Content uploaded by Lee A. Kirkpatrick
Author content
All content in this area was uploaded by Lee A. Kirkpatrick on Nov 19, 2019
Content may be subject to copyright.
Religious Fundamentalism as a Predictor
of Prejudice: A Two-Component Model
BRIAN LAYTHE
DEBORAH G. FINKEL
ROBERT G. BRINGLE
LEE A. KIRKPATRICK
The present study aims to determine whether the empirical relationship between religious fundamentalism and
prejudice can be accounted for in terms of the mutually opposing effects of Christian orthodoxy and right-
wing authoritarianism using multiple regression. Three separate samples (total n =320) completed measures of
religious fundamentalism, right-wing authoritarianism, Christian orthodoxy, ethnic prejudice, and homosexual
prejudice. Consistent with previous research, fundamentalism (1) was essentially unrelated to ethnic prejudice
when considered alone; (2) was positively related to ethnic prejudice when orthodoxy was statistically controlled;
and (3) was negatively related to ethnic prejudice when authoritarianism was statistically controlled. Finally,
when both authoritarianism and orthodoxy were controlled simultaneously, fundamentalism was again unrelated
to prejudice, whereas orthodoxy was negatively related and authoritarianism positively related. In contrast,
fundamentalism was a significant positive predictor of prejudice against gays and lesbians irrespective of whether
authoritarianism and/or orthodoxy were statistically controlled.
Religion is an extremely complex and multifaceted phenomenon, as reflected by the existence
of the enormous number of psychometric scales that have been developed over the years to
measure various aspects of it (Hill and Hood 2000). Because the various dimensions or aspects
of religiousness tend to be intercorrelated or confounded with one another, it is notoriously
difficult to tease apart the effects of different dimensions in assessing the empirical relationships
between religiosity and other variables. An unfortunate consequence of this is that the research
literature contains numerous examples of inconsistent and contradictory results with respect to
such relationships.
One prominent example of this problem in the research literature concerns the relationship
between religious belief and prejudice. This is a particularly vexing issue because it has long
been observed that different aspects of religiousness may relate to prejudice not only to varying
degrees, but in opposite directions. In an oft-quoted passage, Allport (1954:444) summarized the
problem by observing:
The role of religion is paradoxical. It makes prejudice and it unmakes prejudice. While the creeds of the great
religions are universalistic, all stressing brotherhood, the practice of these creeds is frequently divisive and bru-
tal. The sublimity of religious ideals is offset by the horrors of persecution in the name of these same ideals.
...Churchgoers are more prejudiced than the average; they are also less prejudiced than the average.
The solution at which Allport eventually arrived in attempting to resolve this paradox con-
cerned variation in the motivations behind people’s religious beliefs. Thus, the most influential
Brian Laythe is a doctoral student in the Department of Psychology, University of New Hampshire, Conant Hall,
10 Library Way, Durham, New Hampshire 03824. Email: blaythe@cisunix.unh.edu
Deborah G. Finkel is a Professor in the Department of Psychology, Indiana University Southeast, 4201 Grant Line Road,
New Albany, IN 47150. Email: dfinkel@ius.edu
Robert G. Bringle is a Professor in the Department of Psychology, Indiana University Purdue University Indianapolis,
402 North Blackford Street, LD 124, Indianapolis, IN 46202-3217. Email: rbringle@iupui.edu
Lee A. Kirkpatrick is an Associate Professor in the Department of Psychology, College of William & Mary, PO Box 8795,
Williamsburg, VA, 23187-8795. Email: lakirk@wm.edu
Journal for the Scientific Study of Religion 41:4 (2002) 623–635
624 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
aspect of Allport’s own work on religion and prejudice has been the distinction in his later writings
concerning religious motivation or “orientation.”Much research with his scales for measuring
intrinsic versus extrinsic (I-E) religious orientations (Allport and Ross 1967) suggests that extrin-
sic religiousness, in which religion is “used”for personal or social benefit, is (weakly) positively
correlated with racism. However, intrinsic religiousness, in which religion is “lived”and rep-
resents a central organizing and motivating factor in one’s life, is essentially uncorrelated with
racism (see Donahue 1985; Fulton, Gorsuch, and Maynard 1999, for reviews). These findings
have led researchers such as Hunsberger (1995) to suggest that the I-E approach “has not lived up
to its promise to resolve [Allport’s] stated paradox that religion seems to both make and unmake
prejudice”(Hunsberger 1995:117).
This failure of religious orientation measures to account for the religion-prejudice relationship
has led researchers to examine other dimensions of religion for alternative explanations. One such
line of research has focused on the role of religious fundamentalism (RF), a style of belief that is
characterized by a militant belief system, a sense of one absolute truth, and a sense of a special
relationship with God (Altemeyer and Hunsberger 1992). Research has shown a consistent positive
relationship between RF and both racism and homosexual prejudice (Altemeyer and Hunsberger
1992; Hunsberger 1995, 1996; Wylie and Forest 1992).
