The Decline Religious Identity - Welcome to the Institute for

S i d G r o e n e m a n & G a r y T o b i nInstitute for Jewish & Community Research, San Francisco 2004

The Decline of

Religious Identity in the United States

Sid Groeneman, Ph.D.Gary Tobin, Ph.D.

Design - Scott Hummel

Institute for Jewish & Community ResearchSan Francisco

The Decline of

Religious Identity in the United States

2 0 0 4

Religious Change in Contemporary America ......................................................................................1

Major Findings ..........................................................................................................................................3

The Civic Benefits of Religion ................................................................................................................5

Trends in Measures of Religiosity (Other than Identification) ..........................................................7

Belief in God ..........................................................................................................................................7

Belief in the Efficacy of Religion............................................................................................................7

The Importance of Religion in One's Life ............................................................................................7

Religious Behavior: Church Membership, Church Attendance, and Prayer ........................................7

Religious Self-Identification ..................................................................................................................10

A New Data Source: The Survey of Heritage and Religious Identification ..................................13

How many non-identifiers? ................................................................................................................13

Religious Origins of Non-Identifiers ..................................................................................................14

Demographics of Non-Identifiers ........................................................................................................15

Gender..........................................................................................................................................15

Age................................................................................................................................................17

Region of the Country ..............................................................................................................18

Race ..............................................................................................................................................20

Marital Status and Children ....................................................................................................20

Socio-Economic Status ..............................................................................................................21

Political Orientation ............................................................................................................................21

Discussion: Interpreting Non-Identification ......................................................................................23

Concluding Thoughts ............................................................................................................................27

Footnotes ..................................................................................................................................................29

TABLE OF CONTENTS

Figure 1: Church Membership by Year..........................................................................................8

Figure 2: Past Week Church Attendance by Year ........................................................................8

Figure 3: “Prayer is an important part of my daily life.” ..........................................................9

Figure 4: Largest Religion Segments............................................................................................13

Figure 5: Religious Origins of Current Non-Identifiers ............................................................14

Figure 6: Percent Who are now Non-Identifiers by Religion of Origin ................................15

Figure 7: Non-Identifiers by Gender............................................................................................16

Figure 8: Non-Identifiers by Age..................................................................................................16

Figure 9: Raised in No Religion and No Religion Now by Age..............................................17

Figure 10: Non-Identifiers by Region ............................................................................................18

Figure 11: Non-Identifiers by Geographic Sub-Region ..............................................................19

Figure 12: Non-Identifiers by Race ................................................................................................19

Figure 13: Percent Married by Age and Religious Identification ..............................................20

Figure 14: Non-Identifiers by Education ......................................................................................21

Figure 15: Political Orientation by Religious Identification ......................................................22

Table 1: Religious Characteristics of Non-Identifiers ..............................................................24

Table 2: A Typology of Religious Identifiers and Non-Identifiers ........................................24

F IGURES & TABLES

1

The Decline of Religious Identity in the U.S.

RELIGIOUS CHANGE IN CONTEMPORARY AMERICA

Discussion of the religious landscape is notnew to American discourse. Is religious iden-tity in long-term decline or in the midst ofrejuvenation? As a people, are we becomingmore secular in belief and behavior? What isthe relationship between, on one hand, psy-chological affiliation with a church or denomi-nation,1 and, on the other, our spirituality andreligious practice? What is the public signifi-cance if we are, in fact, moving unequivocallytoward secular thinking and away from reli-gion-based belief and understanding?

Much of the debate about religious change,according to noted observer Wade Clark Roof,turns on the impact of the tumultuous 1960sand 1970s, and the lasting effects of that peri-od's social and cultural transformations onreligious beliefs, practices, and institutions.The Vietnam War and outbreaks of urbanunrest had radically changed the mood of thecountry. By the 1960s, deep disaffection withthe social order had given rise to a countercul-ture that embraced Eastern mysticism and thepolitics of the new left. One position holdsthat those decades produced a profound rup-ture in the nation's moral and religious fabric,from which we have not yet recovered — thatit caused a turning away from God, from allbut the most superficial religious involve-ments, and that it produced at best a narcissis-tic, inward-looking spirituality.2

A second group of observers views thosechanges as temporary and sees most of thelarge cohort of boomers who came of age dur-ing those decades as having largely returnedfrom youthful experimentation and the cultur-al excesses of that time. They point to thegrowth in Evangelical and Charismaticchurches and to the more conservative moral

and religious mood that now prevails. Justlike others before them, with passage intomiddle age, and establishment of family,neighborhood, and work-related ties, theboomer generation will re-connect with reli-gious institutions.

A third, middle position acknowledges that aspiritual awakening has taken place but that itwill not lead to return to conventional reli-gious life. According to Roof, his positionemphasizes:

… that the age of strong religious hege-monies in the modern world is over; [and]that cognitive, moral, and religious plural-ism is now a reality in the everyday livesof Americans … Skeptical of establishedinstitutions and highly subjective in theirapproaches to religion, the post-World WarII generations choose for themselves whatto believe and what to practice … [and]are far less concerned about orthodoxiesand inherited faith traditions than weretheir parents … Objective truth losesmuch of its persuasiveness, and consisten-cy of belief and practice becomes lessimportant than coherence … Religion hasnot been abandoned but is expressed in amood, style and discourse strikingly dif-ferent from that of a half-century ago.3

Roof agrees with much of this third positionabout the changing nature of spiritual con-cerns — that “even if not always defined intraditional religious language, [they] will findwidespread expression and are leading tomajor realignments of people and institu-tions.”4

Compared to even 15 years ago, fewerAmericans today espouse a religious identity

Institute for Jewish & Community Research

2

— at least as conventionally ascertainedthrough survey questioning. Based on thedata to be presented, this change, we believe,is indisputable. Supporting evidence will bepresented later. But, as our analysis will alsosuggest, the meaning of religious identity(and non-identity) is increasingly murky.Does growing unwillingness to claim a reli-gious identity — to respond “None” to thestandard survey questions (“What is your reli-gion?”, “What is your religious preference?,”“Are you Roman Catholic, Protestant, Jewish…?”) — represent a broad rejection of religion,i.e., a distinct indicator of secularization? Or,is the meaning more nuanced or more uncer-tain? Could it instead signal that the appro-priate conceptual categories and methodolo-gies we use are increasingly inadequate to thetask, given the more expansive and multifac-

eted nature of contemporary religion? That is,does it reflect the limitations of our tools asresearchers?

In this report we examine not religious practicelike church attendance or membership butrather religious identity. We find that the reli-gious identity of Americans has shifted inpotentially significant ways, and the implica-tions of these shifts may be felt for genera-tions to come in politics, philanthropy, andother areas of civic life seemingly disconnect-ed from religion itself.

Before delving into the data on religious self-identification, it is worthwhile first to addresswhy this matters, and, second, to reviewtrends in other indicators of Americans' reli-giosity.

The Decline of Religious Identity in the United States

3

Approximate ly one of every s ix Amer i canshas no re l ig ious ident i ty

Sixteen percent of United States adults eitherfail to place themselves in any denominationalcategory (answering “none” or ”no religion”),or they describe themselves as secular,humanist, ethical-culturalist, agnostic, or athe-ist.

Ind iv idua ls who ident i fy with no re l ig ion area growing popu lat ion

Based on a review of survey evidence, theproportion of non-identifiers appears to havegrown substantially in the last 10-12 years.

The non-re l ig ious ly ident i f ied make up theth i rd largest group in the country

The two largest groups are Catholics andBaptists. Those non-religiously identified arevirtually tied with Baptists as the secondlargest group since the difference in estimatedsize between them is within “sampling error.”

Those ra ised in no re l ig ion are most l ike ly tonot ident i fy with a re l ig ion

About 1 of every 9 Americans who was raisedin some religion now identify with no reli-gion; nearly three-quarters of those with noreligious upbringing are current non-identi-fiers.

Being ra ised in more than one re l ig ion maylead to no re l ig ion

Those raised in multiple religious traditionsare more than twice as likely to be non-identi-fiers as adults than those raised in a singlereligion.

