Full text of DOI:10.1207/S15324796ABM2401_07

Correlates of Self-Perceptions of Spirituality in American Adults
Leila Shahabi, B.A. and Lynda H. Powell, Ph.D.
Rush-Presbyterian-St. Luke’s Medical Center
Marc A. Musick, Ph.D.
University of Texas-Austin
Kenneth I. Pargament, Ph.D.
Bowling Green State University
Carl E. Thoresen, Ph.D.
Stanford University
David Williams, Ph.D.
University of Michigan
Lynn Underwood, Ph.D.
Fetzer Institute
Marcia A. Ory, Ph.D.
National Institute on Aging
ABSTRACT
tual and religious may be at particularly low risk for morbidity
and mortality based on their good psychological status and on-
going restorative activities.
To advance knowledge in the study of spirituality and physi-
cal health, we examined sociodemographic, behavioral, and at-
titudinal correlates of self-perceptions of spirituality. Partici-
pants were a nationally representative sample of 1,422 adult
respondents to the 1998 General Social Survey. They were
asked, among other things, to rate themselves on the depth of
their spirituality and the depth of their religiousness. Results in-
dicated that, after adjustment for religiousness, self-perceptions
of spirituality were positively correlated with being female (r =
(
Ann Behav Med
2002, 24(1):59–68)
INTRODUCTION
The possibility that religiousness or spirituality is a protec-
tive factor that promotes good physical health has led to a prolif-
eration of epidemiological and psychological research in this
area. Two recent reviews of the literature on religiousness and
spirituality and physical health (1,2) came to the same conclu-
sion: The strongest and most consistent evidence for any rela-
tion is in the link between church or service attendance and
all-cause mortality. Those who attend church or services once a
week or more frequently have a 25% reduction in risk of death
from any cause. This reduction is independent of demographic
factors, physical health status, physical function, social support,
depression, and healthy risk-factor profiles (2). Because this as-
sociation cannot be accounted for by these potential confound-
ers, the responsible factors remain unknown. Of particular inter-
est is whether all churchgoers are equally protected or whether
there is are subgroup members who, if they could be identified,
are at particularly low risk.
.07, p < .01), having a higher education (r = .12, p < .001), and
having no religion (r = .10, p < .001) and inversely correlated
with age (r = –.06, p < .05) and being Catholic (r = –.08, p <
.01). After adjustment for these sociodemographic factors,
self-perceptions of spirituality were associated with high levels
of religious or spiritual activities (range in correlations =
.12–.38, all p < .001), low cynical mistrust, and low political
conservatism (both r = –.08, p < .01). The population was di-
vided into 4 groups based on their self-perceptions of degree of
spirituality and degree of religiousness. The spiritual and reli-
gious group had a higher frequency of attending services, pray-
ing, meditating, reading the Bible, and daily spiritual experi-
ence than any of the other 3 groups (all differences p < .05) and
had less distress and less mistrust than the religious-only group
(
p < .05 for both). However, they were also more intolerant than
People attend church or services for a variety of reasons.
Some may have intrinsic motivations and live their religion by
absorbing it and finding ways to practice the moral precepts
taught (3). Others may go for more extrinsic reasons and use
their religion as a means of socialization, fulfilling an obliga-
tion, or gaining status within an organization (3). Those with in-
trinsic motivations may be a particularly low risk group if their
moral development fosters high levels of psychosocial protec-
tive factors including, for example, altruism and forgiveness
rather than hostility, joy rather than anger, hope and faith rather
than hopelessness and depression, and restorative practices such
either of the nonreligious groups (p < .05 for both) and similar
on intolerance to the religious-only group. We conclude that
sociodemographic factors could confound any observed associ-
ation between spirituality and health and should be controlled.
Moreover, individuals who perceive themselves to be both spiri-
Reprint Address: L. H. Powell, Ph.D., Department of Preventive Medi-
cine, Rush-Presbyterian-St. Luke’s Medical Center, Suite 470, 1700
West Van Buren, Chicago, IL 60612.
©
2002 by The Society of Behavioral Medicine.
59

6
0
Shahabi et al.
Annals of Behavioral Medicine
as prayer and meditation rather than chronic sympathetic
arousal. Efforts to identify those with a more intrinsic religious-
ness have resulted in such conceptualizations as private reli-
gious practices (e.g., reading the Bible, frequent prayer, listen-
ing to religious radio) and depth of religiousness (e.g., deriving
strength and comfort from religion, self-perceptions of how
deeply religious one is). However, there are, at best, weak links
between these constructs and physical health (2), signaling the
need for better measures of intrinsic religiousness.
Spirituality is an appealing construct because it connotes an
intrinsic characteristic that may or may not be induced by reli-
giousness. Whereas religiousness tends to be defined by organi-
zational beliefs, spirituality tends to be defined in personal, ex-
periential terms (4). The personal nature of spirituality, however,
creates problems for its scientific study. There is little agreement
on what spirituality is, how to measure it, and the extent of its
overlap with religiousness. Studies that have examined the
health implications of spirituality have used a variety of mea-
sures that assess a variety of constructs, which, for the most part,
have failed to find any links to physical health (2).
used in numerous studies aimed at the investigation of social
characteristics of the American adult public. Its main areas of
interest are social in nature, including socioeconomic status, so-
cial mobility, social control, family, race relations, civil liber-
ties, and morality. It has been shown to be both reliable and rep-
resentative of the adult population of the United States.