One explanation that has been offered in turn for this RF-prejudice relationship is that it is
a byproduct of the effects of right-wing authoritarianism (RWA), a constellation of psychologi-
cal constructs related to authoritarian aggression, authoritarian submission, and conventionalism
(Altemeyer 1981, 1988; Hunsberger 1995). Several studies have shown that right-wing authoritari-
anism is more strongly associated with prejudice than is fundamentalism, and that fundamentalism
and authoritarianism are strongly correlated with each other (Altemeyer and Hunsberger 1992;
Hunsberger 1996; Wylie and Forest 1992). The role of RWA might also explain why prejudice
is predicted inversely from a quest orientation toward religion (Batson and Ventis 1982; Batson,
Shoenrade, and Ventis 1993). The open, questioning nature of the quest orientation in many ways
appears to reflect the antithesis of an authoritarian approach to religion (Hunsberger 1995).
The idea that authoritarianism is largely responsible for the positive association of fundamen-
talism and prejudice is consistent with the results of several studies. For example, Altemeyer and
Hunsberger (1992) found that statistically controlling for RWA using partial correlation techniques
reduced fundamentalism-prejudice correlations to nonsignificant levels, but not vice versa. Simi-
larly, Wylie and Forest (1992) found that RWA added significantly to a multiple regression equation
in predicting racism and homophobia, while religious fundamentalism did not. However, this latter
analysis included a large number of other variables, and the sample size (n=75) was rather small.
To further examine these relationships, Laythe, Finkel, and Kirkpatrick (2001) conducted
multiple regressions with fundamentalism and authoritarianism as predictors of racism and ho-
mosexual prejudice. Results showed that whereas RWA emerged as a potent positive predictor
of both forms of prejudice, fundamentalism became a significant inverse predictor of racism
(though it remained a positive predictor of homosexual prejudice). To replicate this finding, re-
gression analyses were performed on the data from Wylie and Forest (1992) and Altemeyer and
Hunsberger (1992), based on the correlations reported in their respective published papers. The
results of these analyses showed a similar pattern: in both data sets, fundamentalism was a sig-
nificant inverse predictor of racism while authoritarianism remained a positive predictor.
These findings not only suggest that RWA is responsible for the positive association between
fundamentalism and ethnic prejudice, but also point directly to Allport’s paradox. That is, reli-
gious fundamentalism specifically appears to both “make and unmake”prejudice as a function of
two distinct elements or dimensions. The militant aspect of fundamentalism, right-wing author-
itarianism, “makes”prejudice, but some other aspect of fundamentalism (that becomes visible
when RWA is statistically partialed out) simultaneously “unmakes”prejudice.
This analysis then leads to the question of what this second component of religious fundamen-
talism might be. It seems to us that what defines RF apart from its general authoritarian orientation
TWO-COMPONENT MODEL OF FUNDAMENTALISM 625
is the content of the religious beliefs themselves. In the research literature this construct has been
referred to as Christian orthodoxy (CO), which is highly correlated with fundamentalism (e.g.,
Kirkpatrick 1993) and has often been confused with it conceptually (Hunsberger 1995). Christian
orthodoxy reflects the content of what is believed rather than (as in the case with fundamentalism)
the way the beliefs are held (Hunsberger 1995). Indeed, Allport’s paradox emerges in part from
the observation that Christian teachings, that is, orthodox beliefs, explicitly proscribe many forms
of prejudice, emphasizing brotherhood and love for one’s neighbors irrespective of race.
This reasoning leads to the prediction that Christian orthodoxy itself should be inversely
related to prejudice. However, this possibility has probably been overlooked because psychome-
tric measures of CO tend to be uncorrelated with prejudice (Altemeyer and Hunsberger 1992;
Kirkpatrick 1993). According to our line of reasoning above, however, these findings are likely a
result of the fact that measures of CO are confounded with fundamentalism (and thus indirectly
with RWA). Therefore, the militant attitudes shared by RF and RWA may confound a potentially
negative relationship between CO and prejudice such that CO measures appear to be uncorrelated
with prejudice.
Consistent with this interpretation, Kirkpatrick (1993) found that CO emerged as a negative
predictor of prejudice when RF was also included in the regression model (i.e., was statistically
controlled). Thus, the near-zero correlation between CO measures and prejudice itself appears to
be the result of two opposing effects: whereas the confounding effect of RWA pushes the CO-
racism correlation in a positive direction, orthodox belief per se may be pushing it in the opposite
direction, leading to an overall correlation near zero. This would explain why Kirkpatrick (1993)
found fundamentalism to become a more positive predictor of racism when CO was controlled,
and why Laythe, Finkel, and Kirkpatrick (2001) found fundamentalism to predict racism inversely
when RWA was statistically controlled.
Although the studies reviewed above have examined various combinations of religion vari-
ables as predictors of prejudice, none includes the crucial combination of RWA, CO, and RF that
is required to tease apart fully these effects in the prediction of prejudice. Kirkpatrick’s (1993)
study included CO and fundamentalism but not RWA; Altemeyer and Hunsberger (1992) and
Wylie and Forest (1992) examined fundamentalism and RWA but not CO. The present study was
therefore designed to provide a critical test of the hypothesis that RWA and CO are the principal
elements within fundamentalism that predict racism. When all three variables are included in a
regression analysis, so that each is statistically unconfounded from the others, we predicted that
(1) RWA will be positively related to racism, (2) Christian orthodoxy will be negatively related
to racism, and (3) fundamentalism, with the effects of RWA and CO removed, will be unrelated
to racism.
With respect to homosexual prejudice, however, it is not clear if the same results will be
observed. Laythe, Finkel, and Kirkpatrick (2001) found that fundamentalism remained a positive
(albeit weak) predictor of homosexual prejudice even when RWA was statistically controlled.