Younger Amer i cans are less re l ig ious ly iden-t i f ied than o lder Amer i cans

Younger adults (under 35) are most likely tobe non-identifiers, and those over 65 are least

likely to be. Religious identification shows asteadily increasing association with age. It isunclear whether this represents a persistentgrowth trend in non-identifiers, or if it reflectsa snapshot in time, with younger people likelyto become more affiliated with religion as theypass through customary life-cycle stages.Non-identification in the United States likelywill continue to increase.

Westerners lead the country in the proport ion of peop le who do not ident i fy

with a re l ig ionThose living in the West region of the countryare much more likely to be non-identifiers(24%) than Americans residing elsewhere(14%). While Southerners are somewhat lesslikely than those living in the Midwest andNortheast to be non-identifiers, the differencesacross the three non-West regions are modestcompared to the contrast between the Westand the rest of the United States.

Men are less l ike ly to ident i fy with a re l ig ious denominat ion

Men are more likely to be non-identifiers thanwomen, 20% versus 13%.

Some peop le who do not ident i fy with a re l ig ion , neverthe less , pract i ce some

form of re l ig ionSizable numbers of those who do not affiliatepsychologically with any religion are, never-theless, occasional or unsettled practitioners.As such, they might sometimes attend reli-gious services, have previously identified reli-giously as adults, or expect to take up a reli-gion sometime in the future. A more com-plete religious profiling requires additionalinformation about religious beliefs and behav-ior.

MAJOR F INDINGS


5

As citizens of a democratic nation, there areseveral reasons why we might be concernedby an ostensible decline in religiosity or, atleast, in some connectedness with organizedreligion. First, apart from religion's role inimparting salutary personal values5 — which,we assume, conveys net societal benefits6 —religious participation is also instrumental inbuilding “social capital.” This is especiallycritical for those on the lower rungs of thesocio-economic ladder or who are otherwisedisadvantaged, such as from racial or ethnicdiscrimination.

Social capital, a powerful concept popularizedby Robert Putnam,7 refers to the social net-works, and the norms of reciprocity and trust-worthiness that arise from them. Meaningfulat the individual and aggregate levels, socialcapital develops through interactions withothers, especially with others who move indifferent circles than oneself. The greater theamount of interaction and, particularly, thegreater the diversity of interactions, thegreater the stock of social capital. Social capi-tal can be thought of simultaneously as both a“private good” and a “public good.” That is,it can help individuals succeed in their per-sonal pursuits as well as carry benefits for thebroader community by facilitating collectiveaction through “connections,” interpersonaltrust, and sense of reciprocal obligation. It isthe latter — the use of social capital to facili-tate coordination and cooperation for mutualbenefits — that has gotten the most attention,partly because it is seen as an antidote to theoft bemoaned decline of community.

So, how is religion relevant to social capital?As Putnam points out, “Faith communities inwhich people worship together are arguablythe single most important repository of socialcapital in America,”8 as judged by the numberwho participate (as in the proportion whoattend worship services, but not only thatform of church-based interaction) and the fre-quency of participation. So, while faith-basedparticipation has declined overall in recentdecades according to Putnam's analysis(though not as much as other forms of socialinteraction — political, civic, work-related,neighborhood-based, and other informal typesof interaction), it nevertheless remains themost capacious source of social capital forma-tion. As such, it has tremendous potential toproduce collective benefits and to contributeto the public good.9

Secondly, in addition to the direct importanceof religion to social capital formation and theconsequent individual and collective benefitsthis can bring, much research also documentsthe positive correlation between church-basedparticipation and various forms of altruisticbehavior like charitable giving (to both reli-gious and secular organizations) and volun-teering one's time to help others.10 Whileskeptics might argue that the direction ofcausality is ambiguous - that altruistic peoplemight be attracted to religious participation -the theoretical connection between religiousinstruction/belief and charitable behavior istoo compelling to dismiss.11

A recent in-depth survey analysis by Tom W.Smith of the National Opinion ResearchCenter reached an unequivocal conclusion.

THE C IVIC BENEFITS OF REL IGION

While finding that demographic variableshave only slight-to-moderate relationshipswith altruistic values and a mixed patternwith empathy and altruistic behavior:

A number of the non-demographic vari-ables do show notable, statistically signifi-cant, and consistent relationships withempathy and altruism [altruistic behavior].In particular, religious involvement (e.g.,attending church and praying) are associ-ated with greater empathy and altruism … 12

Thirdly, newer cross-national research pin-points a connection between certain religiousbeliefs and economic growth. The beliefs thatseem to matter, i.e. produce growth, are thoseconcerned with heaven, hell, and an afterlife:

Convictions such as a belief in heaven orhell, might affect individuals by creatingperceived punishments and rewards thatrelate to “good” and “bad” lifetime behav-ior — which in turn can influence behav-iors like thrift, work ethic, and honesty

that contribute to economic growth … Theoverarching conclusion of this research isinescapable: A state's religiosity hasimportant influences on its economic per-formance.13

Whether these cross-national findings applysimilarly to over-time change in one countryis unknown, though the evidence is intriguingand suggestive.

To summarize, considerable evidence docu-ments a linkage between religious beliefs andbehavior, on the one hand, and a variety ofsocietal benefits. Some researchers wouldconclude even more strongly that they producepositive public outcomes. If trends in the rele-vant religion indicators are downward, itcould signify a diminution in these goods. Inthis sense, then, there is reason to believe thatreligion matters for civil society, and that adecline in forms of religious attachmentwould — or, at least, could — have negativeconsequences.


6


7

It is useful to place religious self-identificationtrends in context by reviewing over-timechanges in other religion indicators. Apartfrom changes in the propensity to report areligion category as a component of one'sidentity (I am [a Catholic, a Hindu, aMethodist, etc.]), what other changes areoccurring in religious beliefs, beliefs aboutreligion, and religion-oriented practices? Wereview those indicators for which timelinesexist stretching back at least several decades:first the cognitive measures, and then severalbehavioral gauges.

Bel ie f in God Belief in God (or some higher power) is themost fundamental religious belief. Across abroad range of surveys taken in recentdecades, 90% of Americans or more have saidthey believe in God or a higher power.14 Moreconvincingly, in the ten times the questionwas asked since 1987 by the Pew ResearchCenter, 87-88% completely or mostly agreedwith the stronger statement, “I never doubtthe existence of God”15 Although not quiteuniversal in the United States, belief in God ismore prevalent here than nearly everywhereelse, and so far it has been virtually invariantover time.

Bel ie f in the E f f i cacy of Re l ig ionNeither has there been much change in per-ceptions of the efficacy of religion, as meas-ured by Gallup's question: “Do you believethat religion can answer all or most of today'sproblems, or that religion is largely old-fash-ioned and out of date?” In 30 Gallup Pollstaken since 1974, the percentage giving theefficacious response — Religion can answer

today's problems — has ranged between 56-68% with no clearly discernable trend overall.The question was asked only once before 1974(in 1957), when fully 82% then thought reli-gion could answer our problems. So, if therehas been any decline in this key belief, itappears to have occurred in the third quarterof the past century.16

The Importance of Re l ig ion in One's L i fe“How important would you say religion is inyour own life?” is another question thatGallup has asked regularly since the early1980s and intermittently before that time. Onthis measure, polls taken from 1990 onwardhad somewhat more Americans respondingthat religion is very important in their lives(average = 59%), compared to an average of55% in the 1980s.17 So, there appears to havebeen a modest rebound in the perceivedimportance of religion in one's personal lifeafter an apparent drop-off during the 1980s.18

Rel ig ious Behav ior : Church Membersh ip ,Church Attendance , and Prayer

Three commonly used behavioral measures ofreligiosity are church membership, attendanceat worship services, and the importance ofprayer in one's daily life. Gallup has askedabout church membership and attendance foralmost 70 years. Surveys taken since 1980consistently place membership in the mid-high 60% range, where it appears to have sta-bilized following a decline (Figure 1).

The percent reporting past-week church atten-dance has varied in recent years between thehigh 30s and low 40s — a narrow interval thathas not changed much since the 1960s. Before

TRENDS IN MEASURES OF REL IGIOSITY (OTHER THAN IDENTIF ICATION)

that decade, attendance had increased brieflyto a peak during the 1950s, following WorldWar II, before falling back to the normal pat-tern of minor fluctuations within the specifiedinterval (Figure 2).