The GSS employs a split-ballot design to collect data. That
is, some questions on the survey are asked of all respondents,
whereas others are asked of a subset only. The choice of respon-
dents to answer subsets is random; consequently, the smaller
subsamples are representative of the larger sample and of the
population as a whole. In the 1998 GSS survey, a special ballot
on religiousness and spirituality was added. This ballot was de-
signed by a group of experts in the field to represent the multiple
dimensions included in the constructs of religion and spirituality
(9). Preliminary psychometric data for each of these scales were
presented earlier (9). The 1,422 individuals who responded to
the special ballot formed the population under investigation.
Variables
Despite the lack of conceptual clarity in the research litera-
ture, individuals appear to have little trouble rating themselves
on their own level of spirituality. Two studies using small sam-
ples of participants selected either because they were medically
ill (5) or because they came from preselected religious back-
grounds (6) described characteristics of individuals who per-
ceived themselves to be spiritual or religious. The “spiritual”
people, in contrast to the “religious” people, were of higher so-
cioeconomic status (5,6), less often from a Catholic background
The aim of this study was to create a sociodemographic, be-
havioral, and attitudinal profile of individuals who perceived
themselves to be spiritual. Although the special ballot included
questions about self-perceptions of spirituality and religious-
ness, public and private religiousness, religious denomination,
daily spiritual experiences, religious history, religious social
support, religious coping, religious beliefs and values, religious
commitment, and forgiveness (9), we limited the religious vari-
ables chosen to keep the focus on the nonreligious descriptions.
Thus, we selected only the behavioral factors (public and private
religiousness), religious denomination, and the daily spiritual
experiences items from the special ballot. Included, in addition,
were attitudinal variables from the larger GSS data set. Table 1
presents a selected group of these variables; their scales; and,
where appropriate, an alpha coefficient that estimates the inter-
nal consistency reliability of the scale.
(
5), more likely to be independent and reject traditional orga-
nized religion (6), and more likely to have a horizontal belief
system (7) where concerns were less with connectedness to God
and more with the interconnectedness of all living things (5,6).
Because these findings were from highly selected samples, their
generalizability to more representative populations is not
known.
The purpose of this study was to identify sociodemo-
graphic, behavioral, and attitudinal correlates of self-percep-
tions of spirituality in a large, nationally representative sample
of American adults. We identified these characteristics by com-
paring and contrasting individuals who considered themselves
to be spiritual with those who considered themselves to be reli-
gious or to be neither spiritual nor religious. By so doing, we
aimed to contribute to a refinement of the concept of spirituality
and to identify potentially confounding factors that should be
controlled in studies examining the role of spirituality on physi-
cal health.
Self-perceptions of spirituality and religiousness. Respon-
dents were asked two questions regarding their overall spiritual-
ity and religiousness: “To what extent to do you consider your-
self a spiritual person?” and “To what extent do you consider
yourself a religious person?” For each question, respondents
rated themselves along a 4-point scale ranging from 1 (very spir-
itual or very religious) to 2 (moderately spiritual or moderately
religious) to 3 (slightly spiritual or slightly religious) to 4 (not at
all spiritual or not at all religious).
A four-level typology was created by dichotomization of
each of the two scales and then combination of the two scales
into a 2 × 2 table. For example, the spiritual and religious group
METHODS
Sample
(n = 742, 52%) was composed of those respondents who an-
Data for the study came from the General Social Survey
GSS) (8) distributed by the Roper Center for Public Opinion
Research. The GSS is an independently drawn sample of per-
sons 18 years old or older who are English-speaking and live in
noninstitutionalized settings in the United States. The survey,
which has been collected annually from 1972 to 1998, has been
swered very or moderately to both the question about spirituality
and the question about religiousness. The spiritual-only group
(n = 143, 10%) was composed of those respondents who an-
swered very or moderately to the spiritual question and slightly
or not at all to the religion question. The religious-only group (n
= 128, 9%) was composed of those respondents who answered
(

Volume 24, Number 1, 2002
Correlates of Spirituality
61
TABLE 1
Variable Descriptions and Reliability Estimates for Scales
Variable
Question
Anchors
a
Religious/Spiritual activity
Service attendance
Private prayer
How often do you attend religious services?
How often do you pray privately in places other
than at church or synagogue?
0 (never) to 8 (several times a week)
1 (never) to 8 (more than once a
day)
—
—
Meditation
Within your religious or spiritual tradition, how
often do you meditate?
How often have you read the Bible in the last
year?
The following questions deal with possible
daily spiritual experiences. To what extent
can you say you experience the following:
1 (never) to 8 (more than once a
day)
1 (less than once a week) to 5
(several times a day)
1 (never or almost never) to 6 (many
times a day)
—
—
Reads the Bible
Daily Spiritual
Experiences
.91
(
(
(
(
a) I feel God’s presence
b) I find strength and comfort in my religion
c) I feel deep inner peace or harmony
d) I desire to be closer to or in union with
God
(
(
e) I feel God’s love for me, directly or
through others
f) I am spiritually touched by the beauty of
creation
Religious subscale
Spiritual subscale
Well-being
Items a, b, d, and e
Items c and f
.91
.70
Psychological distress
In the past 30 days how often did you feel: (a)
So sad nothing could cheer you up, (b)
1 (none of the time) to 5 (all of the
time)
.85
nervous, (c) restless or fidgety, (d) hopeless,
(e) that everything was an effort, and (f)
worthless.