Although modern Christianity explicitly proscribes racism (Batson, Shoenrade, and Ventis 1993),
it does not necessarily proscribe prejudice against gays and lesbians. Although RWA is again
expected to emerge as a strong positive predictor of homosexual prejudice, it is less clear what
the respective effects of CO and fundamentalism will be when RWA (and each of the other two)
are statistically controlled.
METHOD
Participants and Procedure
Participants were recruited from introductory psychology and sociology classes from two
midwestern universities (Sample 1: 18 men, 67 women, total n=89; Sample 2: 44 men,
85 women, total n=129), and one southeastern university (Sample 3: 56 men, 43 women,
626 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
total n=100), comprising an aggregate sample of 318. Students participated on a voluntary basis
or for extra credit and completed the questionnaires during their regularly scheduled classes.
Measures
Four measures were used from Altemeyer and Hunsberger (1992): the Religious Fundamen-
talism Scale (20 items), the Manitoba Prejudice Scale1(19 items), the Right Wing Authoritarian-
ism Scale (30 items), and the Attitudes Towards Homosexuals Scale (12 items). A 24-item scale
designed by Fullerton and Hunsberger (1982) to measure Christian orthodoxy was also included.
In Samples 1 and 2, all measures were presented on a nine-point Likert-style scale anchored at 1
(strongly disagree), 5 (neutral), and 9 (strongly agree); participants in Sample 3 used five-point
response scales. All scales evinced high internal consistency reliability, with all within-sample
alpha coefficients greater than 0.85.
RESULTS
Each analysis presented in this section was conducted separately within each sample, as well
as on an aggregated total sample. Because the response format was different across samples (i.e.,
five-point vs. nine-point scales), the aggregate sample was constructed by first converting all
variables to zscores separately within each sample, and then combining the z-score variables into
an aggregate data set. One incidental advantage of this procedure is that because each zvariable
has the same mean (i.e., zero) within each sample, mean differences between samples on any of
the variables cannot influence or confound the results. The standardized beta (β) coefficients for
predictor variables therefore represent the average within-sample coefficients for each predictor.
In the tables we present results of each analysis separately for each sample, as well as for the
aggregate sample. In the text, however, all references to results refer to those for the aggregate,
total sample unless explicitly indicated otherwise.
Zero-Order Correlations
Zero-order correlations among the variables, which are presented in Table 1, were consistent
with previous research (Altemeyer and Hunsberger 1992; Laythe, Finkel, and Kirkpatrick 2001;
Wylie and Forest 1992). The three predictor variables of interest—authoritarianism, orthodoxy,
and fundamentalism—were all strongly intercorrelated with one another (rs=0.51 to 0.72). In
particular, the correlation between fundamentalism and RWA was extremely high in all samples
(r=0.72).
With respect to correlations of the prejudice measures, RWA was correlated with racism.
Fundamentalism was also correlated to racism, but to a lesser degree than RWA. However,
both fundamentalism and RWA were very strongly correlated with homosexual prejudice. Or-
thodoxy was only weakly correlated with prejudice variables, negatively with racial prejudice
(r=−0.10,p<0.10), and positively with homosexual prejudice.
Predicting Prejudice from Fundamentalism and Authoritarianism
The first set of analyses was performed to replicate the results reported by Laythe, Finkel,
and Kirkpatrick (2001), in which fundamentalism was shown to emerge as an inverse predictor
of racial prejudice when RWA was statistically controlled. Details are presented in Table 2.
Consistent with Laythe, Finkel, and Kirkpatrick (2001), the regression equation predicting racial
prejudice revealed RWA to be a significant positive predictor (β=0.54), but fundamentalism to
be a significant negative predictor (β=–0.26). In the second equation, RWA was again a strong
TWO-COMPONENT MODEL OF FUNDAMENTALISM 627
TABLE 1
ZERO-ORDER CORRELATIONS AMONG PREJUDICE MEASURES
AND PREDICTOR VARIABLES
Variable RWA FUND ORTH RPREJ
FUND a. 0.70∗∗
b. 0.70∗∗
c. 0.77∗∗
d. 0.72∗∗
ORTH a. 0.46∗∗ a. 0.59∗∗
b. 0.43∗∗ b. 0.47∗∗
c. 0.63∗∗ c. 0.67∗∗
d. 0.51∗∗ d. 0.58∗∗
RPREJ a. 0.31∗∗ a. 0.08 a. −0.15
b. 0.30∗∗ b. 0.07 b. −0.29∗∗
c. 0.45∗∗ c. 0.25∗c. 0.14
d. 0.35∗∗ d. 0.13∗d. −0.10++
HPREJ a. 0.57∗∗ a. 0.51∗∗ a. 0.10 a. 0.42∗∗
b. 0.57∗∗ b. 0.59∗∗ b. 0.26∗b. 0.33∗∗
c. 0.53∗∗ c. 0.42∗∗ c. 0.18 c. 0.76∗∗
d. 0.56∗∗ d. 0.51∗∗ d. 0.17∗∗ d. 0.50∗∗
Note:RWA=right-wing authoritarianism; FUND =religious fundamentalism; ORTH =Christian ortho-
doxy; RPREJ =racial prejudice; HPREJ =homosexual prejudice. Within each cell, a =Sample 1 (n=
129); b =Sample 2 (n=89); c =Sample 3 (n=100); and d=aggregate sample (n=318).