The Pew Research Center has askedAmericans in eleven polls since 1987 howmuch they agree or disagree with the follow-ing statement: “Prayer is an important part of

my daily life.” If anything, there has been aslight up-tick in daily prayer over the courseof this series (Figure 3).

While the accuracy and meaning of the firsttwo measures have been disputed — (1) the concept of “membership” applies more clearlyin some churches than others and (2) atten-dance is widely believed to be over-reported— these methodological objections are beyond


8

Exhibit 1: ChurchMembership

50

55

60

65

70

75

80

Pre-1950s19521965197619791982-851987-9019911992199319941995199619971998199920002001200220032004

Perce

nt

F igure 1Church Membersh ip by Year

(Source : The Ga l lup Organ izat ion)

Exhibit 2: Past Week Church Attendance(Source: Gallup Polls)

30

35

40

45

50

55

60

1939-401950-581962-691972-791981-851987-9019911992199319941995199619971998199920002001200220032004

Perce

nt

F igure 2Past Week Church Attendance by Year



9

the scope of this paper. Besides, since ourfocus is on change rather than absolute esti-mates at a particular point in time, any biasesshould be constant over time and, thus,should not matter.

To summarize this review of religious trends:

1. Belief in the existence of God and theimportance of daily prayer is wide-spread and largely unchanging.

2. Belief in the efficacy of religion mighthave declined during the third quarterof the last century, but has shown nopersistent directional trend since thattime.

3. The perceived importance of religion inone's life rebounded modestly duringthe 1990s, after a drop-off in the 1980s.

4. Church membership and attendanceboth appear to have declined somewhatsince 1960 (after post-World War IIincreases), but most of the decline tookplace in the 1960s and 1970s. Ratesshow limited changes in the past 25years.

Overall, there is scant evidence of muchchange in these measures of religious beliefsand behavior during the past quarter century.

Exhibit 3: "Prayer is an important part ofmy

20

30

40

50

60

70

80

90

1987 1988 1990 1991 1994 1997 1999 2002 2003

Perce

nt

Total Agree Completely Agree Mostly Agree

F igure 3“Prayer i s an important part of my da i ly l i fe .”

(Source : Pew Research Center for the Peop le and the Press)

[ ]

Religious identification has long been taken asa sign of belongingness to some faith tradi-tion:

Ever since Will Herberg's well-knownCatholic-Protestant-Jew, written in the mid-1950s, Americans have been singled outfor their high levels of religious identifica-tion, even if they are religiously indifferentin other respects. People might have ashallow level of faith and weak institution-al commitments, yet maintain loyalty to areligious community as a means of affirm-ing both a religious identity and theAmerican Way of Life … In a dynamicsociety, the scope and direction of suchpreferences are an important index of whatis happening religiously and culturally atany given time.19

While the data about changes in religiousbelief and commitment reviewed above revealsome evidence of change, mostly in the 1960sand 1970s, they also show signs of impressivestability and virtually no noticeable move-ments since the early 1990s. The picture isdifferent for the tendency to identify oneselfreligiously, which has declined significantly inrecent years after many decades where theproportion choosing a religion was high andunchanging. As we will show, Americanstoday are more likely than ever before toanswer “None” when asked in surveys tospecify their religion or religious preference.

Other estimates of religious identification (andnon-identification) come from the AmericanReligious Identification Survey (ARIS), basedon 50,281 interviews conducted in 2001, whichis intended as a sequel to the National Surveyof Religious Identification (NSRI), taken in

1990.20 Comparing the results reveals a sharpincrease in the number of United States adultsspecifying “no religion” or “none” (or atheist,agnostic, Humanist, Ethical Culture, or secu-lar)21 in response to the open-ended question,“What is your religion (if any)?” The propor-tion giving these responses rose from 8.2% to14.1%. If respondents refusing to answer areexcluded from the base of the percentages, therecalculated figures become 8.4% (1990) and15.0% (2001).

Although the magnitude of this change innon-identification is probably exaggeratedsomewhat because of the addition of “if any”in the wording of the 2001 survey question,for the same reason it is also likely that themore recent estimate is more accurate. As theARIS researchers have duly acknowledged:By not presuming a substantive response, the“if any” appendage makes it easier forrespondents with no religious identity/prefer-ence to answer “none.”

Survey researchers have long recognized thatinterview responses can be affected by percep-tions of “social desirability.” While accept-ance of diversity in recent decades hasincreased with regard to most ethnic and “life-style” categories, the one group that is stilllooked upon with considerable suspicion bymost Americans are those who choose to affil-iate with no religious faith tradition.22 Theimplication is that survey questions presum-ing the respondent to have some religiousidentity or preference by not offering a“None” option, or otherwise making it appearunacceptable to give that answer (e.g., by not


10

RELIGIOUS SELF- IDENTIF ICATION


11

appending “if any”), likely convey a subtlepressure to select some religion, thus biasingthe distribution of responses in the directionof under-estimating the extent of non-identifi-cation in the United States

The National Survey of Religion and Ethnicity(NSRE) is a second recent large-scale surveywhich contains the same identification ques-tion as the ARIS: “What is your religion, ifany?” The NSRE, conducted in parallel withthe National Jewish Population Survey(NJPS), consists of a national random-digit-dial sample of 4,027 interviews.23 This surveyproduced a slightly lower estimate of non-identifiers than the ARIS, 12.3%, but onewhich, nevertheless, represents an increase inthe number of Americans who claim no reli-gious identity, when compared with data froma decade earlier.24

The NSRE/NJPS religion estimates have beencriticized because of the positioning of the keyidentification question at the very beginningof the interview, before a sense of trust andrapport between interviewer and respondentcould be established. Given that the firstquestion in the interview signaled that thetopic is “religion,” it is quite possible that adisproportionate number of non-identifiersrefused to participate, perceiving the contentto be irrelevant or uninteresting. If true, thiswould produce a sample biased in favor ofthose with some religious affiliation, makingthe 12.3% figure is an under-estimate.

Another survey report, released in 2004, alsofinds a significant increase in the number ofreligion non-identifiers. The 2002 GeneralSocial Survey conducted by the NationalOpinion Research Center at the University ofChicago reports the percentage of Americanssaying they have no religion jumped from

nine percent (1992) to nearly 14%. The latestwave of this biennial series of in-person sur-veys, well-known to social researchers, isbased on a nationally representative sample of2,765 adults.

Gallup Polls, which use a differently wordedquestion emphasizing “religious preference,”seem to indicate less change over the sametime period as well as a lower current propor-tion of “no religion” / “none” responses.Unfortunately, Gallup changed the questionwording several times, complicating analysis.Since mid-2000, they included the “if any”phrasing in their primary version of the ques-tion:

“What, if any, is your religious preference— Protestant, Roman Catholic, Jewish,Mormon, Muslim, or an Orthodox religionsuch as Greek or Russian Orthodox?”

Before that time, the question did not contain“if any.” Moreover, in the early 1990s theyadded the reference to Orthodox religions inthe stem of the question. Further confound-ing analysis, they sometimes break out thepercentage of respondents who answer“Atheist” and “Agnostic,” and sometimes notin the information they make available to thepublic. And, they do not present separatecodes for “Ethical Culture,” “Humanist,” and“Secular,” or explain how those responses arecoded.

Since the middle of year 2000, when Gallupbegan using the “if any” wording, the per-centage of none/Atheist/Agnostic responseshas ranged widely in 32 of their surveys from7%-17%, though with most readings nearerthe lower end of the range. This substantialand unpatterned volatility over a short periodof time causes one to be suspicious of themeasure's reliability. Notwithstanding thesemethodological considerations, Gallup's “reli-


12

gious preference” questioning places the cur-rent number of non-identifiers at 10-11%when the recent estimates are averaged.

One can only speculate that the ARIS/NSREand Gallup questions must convey differentmeanings. It seems plausible that somerespondents find it easier to state a preferencefrom among a set of offered choices than toname a religion that categorically defines one-

self (“I am [NAME OF RELIGION]”) — possi-bly because the latter carries a stronger senseof commitment or invariance, whereas “pre-ferring” (as opposed to “being”) implies a lesscertainty/clarity about one's beliefs andbehavior. The same reasoning might helpexplain why the Gallup trend data shows lessgrowth in the number of non-identifiers. (Orshould they be labeled “non-preferrers?”)


