The Doctorate Granting Institution of the City University of New York 365 Fifth Avenue, New York City
The Graduate Center
Doctoral Programs
Other Programs& Research Centers About the Graduate Center Prospective Students Current Students
Faculty Web
Student Web
Mina Rees Library
Administrative Offices
Home Overview
  About Our Faculty
  Directory of Doctoral Faculty
  Recently Appointed GC Faculty
  Faculty Books
  Faculty Activities
  Faculty Research

  For Faculty
  Office of the Provost
  Faculty Development Program
  365 Fifth Newsletter
  Mina Rees Library
  Information Technology
  Academic Calendar
  Building Access
  Faculty Web
  Virtual Bookstore
  Dining Commons
  Sexual Harrassment Policy [pdf]

spacer Middle States Accreditation Self-Study
spacer The Graduate Center is accredited by the Middle States Association.

[A .pdf copy of this study is available here. - Adobe Acrobat or Reader required]


Innovations Between Nsri 1990 And Aris 2001
Coverage Of Religious Groups
Religion And Identity
Religion And Ethnicity
Key Findings
Methodological Appendix
About the Authors

American Religious Identification Survey


- Religious Identification Among American Adults
- Religious Institutional Membership
- Religious or Secular Outlook
- Religious Switching
- Marital Status Among Selected Religious Groups
- Mixed Religion Households
- Age and Gender Patterns Among Selected Religious Groups
- Race and Ethnicity Among Selected Religious Groups
- Political Party Preference Among Selected Religious Groups
- State by State Distribution of Selected Religious Groups

1. Religious Identification Among American Adults

The first area of inquiry in ARIS 2001 concerns the response of American adults to the question: "What is your religion, if any?" This question generated more than a hundred different categories of response, which we classified into the sixty-five categories shown in Exhibit 1 below.

In 1990, ninety percent of the adult population identified with one or another religion group. In 2001, such identification has dropped to eighty-one percent.

Where possible, every effort was made to re-create the categories respondents offered to the nearly identical question as in the NSRI 1990 survey.

As is readily apparent from the first Exhibit below, the major changes between the results of the 1990 survey and the current survey are:

a. the proportion of the population that can be classified as Christian has declined from eighty-six in 1990 to seventy-seven percent in 2001;
b. although the number of adults who classify themselves in non-Christian religious groups has increased from about 5.8 million to about 7.7 million, the proportion of non-Christians has increased only by a very small amount - from 3.3 % to about 3.7 %;
c. the greatest increase in absolute as well as in percentage terms has been among those adults who do not subscribe to any religious identification; their number has more than doubled from 14.3 million in 1990 to 29.4 million in 2001; their proportion has grown from just eight percent of the total in 1990 to over fourteen percent in 2001 [note 5];
d. there has also been a substantial increase in the number of adults who refused to reply to the question about their religious preference, from about four million or two percent in 1990 to more than eleven million or over five percent in 2001.

Exhibit 1 provides the most comprehensive profile of religious identification among the U.S. adult population today and compares the current pattern of identification with what the pattern was in 1990 [note 6].

[Click on image for enlarged version.]

[Click on image for enlarged version.]

NOTE: All figures in Exhibit 1 are rounded.

As is evident from Exhibit 1, with respect to religious self-identification, approximately ninety percent of America's adults are clustered in twenty-two groups. Therefore, the remainder of the analysis in this report focuses on the distribution of adults across these twenty-two groups

2. Religious Institutional Membership in Selected Major Religious Groups

Closely akin to religions group identification in the minds of most people is membership in or affiliation with a place of worship. Indeed, in his classic definition of religion, the nineteenth century sociologist Emile Durkheim characterized religions as systems of belief that unite a group of adherents into common modes of worship, which in turn are organize adherents into churches (or synagogues, temples, mosques or whatever else a group may chose to call the place in which a group of kindred spirits come together to celebrate, worship and recognize the commonality of their beliefs) [note 7].

More than half (54%) of the adult population in America reside in a household where either they themselves or someone else belongs to a church, or temple, synagogue or mosque or some other type of place of worship. To be sure, the significance of membership (its importance, its criteria, and even its definition) varies greatly from one denomination or faith to another. This study is not in position to evaluate the meaning or importance of religious institutional membership for particular groups.

