Download full-text PDF

Negative Stereotypes Cause Christians to Underperform in and Disidentify With Science

Article · December 2015with3,890 Reads
DOI: 10.1177/1948550615598378
Abstract
Despite Christians being a religious majority in the U.S., relatively few pursue higher education and careers in science. Our studies show that stereotypes about Christians being less competent in science than other groups are recognized by both Christians and non-Christians, and are openly endorsed by non-Christians (Study 1). Our studies further demonstrate that when these stereotypes become salient, Christians are less interested in and identified with science (Study 2), and underperform on science-relevant tasks (Studies 3-5), compared to non-Christians. Even subtle contextual cues that bear more or less relevance to science are sufficient to compromise Christians’ scientific task performance, particularly among the highly religious (Study 5). When these stereotypes are explicitly removed, however, performance differences between Christians and non- Christians disappear. These results suggest that Christians’ awareness of the negative societal stereotypes about their group’s scientific competence may be partially responsible for the underperformance and underrepresentation of Christians in scientific fields.
Figures
Article
Negative Stereotypes Cause Christians
to Underperform in and Disidentify
With Science
Kimberly Rios
1
, Zhen Hadassah Cheng
2
, Rebecca R. Totton
1
,
and Azim F. Shariff
2
Abstract
Despite Christians being a religious majority in the United States, relatively few pursue higher education and careers in science.
Our studies show that stereotypes about Christians being less competent in science than other groups are recognized by both
Christians and non-Christians and are openly endorsed by non-Christians (Study 1). Our studies further demonstrate that when
these stereotypes become salient, Christians are less interested in and identified with science (Study 2) and underperform on
science-relevant tasks (Studies 3–5), compared to non-Christians. Even subtle contextual cues that bear more or less relevance to
science are sufficient to compromise Christians’ scientific task performance, particularly among the highly religious (Study 5).
When these stereotypes are explicitly removed, however, performance differences between Christians and non-Christians
disappear. These results suggest that Christians’ awareness of the negative societal stereotypes about their group’s scientific
competence may be partially responsible for the underperformance and underrepresentation of Christians in scientific fields.
Keywords
religiosity, scientific performance, negative stereotypes, social identity threat, stereotype threat
Although over 90%of Americans report believing in God
(Gallup, 2011), religious believers account for only 25%of
natural science faculty at elite American universities
(Calhoun, Aronczyk, Mayrl, & VanAntwerpen, 2007). In gen-
eral, there are lower proportions of religious believers among
science than humanities faculty (Gross & Simmons, 2009). What
explains these discrepancies? Given concerns about the dearth of
American students who ultimately pursue scientific careers
(Moss-Racusin et al., 2012), the lack of diversity of students
and faculty in the sciences (Committee on Equal Opportuni-
ties in Science and Engineering, 2014), and America’s low
scientific literacy (on which religious believers trail nonbelie-
vers; Sherkat, 2011), understanding the reasons is critical.
Previous psychological explanations have focused on two
differences between religious believers and nonbelievers: intel-
ligence and intuitive versus analytical thinking styles. For
instance, a recent meta-analysis found that religiosity and intel-
ligence test performance are inversely correlated (Zuckerman,
Silberman, & Hall, 2013). Other studies point to a disjuncture
between thinking styles that involve faith and intuition on one
hand and scientific and analytical thinking on the other hand.
Consistent with this disconnection, religious believers use
intuition more than do nonbelievers (Pennycook, Cheyne, Seli,
Koehler, & Fugelsang, 2012; Shenhav, Rand, & Greene, 2012).
In addition, experimentally inducing people to adopt an analy-
tical mind-set decreases subsequent religious belief (Gervais &
Norenzayan, 2012). Therefore, religious believers’ underrepre-
sentation in science is often explained as a product of their pre-
sumed lower intelligence, an incompatibility between intuitive
and analytical thinking styles, or both.