13

The data source for our primary analysis isthe national survey of Heritage and ReligiousIdentification (HARI), conducted for theInstitute for Jewish & Community Researchduring the second half of 2001 and the firsthalf of 2002 (n=10,204 interviews).25 The pur-poses of the HARI survey included profilingthe United States population in terms of eth-nic/cultural group identification and religion,estimating the size of the Jewish population,examining the phenomena of religion switch-ing and the practice of multiple religions.Most of the survey items thus dealt withethno-cultural group identification or withreligious identification, background, and prac-tice of self and family.

Supplementing the ARIS/NSRI surveys, theHARI data provides further documentation of

the trend toward religious non-identificationand allows us to examine this dimension ofreligiosity more deeply by profiling the “noreligion” segment — 1,588 respondents — interms of their religious identity when growingup, their spouse's religious identity, pastchurch-based participation, future intentions,and demographics. By facilitating compar-isons with their counterparts on these dimen-sions, this data on non-identifiers provides arich and in-depth source of information todescribe and understand non-identifiers.

How many non- ident i f ie rs?

Fully 16% of the HARI sample answeredeither “none” when asked how they identifythemselves or gave one of the other answersclassified equivalently (atheist, agnostic, secu-lar, Humanist, or Ethical Culture):26

Exhibit 4: Largest Religion Segments(Source: IJCR, HARI Survey)

23.6

16.8 16.1

7.1 6.34.6 3.3 2.8 2.6 1.8 1.7 1.3 1.3

0

5

10

15

20

25

30

Catholic

Baptis

t

NONE

"Chris

tian"

Meth

odist

Luther

an

Pentac

ostal/

Charism

atic

Non-/In

ter-D

enom

inati

onal

Presb

yterian

Church of C

hrist

Episcopali

an

Morm

on/Latter

Day

Sain

ts

Jewish

Perce

nt

F igure 4Denominat iona l Ident i ty

(Source : Inst i tute for Jewish & Communi ty Research , Her i tage and Re l ig ious Ident i f i cat ion Survey)

A NEW DATA SOURCE: THE SURVEY OF HERITAGE AND RELIGIOUS IDENTIF ICATION


14

Do you now consider yourself [INSERTRELIGION RAISED IN (FROM PREVI-OUS ANSWER); IF NO RELIGIOUSUPBRINGING, READ:] Catholic,Protestant, Jewish, Muslim, some otherreligion, more than one religion, ornone]?27

This new figure — approximately one ofevery six adults — suggests that the growth inthe proportion of Americans having no reli-gious identity might be further advanced andproceeding faster than previously believed. Italso makes non-identifiers the third largestreligion-defined segment in the United Statesafter Catholics and Baptists, and only barelybehind Baptists (Figure 4).

As noted earlier, this 16% figure representsnearly a doubling of the ARIS study estimatearrived at one decade earlier, perhaps a bitless of a change if the 1990 ARIS estimate issomewhat understated because “if any” wasnot in the 1990 question wording. Neverthe-less, this methodological difference in meas-urement cannot possibly explain all of the dif-ference between the 1990 ARIS estimate andour 2002 HARI survey estimate. It is our view

that most of this change represents realgrowth in the number of Americans who viewthemselves as belonging to no religious group.The large-sample 2001 ARIS analysis, whichderives an estimate of non-identifiers almostas high as ours, supports this conclusion, asdoes the trend documented in the NORC’s lat-est General Society Survey.

Rel ig ious Or ig ins of Non- Ident i f ie rs

The largest number of non-identifiers (35%)were not raised in any religion as a child.Nevertheless, 62% of them had religious train-ing or background (Figure 5). This compareswith virtually all (97%) of respondents desig-nating a current religion. The fact that wellover half of current non-identifiers were raisedin some religion indicates substantial attritionin the proclivity to claim a religious affiliation.Movement in the opposite direction — fromno religion to a current identification — is lessextensive (26% of those raised in no religionnow identify with some religious category).This opposite-direction movement — acquir-ing a religious identity after having none inchildhood — is not nearly large enough to off-

Exhibit 5: Religious Origins of Current Non-Identifiers(Source: IJCR, HARI Survey)

None 35%

Catholic22%

ConservativeProtestant

19%

Liberal/MainlineProtestant

16%

MultipleReligions 3%

All Other 2%

Don'tknow/

F igure 5Rel ig ious Or ig ins of Current Non- Ident i f ie rs


Don’t know/refused 3%


15

set the elimination of identity28 among thosewith a religious upbringing. Among adultsalive today, our calculation produces a net lossof about 8% in the number having a religiousidentity relative to when they were young (nofigure).

Former Catholics (22%), former ConservativeProtestants (19%), and former Liberal/Mainline Protestants (16%) make up most ofthe remaining non-identifiers. Combiningthese groups shows that over half of current“nones” in the United States are formerChristians.

Figure 6 reverses the percentaging to computehow many from each tradition when younglater repudiated all religious identification (or,for those with no religion at youth, remainedunaffiliated). Nearly three-quarters of thoseraised in no religion remain non-identifierstoday. Among persons brought up in a singlereligious faith, between 9-13% have becomenon-identifiers. Most interesting is the highpercentage of Americans raised in two or

more religious traditions, fully 26% of whomhave relinquished all religious identity. Thefinding is consistent with the popular theorythat children raised in mixed religious tradi-tions are less likely to maintain any religiousorientation as adults. No other category oforigin (including “Jewish” — included in “AllOther”) indicates a loss of nearly that magni-tude.

Demograph i cs of Non- Ident i f ie rs

Just as non-identifiers' religious origins exhibitpatterning, their demographic characteristicsare also far from randomly distributed in thepopulation. Instead, our analysis demon-strates that they are disproportionately male,younger, unmarried, less likely to have chil-dren if married, living in the West region ofthe country, and Asian-American.

Gender

It is a sociological truism that women, onaverage, are more “religious” than men. AsBarry Kosmin and Seymour Lachman note,

Exhibit 6: Percent of Religious Origin Group ThatBecame Non-Identifiers (Source: IJCR, HARI Survey)

72

26

13 12 11 9

0

10

20

30

40

50

60

70

80

No religion Multiple religions Liberal/MainlineProtestant

Catholic Other ConservativeProtestant

Tradition raised

Perce

ntF igure 6

Percent Who are now Non- Ident i f ie rs by Re l ig ion of Or ig in(Source : Inst i tute for Jewish & Communi ty Research ,

Her i tage and Re l ig ious Ident i f i cat ion Survey)


16

“the lay and professional literature have con-sistently shown what ministers and parish-ioners have observed: that women are morelikely than men to join religious organizationsand participate actively. It appears thatChristianity is especially associated withfemale spirituality.”29 Our data indicate thatwomen in contemporary America are also sig-nificantly more likely than men to possess areligious identity (Figure 7). The relationshipbetween gender and reporting a religious

identity is robust, holding within all cate-gories of education, income, and age. So, it isnot just a case of women having lower socio-economic status than men, for example, thatdrives the statistical association — 85% offemale college graduates are identifiers com-pared to 78% of males — or that women tendto be slightly older than men, on average.(Even among the oldest age segments, womenare 5-6 percentage points more likely to beidentifiers than are men.)

Exhibit 8: Non-Identifiers by Age(Source: IJCR, HARI Survey)

2523

1615

12

8

5

0

5

10

15

20

25

30

18-24 25-34 35-44 45-54 55-64 65-74 75+

Perce

nt

F igure 8Non- Ident i f ie rs by Age


Exhibit 7: Non-Identifiers byGender

20

13

0

5

10

15

20

25

Males Females

Perce

nt

F igure 7Non- Ident i f ie rs by Gender



17

Age

Age is an even stronger predictor of religiousnon-identification. Figure 8 shows a steepand perfectly monotonic relationship betweenage and reporting “no religion” / “none”which ranges from only five percent, amongthe oldest segment of the population 75 andover, to 25% of 18-24 year-olds. In a singlecross-sectional measurement, such as theHARI survey, it is not possible to calculateprecisely or conclusively how much of thesedifferences in religious non-affiliation are dueto generational changes affecting specific agecohorts more than others — and which aremore likely to persist — and how much aredue to life-cycle changes. For example, aspeople reach certain stages of life (gettingmarried and raising young children), they aremore likely to join a church and assume a cor-responding religious identity. It is also notclear from the data how much of the totalobserved decline in identification hasoccurred among members of each age group.