∗∗ p<0.01; ∗p< 0.05; +p<0.10.
TABLE 2
REGRESSIONS OF PREJUDICE MEASURES ON AUTHORITARIANISM
AND FUNDAMENTALISM
Authoritarianism Fundamentalism
Dependent Variable βtβtR
2adj.
Racial prejudice
Sample 1 0.496 4.26∗∗ −0.270 −2.32∗0.118
Sample 2 0.4840 3.45∗∗ −0.272 −1.94+0.105
Sample 3 0.636 4.58∗∗ −0.241 −1.74+0.211
Aggregate 0.5350 7.16∗∗ −0.258 −3.45∗∗ 0.148
Homosexual prejudice
Sample 1 0.403 3.99∗∗ 0.230 2.28∗0.335
Sample 2 0.321 2.74∗∗ 0.364 3.12∗∗ 0.384
Sample 3 0.497 3.72∗∗ 0.046 0.34 0.269
Aggregate 0.399 6.02∗∗ 0.219 3.30∗∗ 0.329
∗∗ p<0.01; ∗p<0.05; +p<0.10.
positive predictor of homosexual prejudice (β=0.40), but fundamentalism also emerged as a
significant and positive predictor (β=0.22).
Predicting Prejudice from Fundamentalism and Orthodoxy
To replicate the findings of Kirkpatrick (1993), where orthodoxy was shown to emerge as
a negative predictor of prejudice when fundamentalism was statistically controlled, a second set
628 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
TABLE 3
REGRESSIONS OF PREJUDICE MEASURES ON ORTHODOXY
AND FUNDAMENTALISM
Orthodoxy Fundamentalism
Dependent Variable βtβtR
2adj.
Racial prejudice
Sample 1 −0.302 −2.82∗∗ 0.258 2.41∗0.050
Sample 2 −0.408 −3.59∗∗ 0.257 2.26∗0.114
Sample 3 −0.051 −0.38 0.280 2.10∗0.042
Aggregate −0.261 −3.88∗∗ 0.280 4.16∗∗ 0.055
Homosexual prejudice
Sample 1 −0.323 −3.56∗∗ 0.705 7.77∗∗ 0.320
Sample 2 −0.015 −0.16 0.594 6.01∗∗ 0.330
Sample 3 −0.196 −1.60 0.558 4.56∗∗ 0.186
Aggregate −0.192 −3.27∗∗ 0.619 10.51∗∗ 0.276
∗∗ p<0.01; ∗p<0.05; +p<0.10.
of regressions parallel to those above was conducted in which RWA was replaced by Christian
orthodoxy as a predictor variable (along with fundamentalism). Results are presented in Table 3.
In these analyses, CO was a significant negative predictor of both forms of prejudice (βs=
−0.26,−0.19) whereas fundamentalism became a significant positive predictor of both forms of
prejudice with CO statistically controlled (βs=0.28, 0.62).
Predicting Prejudice from All Three Variables
The preceding analyses showed that fundamentalism emerges as a positive predictor of both
forms of prejudice when Christian orthodoxy is statistically controlled, but an inverse predic-
tor when RWA is statistically controlled. Both these findings are consistent with our hypothesis
that the relationship between fundamentalism and prejudice is a function of the two opposing
influences of CO and RWA. When either of these components is held constant, the remaining
variance in fundamentalism reflects the influence of the other. The critical test of our hypothesis,
however, concerns the question of what happens to the fundamentalism-prejudice relationships
when both CO and RWA are controlled simultaneously. According to our two-process model, CO
should again emerge as a strong inverse predictor of racism and RWA a strong positive predictor.
If these two variables can completely account for the fundamentalism-racism relationship, fun-
damentalism should be unrelated to racism with CO and RWA controlled.
Thus, the measures of Christian orthodoxy, right-wing authoritarianism, and fundamentalism
were entered simultaneously as predictor variables in separate regression equations for racism and
homosexual prejudice. Results are shown in Table 4. Consistent with the previous analyses, CO
was a significant negative predictor of racism (β=−0.34) and RWA was a significant positive
predictor of racism (β=0.59). With the positive effects of RWA and the negative effects of CO
both statistically controlled, the remaining variance in fundamentalism was no longer related to
racism (β=−0.10, ns).
Results for the prediction of homosexual prejudice followed the same pattern with respect to
RWA and CO: RWA was a significant positive predictor (β=0.44) and CO a significant negative
predictor (β=−0.25). However, unlike the equation predicting racism, fundamentalism also
contributed significantly and positively to the prediction of homosexual prejudice (β=0.33).
TWO-COMPONENT MODEL OF FUNDAMENTALISM 629
TABLE 4
REGRESSIONS OF PREJUDICE MEASURES ON AUTHORITARIANISM,
FUNDAMENTALISM, AND ORTHODOXY
Authoritarianism Fundamentalism Orthodoxy
Dependent Variable βtβtβtR
2adj.