We propose an additional, novel factor: Negative societal
stereotypes about the scientific competency of Christians (who
comprise 94%of religious Americans; Gallup, 2012), and about
the perceived incompatibility between Christianity and science,
may cause Christians to disengage from and underperform in
scientific disciplines. The detrimental effects of negative stereo-
types on academic identification and performance in have been
documented across many groups, including women in science
(Cheryan, Plaut, Davies, & Steele, 2009; Murphy, Steele, &
Gross, 2007), African Americans in academics (Steele & Aron-
son, 1995), and low socioeconomic status (SES) individuals in
higher education (Stephens, Fryberg, Markus, Johnson, & Cov-
arrubias, 2012). In these studies, increasing the salience of one’s
group membership or the stereotype itself exacerbates the
1
Department of Psychology, Ohio University, Athens, OH, USA
2
Department of Psychology, University of Oregon, Eugene, OR, USA
Corresponding Author:
Kimberly Rios, Department of Psychology, Ohio University, 219 Porter Hall,
Athens, OH 45701, USA.
Email: rios@ohio.edu
Social Psychological and
Personality Science
1-9
ªThe Author(s) 2015
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1948550615598378
spps.sagepub.com
by guest on November 11, 2015spp.sagepub.comDownloaded from
stereotyped behaviors. For example, African Americans asked to
indicate their ethnicity on a demographic questionnaire subse-
quently underperform on standardized test questions and disiden-
tify with academics (Steele & Aronson, 1995). Additionally,
women underperform on mathematics tests when led to believe
there are gender differences in math ability (Dar-Nimrod &
Heine, 2006; Spencer, Steele, & Quinn, 1999), and they demon-
strate reduced scientific interest and performance when exposed
to cues suggesting that women do not belong in science (Cheryan
et al., 2009; Inzlicht & Ben-Zeev, 2000; Murphy et al., 2007).
These stereotypes need not be personally endorsed for such
effects to emerge. Indeed, the mere awareness that others may
endorse the stereotypes is sufficient to undermine academic per-
formance and identification (Steele, 1997). Furthermore, the
impact of negative stereotypes has been shown even for majority
groups that are not generally stigmatized in society as a whole.
European American men, for instance, are susceptible to under-
performance in mathematics when compared to Asian Ameri-
cans (Aronson et al., 1999) and in athletics when compared to
African Americans (Stone, Lynch, Sjomeling, & Darley, 1999).
If Christians disengage from and underperform in science
due to their perception of negative stereotypes about Christians
and science, these stereotypes may ultimately deter them from
scientific disciplines and careers, thereby perpetuating the orig-
inal stereotypes. Critically, however, when these negative
stereotypes are removed, Christians may identify as much with
science and perform as well on science-relevant tasks as non-
Christians, just as women exhibit equivalent scientific identifi-
cation and performance to men when stereotypic cues about
gender and science are removed (e.g., Cheryan et al., 2009;
Murphy et al., 2007; Spencer et al., 1999).
Across five studies, we tested both Christians’ and non-Chris-
tians’ awareness of negative stereotypes about Christians in sci-
ence (Study 1) as well as the impact of such stereotypes on
scientific identification (Study 2) and performance (Studies 3–5).
We focused on Christians both because they constitute the
majority of religious believers in the United States (Gallup,
2012) and because some preliminary evidence suggests that
Christians—unlike other religious groups (e.g., Jews and Mus-
lims)—tend to be stereotyped as less competent than warm
(Fiske, Cuddy, Glick, & Xu, 2002).
1
We sought to collect data
from approximately 200 participants in Study 1 (which tested
stereotype content) and 30 participants per cell in Studies 3–4
(which tested effects of stereotypes on Christians’ performance;
equivalent sample sizes were obtained in similar studies: Dar-
Nimrod & Heine, 2006; Mrazek et al., 2011). In Studies 2 and
5, we stopped data collection at the end of the semester.
Study 1
Method
Participants
Two hundred and two U.S. residents (102 women; M
age
¼
34.95, SD ¼12.09; 99 Christians, 103 non-Christians),
recruited from Amazon’s Mechanical Turk (mTurk) website,
participated online in exchange for US$.50. Thirty-three parti-
cipants were omitted for failing a basic attention check item,
leaving 169 in the final sample.