One way to begin to sort out the relative influ-ence of these various causes is to examine the

proportion of non-identifiers in each age seg-ment who have a religious origin, that is,whether they were raised in some religion ornot. The narrowing gap between the two timeseries in Figure 9 suggests that the rate ofgrowth in the proportion of non-identifiers byage cohort is not constant. Instead, the leastamount of change has occurred among thetwo oldest cohorts (65-74 and 75+), with con-siderably more apparent in the other cohorts— among those born after the middle-to-late1930s and coming of age as adults after themiddle-to-late 1950s.

The growth in non-identifiers since youthincreases steadily as one moves from the old-est group to the youngest group, as indicatedby a widening gap between the two lines.Although one might reasonably expect theopposite trend, as progressively more timehas elapsed since youth in the older cohorts,thus allowing more time for change to occur,the overall pattern seems to be acceleration inthe growth of non-identifiers as new agecohorts enter adulthood: Not only areyounger cohorts today less likely to start out(as youth) with less religious background, but

Exhibit 9: Raised in No Religion and No ReligionNow

0

5

10

15

20

25

30

18-24 25-34 35-44 45-54 55-64 65-74 75+

Perce

nt

Raised no religion No religion now

F igure 9Raised in No Re l ig ion and No Re l ig ion Now by Age


[ ]


18

those that do have some religious upbringingare also more likely to become non-identifiers.The data indicate that the observed change inthe direction of increasing religious non-iden-tity, while applying to all population agecohorts to some extent, least characterizes per-sons 65 and older.

Without additional evidence it is hazardous toextrapolate the trend. However, the perfectage monotonicity of the patterns — both theproportion of persons with no religious identityas well as in the amount of growth in non-iden-tity by cohort from youth to the present — issuggestive. The data are inadequate foraddressing the extent to which life-cycleeffects play a role (they cannot tell us whenchanges occurred within the cohorts betweenyouth and the time of the survey), though it isplausible that some regularities in identifying,or not identifying, with a religious group per-tain as one moves from adolescence intoyoung adulthood into middle age and, finally,into senior citizenship. Testing that hypothe-sis conclusively would require longitudinaldata (multiple measurements on the samesample over time) on more than one cohort.

Region of the Country

The American West has always been a pio-neering, free-wheeling region where unortho-dox and experimental ways of life are com-mon and tolerated. Many new sects anddenominational splinter groups originated inthe West or moved there seeking religiousfreedom. It should come as no surprise, then,that the largest proportion of non-identifiers,conceptualized perhaps as an extreme form ofreligious unorthodoxy, should also reside inthe West. As Figure 10 shows, the incidenceof religious non-identity is markedly higher inthe West region than anywhere else.30

Sociologist Mark Shibley observes:

The West is a vast place geographically,which has made it harder for religiousinstitutions to pervade the landscape andcorner the market ... There's a sense ofspace — an openness — in this culture.31

Others point to the fact that people in theWest, especially on the West Coast, tend to befrom somewhere else — and that they tear uproots when they move to the West, includingtheir religious roots.32

Exhibit 10: Non-Identifiers byRegion

16 1513

24

0

5

10

15

20

25

30

Northeast Midwest South West

Perce

nt

F igure 10Non- Ident i f ie rs by Reg ion



19

Breaking the four standardized United StatesCensus regions into their smaller divisionalcomponents reveals that the Pacific coastalarea has the highest percent of “no religion”inhabitants — fully one-quarter of the popula-tion (Figure 11). Also evident is the significantcontrast in the Northeast region between New

Englanders — which, at 21% no religious affil-iation, is significantly higher than the nationalaverage — and those in the Mid-Atlantic area,which has a non-identifier incidence close tothe national average. Not coincidentally, thePacific states and New England also tend tobe the most liberal areas of the country. (The

Exhibit 11: Non-Identifiers by Geographic Division(Source: IJCR HARI Survey)

1513

21

14

22

25

1311

14

0

5

10

15

20

25

30

East-N

orth C

entra

l

Wes

t-North

Cen

tral

New E

ngland

Mid

dle Atla

ntic

Mountai

n

Pacifi

c

Wes

t-South

Cen

tral

East-S

outh C

entra

l

South

Atla

ntic

Perce

ntF igure 11

Non- Ident i f ie rs by Geograph i c Sub-Reg ion(Source : Inst i tute for Jewish & Communi ty Research ,

Her i tage and Re l ig ious Ident i f i cat ion Survey)

Exhibit 12: Non-Identifiers by Race(Source: IJCR, HARI Survey)

17

11

24

20

0

5

10

15

20

25

30

White African American Asian Native American

Race

Perce

nt

F igure 12Non- Ident i f ie rs by Race



20

end of this section documents at the individ-ual level of analysis the linkage between non-identification and political liberalism.)

Race

African-Americans are less likely than otherracial groups to have no religious affiliation(Figure 12). The long-time prominence of thechurch in African-American communities iswell-known. Less well established in the soci-ology of American religion is the high inci-dence of non-identification among Asian-Americans (24%).33 Native-Americans alsoappear to have a higher-than-average rate ofnon-identification, but this difference is notstatistically significant.

Mari ta l Status and Ch i ldren

In every age bracket through 55-64, non-iden-tifiers are significantly less likely to be mar-ried than identifiers (Figure 13). At the sametime, through age 74, non-identifiers are morelikely to be living unwed with a partner (not

shown in figure), but for persons 35 and olderthe incidence of living unmarried with a part-ner is low for both segments and the differ-ences are small. Having a religious identityseems consistent with traditional marriageand incompatible with living unmarried witha partner. It is also the case that married non-identifiers are less likely to have children thanmarried individuals with a religious affilia-tion.34

To what extent are couples homogeneous withrespect to having no religious identity? Inother words, do the non-religiously like-mind-ed tend to attract one another? Among mar-ried/partnered couples, if one partner hassome religious identity, the other partner isalso likely to be an identifier (91%). However,if one partner has no religious identity, thenthe chances of the other partner having anidentity are the same as his/her being a non-identifier: 48% yes, 48% no, and four percentnot sure. Given that the overall populationincidence of “no religion” is 16%, this indi-cates a strong tendency for clustering to occur.

Exhibit 13: Percent Married by Identity Segment by Age (Source: IJCR, HARI Survey)

16

58

6872

56

8

47

57 57 55 56

69

0

10

20

30

40

50

60

70

80

18-24 25-34 35-44 45-54 55-64 65+

Perce

nt

Identifiers Non-Identifiers

F igure 13Percent Marr ied by Age and Re l ig ious Ident i f i cat ion


[ ]

Soc io-Economic Status

Contrary to what some might anticipate, edu-cation and income show little correlation withreligious identity rates. Household income isentirely unrelated. Education is only modest-ly related, as persons who have not been tohigh school are less likely than others to benon-identifiers, while those with graduate orprofessional degrees are slightly more inclinedto report no religion (Figure 14). But, exceptfor those in the lowest education bracket(grade school or less), the differences are mod-est. Although greater levels of education arethought to expose a person increasingly torational and scientific modes of thought and

discourse, it has only modest impact on thepropensity to claim a religious affiliation.

Pol i t i ca l Or ientat ion

Analysis of Gallup Poll data shows that non-identifiers are much more likely to say theyare moderate or liberal than are those with areligious preference and less likely to describethemselves as conservative (Figure 15). Thisfinding is consistent with the well-document-ed tendency for religious Americans to bemore politically conservative. Among reli-gious identifiers, there are more than twice asmany conservatives as liberals. Among non-identifiers, the ratio is 1.5 to 1 in favor of liber-als.35


21

Exhibit 14: Non-Identifiers by Education(Source: IJCR, HARI Survey)

11

17

15

17 17

19

0

2

4

6

8

10

12

14

16

18

20

22

Grade school orless

Some highschool

High schoolgraduate

Some college 4-year collegedegree

Graduate/professional

degree

Perce

nt

F igure 14Non- Ident i f ie rs by Educat ion



22

Figure 15: Political Orientation by Religious Identification(Source: The Gallup Organization, 2002)

0

10

20

30

40

50

60

70

80

90

100

Identifiers Non-identifiers

Perce

nt

Liberal Moderate Conservative

F igure 15Pol i t i ca l Or ientat ion by Re l ig ious Ident i f i cat ion


[ ]

41

41

17

43

20

35


23

What does it really mean to have no religiousidentification — to answer “none” whenasked to specify one's religion or religiouspreference? To what extent does the answer-er's non-affiliation extend to other dimensionsof secularity? Are non-identifiers, in fact, aworld apart when it comes to matters of con-ventional (and non-traditional) religiousbeliefs, behavior, and spirituality?