Cynical mistrust
Do you think most people would try to take
advantage of you if they got a chance or
would they try to be fair?
1 (would try to be fair) or 2 (would
take advantage of you)
.67
Generally speaking, would you say that most
people can be trusted or that you can’t be too
careful in dealing with people?
1 (most people can be trusted) or 2
(can’t be too careful)
Would you say that most of the time people try
to be helpful or that they are mostly just
looking out for themselves?
1 (try to be helpful) or 2 (just look
out for themselves)
Self-rated health
Would you say your own health, in general, is
excellent, good, fair, or poor?
1 (poor) to 4 (excellent)
—
—
Beliefs and values
Conservatism
I’m going to show you a 7-point scale on which
the political views that people might hold are
arranged from extremely liberal to extremely
conservative. Where would you place
yourself on this scale?
1 (extremely liberal) to 7 (extremely
conservative)
Fatalism
Justice
Do you agree or disagree? There is little that
people can do to change the course of their
lives.
Some people say that people get ahead by their
own hard work; others say that lucky breaks
or help from other people are more important.
Which do you think is the most important?
Do you agree or disagree? In my opinion, life
does not serve any purpose.
1 (strongly disagree) to 4 (strongly
agree)
—
1 (luck most important) to 3 (hard
work most important)
Nihilism
1 (strongly disagree) to 4 (strongly
agree)
—
(
continued)

6
2
Shahabi et al.
Annals of Behavioral Medicine
TABLE 1 (Continued)
Question
Variable
Anchors
a
Intolerance (total score)
Atheists subscale
.87
.71
There are always some people whose ideas are
considered bad or dangerous by other people.
For instance, somebody who is against all
churches and religion …
(
a) If such a person wanted to make a speech
in your (city/town/community) against
churches and religion, should he be
allowed to speak or not?
1 (yes, allowed to speak) or 2 (not
allowed)
(
(
b) Should such a person be allowed to teach
in a college or university, or not?
1 (yes, allowed to teach) or 2 (not
allowed)
1 (not favor) or 2 (favor)
c) If some people in your community
suggested that a book he wrote against
churches and religion should be taken out
of your public library, would you favor
removing this book, or not?
.71
Racists subscale
Or consider a person who believes that Blacks
are genetically inferior. (same set of
(same anchors as Atheists subscale)
(same anchors as Atheists subscale)
.74
.79
follow-up questions as Atheists subscale)
Consider a person who advocates doing away
with elections and letting the military run the
country. (same set of follow-up questions as
Atheists subscale)
Militarists subscale
Communists subscale
Homosexuals subscale
Now, I should like to ask you some questions
about a man who admits he is a communist.
(same anchors as Atheists subscale)
(same anchors as Atheists subscale)
.73
.79
(same set of follow-up questions as Atheists
subscale)
And what about a man who admits that he is a
homosexual? (same set of follow-up
questions as Atheists subscale)
slightly or not at all to the spiritual question and very or moder-
ately to the religious question. The neither spiritual nor religious
group (n = 409, 29%) was composed of those respondents who
answered slightly or not at all to both the spiritual and religious
questions.
ans, Methodists), and liberal (e.g., Episcopalians, Presby-
terians).
Religious or spiritual activity. Several variables that as-
sessed common religious behaviors were included to assess ex-
tent of religious activity. Church or service attendance repre-
sented the extent of public religious activity. Religious and
spiritual practices included private prayer, meditation, and Bible
reading. The Daily Spiritual Experience Scale (9) is a six-item
scale that measures the frequency with which the respondent has
spiritual experiences on a day-to-day basis. This scale has items
that represent spiritual experiences phrased in religious termi-
nology (Religious subscale α = .91) and items that represent
spiritual experiences that do not require religious terminology
Sociodemographic variables. The demographic indicators
included age, sex, ethnicity, city size, geographic location, and
marital status. Indexes of socioeconomic status were education
and family income. The coding for these variables is presented
in Table 2.
Religious denomination. Because respondents who identi-
fied themselves in different ways along the spiritual–religious
continuum may have clustered along the lines of particular reli-
gious faiths, we included several dichotomous measures of reli-
gious affiliation. Respondents who stated that their religious
preference was Catholic or no religion were coded as such. Re-
spondents who stated that their preference was Protestant were
then asked to indicate their denomination. With the classifica-
tion scheme designed by Roof and McKinney (10), Protestant
denominations were classified into three groups: conservative
(Spiritual subscale α = .70).
Well-being. We include three measures of well-being. The
first, Psychological Distress, is a six-item index designed to as-
sess depressive and anxious feelings over the month preceding
the survey. The second measure is a three-item scale measuring
cynical mistrust of the motives of others (11). The third measure
is a standard item assessing self-rated health. Although
self-rated health measures contain information about both phys-
(
e.g., Southern Baptists, Pentecostals), moderate (e.g., Luther-

Volume 24, Number 1, 2002
Correlates of Spirituality
63
TABLE 2
TABLE 2 (Continued)
Description of Sociodemographic Factors and Religious
Activity of the Special Spirituality/Religion Panel
of the 1998 GSS Population
Variable
Less than once a year
n
%
151
274
10.6
19.3
Never
Variable
n
%
Frequency of prayer
Several times a day
Once a day
Several times a week
Once a week
Less than once a week
Never
Frequency of meditation
More than once a day
Once a day
A few times a week
Once a week
A few times a month
Once a month
Less than once a month
Never
349
421
201
108
308
24
24.5
29.6
14.1
7.6
21.7
1.7
Sample size
Sociodemographic factors
Female
1,422
775
54.5
Ethnicity
White
African American
Other
Agea
1,124
194
104
79.0
13.6
7.3
123
187
129
68
60
46
8.6
13.2
9.1
4.8
4.2
3.2
6.6
46.0
45.6
17
City size
City
Suburb
496
526
75
194
131
34.9
37.0
5.2
13.7
9.2
Unincorporated area
Town or village
Open area (< 2,499)
Geographic location
New England
Middle Atlantic
South Atlantic
North Central
South Central
Mountain
94
654
70
225
260
339
256
91
4.9
15.8
18.3
23.8
18.0
6.4
Note. GSS = General Social Survey.