Procedure and Materials
After providing demographics and their religious affiliation,
participants rated four groups—atheists, Christians, Jews, and
Muslims—in random order on societal stereotypes. Specifi-
cally, they rated on a scale from 3 (e.g., less competent than
the average person) to 3 (e.g., more competent than the aver-
age person) whether the group was stereotyped to be higher
or lower than the average person in terms of overall compe-
tence, competence at science, trust in science, and warmth.
To assess whether participants actually endorsed societal
stereotypes about the four groups, participants subsequently
rated their personal beliefs about each group using the same
scales and characteristics.
Results and Discussion
One sample t-tests revealed that participants perceived Chris-
tians to be stereotyped as low in scientific competence and trust
in science compared to the scale midpoint (ps<.001).How-
ever, they perceived no stereotype about Christians’ general
competence (p> .250) and perceived Christians to be stereo-
typed as high in warmth (p< .001), suggesting that the negative
stereotypes were specific to science (see Table 1 and Figure 1).
Notably, both Christians (M
competence
¼.65, SE ¼1.66,
p< .005; M
trust
¼.92, SE ¼1.71, p< .001) and non-
Christians (M
competence
¼1.31, SE ¼1.30, p< .001;
M
trust
¼1.89, SE ¼1.26, p< .001) recognized the societal
stereotypes of Christians as low in competence in and trust of
science. Paired samples t-tests comparing Christians to each
of the target groups revealed that Christians were also per-
ceived to be stereotyped as lower in scientific competence (ts<
2.75, ps < .001) and trust in science (ts<3.16,ps < .010)
than Jews, Muslims, and atheists.
Regarding personal beliefs about Christians, Christian
participants believed their own group to be as competent in
(M¼0.14, SE ¼1.47, p> .250) and trusting of science
(M¼0.18, SE ¼1.69, p> .250) but more generally compe-
tent (M¼0.68, SE ¼1.42, p< .001) and warmer (M¼1.04,
SE ¼1.59, p< .001) than the average person. Non-
Christians, by contrast, personally believed Christians were
less competent in (M¼0.84, SE ¼1.31, p< .001) and trust-
ing of science (M¼1.46, SE ¼1.34, p< .001), less generally
competent (M¼0.34, SE ¼1.11, p¼.004), and equally
warm (M¼0.02, SE ¼1.24, p> .250) compared to the average
person (see Figure 2).
Study 2
Study 1 shows the general awareness of negative stereotypes
about Christians in science. Furthermore, non-Christians
2Social Psychological and Personality Science
by guest on November 11, 2015spp.sagepub.comDownloaded from
personally believe that Christians are inferior at and distrustful
of science, suggesting that the stereotypes are both pervasive
and socially acceptable. In Study 2, we tested the possibility
that these stereotypes may influence Christian college students’
own feelings about science, just as women’s and minorities’
feelings about science can be affected by stereotypes about
their groups (Cheryan et al., 2009; Murphy et al., 2007).
Method
Participants
One hundred psychology undergraduates (39 men, 61 women;
M
age
¼19.36, SD ¼1.88; 62 Christians, 35 non-Christians,
3 unspecified) completed a laboratory study for credit. Partici-
pants reported their religious affiliation in a prescreening sur-
vey. Three participants who suspected that the news article
was not real and three participants whose religious affiliation
was unspecified were dropped from the analyses as was one
outlier with an extreme Cook’s Dscore of .10 (5 SD above the
mean). The remaining 93 participants were randomly assigned
to the high-threat (n¼34), low-threat (n¼28), or no-article
(n¼31) condition.
Procedure and Materials
The study was described as assessing the relationship between
identity, interests, and abilities. Participants in the high-threat
(low-threat) condition first read an article allegedly published
by the local newspaper, presented as ‘‘background information.’
The article described the results of a bogus poll, suggesting that
most students at the university (78%) believed Christians were
bad (good) at science. Participants in the no-article condition
received the dependent measures without reading an article.
Next, participants completed a 20-item self-reported mea-
sure of their identification with science, adapted from Marsh
and O’Neill (1984; e.g., ‘‘I have never been very excited about
science’’ [reverse-coded], ‘‘I am quite good at science’’).
Results and Discussion
We predicted that Christians would identify less with science
than non-Christians after reading that Christians were
Table 1. Reported Stereotypes and Personal Beliefs by Group, Study 1.