First off, as we have already seen, one wouldbe mistaken to assume that this religious com-ponent of one's identity is unchanging. Alarge number have abandoned their religiousidentity, a smaller number who had no reli-gious upbringing have adopted one, manyothers have switched affiliation from onechurch or denomination to another — some ofthem multiple times.36 As the ARISresearchers describe, identity today is poly-morphous and in constant flux:

… in an environment where individualsmay hold multiple notions of self, andhold membership in multiple, non-contin-uous communities and associations, estab-lishing any fixed notions of identity isproblematic. One of the hallmarks of con-temporary American society in particularis that individuals can lay claim to a vari-ety of identities … with varying commit-ment to each … In such an environment,it becomes difficult to speak of anyone'sidentity as a permanent fixture of the self.37

Given the fluidity of religious identificationand the fact that it is but one of numerouscomponents of one's sense of self, it is unlike-ly that the non-identifier segment of the popu-lation is entirely distinctive, even on mattersof religion.

Indeed, analysis of the ARIS data reveals thateight percent of the “no religion” groupdescribe their outlook as “religious” andanother 28% describe it as “somewhat reli-gious.” (This compares with 43% “religious”and another 43% “somewhat religious,” forUnited States adults who profess some reli-gion.) Fully 67% of them believe that Godexists, and 68% of those believers agree that“God performs miracles” — clearly a reli-giously grounded belief.38

In the HARI data, we find similar evidencethat current non-identifiers are far fromadamantly, persistently, or consistently secu-lar: At least 26% of them say that they willdefinitely (4%) or probably (22%) take up areligion sometime in the future, and anothernine percent do not know (Table 1 below:sum of rows C, D, G, and H)39. More thanone-third of them (36%) had attended a reli-gious worship service other than a weddingor funeral sometime in the past year (Table 1:sum of rows D, E, F, and H) — 15% had doneso within the past month. (Nine percent ofnon-identifiers reported that they typicallyattended religious services at least once amonth.) Fully one-half of non-identifiers(51%) reported practicing a religion sometimein the past (Table 1: sum of rows B, C, D, andE).

Only 45% are what might be called “hard-core” in their current and expected non-reli-gion (rows A + B). In contrast, the data analy-sis indicates that 55% of those reporting noreligious identity are either presently engag-ing in some form of manifestly religious

DISCUSSION: INTERPRETING NON-IDENTIF ICATION


24

behavior or say they intend to do so, as pre-sented in the highlighted rows C-H in Table1.40 As a group, their “secularity” is hardlypure in this sense. That two-fifths of them(41%) expect to adopt some religion in thefuture or are undecided — only one-tenth ofthem being uncertain — indicates an opennessto change among a sizeable minority of thecurrently unaffiliated as well as further evi-dence that the “non-identifier” label is notalways capturing a fixed trait. The latterpoint is underscored by the relative youthful-ness of non-identifiers: 43% are younger than35 and fully two-thirds are less than 45.

While non-identifiers are demographicallydifferent than their counterparts — and obvi-ously join churches and participate less infaith-based activities — they are not as dis-tinctive a group as might be expected. Evenin terms of basic religious tenets, it is moreaccurate to describe them as formally unaffili-ated than as non-believers, although a sizeableminority of non-identifiers do reject notions ofGod's existence.

The foregoing suggests a 2 X 2 mapping(Table 2).

Tab le 1Rel ig ious Character i s t i cs of Non- Ident i f ie rs

A. No past religion, no current attendance, no religion in future

B. Past religion, no current attendance, no religion in future

C. Past religion, no current attendance, possible religion in future

D. Past religion, current attender, possible religion in future

E. Past religion, current attender, no religion in future

F. No past religion, current attender, no religion in future

G. No past religion, no current attendance, possible religion in future

H. No past religion, current attendance, possible religion in future

26%

19

11

11

10

9

8

6

Some Re l ig ious Group Ident i ty(“Ident i f ie rs”)

No Re l ig ious Group Ident i ty(“Non- Ident i f ie rs”)

Re l ig ious Part i c ipants :

“Traditionals”(70%)

“Non-Rejectors” (6%)

Non-Part i c ipants :

“Absentees” (13%)

“Seculars”(11%)

Tab le 2A Typo logy of Re l ig ious Ident i f ie rs and Non- Ident i f ie rs


25

“Traditionals,” by far the largest populationgroup, affiliate psychologically with somechurch/denomination and practice their reli-gion by participating in worship services atleast occasionally.41 “Seculars” represent theopposite extreme, neither identifying with anyreligion nor attending services. They make upabout one in nine Americans. But evenamong these non-identifying, non-participat-ing Seculars, 29% expect to take up some reli-gion in the future and another 7% of them didnot know. “Absentees” (who are 13% of thepopulation) maintain a religious identificationbut, like Seculars, rarely or never attend wor-ship services. “Non-Rejectors” are the small-est segment at six percent. They are the oneswith no fixed religious identity but neverthe-less sometimes participate in religious servic-es. Non-Rejectors might be the most interest-ing segment for further study.

While it is beyond the scope of this research toexplore the differences between “Traditionals”and “Absentees,” in keeping with the focus ofthis paper it might be illuminating to pointout what distinguishes “Non-Rejectors” from“Seculars” among Non-Identifiers. In fact, thedifferences that divide the two sub-segmentsare narrower than what distinguishes theaggregate of the two (non-identifiers in total)from identifiers. Apart from Non-Rejectorsattending religious services more often —although not nearly as often as religious iden-tifiers — and being more open to taking up areligion sometime in the future, there are onlya few differences in the respective profiles ofthe Non-Rejector and Secular varieties of non-identifiers:

• Non-Rejectors are more likely thanSeculars to be in the youngest age group(18-24), more likely to be single, but alsomore likely to have minor-age children;

• Non-Rejectors are more likely thanSeculars to be African-American and lesslikely than Seculars to be Caucasian; and

• Non-Rejectors are more likely thanSeculars to live in the South region andless likely to live in the West.

Although non-identifiers, especially Non-Rejector non-identifiers, are not as differentiat-ed from the rest of society or outside themainstream as certain religious conservativeslike to claim, they are nonetheless sufficientlydifferent and proliferating in number forobservers to try to understand them and thephenomenon they represent.

So what accounts for their growth, and whatdoes it mean? Some have argued that thegrowth in non-identifiers connotes much lessthan a full-fledged abandonment of religion.One line of thinking, propounded by Berkeleysociologists Michael Hout and Claude Fischer,holds that the increase in “nones” is a reactionto the emerging activism of Christian conser-vatism in American politics:

In the 1990s, many people who had weakattachment to religion and either moderateor liberal political views found themselvesat odds with the conservative politicalagenda of the Christian Right and reactedby renouncing their weak attachment toorganized religion.42

As we have seen, non-identifiers are indeedless conservative. Such a politically basedresponse might account for some of thegrowth in non-identifiers, but this explanationis too ad hoc and simplistic to be the mainreason. It seems just as plausible that identi-fiers put off by the Christian Right mightchannel their anti-conservative energies with-in the more liberal religious denominations.Moreover, the act of renouncing all affiliation,


26

however weakly held before, is itself worthyof note.

Other students argue that because the natureof religious belief and practice in the UnitedStates today is not well encapsulated by theconventional set of categories and denomina-tions, the standard survey approaches do notaccommodate the newer forms of identity,often non-institutionalized, non-traditional, ormulti-category hybrids.43 In practice, theabsence of appropriate survey responseoptions and/or failure to probe for in-depthdescriptions results in some fraction ofrespondents erroneously being countedamong the “no religion” group.

This methodological consideration might helpaccount for some of the observed increase innon-identification. To be sure, however, itcannot explain all of the observed increasebetween the 1990 NSRI survey and its 2001ARIS follow-up (which employed the samequestions and coding procedures), or betweenthe NSRI and HARI surveys.

