-34 year olds, 29 percent of whom prefer no religion, and 35-to-44 year

olds, 19 percent of whom prefer no religion. The younger age groups are changing significantly

faster than older groups, as we see when we compare the rate of change between 1990 and 2012.

Having no religious preference increased 22 points among 18-24 year olds but less than 10 points

among people 55 years old and over. We suspect that these age differences will not diminish

as the people in them age. Instead we see them as persisting generational differences that are

likely to characterize these collections of people throughout their life course (see our earlier work

listed below under “references” for statistical calculations that demonstrate the degree to which

age patterns observed through 2000 reflect generational differences rather than the effects of aging

per se).

The preference for no religion varies modestly across educational levels from 16 percent of

high school dropouts to 24 percent of people with advanced degrees. The differences by education

were larger in 2012 than in 1990 when there was no significant pattern by education.

Regional variation in stating no religious preference is both large and growing. In 2012, people

in the Mountain and Pacific regions as well as those in Northeastern states more often answered

no religion than did people in the Midwest, while the southerners expressed the most religious

attachment. The range was from 28 percent in the Mountain states to 15 percent in the South.

In contrast, two decades ago, respondents in the Pacific states stood out as substantially different

from other Americans; 15 percent had no religious preference in 1990 compared with insignificant

variation around an average of 6 percent in the rest of the country. The geographic spread of

non-affiliation is little noted and not well understood in academic research on the issue.

Political liberals are not only significantly more likely to have no religious affiliation than other

Americans, their distinctiveness grew rapidly after 1990. Forty percent of liberals expressed no

religious preference — twice the national rate of 20 percent and 25 percentage points higher than

in 1990. Political conservatives, on the other hand, have registered only the slightest drift away

from organized religion, increasing from 5 percent to 9 percent preferring no religion. Political

moderates are in between with respect to religious preference just as they are with respect to

politics. Hout and Fischer wrote extensively about this trend in their 2002 article in the American

Sociological Review, pointing to politics and generational change as the keys to understanding the

trend as it stood in 2000. We expressed the view that a growing alliance between the leadership of

3

conservative religious denominations and politicians promoting a conservative social agenda was

pushing political liberals from conservative denominations away from organized religion. We plan

to extend our work in light of the new evidence that the trend continues.

Is religious belief declining too?

The decline in affiliation invites speculation that religious belief is also declining. The GSS data

on religious beliefs suggests otherwise. The clearest, most direct summary of religious belief is

belief in God. Table 2 shows the trend.

Because people make many subtle distinctions in belief, the GSS generally avoids a simple yes-

or-no question about believing in God and uses instead one that dates from the 1960s. People are

asked to read six responses from a printed card and pick the one that comes closest to expressing

what they believe about God. The left-hand side of Table 2 shows the options in the exact wording

the GSS uses for the survey. In 2012 only 3 percent of Americans chose the first answer “I do

not believe in God.” That was an insignificant change from the 2 percent who chose that response

in 1991 (the question was not asked in the 1990 GSS). Nor is there a significant change in the

next, quite skeptical response “I don’t know if there is a God and I don’t believe there is any way

to find out.” More Americans now believe in a higher power than was the case in 1991 — 12

percent in 2012 compared with 7 percent in 1991. That increase is statistically significant. The

next two responses express belief although they place conditions on the statement of belief. They

did not change significantly from 1991 to 2012. Last on the card is the response “I know God

really exists and I have no doubts about it.” A majority of Americans in 2012 as in 1991 chose this

unambiguous statement of faith, although the decrease from 64 percent to 59 percent is statistically

significant.

In 1965, according to our calculations from the original data records from a NORC survey con-

ducted that year, 1 percent said they did not believe in God, another 2 percent chose the skeptical

second response, and 77 percent said they believed in God without doubt.

In short, the certainty of believing in God decreased more between 1965 and 1991 than since,

while preference for no religion barely changed from 1965 to 1990, then almost tripled since

1991. This asymmetrical timing of changes indicates that the connection between faith in God and

4

identifying with an organized faith, if there is one, is far from simple. Unchurched believers still

far outnumbered completely secular people in 2012.

Changes among denominations

We can study the implications of these changes for religious denominations in several ways, but

one of the most accessible ways is to compare peoples answers about their current religion to their

answers about what religion (if any) they were raised in. Those comparisons are in Table 3.3

The GSS asks people who expressed a religious preference “What specific denomination is

that, if any?” People have named over 300 specific denominations in reply. Scholars classify those

responses in various ways. Here we distinguish conservative Christian responses like Baptist,

Christian, Pentecostal, and Church of God in Christ from other Protestant denominations such

as Episcopal, Methodist, Lutheran, and Presbyterian among the Protestants; we also distinguish

Catholics and Jews from other religions. The same classification applies to religious origins and

current religions.