Christians
Atheists Jews MuslimsMean (SD)
Stereotypes
Competence in science 1.00 (1.52)*** 1.26 (1.46)*** 0.59 (1.46)*** 0.59 (1.56)***
Trust of science 1.44 (1.56)*** 1.78 (1.47)** 0.37 (1.45)** 0.98 (1.48)***
General competence 0.06 (1.46) 0.20 (1.50)y1.07 (1.41)*** 0.56 (1.41)***
Warmth 0.75 (1.64)*** 0.95 (1.56)*** 0.31 (1.52)** 1.49 (1.43)***
Personal beliefs
Competence in science 0.38 (1.47)*** 0.96 (1.47)*** 0.49 (1.15)*** 0.22 (1.28)**
Trust of science 0.85 (1.64)*** 1.58 (1.53)*** 0.37 (1.21)*** 0.53 (1.27)***
General competence 0.14 (1.36) 0.46 (1.31)*** 0.69 (1.16)*** 0.15 (1.19)y
Warmth 0.50 (1.50)*** 0.12 (1.38) 0.12 (1.24) 0.51 (1.38)***
Note. N ¼169.
**p< .01; ***p< .001;
y
p< .10.
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
Competence
in science
Trust in science General
competence
Warmth
Perceived Stereotype Score (-3 to +3)
Christians
Atheists
Jews
Muslims
Figure 1. Reported stereotypes associated with each group, Study 1
(error bars represent standard errors of means).
-2
-1.5
-1
-0.5
0
0.5
1
1.5
Competence in
science
Trust in
science
General
competence
Warmth
Endorsed stereotype (-3 to +3)
Christians abou
t
themselves
Non-Christians
about Christians
Figure 2. Personal beliefs about Christians by group, Study 1.
Rios et al. 3
by guest on November 11, 2015spp.sagepub.comDownloaded from
stereotyped as bad at science but not after reading that Chris-
tians were stereotyped as good at science. We had no a priori
predictions about the control (no-article) condition.
A 3 (Condition: High-Threat vs. Low-Threat vs. No-Article)
2 (Religious Identity: Christian vs. Non-Christian) analysis
of variance revealed that Christians (M¼4.26, 95%confidence
interval [CI] ¼[4.06, 4.46]) reported weaker identification
with science than did non-Christians overall (M¼4.67, 95%
CI ¼[4.41, 4.94]), F(1, 87) ¼6.19, p¼.015, Zr
2
¼.07. How-
ever, this main effect was qualified by a two-way interaction,
F(2, 87) ¼3.59, p¼.032, Zr
2
¼.08. Simple effects tests indi-
cated that in the high-threat condition, Christians (M¼3.98,
95%CI ¼[3.67, 4.30]) identified significantly less with sci-
ence than non-Christians (M¼4.80, 95%CI ¼[4.34, 5.26]),
F(1, 87) ¼8.55, p¼.004, Zr
2
¼.09. This difference also
emerged in the control condition (Christians: M¼4.08, 95%
CI ¼[3.73, 4.43]; non-Christians: M¼4.72, 95%
CI ¼[4.28, 5.16]), F(1, 87) ¼5.21, p¼.025, Zr
2
¼.06. In the
low-threat condition, there was no significant difference
between Christians (M¼4.72, 95%CI ¼[4.28, 5.16])
and non-Christians (M¼4.50, 95%CI ¼[4.02, 4.97]),
F(1, 87) ¼.46, p> .250, Zr
2
¼.01 (see Figure 3).
Study 3
Study 2 suggests that Christians’ science identification tends to
suffer unless the antiscience stereotype is explicitly removed—
that is, unless Christians are given information suggesting that
they are just as competent in science as other groups. Perhaps
because of the social acceptability of expressing negative
stereotypes about Christians in science (demonstrated in Study
1), Christians may experience a default state of feeling that sci-
ence is incompatible with their religious identity, similar to
low-SES students in higher education (Stephens et al., 2012).
Notably, however, Christians identify as much with science
as non-Christians when reassured that others do not endorse the
negative stereotypes. We next tested the consequences of these
stereotypes for Christians’ scientific task performance.