A similar viewpoint denying that true changehas occurred is espoused by Robert Fuller,author of Spiritual But Not Religious. Fullerclaims that there have always been manymore people with no religious identification,but until recently they have been reluctant to

admit it. But lately, a sufficiently large criticalmass of non-identifiers has emerged, creatinga social reference group and making it morecomfortable for people to now say they haveno religion.44 Roof too recognizes this devel-opment, pointing to a new and stable “cultureof non-affiliation” making defection and/orthe rejection of religious identity “an increas-ingly acceptable alternative in American socie-ty.”45

This hypothesis would be difficult to test. Inone important sense though, it does not mat-ter, as some — probably most — of the newmembers of the critical mass making up the“no religion” reference group are surely trueconverts to “no religion” — not just closetnon-identifiers who have finally come out.Moreover, the very willingness to expressone's non-identity and associate “publicly”with other unaffiliateds is itself significant.

None of these explanations seems convincingenough to rebut the conclusion that thegrowth in non-identification is, to a significantdegree, a genuine movement away from reli-gion as historically understood in this country.Though in some cases temporary, and in mostcases not an abandonment of spirituality, webelieve that the numeric trend represents asubstantive alteration in the American reli-gious landscape.


27

This paper has documented a marked declinesince 1990 in the number of Americansespousing a religious identity, from less thanone-tenth of the public to nearly one in everysix adults today. Far from being randomlydistributed in the population, the data portraynon-identifiers as more likely than thoseexpressing a religious affiliation to be young,male, living in the West and New England,moderate-to-liberal politically, and unmarried.

Although not all non-identifiers are thorough-ly secular in their outlook or behavior, theyare certainly far less likely than identifiers tohold religious beliefs and engage in tradition-al religious behaviors such as participating inworship services. While the empirical evi-dence presented is hardly definitive on thispoint (let alone the causal relationships), thedecline in identification might be accompa-nied by a decline in civic altruism, sociallyconstructive attitudes, economically produc-tive beliefs, and social capital formation. Ifsimilar patterns were observable on otherindicators of religious belief and practice, wewould feel more confident in projectingimpacts. But because the other trend meas-ures were stable or mixed in direction, it is notpossible to be unequivocal on the interpreta-tion and implications.

Wade Clark Roof concurs on the ambiguity ofthe meaning of the changes in contemporaryreligion:

Boundaries that once separated one faithtradition from another are now oftenblurred; religious identities are malleableand multifaceted, often overlapping sever-al traditions. Trends and events stretchour imagination, even as we try to predict

the direction of religious change. Someindicators point to institutional religiousdecline, others to a profound spiritual fer-ment, confusing the picture and making itdifficult to describe what is really happen-ing.46

A lingering question for this research iswhether the decline in psychological affilia-tion with any church or denominationalorganization will persist, level off, or reversecourse. Will the decline in religious identifica-tion observed in this research continue orprove to be a temporary phenomenon?

If the question is meant to refer to the tradi-tional, established churches and faith systemsthat have been with us for some time, theanswer is very likely to be yes, it is likely tocontinue. But if we expand the concept of“religion” to include the increasingly popularforms such as New Age religions, Eastern-Western blends, multi-stranded hybrids, the“small-group movement,” pseudo-scientificspiritual formulations, and other types, thenmaybe no.

While this report has focused on the psycho-logical aspects of religious identity, moreresearch is needed to examine some of thebeliefs and behaviors not traditionally associat-ed with church attendance but that may beaffiliated with spirituality, such as specificbeliefs about religion and God, civic practicesthat correlate with religious ones, and non-tra-ditional definitions of religion, spirituality,and beliefs.

The basic character of American society virtu-ally guarantees that religious identity, like

CONCLUDING THOUGHTS

other dimensions of our concept of self, willcontinue to evolve and be transformed:

Religious identity … can be difficult topreserve unchanged even for those who

live in the most traditional of societies.But it can be especially difficult to main-tain when faced with the individualismand mobility of American life.47


28


29

FOOTNOTES

1 For smoother presentation, we use the term “church” in this paper to be an all-inclusive refer-ence to all religions. It is meant to incorporate synagogues, mosques, temples, etc.2 Wade Clark Roof, The Spiritual Marketplace (Princeton, NJ: Princeton University Press, 1999),pp.113-115.3 Ibid., pp.114-115.4 Ibid., pp.115.5 Robert Barro sees religion influencing individual traits such as honesty, work ethic, thrift, andopenness to strangers through the build-up of “spiritual capital” - a concept which includesformal education through organized religion as well as influences from family and social inter-actions (“Spirit of Capitalism: Religion and Economic Development,” Harvard InternationalReview [Winter, 2004]). Spiritual capital is not to be confused with “social capital,” as describedin the text.6 We do not wish here to be drawn into the argument over whether organized religion is moreof a force for good or for evil, though, with the recent growth in religion-inspired terrorism andviolence, it is tempting to dismiss its positive effects as, on the whole, insufficient to counter thenegatives.7 Robert D. Putnam, Bowling Alone: The Collapse and Revival of American Community (New York:Simon & Schuster, 2000).8 Ibid., p.66.9 This fact is doubly significant for those concerned about inequalities in American life becausereligion oriented interaction is the one source of social capital that is not highly stratified.Religious involvement provides one effective resource enhancing vehicle for the poor and dis-advantaged that is not disproportionately available to the better off. 10 The Giving and Volunteering in the U.S. series by Independent Sector is perhaps the bestsource. A recent edition is co-authored by Virginia A. Hodgkinson and Murray S. Weitzman(Washington DC: Independent Sector, 1996). 11 Alan Wolfe is one observer who demurs, presenting a more mixed summary of the empiricalliterature. See The Transformation of American Religion (The Free Press, 2003), pp.151-54.12 Tom W. Smith, “Altruism in Contemporary America: A Report from the National AltruismStudy,” unpublished report prepared for the Fetzer Institute, (July, 2003), p.13.13 Barro, op cit.14 George Gallup, Jr. and D. Michael Lindsay, Surveying the Religious Landscape: Trends in U.S.Beliefs (Harrisburg, PA: Morehouse Publishing, 1999), p.23. Unless otherwise specified, all datareported in this paper comes from nationally representative surveys.15 Pew Research Center for the People and the Press web site: http://people-press.org .16 Gallup Polls. There appears to have been one modest and short-lived upward blip in thismeasure, observable in a half-dozen polls between mid-1997 and mid-2000, when the averageefficacious response was 66%, compared to an average close to 60% both before and after thatperiod.17 In the few readings before the 1980s, the percentage answering “very important” was higher:70% (1965) and 75% (1952).


30

18 As with other religion-related measures, the contrast with Europe is striking: In the PewResearch Center's Global Attitudes Project survey conducted in May 2003, 59% of Americansregarded religion as important in their lives, compared to 11% of the French, 21% of Germans,27% of Italians, 33% of the British, and 35% of Poles (http://people-press.org/reports/dis-play.php3?ReportID=185). 19 Roof, p.123.20 Both surveys were designed and analyzed by Egon Mayer, Barry A. Kosmin, and ArielaKeysar of the Graduate Center of the City University of New York. The 1990 NSRI sample ismore than twice as large. Documentation and further detail is available at:http://www.gc.cuny.edu/studies/aris_index.htm.21 We and the ARIS researchers consider these answers as equivalent to having no religiousidentity as it is traditionally understood.22 Wolfe, p.248. 23 It was conducted in conjunction with the 2000-2001 National Jewish Population Survey byUnited Jewish Communities, the association of local American Jewish federations and inde-pendent communities. The NSRE sample is nationally representative on religious identificationwhen combined and properly weighted with the NJPS sample data.24 The NORC researchers note that much of this decline comes from former Protestants, theonce dominant broad religious category which now constitutes a bare majority (52%) in theU.S. and whose numbers are expected to further decline. (University of Chicago News Officepress release, July 20, 2004: http://www-news.uchicago.edu/releases/04/040720.protestant.shtml).25 The survey was conducted using random-digit-dialing. At least ten contact attempts weremade before sample replacement, including refusal conversion attempts. Sampling reliabilityfor estimates based on the total sample of 10,204 is ±1.0 points; for estimates based on subsetsof the total, confidence intervals are wider: For example, for the 1,523 non-identifiers, the sam-pling reliability is ± 2.5 points.26 Carried out to one decimal place, the figure is 16.1%. If those refusing to answer (or saying“Don't know”) are excluded from the base, the number increases to 16.4%. Henceforth, all per-centages include “don't know” and “refused” in the base unless otherwise specified.27 Appropriate follow-up questions were asked to pinpoint the respondent's current religiousidentity if (1) their current religion was not the same as the religion they were raised in, (2) tospecify which Protestant denomination, or (3) to specify which “other” religion.28 While it might not be literally true that all those brought up in a religion “identified” with itat the time of their youth, we consider this more of a semantic than substantive problem.29 Barry A. Kosmin and Seymour P. Lachman, One Nation under God: Religion in ContemporaryAmerican Society (New York: Crown Trade Paperbacks, 1993), pp. 210-211.30 It is not just that there are more young adults in the West driving this finding because the pat-tern holds when age is controlled.31 Quoted by Don Lattin, San Francisco Chronicle Religion writer, in “Living the Religious Life ofa None: Growing Numbers Shed Organized Church for Loose Spiritual Sensibility,”SFGate.com, (December 4, 2003).32 Ibid.