Racial prejudice
Sample 1 0.524 4.67∗∗ −0.089 −0.72 −0.336 −3.38∗∗ 0.185
Sample 2 0.584 4.56∗∗ −0.113 −0.86 −0.487 −4.68∗∗ 0.280
Sample 3 0.688 4.86∗∗ −0.152 −1.01 −0.193 −1.57∗∗ 0.223
Aggregate 0.593 8.20∗∗ −0.104 −1.35 −0.337 −5.45∗∗ 0.219
Homosexual prejudice
Sample 1 0.432 4.53∗∗ 0.419 3.97∗∗ −0.351 −4.15∗∗ 0.411
Sample 2 0.333 2.80∗∗ 0.384 3.16∗∗ −0.060 −0.63 0.379
Sample 3 0.584 4.39∗∗ 0.192 1.37 −0.317 −2.74∗∗ 0.315
Aggregate 0.443 6.80∗∗ 0.333 4.81∗∗ −0.249 −4.47∗∗ 0.367
∗∗ p<0.01; ∗p<0.05; +p<0.10.
TABLE 5
REGRESSION OF FUNDAMENTALISM ON AUTHORITARIANISM
AND ORTHODOXY
Authoritarianism Orthodoxy
Sample βtβtR
2adj.
Sample 1 0.543 8.40∗∗ 0.346 5.35∗∗ 0.579
Sample 2 0.607 7.32∗∗ 0.207 2.50∗0.509
Sample 3 0.567 7.31∗∗ 0.316 4.07∗∗ 0.639
Aggregate 0.571 13.56∗∗ 0.296 7.03∗∗ 0.581
∗∗ p<0.01; ∗p<0.05; +p<0.10.
Predicting Fundamentalism from Christian Orthodoxy and Authoritarianism
Finally, a regression was performed to verify that authoritarianism and Christian orthodoxy
function as two of the major components of fundamentalism. These results are presented in Table 5.
Results show that both variables contributed significantly and independently to the prediction of
fundamentalism scores (βs=0.57 for RWA and 0.30 for CO). Overall, the two variables predicted
nearly 60 percent of the variance in fundamentalism according to the adjusted R2coefficient.
DISCUSSION
The present study was conducted to assess three sets of questions designed to test a two-
component model of religious fundamentalism and its empirical relationship to prejudice. First,
does the fundamentalism-racism relationship become negative when RWA is statistically con-
trolled (as in Laythe, Finkel, and Kirkpatrick 2001), and does this relationship become more
positive when Christian orthodoxy is statistically controlled (as in Kirkpatrick 1993)? Second,
if RWA and CO are statistically controlled simultaneously, is the fundamentalism-racism rela-
tionship eliminated? Third, are right-wing authoritarianism and Christian orthodoxy themselves
independent and additive predictors of fundamentalism? The answer to all three questions was
630 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
“yes.”In general, the results support a two-component model of fundamentalism as a predictor
of racial prejudice.
Predicting Racial Prejudice
As predicted, controlling for RWA resulted in a negative predictive relationship between
fundamentalism and racism (as in Laythe, Finkel, and Kirkpatrick 2001), whereas controlling
instead for CO resulted in a positive predictive relationship between fundamentalism and racism
(as in Kirkpatrick 1993). When the effect of one component of fundamentalism is partialed out,
what remains of fundamentalism reflects the effect of the other component.
The next question was whether fundamentalism retains any unique predictive power of its
own after the effects of both RWA and CO are removed. As predicted, in a multiple regression
containing all three predictors, (1) RWA emerged as a significant positive predictor of racism,
(2) CO emerged as a significant negative predictor, and (3) fundamentalism was a nonsignificant
predictor. Consistent with our proposed model, the mutually opposing effects of RWA and CO
appear to account adequately for the fundamentalism-racism relationship.
It has been more than 50 years since Allport (1950) first pointed to the paradox that reli-
giousness was positively correlated with racism despite the fact that Christian teaching explicitly
proscribes racial prejudice. Our results suggest that the answer to this paradox may be much sim-
pler than the last several decades of conflicting research would suggest. Adherence to orthodox
Christian belief per se is indeed inversely related (albeit weakly) to racial prejudice, just as most
Christians, religious leaders, and researchers (including, no doubt, Allport himself) might have
hoped. The problem is that adherence to Christian orthodox belief—or, at least, psychometric
instruments designed to measure it—is empirically associated with authoritarianism, which has
the opposite effect of increasing racism. The high correlations among the variables mean that
people who are highly orthodox or fundamentalistic, but simultaneously low on authoritarianism,
are relatively rare; however, our results suggest that such individuals might well be the least prej-
udiced of all. These results further suggest that if the effects of authoritarianism could somehow
be eliminated, orthodox Christian belief and even Christian fundamentalism would be inversely
associated with racism, consistent with its explicit message.
It is important to note that these conclusions may be restricted to populations in which
prevailing attitudes and beliefs proscribe racial or ethnic prejudice. According to Allport (1954),
prejudiced thinking is largely dependent on the beliefs proscribed by the group an individual
belongs to. Therefore, Christian orthodoxy might well be positively related to racial or ethnic
prejudice in populations in which local beliefs fail to proscribe, or even encourage, racism (see
Batson, Shoenrade, and Ventis 1993; Coward 1986; Hunsberger 1995). For example, Griffin,
Gorsuch, and Davis (1987) found intrinsic religious orientation to be positively correlated with
prejudice against Rastafarians among Seventh-Day Adventists in St. Croix, where such prejudice
was not generally regarded as socially undesirable. Similar results might well have been found
for Christian orthodoxy had it been measured by Griffin, Gorsuch, and Davis (1987) instead of
intrinsic religiosity (with which orthodoxy generally is very highly correlated).