a
These values are mean and standard deviation, respectively.
ical and mental health (12,13), they have been found to predict
mortality (14–16) and coronary heart disease (17).
Pacific
181
12.7
Marital status
Married
Widowed
Divorced/Separated
Never married
Education
Less than high school
High school
Associate/Junior college
Bachelor’s degree
Graduate degree
Family income
Less than $10,000
674
142
258
348
47.4
10.0
18.2
24.5
Beliefs and values. This domain is measured with four sin-
gle-item indicators. The first, Conservatism, is an indicator of
the political leanings of the respondent, where high scores de-
note political conservativism and low scores denote political lib-
eralism. The second, Fatalism, measures the degree to which re-
spondents believe that they can control their destiny. The third,
Justice, measures beliefs about success being a function of hard
work or lucky breaks. The fourth, Nihilism, measures beliefs
about whether life has any purpose.
228
747
106
228
107
16.0
52.5
7.5
16.0
7.5
159
184
401
399
156
123
12.7
12.9
28.2
28.1
11.0
8.6
Intolerance. Because religious activity has been linked to
intolerance toward certain groups (18), it is possible that respon-
dents will vary in levels of intolerance based on their religious
position. The GSS Intolerance Scale is composed of a total score
and subscales scores that measure intolerance toward each of
five different groups: atheists, racists, militarists, communists,
and homosexuals. For each group, three questions were posed
about whether a member of the specific subgroup should be al-
lowed to make a speech, teach in a college or university, or have
his or her books in a public library.
$
$
$
$
10,000–$19,999
20,000–$29,999
30,000–$59,999
60,000–$89,999
>
$90,000
Religious beliefs and activity
Religious affiliation
Protestant
774
365
26
54.4
25.7
1.8
Catholic
Jewish
Other
60
4.6
None
192
13.5
Statistical Analyses
Religious attendance
Several times a week
Every week
Nearly every week
About once a month
Several times a year
The analytic goal of this study was to identify correlates of
self-perceptions of spirituality. Toward this end, analyses fol-
lowed several steps. First, to describe the population under
study, frequency distributions for selected variables were deter-
mined. Second, to identify correlates of self-perceptions of spir-
ituality and compare them to self-perceptions of religiousness,
partial correlations between spirituality and religiousness and
115
242
203
115
302
8.1
17.0
14.3
8.1
21.3
(
continued)

6
4
Shahabi et al.
Annals of Behavioral Medicine
the full set of demographic, behavioral, and attitudinal variables
were calculated. To ensure that these correlations were not con-
founded by extraneous factors, they were determined first after
the other spiritual/religious variable was partialled out and then
after the other spiritual–religious variable and the full set of
sociodemographic correlates were partialled out.
variables. In general, there was a wide range of significant corre-
lations (range = .07–.52), and those that were weak (e.g., .07)
were nonetheless significant because of the large sample size.
Among the demographic factors, self-perceptions of spirituality,
after adjustment for religiousness, were associated with being
female, younger, and better educated. In contrast, self-percep-
tions of religiousness, after adjustment for spirituality, were as-
sociated with being in an ethnic minority, older, and less well
educated; living in a smaller town; and living in the South.
Self-perceptions of spirituality were inversely correlated with
being Catholic and positively correlated with having no religion,
whereas self-perceptions of religiousness were correlated with
being a conservative or moderate Protestant and inversely corre-
lated with having no religion.
After adjustment for these sociodemographic differences,
self-perceptions of spirituality were correlated significantly
with all of the religious activities. However, these correlations
were weaker than the corresponding correlations with religious-
ness. The only exceptions to this were for meditation and the
Spiritual subscale of the Daily Spiritual Experience Scale,
which were both more strongly related to self-perceptions of
spirituality (rxy = .28 and .38, respectively) than they were to
self-perceptions of religiousness (rxy = .09 and .20, respec-
tively). Among the indexes of well being, self-perceptions of
spirituality were inversely associated with cynical mistrust (rxy =
Next, we moved to an examination of the correlates of the
four-level spiritual–religious typology in which information
about both self-perceived spirituality and self-perceived reli-
giousness was combined. To examine sociodemographic corre-
lates of the four-level typology, we assumed that the direction of
causality was that these factors predict placement on the
typology rather than the reverse. Consequently, we modeled the
association by regression of the typology (the dependent vari-
able) on the sociodemograhic factors (the independent vari-
ables) using multinomial logistic regression. This procedure is
akin to logistic regression in that it computes the likelihood of
being in one category of the dependent variable versus another,
based on levels of the independent variables. However, it differs
from logistic regression in that it allows the dependent variable
to be more than two categories. Because our greatest interest is
in understanding spirituality, we calculated the odds for these
analyses with the spiritual and religious group as the referent
and compared it to the other three groups.