31

33 It could be that this is partly a methodological artifact - a result of first question on religiousidentity administered to respondents mentioning Catholic, Protestant, Jewish, and Muslim, butnot traditional Asian religions such as Buddhism, Hinduism, etc. But “Other religion” wasexplicitly offered as a choice, and those selecting it were asked what other religion in a follow-up question. So, we believe that this approach is unlikely to have inflated the number ofAsian-American non-identifiers by more than a few percentage points. Similar logic applies tothe higher than average rate of non-identity among Native Americans.34 Between 25-54 years of age (the general age span for having/raising children under 18), 65%of married identifiers have children under 18, compared to 58% of married non-identifiers. Thesame generalization, in fact, applies to all non-singles under 55 (not just people currently mar-ried): 60% of identifiers have children under 18, compared to 53% of non-identifiers. Perhapsidentifiers are more mindful of biblical injunction to “be fruitful and multiply.”35 Steve Hanway, Gallup Poll Tuesday Briefing Correspondent, April 22, 2003. This analysis isbased on 14,928 interviews conducted in 2002. The HARI survey did not include questions onpolitical party preference or ideology.36 The highly regarded sociologist of religion Robert Wuthnow believes that 20-30% ofAmericans practice a faith other than the one in which they were raised - cited in StevenWaldman, “The Politics of Piety” Washington Post (January 11, 2004). The 2001 ARIS studyderived a smaller estimate - 16%.37 Egon Mayer, Barry A. Kosmin, and Ariela Keysar, “American Jewish Identity Survey 2001,”p.31.38 Ariela Keysar, Egon Mayer, and Barry A. Kosmin, “No Religion: A Profile of America'sUnchurched,” Public Perspective (January-February, 2003), pp.28-32.39 The numbers do not appear to add to 35% exactly because of rounding.40 This includes those are unsure about future plans to take up some religion. Excluding thosewho are uncertain about their religious future, the proportion drops to 51%.41 They attended worship services at least once within the past year (other than a wedding,Bar/Bat Mitzvah, or funeral).42 Quoted in Lattin, op cit.43 Roof, op cit, p.207.44 Robert C. Fuller, Spiritual But Not Religious: Understanding the Unchurched in America, (NewYork: Oxford University Press, 2001); quoted in Lattin, op cit.45 Roof, p.125, quoting C. Kirk Hadaway and Penny Long Marler, “All in the Family: ReligiousMobility in America,” Review of Religious Research, Vol. 35 (December, 1994), p.104.46 Roof, op cit, p.4.47 Wolfe, op cit, p.215.


32

Dr. Gary A. Tobin is president of the Institutefor Jewish & Community Research and is alsodirector of the Leonard and MadlynAbramson Program in Jewish Policy Researchat the University of Judaism in Los Angeles.He earned his Ph.D. in City and RegionalPlanning from the University of California,Berkeley. He was the director of the Mauriceand Marilyn Cohen Center for Modern JewishStudies at Brandeis University for fourteenyears. Prior to joining Brandeis, Dr. Tobinspent eleven years on faculty at WashingtonUniversity, St. Louis.

Dr. Tobin has worked extensively in the areaof patterns of racial segregation in schoolsand housing. He is the editor of two volumesabout the effects of the racial schism inAmerica, What Happened to the Urban Crisis?and Divided Neighborhoods.

He has been a consultant in planning, demog-raphy, and philanthropy with hundreds ofnon-profits, including federations, syna-gogues, Jewish community centers, founda-tions and others.

Dr. Tobin is the author of numerous books,articles, and planning reports on a broadrange of subjects. His books include JewishPerceptions of Antisemitism, Rabbis Talk AboutIntermarriage and Opening The Gates: HowProactive Conversion Can Revitalize The JewishCommunity. Dr. Tobin is now completing abook entitled, Anti-Israelism & Anti-Semitismin America's Educational Systems. He is cur-rently involved in research concerning, phi-lanthropy, racial and ethnic diversity in theJewish community, and anti-Semitism.

GARY A. TOBIN, PH.D.

S ID GROENEMAN, PH.D.

Sid Groeneman is a senior research associateat the Institute for Jewish & CommunityResearch. He is also the head of GroenemanResearch & Consulting, an independent sur-vey/marketing research practice in Bethesda,Maryland (www.groeneman.com). Dr.Groeneman’s main areas of expertise are pub-lic opinion, consumer research, and surveymethodology. He is also an experienced focusgroup moderator and analyst.

During the 1980s and 1990s, he was aResearch Project Director and AccountManager for the Washington DC office ofMarket Facts (now Synovate), where he man-aged a more than 120 survey studies for gov-ernment agencies, Fortune 500 corporations,advertising and public relations firms, andnon-profit organizations such as research

institutes, trade associations, labor unions,philanthropies, and membership groups.

Dr. Groeneman has participated in severalmajor research projects for the IJCR includingsurveys (as well as subsequent reports andpress releases) about attitudes toward Israel,international terrorism, anti-Semitism, the sizeof the United States. Jewish population, andreligious identity in the United States. He isthe project manager of research on mega-giftsin American philanthropy.

He completed his undergraduate studies atthe University of Michigan and received aPh.D. in political science from the Universityof Minnesota. He lives in Bethesda, Marylandwith his wife, Beth, and teenage son, Michael.


33

Miss ion

The Institute for Jewish & CommunityResearch, San Francisco, is an independentnon-profit dedicated to the growth of theJewish community. The Institute serves as anational and international think tank provid-ing policy research to the Jewish communityand society in general. We design and devel-op initiatives that help build a more vibrantand secure Jewish community.

Pro jec t Summar ies

We educate both the public and opinion lead-ers through publications and conferences infour areas:

Demography and Re l ig ious Ident i ty

The Jewish people face serious demographicchallenges. It is not easy to clearly identifywho is a Jew and who is not. How does theJewish community adjust to significantchanges in religious belief and identity?

The Growth and Vi ta l i ty of JewishPeop lehood

Organizational, cultural and ideological bar-riers to growth have developed. How canthe Jewish community be more positive,open and welcoming?

The Secur i ty of the Jewish Communi ty

Anti-Semitism and anti-Israelism inAmerica and abroad have seen a dramaticincrease in recent years. How is currentanti-Semitism different than in the past andwhat are the appropriate institutionalresponses?

Phi lanthropy

American philanthropy, both Jewish andgeneral, set the standard for giving in theworld. What are the most important trendsregarding both foundations and individualdonors and how do they facilitate or inhibitpositive societal change?

THE INSTITUTE FOR JEWISH & COMMUNITY RESEARCHDr. Gary A. Tob in , Pres ident

www.jewishresearch.org

© 2004 Institute for Jewish & Community Research

For more information contact:Institute for Jewish & Community Research3198 Fulton StreetSan Francisco, Cal ifornia 94118p + 415 386 2604f + 415 386 2060e + [email protected]

The Decline Religious Identity - Welcome to the Institute for

Documents

Transcript of The Decline Religious Identity - Welcome to the Institute for