Predicting Homosexual Prejudice
Many of the results in the prediction of prejudice against gays and lesbians paralleled those
for racial prejudice. Right-wing authoritarianism was positively correlated with homosexual prej-
udice and remained a strong positive predictor in all regression equations, irrespective of which
other predictor(s) were included, just as in the racism analyses. Although Christian orthodoxy
was (modestly) positively correlated with homosexual prejudice, in contrast to its small inverse
correlation with racism, CO again emerged as a significant negative predictor when fundamental-
ism and/or authoritarianism were controlled. In this sense, orthodox Christian belief per se, when
TWO-COMPONENT MODEL OF FUNDAMENTALISM 631
empirically disentangled from the confounding influences of authoritarianism and fundamental-
ism, appears to be a factor that “unmakes”prejudice against gays and lesbians much as it does
with racial prejudice.
On the other hand, the results for religious fundamentalism as a predictor of homosexual
prejudice differed from those with respect to racism. First, the zero-order correlation of funda-
mentalism with homosexual prejudice was very strong and positive, in contrast to its near-zero
correlation with racism. Second, and more important with respect to our model, fundamental-
ism remained a significant positive predictor of homosexual prejudice in all regression analyses,
even when authoritarianism was statistically controlled. Evidently, fundamentalism (at least as
measured in this study) contains some third component independent of both CO and RWA that is
itself positively related to homosexual prejudice.
The key to understanding the relationship between homosexual prejudice and fundamen-
talism may lie in the content of the particular beliefs that are being held in a fundamentalistic
way. With the exception of certain extremist groups such as white supremacists, few Christian
fundamentalists find a theological basis for racial prejudice in scripture. In contrast, many bibli-
cal literalists see homosexuality as a moral abomination explicitly proscribed by scripture. Such
specific theological beliefs would be expected to contribute independently to negative attitudes
about gays and lesbians above and beyond the general antipathy toward out groups associated
with authoritarianism. That is, authoritarians tend to be prejudiced against gays and lesbians, but
authoritarians who additionally hold specific religious beliefs about the immorality of homosex-
uality are more prejudiced still. This is essentially the same argument made by Fulton, Gorsuch,
and Maynard (1999), stated in reverse. These authors suggested that although fundamentalists’
antipathy toward homosexuals was in part a function of their theological beliefs, it was “in excess
of what is required by their ideology”(1999:14). From our perspective, the “excess”is explained
by the correlated factor of right-wing authoritarianism, which, according to our results, indepen-
dently contributes to homosexual prejudice above and beyond the effects of fundamentalism (as
well as vice versa).
Another perspective on the problem is suggested by the ambivalent attitude toward gays
and lesbians held by many modern Christian churches. On the one hand, orthodoxy teaches
that homosexuality is a sin; on the other hand, Christians are encouraged to “hate the sin, but
love the sinner.”This again contrasts with the case of racism, which is proscribed almost uni-
versally in modern Christian churches (Batson, Shoenrade, and Ventis 1993). Thus, Christian
orthodoxy per se is related to the tolerant, love-the-sinner half of the pronouncement, whereas
fundamentalism per se is associated more strongly with the hate-the-sin half with respect to
homosexuals.
Methodological Issues: Measurement and Sampling
An obvious question arises about our results due to our use of college students as research par-
ticipants, namely, whether our results (and thus our model) are generalizable to other, nonstudent
populations. This is admittedly a potential limitation of our research and conclusions. However,
our use of college-student samples confers several useful advantages as well. The issues involved
warrant some discussion, particularly with respect to a topic such as the religion-prejudice re-
lationship that has been widely studied by researchers from multiple disciplines using diverse
research methodologies.
First, much previous psychological research on the religion-prejudice relationship has sim-
ilarly employed college samples. This is important because our goal is not only to present new
findings regarding the relationships between prejudice and our various predictor variables, but
also to suggest that many previous findings—particularly those regarding the apparent role of
religious fundamentalism in contributing to prejudice—might be usefully reinterpreted in light
of these insights. Had we examined a qualitatively different population, it would remain an open
632 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
question as to whether our explanation is applicable to previous research findings or is specificto
whatever population was studied instead.
Second, although our college-sample methodology admittedly suffers from limited general-
izability, it confers important advantages with respect to psychometrics. Proper measurement of
complex psychological constructs such as authoritarianism and social attitudes requires multiple-
item scales to ensure reasonable levels of reliability and validity. This results unavoidably in
lengthy questionnaires, which would be extremely difficult and resource-intensive to administer
in a large-scale study with a nationally representative sample. Questionnaires in large-scale stud-
ies, such as the widely used General Social Survey (GSS), typically are limited to only one or
a few questions on a given topic in order to keep questionnaire size sufficiently small to obtain
high levels of cooperation and participation. Moreover, researchers mining these data archives
subsequently must often settle for proxy measures not well designed for their own particular
purposes. This kind of methodology therefore tends to sacrifice measurement in favor of gen-
eralizability, whereas our methodology sacrifices generalizability for quality of measurement.
Both methodological approaches are clearly necessary in light of their different advantages and
disadvantages and, ideally, should inform each other. For example, our findings might provide a
basis for including a smaller number of items in a future General Social Survey to examine the
generalizability of our results.