Method
Participants
One hundred and eighty-three mTurk workers (103 men,
80 women; M
age
¼33.5, SD ¼11.7; 71 Christians, 112 non-
Christians) participated in exchange for US$.50. Eleven parti-
cipants were omitted: one for taking the study twice, three for
completing the study in 2 min or less, five for correctly gues-
sing the hypothesis, and two statistical outliers with Cook’s
Dscores above .035 (more than 3 SD above the sample mean).
The remaining 172 participants were retained.
Procedure and Materials
Participants were randomly assigned to either a high-threat
condition, in which they read a paragraph stating that Chris-
tians perform worse on scientific reasoning tasks than non-
Christians (n¼91), or a low-threat condition, in which they
read a paragraph stating that no performance differences
between Christians and non-Christians exist (n¼81). The
paragraphs were presented in the form of background informa-
tion (see Spencer et al., 1999).
Next, to measure performance on a task that participants
associated with scientific ability, all participants completed
an alleged ‘‘scientific reasoning test,’’ which involved indicat-
ing whether 15 syllogisms (i.e., sets of premises and conclu-
sions) reflected good or poor reasoning (e.g., All ghosts are
electrified. No cats are electrified. Therefore, no ghost is a cat;
Markman, Lindberg, Kray, & Galinsky, 2007). The total num-
ber of correct solutions served as the dependent measure.
Finally, in this study as well as in subsequent studies, partici-
pants completed a demographic survey and suspicion probe.
Results and Discussion
We predicted that being told there are differences in scientific
reasoning ability between religious groups would lead
Christian participants to solve fewer syllogisms correctly than
non-Christian participants. We expected Christians and non-
Christians to perform comparably when told that no differences
in scientific reasoning ability exist. We included participant
age as a covariate because there were age differences between
Christian and non-Christian participants (i.e., Christians were
older than non-Christians), F(1, 168) ¼13.50, p< .001 and
(marginally) between conditions (i.e., high-threat participants
were older than low-threat participants), F(1, 168) ¼3.35,
p¼.069.
2
A two-way analysis of covariance (ANCOVA) controlling
for age revealed an interaction between Christian identity and
threat condition, F(1, 167) ¼3.63, p¼.059, Zr
2
¼.02,
which—though marginal—was consistent with our hypotheses.
Specifically, Christians (M¼9.71, 95%CI ¼[8.92, 10.49])
underperformed relative to non-Christians (M¼11.59, 95%
CI ¼[11.01, 12.16]) in the high-threat condition, F(1, 167) ¼
14.26, p< .001, Zr
2
¼.08 but performed as well as non-
Christians in the low-threat condition (M
Christians
¼10.89, 95%
1
2
3
4
5
6
High threat No article
Condition
Low threat
Identiication with science score
Christians
Non-Christians
Figure 3. Identification with science as a function of threat condition
and Christian identity, Study 2.
4Social Psychological and Personality Science
by guest on November 11, 2015spp.sagepub.comDownloaded from
CI ¼[10.11, 11.68]; M
non-Christians
¼11.44, 95%CI ¼[10.81,
12.07]), F(1, 167) ¼1.12, p> .250, Zr
2
¼.007 (see Figure 4).
Additionally, there was a significant main effect of religious
identity (non-Christians outperformed Christians), F(1, 167)
¼11.08, p¼.001, Zr
2
¼.06.
Although Study 3 suggests that Christians’ awareness of
negative stereotypes about their group compromises science-
related task performance, it is unclear whether this perfor-
mance difference emerges even when the stereotypes are not
made salient or only when Christians read information impugn-
ing their group’s scientific ability. That is, without receiving
any information about religious differences in performance,
would the mere description of a task as measuring ‘‘scientific
reasoning’’ have been sufficient to elicit Christians’ underper-
formance relative to non-Christians? To address this question,
we recruited a separate sample of 213 mTurk workers
(117 Christians) to complete the syllogisms without reading
any background information. Similar to the high-threat condi-
tion in Study 3, a one-way ANCOVA controlling for age
revealed that Christians (M¼9.97, 95%CI ¼[9.55, 10.38])
solved (marginally) fewer syllogisms correctly than did non-
Christians (M¼10.51, 95%CI ¼[10.05, 10.97]), F(1, 210)
¼2.94, p¼.088, Zr
2
¼.01.