On the other hand, given that about eighty percent of adults identify with some religious group, there appears to be a considerable gap between "identification" with a religion and reported "membership" or "belonging" to a an institutional embodiment of that faith community. That difference between religious identification and belonging could well contain the seeds of a potent cultural shift in which religion means something quite different to those who adhere to one from those who see themselves as the institutional custodians of one.

More than thirty years ago, the sociologist Thomas Luckmann anticipated the emergence of an increasingly de-institutionalized form of religious identification in an incisive analysis of modern religious life, The Invisible Religion. In that work he concluded: "The modern sacred cosmos legitimates the retreat of the individual into the 'private sphere' and sanctifies his (or her) subjective autonomy." [note 8]

Luckmann's analysis notwithstanding, aggregated survey data from the General Social Survey 1972-1994 showed a persistence of church membership among a somewhat larger percentage of U.S. adults than found in the current study. Among a nationally representative sample of 1,481 American adults surveyed in by GSS between the early 1970s and the early 1990s, 61% had indicated membership in a church.

The decade of the nineties appears to have been a period in which religious institutional membership slid, underscoring what Luckmann described as the rise of "invisible religion."

Exhibit 2 below describes the varied pattern of religious institutional membership among the twenty-two largest religious groups - including "no religion," which is the choice made by a very large number. Except where otherwise noted, we have limited our analyses to these twenty-two groups, which encompass nearly 190 million adults or nearly 92% of the adult population.

[Click on image for enlarged version.]

As Exhibit 2 illustrates, there are notable differences between various religious groups with respect to the relationship between identification and affiliation. For example, 68% of those identifying themselves as Lutheran report church membership, while only 45% of those who describe themselves as Protestant (without a specific denominational identification) report church membership. Nearly 68% of those identifying with the Assemblies of God report church membership. Church membership is reported by 59% of Catholic adults. About 53% of adults who identify their religion as Jewish or Judaism report temple or synagogue membership. Among those calling themselves Muslim or Islamic, 62% report membership in a mosque.

Perhaps, it will come as no surprise to religious leaders, but nearly 20% of adults who describe themselves as atheist or agnostic also report that either they themselves or someone else in their household is a member of a church, temple, synagogue, mosque or some other religious institution. On the other hand, nearly 40% of respondents who identified with a religion indicated that neither they themselves nor anyone else in their household belongs to a church or some other similar institution. It is this group in particular that best exemplifies the notion of "invisible religion" first proposed by Luckmann.

The obvious difference between the percentage of the total adult population that identifies with one or another religion and the percentage that report living in a household where either they themselves or someone else is a member of an organized religious body draws attention to the difference between identification as a state of heart and mind and affiliation as a social condition.

The difference in the proportions between identification and affiliation in each group draws attention to the possible differences in the value and meaning attached to affiliation within various religious movements. For example, it is instructive to note that among adults identifying themselves as Buddhist, just 28% report affiliation with a temple. Among adults identifying themselves with "native American religion," affiliation with a church or temple or some other religious institution is just 16%.

Differences between the percentages of identification and affiliation also draw attention to differences in meaning associated with religion itself. For some, religious identification may well be a social marker as much as a marker designating a specific set of beliefs. For others, it may be a reflection of a community or family anchor point to one's sense of self. For other still, it may simply be the "gut response" evoked by the question, "What is your religion, if any?" without any wider emotional, social or philosophical ramifications.

This survey made no attempt to define for people what the meaning of any religious identification might be. Rather, it sought to detect what those identifications might mean for those who claim them. The survey went beyond the simple questions of self-labeling and institutional membership to inquire about a number of key questions such as general outlook (weltanschauung) and beliefs with respect to God.

3. Religious or Secular Outlook Among American Adults

Apart from identification with one or another of a wide range of religions, ARIS 2001 sought to determine whether and to what extent American adults consider their outlook on life to be essentially religious or secular.