In the final stage of the analyses, we posited that other fac-
tors, such as behaviors and attitudes, were associated with the
spirituality–religiousness typology. Consequently, we modeled
these associations by generating mean scores of the attitudi-
nal–behavioral items based on levels of the typology. Given that
levels of the typology were influenced by sociodemographic
confounders, we adjusted first for these factors. All analyses
were completed with the PROC CORR, PROC GLM, and
PROC CATMOD procedures within the SAS software package.
–.08), and self-perceptions of religiousness were associated
with better self-rated health (rxy = .07). Among the beliefs and
values, spirituality was inversely related to being conservative
(
rxy = –.08), in contrast to religiousness, which was positively re-
lated to being conservative (rxy = .14). Spirituality was unrelated
to being intolerant, in contrast to religiousness, which was asso-
ciated with the Intolerance total score (rxy = .11) and intolerance
on the Atheist, Militarist, and Homosexual subscales (rxy = .12,
.10, and .13, respectively).
RESULTS
Our next step in the analyses was to examine the correlates
of self-perceptions when information about spirituality and reli-
giousness was combined to form the four-group typology. Table
Table 2 presents a description of the 1,422 respondents that
were evaluated in the spirituality–religion panel of the 1998
GSS. The average age was 45.6 years, with 68% of participants
falling within the approximate range of 28 to 62. The population
was 54.5% female, 79% White, and 47.4% married, and 16%
had less than a high school education. Approximately 34.9%
lived in an urban area, and 36.3% lived in the South. These char-
acteristics essentially mirrored those of the overall United
States, where according to the 2000 census or the 2000 Current
Population Survey (19), 50.9% were female, 75.1% were White,
4
focuses on the sociodemographic variables and presents the
independent odds of having each of the variables, with the spiri-
tual and religious group as the referent. Relative to this referent,
the spiritual-only group had an 80% lower odds of being minor-
ity, a 46% lower odds of being from the South, and a 52% lower
odds of being married. In addition, the spiritual-only group was
younger and better educated. These demographic differences
were less extreme when the religious-only group was compared
to the referent and more extreme when the neither spiritual nor
religious group was compared to the referent. The only excep-
tion to this was the educational level of the neither spiritual nor
religious group, which was lower than the referent and consider-
ably lower than the spiritual-only group.
Table 5 presents the behavioral and attitudinal profiles of
each of the spiritual–religious subgroups, after the socio-
demographicconfounderswerecontrolled.Thespiritualandreli-
gious group engaged in more religious or spiritual activity than
anyoftheotherthreegroups.Thisincludedmorefrequentservice
attendance, private prayer, meditation, Bible reading, and daily
5
2.9% were married, and 21.5% had less than a high school edu-
cation.
With regard to religious activities, 80% of the sample was
either Protestant (54.5%) or Catholic (25.5%), and 13.5% re-
ported that they had no religion. Approximately 31.6% attended
church or services nearly every week or more, 54.1% prayed at
least once a day, and 21.8% meditated at least once a day. This
was comparable to religious activity in the U.S. population (20).
Self-perceptions of spirituality and of religiousness were
highly correlated (rxy = .63). Table 3 presents correlations be-
tween these self-perceptions and the entire set of independent

Volume 24, Number 1, 2002
Correlates of Spirituality
65
TABLE 3
Correlates of Self-Perceptions of Spirituality and Religiousness
Spirituality
Religiousness
Adjusted for
Religiousness
Multiple
Adjustments
Adjusted for
Spirituality
Multiple
Adjustments
a
b
Variable
Sociodemographic factors
Female
Minority
Age
City size
South
.07**
.03
–.06*
.02
—
—
—
—
—
.02
—
—
.05
—
—
—
—
—
.07*
—
—
.09***
.17***
–.07*
.10***
.11***
–.11***
–.02
c
.02
Married
Education
–.02
.12***
–.02
Family income
Denomination
Conservative Protestant
Moderate Protestant
Liberal Protestant
Catholic
–.03
.01
.00
–.08**
.10***
—
—
—
—
—
.16***
.08**
.03
.06
–.42***
—
—
—
—
—
No religion
Religious/Spiritual activity
Service attendance
Private prayer
.12***
.18***
.27***
.19***
.35***
.29***
.37***
.12***
.20***
.28***
.18***
.34***
.28***
.38***
.44***
.43***
.16***
.32***
.48***
.52***
.25***
.34***
.30***
.09**
.26***
.40***
.43***
.20***
Meditation
Reads the Bible
Daily spiritual experiences
Religious subscale
Spiritual subscale
Well-being
Psychological distress
Self-rated health
Cynical mistrust
Beliefs and values
Conservatism
–.02
.00
–.05
.00
–.03
–.08**
.01
.00
.04
–.01
.07*
.05
–.09***
–.03
–.08**
.00
.19***
.04
.14***
–.03
Fatalism
Nihilism
Justice
–.06*
.00
–.02
.01
–.09**
.01
–.09**
–.02
Intolerance
Total score
–.05
–.05
–.01
–.08*
–.02
–.02
.01
.01
.03
–.03
.03
.04
.21***
.21***
.13***
.18***
.10**
.11**
.12***
.04
.10**
.03
Atheists subscale
Racists subscale
Militarists subscale
Communists subscale
Homosexuals subscale
.19***
.13***
a
b
Adjusted for religiousness, sex, age, race, education, income, city size, region, and religious affiliation. Adjusted for spirituality, sex, age, race, education,
c
income, city size, region, and religious affiliation. Larger values indicate larger cities.