Third, the fact that college students are a relatively homogeneous population confers other
methodological advantages as well. Although sample homogeneity clearly reduces external va-
lidity (i.e., generalizability), it tends to enhance internal validity. One powerful way to control
for correlated effects of potential confounding variables is to hold such variables constant, as
in rigorously controlled experimental designs: any variable with little or no variance in a sam-
ple cannot provide an alternative explanation for the pattern of results obtained. In nonexperi-
mental research with heterogeneous samples it is common to use statistical methods to control
for potential confounds, but these methods are always only approximate. In college samples,
age and a variety of other demographic characteristics are essentially held constant, thus lim-
iting the possible confounds and making it unnecessary to statistically control for such factors
as age.
In addition, we strongly suspect that the homogeneity of our samples probably renders our
findings somewhat conservative: if anything, we would expect the patterns of findings to be
even stronger in more heterogeneous populations. Relative to the general (U.S.) population,
college students on average tend to be less authoritarian, less religiously fundamentalistic, and
less prejudiced. This produces a statistical range-restriction problem that is likely to have the
effect of reducing the magnitudes of the correlations obtained among such variables. The fact that
we found such strong results despite restricted range is encouraging, and it seems reasonable to
expect that even stronger relationships would be observed in more heterogeneous samples.
Finally, it is worth noting that there is no obvious reason to expect the complex patterns
of relationships among variables to differ across populations. For some kinds of parameters,
considerable variability across populations is likely: for example, there are many reasons to
expect that the means of our college samples for variables such as authoritarianism and prejudice
would be poor estimates of the means of the total U.S. population. However, it is much less
obvious why the covariation between these variables, as indexed by correlation or regression
coefficients, should vary greatly across populations. To illustrate, consider the fact that although
the mean height and mean weight of student athletes probably both differ substantially from the
respective means of nonathletes, the correlation between height and weight is probably about the
same within both populations: within either group, taller people tend to be heavier on average
and shorter people lighter. This does not prove that results similar to ours would be found in other
populations, of course; the point is merely that just because a sample is not representative of other
populations in terms of means of variables does not imply necessarily that it is nonrepresentative
with respect to relationships between variables.
TWO-COMPONENT MODEL OF FUNDAMENTALISM 633
Conclusions
Several new findings in the current study help to clarify an understanding of the relationship
between religion and prejudice. These findings show that the additive effects of two distinct
components of fundamentalism, Christian orthodoxy and right-wing authoritarianism, can account
for varying results in the research literature. Whereas Christian orthodoxy provides consistent
negative predictive power in regression of prejudice variables, Hunsberger (1995) appears to have
been correct in his conclusion that “fundamentalism might be viewed as a religious manifestation
of right wing authoritarianism”(1995:120). It is the militant attitudes held by authoritarians that
consistently predict racist attitudes when other variables are controlled.
The current research also sheds doubt on the viability of religious fundamentalism as a
predictor variable of racist tendencies. Our findings clearly suggest that the opposing forces of
right-wing authoritarianism and Christian orthodoxy account for the lion’s share of the variance
in prejudice, and that fundamentalism per se adds no unique predictive power of its own for racial
prejudice. In contrast, it does play a role in prejudice against gays and lesbians.
As noted by Hunsberger (1995), empirical research on the religion-prejudice relationship has
been dominated largely by Allport’s intrinsic-extrinsic distinction, but with little success. Empir-
ical measures of these dimensions, beginning with the scales first developed by Allport and Ross
(1967), have failed to shed much useful light on the subject. Hunsberger (1995) concluded that re-
ligious fundamentalism, in connection with the more general trait of right-wing authoritarianism,
is a much more useful construct than intrinsic-extrinsic religious orientation for understanding
religion’s relationship to prejudice. Our research suggests that another dimension of religiosity,
related to Christian orthodoxy, may also play an important role. In Allport’s (1950) terms, the
aspect of religion responsible for “making”prejudice appears to be right-wing authoritarianism,
whereas the aspect that “unmakes”prejudice is related to Christian orthodoxy.
Indeed, it seems to us that Allport was closer to the answer to the religion-prejudice riddle in
1954, prior to his formalization of intrinsic versus extrinsic dimensions of religious orientation.
In The Nature of Prejudice (1954:453) he wrote:
Belonging to a church because it is a safe, powerful, in-group is likely to be the mark of an authoritarian character
and to be linked to prejudice. Belonging to a church because its basic creed of brotherhood expresses the ideals
one sincerely believes in, is associated with tolerance. Thus, the “institutionalized”religious outlook and the
“interiorized”religious outlook have opposite effects in the personality.
Although it is easy to see the seeds of Allport’s later constructs of extrinsic and intrinsic
religious orientations here, there are two key ideas in this passage that have been lost in sub-
sequent research on these dimensions. First, the “institutionalized”dimension described here is
characterized specifically by ingroup-outgroup thinking and authoritarianism, in contrast to later
work (and psychometric measures) in which the extrinsic orientation is interpreted as reflecting
ulterior motives for religiousness in general, such as the maintenance of social relationships and
the desire for comfort and support from prayer (for a review of these distinctions, see Kirkpatrick
and Hood 1990). Second, in contrast to subsequent interpretations of the intrinsic orientation, the
“interiorized”dimension described here focuses not only on the devoutness of genuine belief,
but also on the specificcontent of that belief—specifically, belief in a “basic creed of brother-
hood.”Our results, in conjunction with those of Laythe, Finkel, and Kirkpatrick (2001) and other
researchers (Altemeyer and Hunsberger 1992; Hunsberger 1995), suggest that measures of right-
wing authoritarianism and of Christian orthodoxy, rather than extrinsic and intrinsic religious
orientation, better represent the two opposing forces in religion’s relationship to racial prejudice.