Detecting people's worldview or outlook with respect to religion is potentially very challenging. Some would argue that it cannot be done at all with the tools of survey research. Yet, much can be gained by asking rather simple questions of a broad and representative spectrum of people. While not much will be learned about any one individual or even a single group, great insights can be gleaned about the mindscape of diversity in the American population as a whole.

To that end, this survey asked respondents the following: "When it comes to your outlook, do you regard yours as Š (1) Secular, (2) Somewhat Secular, (3) Somewhat Religious or (4) Religious?" Respondents were also permitted to indicate they were unsure or a little of both.

Ninety-three percent of survey respondents were able to reply to this question without much difficulty. In all, sixteen percent (16%) described their outlook as secular or somewhat secular, while seventy-five percent (75%) described their outlook as religious or somewhat religious. Just one percent said they were "a little of both" and two percent said they were unsure. Five percent declined to answer the question.

The question yielded the distribution shown below in Exhibit 3, which indicates that at least ten percent of the population clearly and unambiguously considers itself "secular" rather than "religious." Another six percent regard themselves as "somewhat secular."

[Click on image for enlarged version.]

Our interviews on the question of outlook, as our questions on other matters of belief, generated a fair amount of ambivalence, which is reflected in the high proportion of respondents who fall into the category of "somewhat," that is "somewhat secular" and "somewhat religious." Certainty apparently is the possession of only a minority - though, to be sure, a larger minority among the religious than among the secular.

More interesting still are some of the demographic characteristics of the adult population, which seem to be associated with the disposition to be more or less secular, or more or less religious in one's outlook. Exhibits 4, 5 and 6 provide a glimpse at some of those associations.

- Women are more likely than men to describe their outlook as "religious."
- Older Americans are more likely than younger to describe their outlook as "religious."
- Black Americans are least likely to describe themselves as secular, Asian Americans are most likely to do so.

[Click on image for enlarged version.]

[Click on image for enlarged version.]

[Click on image for enlarged version.]

4. Religious Switching Among Selected Religious Groups

More than thirty-three million American adults, about 16% of the total U.S. adult population report that they have changed their religious preference or identification. Perhaps, this phenomenon of "religion switching" is a reflection of a deeper cultural phenomenon in contemporary America. In the early 1990s, the sociologist Wade Clark Roof described the increasingly middle-aged baby boomers as a "generation of seekers." [note 9] However, the 1990s were also a period of great immigration and great economic boom. Therefore, the religious life of the nation has been influenced by social forces that are wider and more varied than simply the aging of the 'boomers.'

As will be seen in the Exhibit below, switching has involved not only the shift of people's spiritual loyalties from one religion to another -- which could reflect some kind of spiritual seeking -- but also, and perhaps more importantly, a dropping out of religion altogether. To be sure, there is no indication in the current data whether the "religious switching" actually occurred in the 1990s or earlier. Surely, for our older respondents the switching very likely had occurred earlier.

Exhibit 7 below describes the patterns of "religion switching" among the twenty-two largest aggregates. As was indicated earlier, taken together these groups constitute about ninety percent of the entire adult population residing in the U.S. currently.

Click here for Exhibit 7

The top three "gainers" in America's vast religious market place appear to be Evangelical Christians, those describing themselves as Non-Denominational Christians and those who profess no religion. Looking at patterns of religious change from this perspective, the evidence points as much to the rejection of faith as to the seeking of faith among American adults. Indeed, among those who previously had no religion, just 5% report current identification with one or another of the major religions.

Some groups such as Mormons and Jehovah's Witnesses appear to attract a large number of converts ("in-switchers"), but also nearly as large a number of apostates ("out-switchers"). It is also interesting to note that Buddhists also fall into this category of what one might call high-turnover religious groups.

5. Marital Status Among Selected Religious Groups

In most people's minds there is a close association between religious belonging and family values, though to be sure that latter concept is often quite vague as to its meaning. For both demographic and sociological reasons, the present study also focused on household structure, marital status and the religious composition of households.

As context for a discussion of the marital status patterns of different religious groups, it should be noted that the U.S. Census reports the following distribution for the marital status of Americans aged fifteen or older.