About one-third of Americans were conservative Christians in 2012, and about one-third were

raised in a conservative Christian tradition. There is lots of flux behind that unchanging number

as new converts to conservative Christian churches offset defections by others, resulting in little

change in the size of the conservative Christian segment within the generation.

The same is more or less true of the mainline Protestant denominations. While 19 percent of

adults in 2012 were mainline Protestants 21 percent were raised in that tradition.

The Catholic church experienced the greatest net exit within the generation. One-fourth of

American adults were Catholic in 2012. If we were to compare that share with the current religion

of Americans in the past, we would think there had been little change. Throughout the last fifty

years, a consistent 25 percent (plus or minus one percentage point) of Americans were Catholic.

Lurking below the surface of that constancy is very important change. All else being equal, the

Catholic share of the U.S. population should be rising, not staying constant. For almost two gener-

3Simply comparing the distribution of Americans across religious denominations in 1990 with the distribution in

2012 can be very misleading. Demographic changes due to differences in fertility, mortality, and immigration can alter

the distributions over time even when there is no underlying religious change. Comparing a sample of people over

their lifetime, as we do here, is a more direct assessment of the changing religious preferences.

5

ations Catholics had the demographic advantages of higher fertility and, just as fertility dropped,

higher immigration began adding more Catholics to the population. These population fundamen-

tals predict that one-third of Americans would be Catholic, all else being equal. And we see that in

the religious origins in Table 3. In 2012, 35 percent of American adults had been raised Catholic.

In the light of this information, we can see that the 24 percent who were currently Catholic in

2012 actually represented a serious loss of 11 percentage points (or about one-third of the pool of

potential members) for the Catholic Church in America.

The Jewish population is very small. But the current share of 1.5 percent matched the share of

origins.

It is hard to say much about the “other religion” category because it combines a very heteroge-

neous collection of faiths even though all together they do not comprise a very large share of the

total.

Finally, while 20 percent of people currently had no religious preference in 2012, only 8 percent

had been raised without a religious preference. The summary data in Table 3 suggest a direct flow

from Catholic origins to no current religious preference. Reality is, of course, more complicated.

Our review of the detailed flows from each origin to each current preference shows that every origin

contributes something to the rise of no preference. About 16 percent of people raised Christian -

— Protestants and Catholics alike — plus those who describe their religious upbringing as “other

religion” currently prefer no religion. That boosts no religion. But people raised with no religion

also develop religious preferences through life; 40 percent of those raised without religion had a

current religion in 2012. The net is the redistribution we show in the table. The dynamics of these

exchanges among denominations are a field of study unto itself.

Conclusions and Future Work

The historic distancing of Americans from organized religion continues to evolve. More Ameri-

cans than ever profess having no religious preference. Their quarrel appears to be with organized

religion, because conventional religious belief, typified by belief in God, remains very widespread

— 59 percent of Americans believe in God without any doubt. Atheism is barely growing; one

percent of Americans positively did not believe in God in 1965, two percent in 1991, and three

6

percent in 2012. Nor is disbelief fueling the trend toward no preference as beliefs changed much

less during these years of institutional defection (1990-2012) than between the 1960s to the 1990s

when religious preferences changed little (and differential birth rates explained the changes that

did occur).

Over the next few months, Hout and Fischer plan to update their 2002 work on the trend toward

no religion, sorting among the hypotheses that have emerged in the academic literature as well as

assessing their own hypotheses about generational succession and political polarization.

Chaves plans to focus on the congregational context of religious participation. His National

Congregations Study is linked to the GSS in that it is an intensive study of the places GSS respon-

dents worship (when they do worship).

About the Data

The General Social Survey (GSS) is a project of NORC, an independent research unit at the Uni-

versity of Chicago, with principal funding provided by the National Science Foundation. It is a

unique and valuable resource that has tracked the opinions and behaviors of Americans over the

last four decades. The GSS is NORC’s longest running project, and one of its most influential.

Except for U.S. Census data, the GSS is the most frequently analyzed source of information in

the social sciences. More than 20,000research publications of many types are based on the GSS;

and about 400,000 students use the GSS in their classes each year. Since 1985, the GSS has taken

part in the International Social Survey Programme (ISSP), a consortium of social scientists from

49 countries around the world. The ISSP asks an identical battery of questions in all countries; the

U.S. version of these questions is incorporated into the GSS.