*
p < .05. **p < .01. ***p < .001.
spiritualexperience. Theyexperiencedsignificantlylesspsycho-
logical distress and cynical mistrust than the religious-only
group. They tended to be more politically conservative and less
nihilistic(e.g., moreinclinedtobelievethatlifehaspurpose)than
the nonreligious groups. They were similar to the religious-only
groupontheirlevelofintolerance,whichwassignificantlyhigher
than that observed in either of the two nonreligious groups.
In contrast, the spiritual-only group was more politically
liberal and more likely to claim to have no religion (34%) than
any other group, including the neither spiritual nor religious
group (26%). They engaged in religious and spiritual activities
but not as frequently as the spiritual and religious group.
Compared to the religious-only group, they engaged in fewer re-
ligious activities (e.g., service attendance, prayer, Bible reading,
Daily Spiritual Experience–Religious subscale) but more spiri-
tual activities (e.g., meditation, Daily Spiritual Experi-
ence–Spiritual subscale; nonsignificant trend, p < .10). They
were more inclined to be nihilistic (e.g., to believe that life has

6
6
Shahabi et al.
Annals of Behavioral Medicine
TABLE 4
cated are all related to good health, they could confound any ob-
served relation between spirituality and health. That is, a finding
that spirituality is associated with good health may be alterna-
tively explained by the younger age, more frequent being fe-
male, or the higher socioeconomic status of those identified as
spiritual. We believe that it is imperative to control for these
sociodemographic factors before making inferences about any
association between spirituality and health.
Second, people who see themselves as spiritual are not a
homogeneous group. When information about spirituality and
religiousness was combined into a four-level typology, it was
clear that spiritual people actually comprised two subgroups
that differ on ethnicity, age, education, marital status, and geo-
graphic region of residence. The larger group, spiritual and reli-
gious, composed 52% of our sample and had a demographic
profile that was characteristic of religious people. The smaller
group, spiritual only, composed only 10% of our sample and had
a demographic profile that was characteristic of religious inde-
pendents (22).
Independent Sociodemographic Correlates of
Spirituality/Religion Typology
Neither
Spiritual Nor
Religious
Spiritual
Religious
Only
a
b
c
Variable
Only
Female
Minority
Age
City size
South
Married
Education
Family income
0.76
0.69
0.51*
1.00
1.01
0.63*
0.90
0.95
1.00
0.59***
0.45***
0.98***
1.06*
0.57***
0.61**
0.91***
1.03*
0.20***
0.98***
1.03
0.54**
0.48***
1.09*
d
e
1.01
Model fit statistics
2
χ
166.78
24
df
2
Pseudo R
N
.10
1,422
It was only when we studied the spiritual people as two dis-
tinct groups that we came on our most important finding: The
spiritual and religious group appeared to be the most intrinsi-
cally religious. This group attended services, prayed, meditated,
read the Bible, and had more daily spiritual experiences than any
other group. In particular, they were higher than the spiri-
tual-only group on their frequency of meditating, feeling deep
inner peace or harmony, and being spiritually touched by the
beauty of creation. Moreover, they were less distressed and less
mistrusting than the religious-only group. They appeared to take
the best from the spiritual side and the best from the religious
side, an observation also made by others (5,6). The only charac-
teristic on which they were not highly functioning was intoler-
ance, which appeared to be a byproduct of religiousness, regard-
less of level of spirituality.
Note. Odds ratios are shown with the spiritual and religious group (n =
7
42) as the referent.
a
b
c
d
e
n = 143. n = 128. n = 409. Larger values indicate larger cities. Larger
values indicate higher incomes.
p < .05. **p < .01. ***p < .001.
*
no purpose) than either one of the two religious groups and were
more tolerant than the religious and spiritual group only.
The religious-only group (those who perceived themselves
to be slightly spiritual or not at all spiritual) was distinctive in
that they were higher on distress and cynical mistrust than the
spiritual and religious group. Thus, within the religious groups,
the self-perception of being very or moderately spiritual was as-
sociated with better psychological status.
The spiritual-only group also appeared to be developing
their spirituality but not as frequently or to the same degree.
They prayed, meditated, and had daily spiritual experiences less
often than the spiritual and religious group but more often than
the neither spiritual nor religious group. The spiritual-only
group has been described as “New Age” (6), but this observation
may have been biased by the large number of individuals from
New Age groups that were involved in that study. Although New
Age characteristics were not measured in our study, we were im-
pressed with how mainstream this group actually appeared. The
main factors that distinguished them were their political liberal-
ism, tolerance of the expression of diverse points of view, and
unwillingness to claim any religious denomination. They appear
to be more like Roof’s (23) description of baby boomers who re-
nounced traditional religion to pursue their own, perhaps more
circuitous, path for spiritual development.
DISCUSSION
To understand better what people mean when they call
themselves spiritual, we analyzed data from a nationally repre-
sentative population of American adults. By comparing, con-
trasting, and integrating self-perceptions of spirituality with
self-perceptions of religiousness, we developed a sociodemo-
graphic, attitudinal, and behavioral profile. From these analyses,
we made several observations that can help to advance our un-
derstanding of self-perceptions of spirituality, a construct that
has been referred to as “obscure” (21) and “fuzzy” (6).