On the other hand, the somewhat different results obtained regarding prejudice against gays
and lesbians serve as an important reminder about the empirical and theoretical complexity of
the religion-prejudice question. First, “prejudice,”like “religion,”is not a unitary construct: the
634 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
question of how religion is related empirically to prejudice depends not only the particular di-
mension or aspect of religion in question—as researchers have acknowledged for some time—but
also the particular form of prejudice. We have focused mainly on the findings for racial prejudice
in this article because it has been such a widely researched topic in psychology of religion for
many years, but future research should examine other forms of prejudice without assuming that
similar results should be expected.
Second, we suspect that some of the crucial differences across domains of prejudice are
highly content-specific. One’s attitudes about gays and lesbians are driven, at least in part, by
specific beliefs about gays and lesbians that are largely independent of one’s specific beliefs about
ethnic minorities (or, for that matter, about women, persons with disabilities, or any other target
group). Some of these attitudes are rooted in specific religious beliefs. Psychologists of religion
have long attempted to conceptualize and measure religion with respect to abstract, content-
independent dimensions such as intrinsic/extrinsic orientation, means/ends/quest orientation, and
fundamentalism. The details of what people actually believe often seem to be overlooked in such
efforts. Our differential results with respect to racial and homosexual prejudice point, we think,
to the importance for religion researchers to attend more carefully to belief content in examining
the relationship between religion and other variables such as prejudice.
ACKNOWLEDGMENTS
The authors thank Brandy Burkett, Christine Petrone, Mindy Phillips, Lila Valencia, and Greg Webster for their
assistance in data collection.
NOTE
1. The Manitoba Prejudice Scale was slightly altered to accurately assess American prejudice. The method used was the
exact procedure used in Laythe, Finkel, and Kirkpatrick (2001), who reported strong overall α(0.88 to 0.93) for all
scales, despite the minor alterations to the scale.
REFERENCES
Allport, G. W. 1950. The individual and his religion. New York: Macmillan.
———. 1954. The nature of prejudice. Cambridge, MA: Addison Wesley.
Allport, G. W. and J. M. Ross. 1967. Personal religious orientation and prejudice. Journal of Personality and Social
Psychology 5:432–43.
Altemeyer, B. 1981. Right wing authoritarianism. Winnipeg: University of Manitoba Press.
———. 1988. Enemies of freedom: Understanding right wing authoritarianism. San Francisco, CA: Jossey-Bass.
Altemeyer, B. and B. Hunsberger. 1992. Authoritarianism, religious fundamentalism, quest, and prejudice. International
Journal for the Psychology of Religion 2:113–33.
Batson, C. D., P. Shoenrade, and W. L. Ventis. 1993. Religion and the individual: A social psychological perspective.
New York: Oxford University Press.
Batson, C. D. and W. L. Ventis. 1982. The religious experience, a social psychological perspective. New York: Oxford
University Press.
Coward, H. 1986. Intolerance in the world’s religions. Studies in Religion 15:419–31.
Donahue, M. J. 1985. Intrinsic and extrinsic religiousness: Review and meta-analysis. Journal of Personality and Social
Psychology 48:400–19.
Fullerton, T. J. and B. Hunsberger. 1982. A unidimensional measure of Christian orthodoxy. Journal for the Scientific
Study of Religion 21:317–26.
Fulton, A. S., R. L. Gorsuch, and E. A. Maynard. 1999. Religious orientation, anti-homosexual sentiment, and fundamen-
talism among Christians. Journal for the Scientific Study of Religion 38:14–22.
Griffin, G. A., R. L. Gorsuch, and A. L. Davis. 1987. A cross-cultural investigation of religious orientation, social norms,
and prejudice. Journal for the Scientific Study of Religion 26:358–65.
Hill, P. C. and R. W. Hood, Jr. 2000. Measures of religiosity. Birmingham, AL: Religious Education Press.
Hunsberger, B. 1995. Religion and prejudice: The role of religious fundamentalism, quest and right wing authoritarianism.
Journal of Social Issues 51(2):113–29.
TWO-COMPONENT MODEL OF FUNDAMENTALISM 635
———. 1996. Religious fundamentalism, right wing authoritarianism, and hostility in non-Christian religious groups.
International Journal for the Psychology of Religion 6:39–49.
Kirkpatrick, L. 1993. Fundamentalism, Christian orthodoxy, and intrinsic religious orientation as predictors of discrimi-
natory attitudes. Journal for the Scientific Study of Religion 32:256–68.
Kirkpatrick, L. and R. W. Hood, Jr. 1990. Intrinsic-extrinsic religious orientation: The boon or bane of contemporary
psychology of religion? Journal for the Scientific Study of Religion 29:442–62.
Laythe, B., D. Finkel, and L. Kirkpatrick. 2001. Predicting prejudice from religious fundamentalism and right-wing
authoritarianism: A multiple regression approach. Journal for the Scientific Study of Religion 40:1–10.
Wylie,L. and J. Forest. 1992. Religious fundamentalism, right wing authoritarianism and prejudice. Psychological Reports
71:1291–98.