The GSS makes high-quality data easily accessible to scholars, students, policy makers, and

others, with minimal cost and waiting. People use the GSS to monitor and explain trends and

constants in attitudes, behaviors, and attributes, to examine the structure and functioning of society

in general as well as the role played by relevant subgroups, to compare the United States to other

societies in order to place American society in comparative perspective and develop cross-national

models of human society.

The GSS has an unusually broad and detailed array of demographic information about each

7

respondent that researchers find valuable in placing subjective responses in social context. These

include basics like sex, age, race, and education, but details about religious denominations (if any),

occupations (if any), marital history, geography, and nativity make the survey particularly useful

for both descriptive and analytical work. GSS questions include such items as national spending

priorities, marijuana use, crime and punishment, race relations, quality of life, and confidence in

institutions. Since 1988, the GSS has also collected data on sexual behavior including number of

sex partners, frequency of intercourse, extramarital relationships, and sex with prostitutes.

Interviews are designed to be conducted in-person, usually at a person’s home. In 2012 83

percent of cases were completed in-person; the rest of the interviews were conducted via telephone.

Over 70 percent of individuals selected for interview participate; this is a much higher response rate

than most surveys achieve. Interviews last an average of 90 minutes. They are usually conducted

between March and June of the year in question. The 2012 survey had a broader interviewing

period; the first interview was March 20th and the last was September 5th; ninety percent of

interviews were between April 1st and August 1st. The 2012 GSS consists of 1,974 completed

interviews and some supplementary information about each case.

The survey was drawn in stages that assure every household in the United States an equal

chance of being included. The probabilities of being selected were updated in 2012 to reflect the

results of the 2010 census. The largest metropolitan areas are drawn into the sample with certainty;

smaller places are included with a probability proportional to the count in the most recent census.

Places are divided into sampling units, addresses are sampled randomly within sampling units, and

one adult is selected randomly at each sampled address that has residents. Initial nonrespondents

are recontacted many times to assure representativeness. Most respondents receive a payment to

participate; $20 is the most common amount.

8

References

Chaves, Mark. 2012. American Religion: Contemporary Trends. Princeton: Princeton University

Press.

Fischer, Claude S., and Michael Hout. 2006. Century of Difference: How America Changed in

the Last One Hundred Years. New York: Russell Sage Foundation.

Hout, Michael, and Claude S. Fischer. 2002. “Why More Americans Have No Religious Prefer-

ence: Politics and Generations.” American Sociological Review 67: 165-190.

9

About the Authors Michael Hout is Natalie Cohen Professor of Sociology and Demography at the University of Cali- fornia, Berkeley. He is one of the principal investigators on the National Science Foundation grant that supports the GSS. His research covers social change quite broadly. He has published almost one hundred scholarly articles and five books. Century of Difference, co-authored with Claude Fischer, exemplifies Hout’s research. Email: [email protected] Phone: (510)643-6874 (Tuesday, Thursday, & Friday; no phone contact Monday or Wednesday)

Claude S. Fischer is professor of sociology at the University of California, Berkeley. Most of his early research focused on urban studies and on social networks, including To Dwell Among Friends (University of Chicago Press, 1982) and Still Connected (Russell Sage Foundation, 2011). In recent years, he has worked on American social history, including his collaboration with Hout Century of Difference, and Made in America (University of Chicago Press, 2010), which analyzes social and cultural change since the colonial era. He is a member of the GSS Board of Overseers. Email: [email protected]

Mark A. Chaves is professor of sociology, religion, and divinity at Duke University. Among other projects, he directs the National Congregations Study (NCS), a wide-ranging survey of nationally representative samples of religious congregations. He is the author of four books including Amer- ican Religion: Contemporary Trends (Princeton University Press, 2011) and Congregations in America (Harvard University Press 2004). The 2012 NCS was conducted by contacting the place of worship of 2012 GSS respondents; analysis of those data is now underway. Chaves served on the GSS Board of Overseers from 2005 to 2012 and chaired it from 2008 to 2012. Email: [email protected]

Acknowledgments and Disclaimer The National Science Foundation (SES-0824618) supported the data collection, coding, and dis- semination. It also provided financial support to Hout, as co-principal investigator on the General Social Survey. We received additional support from the Berkeley Population Center (NICHD R21 HD056581), and the Institute for the Study of Societal Issues at the University of California, Berkeley. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Science Foundation, the National Institute of Child Health and Human Development, or the National Institutes of Health.