First, people who see themselves as spiritual have a differ-
ent sociodemographic profile than those who see themselves as
either religious or nonreligious. Self-perceptions of spirituality
are associated with being younger, female, and highly educated.
This is in contrast to self-perceptions of religiousness, which are
more likely to be associated with being older, in an ethnic mi-
nority, less well educated, and living in the South. Furthermore,
compared to those who see themselves as nonspiritual and non-
religious, those who perceive themselves to be both spiritual and
religious differ on each of the sociodemographic factors studied.
Because such factors as being younger, female, and highly edu-
One mechanism by which religion or spirituality may en-
hance physical health is via salutary psychosocial and
psychophysiological status (24). As such, these data suggest the
hypothesis that a subgroup of healthy individuals, or of healthy
service attenders, who may be at particularly low risk for mor-
bidity and mortality is the group who perceive themselves to be

Volume 24, Number 1, 2002
Correlates of Spirituality
67
TABLE 5
Comparison of Adjusted Mean Levels of Selected Variables by Spirituality and Religiousness Typology
Group 4:
Group 1:
Spiritual and
Religious
Group 3:
Religious
Only
Neither
Spiritual Nor
Religious
Group 2:
Spiritual Only
Variable
Sociodemographics (%)
Married
51a,c
39b
51c
41
Denomination (%)
Conservative Protestant
Moderate Protestant
Liberal Protestant
Catholic
29a,c
19
16
13
04
22
34
22
19
07
21
14
05
23
26
07
b
b
c
24
a,c
35
b,c
c
No religion
04
02
Religious/Spiritual activity
Service attendance
Private prayer
a,b,c
2.26b
c
4.86
3.79
5.67
2.61
1.93
3.85
3.83
3.88
1.82
3.62
2.08
1.50
2.70
2.52
3.07
a,b,c
b,c
c
6.60
4.71
3.55
1.71
3.46
3.16
4.03
a,b,c
b,c
Meditation
4.22
2.79
4.61
4.63
a,b,c
b
c
Reads the Bible
Daily spiritual experiences
Religious subscale
Spiritual subscale
Well-being
Psychological distress
Self-rated health
Cynical mistrust
Beliefs and values
Conservatism
a,b,c
b,c
c
a,b,c
b,c
c
a,b,c
c
c
4.57
1.99b
3.10
2.03
2.96
1.90
2.14c
3.11
2.17
1.96
3.02
2.05
b
1.97
4.17a,c
1.93
3.60b,c
1.89
4.13
1.83
1.49
3.93
1.98
1.72
2.55
Fatalism
Nihilism
Justice
a,c
b
c
1.48
2.53
1.72
2.59
2.60
Intolerance
Total score
a,c
c
3.8
3.5
4.5
2.9
2.4
3.9
3.9
3.7
1.7
3.5
3.6
2.9
2.3
3.3
3.3
3.7
1.7
a,c
b
c
Atheists subscale
Racists subscale
Militarists subscale
Communists subscale
Homosexuals subscale
c
3.8
3.8
4.2
1.9
a,c
4.0
a,c
4.2
a,b,c
2.7
Note. Means are adjusted for sex, age, race, education, income, city size, and region. Group differences indicate significant differences (p < .05) between
the groups as indicated by the following superscripts: a = different than Group 2; b = different than Group 3; c = different than Group 4. Group 1, n = 742;
Group 2, n = 143; Group 3, n = 128; Group 4, n = 409.
both spiritual and religious. They have good psychosocial status
by virtue of having low levels of distress and cynical mistrust,
two risk factors for mortality (25–27), and potentially good
psychophysiological status by virtue of having high levels of po-
tentially restorative activities (e.g., service attendance, prayer,
meditation, and daily experiences of comfort, peace, and awe),
which may promote ongoing shifts from sympathetic arousal to
parasympathetic relaxation or lower levels of cortisol secretion
mately 5% of the tests would be significant due to chance
alone. However, we place validity in our findings because of
their consistency with those of two similar studies conducted
on highly select samples (5,6). We believe these studies com-
plement ours and should be referred for more in-depth de-
scriptions of attitudes and behaviors of people who perceive
themselves to be spiritual.
In summary, the inherently personal nature of spirituality is
perhaps its greatest strength but also its greatest weakness, at
least as far as scientific study is concerned. Its strength lies in the
possibility that it reflects the internalization of virtues advocated
across numerous religious traditions and in the promise such a
conceptualization offers for establishing a relation to physical
health. Its weakness is that the wide variety of conceptualiza-
tions and assessment options limit comparisons across studies
(
see Ironson et al., this issue).
The strength of this study is its large, representative sam-
ple size. However this large sample is limited because even
weak associations become statistically significant and more
in-depth description of attitudes and behaviors is not possi-
ble. Moreover, a large number of statistical tests were con-
ducted, and working at .05 level of significance, approxi-

6
8
Shahabi et al.
Annals of Behavioral Medicine
and leave unanswered questions about consistency in research
findings.
tional Institute of Aging Working Group. Kalamazoo, MI:
Fetzer Institute, 1999.
10) Roof WC, McKinney W: American Mainline Religion: Its
Changing Shape and Future. New Brunswick, NJ: Rutgers Uni-
versity Press, 1987.