Married115,580,691 54%
Single, never married58,049,225 27%
Separated4,795,275 2%
Divorced21,365,741 10%
Widowed13,887,524 7%
TOTAL213,678,456 100%
Source: USCensus QT-02 Profile of Selected Social Characteristics: 2000 (American Fact Finder)

Because ARIS 2001 has defined its survey population as "adults 18 or over" its distribution varies slightly from that of the US Census, which recorded marital status information for all people aged fifteen or older. In addition, as the fact box below shows, ARIS also included an additional category for "single, living with partner." It also recorded those who refused to supply marital status information.

ARIS 2001 FACT BOX 2 (Weighted Estimate)

Married122,053,785 59%
Single, never married40,914,395 20%
Single, living with partner11,101,951 5%
Separated3,431,149 2%
Divorced15,005,207 7%
Widowed12,502,674 6%
Refused info2,959,032 1%
TOTAL207,968,192 100%

Exhibit 8 below draws attention to the variations among the different religious groups with regard to household structure.

Click here for Exhibit 8

The data in Exhibit 8 underscore the accuracy of conventional wisdom in the main: those who identify with one or another of the main religious groups are considerably more likely to be married than those who have no religion. Particularly the "no religion" group was far more likely to be either single, never married or single, living with a partner than any other group. Indeed, the "no religion" group shows the lowest incidence of marriage (just 19%) of all twenty-two groups. In sharp contrast, those identifying with the Assemblies of God or Evangelical/Born Again Christians show the highest proportions married, 73% and 74% respectively.

The percent currently divorced or separated varies considerably less, from a low of six percent (Jehovah's Witnesses) to a high of fourteen percent (Pentecostals).

In Exhibit 9 the study looks at the patterns of divorce and separation between 1990-2001 across the twenty-two religious self-identification groups. While this comparison offers no dramatic changes over the past eleven years, it does underscore the constancy of most of the patterns.

Click here for Exhibit 9

6. Mixed Religion Families Among Selected Religious Groups

Much as normative marriage patterns serve as a sociological buttress to traditional religious identification and belonging, they may also mask underlying change. As we noted earlier, ARIS2001 shows substantial shifts toward secularism among a large number of American adults.

Therefore in this section of the report we look at the incidence of marriage across religious lines. We should add that ARIS2001 is the first national survey that has looked at the religious composition of marriage and domestic partners in large enough numbers to be able to make generalizations among different groups. Because of the size of our sample and the nature of our questions, this survey has generated a wealth of data that will require much further mining with regard to issues pertaining to interfaith households.

ARIS2001 found that of all households that contained either a married or domestic partner couple, 22% reported a mixture of religious identification amongst the couple. At the low end there are the Mormon adults who are found in mixed religion families at 12% and such other groups as Baptists, those adhering to the Churches of Christ, Assemblies of God, the Evangelicals and those adhering to the Church of God (all at about 18%). At the high end we find the Episcopalians at 42% and Buddhists at 39% living in mixed religion families. In all, about 28 million American married or otherwise "coupled" adults live in a mixed religion household.

Click here for Exhibit 10

7. Age and Gender Patterns Among Selected Religious Groups

It is difficult to overestimate the importance of age and sex either in the life of the individual or in the life of any group. Personal outlook is often deeply influenced by these two rather obvious personal attributes. The future of a group is also often shaped by the relative distribution of the old and the young and the relative proportions of males and females. Therefore Exhibits 11 and 12 explore these demographic patterns in the current survey, and for comparison purposes in NSRI 1990.

[Click on image for enlarged version.]

[Click on image for enlarged version.]

As in 1990 so too in the current study, the Buddhist and Muslim population appears to have the highest proportion of young adults under age thirty, and the lowest percentage of females. A number of the major Christian groups have aged since 1990, most notably the Catholics, Methodists, and Lutherans. Congregationalist/United Church of Christ and Presbyterian adherents show an older age structure with three times as many over age 65 as under age 35. Baptists also have fewer young adults than they had in 1990. Among Jews the ratio of the over-65 to those under-thirty has shifted from nearly even in 1990 to about 2:1 in the current study. It should be noted, again, that this survey has focused only upon adult adherents. The observations about age structure do not include the children who may be present in the household of adult adherents.