10

Table 1. Percentage with no religious affiliated by year and selected characteristics: Adults, United States, 1990 and 2012 Table 1. Percentage with no religious affiliated by year and selected characteristics: Adults, United States, 1990 and 2012 Table 1. Percentage with no religious affiliated by year and selected characteristics: Adults, United States, 1990 and 2012 Table 1. Percentage with no religious affiliated by year and selected characteristics: Adults, United States, 1990 and 2012

Subgroup 1990 2012 Change Total 7.7% 19.7% 12.0%

Women 6.0% 16.4% 10.4% Men 9.9% 23.6% 13.7%

Whites 8.3% 20.7% 12.4% African Americans 3.1% 16.8% 13.7% Mexican Americans 4.4% 14.1% 9.7%

18-24 year olds 9.7% 32.0% 22.3% 25-34 year olds 10.1% 28.6% 18.5% 35-44 year olds 8.0% 18.8% 10.8% 45-54 year olds 5.9% 16.0% 10.1% 55-64 year olds 7.1% 14.7% 7.5% 65-74 year olds 5.2% 11.9% 6.7% 75 year olds and up 3.8% 7.1% 3.3%

Less than high school 9.0% 16.0% 7.0% High school diploma 5.5% 16.7% 11.1% Some college 7.1% 21.0% 13.9% College degree 12.5% 22.9% 10.5% Advanced degree 8.2% 24.2% 16.0%

Northeastern states 7.4% 23.9% 16.5% Midwestern states 6.3% 18.7% 12.4% Southern states 5.6% 14.6% 9.1% Mountain states 4.3% 27.8% 23.5% Pacific states 14.6% 24.9% 10.3%

Political liberals 14.6% 39.6% 25.0% Political moderates 6.1% 17.9% 11.8% Political conservatives 5.0% 9.1% 4.1%

Source: General Social Surveys, 1990 and 2012.Source: General Social Surveys, 1990 and 2012.Source: General Social Surveys, 1990 and 2012.Source: General Social Surveys, 1990 and 2012.

11

Table 2. Beliefs about God by Year: Adults, United States, 1991 and 2012 Table 2. Beliefs about God by Year: Adults, United States, 1991 and 2012 Table 2. Beliefs about God by Year: Adults, United States, 1991 and 2012 Table 2. Beliefs about God by Year: Adults, United States, 1991 and 2012

Which statement comes closest to expressing what you believe about God? 1991 2012 Change

I don't believe in God 2.2% 3.1% 0.9%

I don't know whether there is a God and I don't believe there is any way to find out 4.1% 5.6% 1.5%

I don't believe in a personal God, but I do believe in a Higher Power of some kind 6.6% 11.6% 5.1%

I find myself believing in God some of the time, but not at others 5.3% 4.2% -1.1% While I have doubts, I feel that I do believe in God 18.0% 16.5% -1.5% I know God really exists and I have no doubts about it 63.9% 59.1% -4.8%

Total 100.0% 100.0%

Number of persons 1,331 1,952 Source: General Social Surveys, 1991 and 2012.Source: General Social Surveys, 1991 and 2012.Source: General Social Surveys, 1991 and 2012.Source: General Social Surveys, 1991 and 2012.

12

Table 3. Religious Denomination of Origin and Current Religious Denomination: Adults, United States, 2012 Table 3. Religious Denomination of Origin and Current Religious Denomination: Adults, United States, 2012 Table 3. Religious Denomination of Origin and Current Religious Denomination: Adults, United States, 2012 Table 3. Religious Denomination of Origin and Current Religious Denomination: Adults, United States, 2012

Religious denomination Origin Current Difference

Conservative Christian (e.g., Baptist, Pentecostal, “Christian,” Church of God in Christ) 30.9% 31.5% 0.6%

Mainline Protestant (e.g., Episcopal, Methodist, Lutheran, Presbyterian) 21.2% 19.0% -2.2% Catholic 35.0% 24.2% -10.8% Jewish 1.5% 1.5% 0.0% Other religion 3.4% 4.2% 0.8% No religion 8.1% 19.7% 11.6%

Total 100.0% 100.0%

Number of persons 1,962 1,962 Source: General Social Surveys, 1991 and 2012.Source: General Social Surveys, 1991 and 2012.Source: General Social Surveys, 1991 and 2012.Source: General Social Surveys, 1991 and 2012.

13