11) Demaris A, Yang R: Race, alienation, and interpersonal mis-
trust. Sociological Spectrum. 1994, 14:327–349.
12) Fylkesnes K, Forde OH: Determinants and dimensions in-
volved in self-evaluation of health. Social Science and Medi-
cine. 1992, 35:271–279.
13) Manderbacka K, Lundberg O, Martikainen P: Do risk factors
and health behaviors contribute to self-ratings of health? Social
Science and Medicine. 1999, 48:1713–1720.
(14) Idler EL: Self-assessed health and mortality: A review of stud-
ies. In Maes S, Leventhal H, Johnston M (eds), International
Review of Health Psychology. New York: Wiley, 1992, 33–54.
15) McCallum J, Shadbolt M, Wang D: Self-rated health and sur-
vival: A 7-year follow-up study of Australian elderly. American
Journal of Public Health. 1994, 84:1100–1105.
16) Dasbach EJ, Klein R, Klein BEK, Moss SE: Self-rated health
and mortality in people with diabetes. American Journal of
Public Health. 1994, 84:1775–1779.
17) Moller L, Kristensen TS, Hollnagel H: Self-rated health as a
predictor of coronary heart disease in Copenhagen, Denmark.
Journal of Epidemiology and Community Health. 1996,
(
We took the approach of going to the general public and
asking them to define for themselves where they believed they
fell along the spirituality continuum rather than relying on our
own preconceived notions about what spirituality is. We gained
trust in the validity of these self-perceptions because they were
consistent with religious or spiritual restorative behaviors and
salutary psychological status. The combination of being both
spiritual and religious is an intriguing concept because it may
suggest that efforts toward spiritual development are maximized
when pursued within the context of a supportive religious envi-
ronment. That is, going each week to a serene place that encour-
ages reflection, listening to sermons that advocate love and ser-
vice to others, and experiencing the joy of joining in song in
praise of faith may provide direct, ongoing, and reinforcing ex-
periences of the virtues enjoyed by highly spiritual people. Con-
versely, it is not inconceivable that people who are already
highly spiritual would seek out religious contexts and, in the
process, develop their religiousness. We look forward to future
studies that test whether individuals who perceive themselves to
be both religious and spiritual are at particularly low risk for
death and disease.
(
(
(
(
(
(
50:423–428.
(
(
(
(
18) Ellison CG, Musick MA: Southern intolerance: A Fundamen-
talist effect? Social Forces. 1993, 72:379–398.
19) United States Department of Commerce: U.S. 1990 Census
data [online]. Available from http://www.census.gov
20) Gallup Organization: 2000 Gallup poll [online]. Available from
http://www.gallup.com
REFERENCES
(
(
(
(
1) McCullough ME, Hoyt WT, Larson DB, Koenig HG, Thoresen
CE: Religious involvement and mortality: A meta-analytic re-
view. Health Psychology. 2000, 19:211–222.
2) Powell LH, Shahabi L, Thoresen CE: Religion and spirituality:
Linkages to physical health. Society for Epidemiologic Re-
search Annual Meeting. Toronto, Ontario, Canada: 2001.
3) Allport GW, Ross JM: Personal religious orientation and preju-
dice. Journal of Personality and Social Psychology. 1967,
21) Hood RW, Spilka B, Hunsberger B, Gorsuch R: The Psychology
of Religion: An Empirical Approach. New York: Guilford,
1996.
(
22) Hayes BC: Religious independents within Western industrial-
ized nations: A socio-demographic profile. Sociology of Reli-
gion. 2000, 61:191–207.
23) Roof WC: A Generation of Seekers: The Spiritual Journeys of
the Baby Boom Generation. San Francisco: Harper, 1993.
24) Levin J: How religion influences morbidity and health: Reflec-
tions on natural history, salutogenesis and host resistance. So-
cial Science and Medicine. 1996, 43:849–864.
5
:432–443.
4) Spilka B, McIntosh DN: Religion and spirituality: The known
and the unknown. American Psychological Association Annual
Meeting. Toronto, Ontario, Canada: 1996.
5) Woods TE, Ironson GH: Religion and spirituality in the face of
illness. Journal of Health Psychology. 1999, 4:393–412.
6) Zinnbauer BJ, Pargament KI, Cole B, et al.: Religion and spiri-
tuality: Unfuzzying the fuzzy. Journal for the Scientific Study of
Religion. 1997, 36:549–564.
7) Davidson J: Glock’s model of religious commitment: Assessing
some different approaches and results. Review of Religious Re-
search. 1975, 16:83–93.
8) Davis JA, Smith TW: General Social Surveys, 1972–1991. Chi-
cago: National Opinion Research Center, 1991.
(
(
(
(
(25) Miller TQ, Smith TW, Turner CW, et al.: A meta-analytic re-
view of research on hostility and physical health. Psychological
Bulletin. 1996, 119:322–348.
(26) Allison TG, Williams DE, Miller TD, et al.: Medical and eco-
nomic costs of psychologic distress in patients with coronary
artery disease. Mayo Clinic Proceedings. 1995, 70:734–742.
(27) Denollet J, Brutsaert DL: Personality, disease severity, and the
risk of long-term cardiac events in patients with a decreased
ejection fraction after myocardial infarction. Circulation. 1998,
97:167–173.
(
(
(
9) Fetzer Institute/National Institute on Aging Working Group:
Multidimensional Measurement of Religiousness/Spirituality
for Use in Health Research: A Report of the Fetzer Institute/Na-