8. Race and Ethnicity Among Selected Religious Groups

Although the ideals faith are supposed unite people across the great chasms carved by race and ethnicity, social scientists have long noted the in a manner of speaking "Sunday morning service is the most segregated hour in America." ARIS2001 addressed the interplay between faith, ethnicity and race by inquiring into each component of those who were surveyed.

Click here for Exhibit 13

Exhibit 13 describes the make-up of each of the twenty-two major religious groups in terms of proportion non-Hispanic White, non-Hispanic Black, Asian or Hispanic or something else. It should be noted that these characterizations were provided by respondents as answers to fairly straight forward objective questions.

- "Would you consider yourself to be White, Black, or of some other race?"
- "Are you of Hispanic origin or background?"

9. Political Party Preference Among Selected Religious Groups

Given the current debates over a wide variety of public policy issues in which religious convictions and principles are thought to be of some consequence, this study sought to determine with generally broad brushstrokes to what extent religious groups might differ with respect to the political party preferences of their adherents. Exhibit 14 below describes that pattern.

Click here for Exhibit 14

To be sure, political party preferences probably fluctuate more than do religious preferences. It is especially difficult to determine from survey data the extent to which political party preferences are influenced by the heat of the most recent elections. Those caveats aside, the data in Exhibit 14 point to some important continuities as well as shifts.

Jews, Muslims, Buddhists and those with no religion continue to have a greater preference for the Democratic party over the Republican - much as they did in 1990. Evangelical or Born Again Christians and Mormons are the most apt to identify as Republicans. Buddhists and those with no religion are most likely to be political independents. In keeping with their theology, Jehovah's Witnesses disavow political involvement.

10. State and Faith

The final section of this report pays due recognition to the fact that America is also the United States - a name which often masks as much diversity as it portrays unity. With respect to religion in particular, states differ considerably in the religious make-up of their populace. That diversity is likely to contribute as much as any other source of social variation to differences in their cultural and political climate.

Despite the growing diversity nationally, some religious groups clearly occupy a dominant demographic position in particular states. For instance, Catholics are the majority of the population in Massachusetts and Maine as are Mormons in Utah and Baptists in Mississippi. Catholics comprise over 40% of Vermont, New Mexico, New York and New Jersey, while Baptists are over 40% in a number of southern states such as South Carolina, Tennessee, North Carolina, Alabama and Georgia.

Historical traces of the Bible belt in the South and an irreligious West are still evident. Those with "no religion" constitute the largest group in Oregon, Washington, Idaho and Wyoming. In contrast, the percentage of adults who adhere to "no religion" is below 10 % in North and South Dakota, the Carolinas, Alabama, Mississippi and Tennessee.

Such religious concentrations might well have significant impact on host of public policy issues as well as on such matters as religious-based philanthropy.

It remains the challenge of further explorations of these and related data to discover the complex ways in which the religious identification patterns of the American populace shapes the culture and fate of the United States.


5 The growth in the "no religion" population appears to be reflecting a patterns that has also been noted widely in England.

6 Barry A. Kosmin & Seymour P. Lachman, One Nation Under God: Religion in Contemporary America (New York: Harmony Books, 1993)

7 Emile Durkheim, The Elementary Forms of Religious Life (New York: Free Press, 1955).

8 Thomas Luckmann, The Invisible Religion (New York: The Macmillan Co., 1967).

9 Wade Clark Roof, A Generation of Seekers: The Spiritual Journeys of the Baby Boom Generation (San Francisco: Harper, 1993)
Search Site
Telephone/Email Search Information Resources Admissions Employment Academic Calendar Home
Building Access | Policies & Procedures | GC Online Services | Outlook Web Access (access your GC Email)
Admissions queries to: | For inquires reqarding this website: Webmaster
The Graduate Center, The City University of New York, 365 Fifth Avenue, New York, NY 10016-4309
General Telephone: 1-212-817-7000, (Toll Free) 1-877-428-6942 more> | Campus Security: ext. 7777
All Contents © 2006 The Graduate Center.

Site Map | About This Site | CUNY Privacy Policy | Content Disclaimers | Copyright Notice | CUNY