Study 4
Study 3 showed that when Christians believe a task is science
relevant, they underperform compared to non-Christians. That
this effect could only be completely mitigated by explicitly
removing the antiscience stereotype about Christians speaks
to the pervasiveness of this stereotype and supports the notion
that Christians may see scientific tasks as incompatible with
their religious identity.
Past research demonstrates that women in male-dominated
environments underperform on math problems but not verbal
problems (on which they experience no negative stereotype;
Inzlicht & Ben-Zeev, 2000). Because Study 1 showed that
Christians perceive themselves as targets of negative
stereotypes about scientific but not general competence, we
predicted that the negative effects of stereotype on performance
should only emerge on tasks that Christians believe to be sci-
ence relevant. Thus, in Study 4, we described the task as either
scientific or not.
We used the remote associates test (RAT) as our dependent
measure because its solution norms are well established (Dorf-
man, Shames, & Kihlstrom, 1996). This allowed us to compare
performance across relatively easy (45–80%solution rate) and
relatively difficult (20–40%solution rate) task items, in order
to investigate the mechanism behind our findings. If Christians’
underperformance is anxiety-driven, the performance differ-
ences should emerge on difficult stereotypic tasks (which are
more likely to trigger arousal) but not easy stereotypic tasks
(on which performance would be preserved and may even see
improvement; O’Brien & Crandall, 2003). However, if Chris-
tians’ underperformance is due to disengagement from science,
performance differences should emerge on both difficult and
easy stereotypic tasks, as stigmatized group members should
see both tasks as incompatible with their identity (see Stephens
et al., 2012).
Method
Participants
One hundred and twenty-eight mTurk workers (55 Christians,
72 non-Christians; 58 men, 69 women; M
age
¼35.2,
SD ¼12.7) participated in exchange for US$.50. Two partici-
pants who took the study twice and two statistical outliers
whose Cook’s Dscores (.15 and .07) fell at least 3 SD above
the sample mean were omitted, leaving 123 in the final sample.
Procedure and Materials
The procedure was identical to Study 3, with two exceptions.
First, to manipulate high or low threat, respectively, we
described the task as measuring scientific reasoning (n¼57)
or ‘‘intuitive thought’’ (n¼66). We also stated that we were
interested in performance differences between Christians and
non-Christians, although (unlike in Study 3 but consistent with
prior research, e.g., Spencer et al., 1999) we did not specify the
expected direction of the differences. Second, rather than sol-
ving syllogisms, participants completed 10 RAT items from
Kray, Galinsky, and Wong (2006), in which they were given
three words (e.g., golf,beans,andenvy) and had to generate
a fourth word related to all of them (green). All items were
randomized.
Results and Discussion
We predicted that Christians would generate fewer correct
solutions than non-Christians in the high-threat (scientific rea-
soning) condition but not in the low-threat (intuitive thought)
condition. A two-way ANCOVA controlling for age revealed
main effects of age, F(1, 118) ¼26.26, p< .001, Zr
2
¼.18,
and Christian identity, F(1, 118) ¼4.10, p¼.045,
Figure 4. Syllogisms solved (out of 15) as a function of threat con-
dition and Christian identity, Study 4.
Rios et al. 5
by guest on November 11, 2015spp.sagepub.comDownloaded from
Zr
2
¼.03, as well as the predicted interaction between
Christian identity and condition, F(1, 118) ¼4.67, p¼.033,
Zr
2
¼.04. Simple effects tests indicated that Christians
(M¼6.09, 95%CI ¼[5.13, 7.05]) performed worse than
non-Christians (M¼7.88, 95%CI ¼[7.08, 8.68]) when told
the task measured scientific reasoning, F(1, 118) ¼8.05,
p¼.005, Zr
2
¼.06, whereas Christians and non-Christians
performed equally well when told the task measured intui-
tive thought (M
Christians
¼6.94, 95%CI ¼[6.08, 7.80];
M
non-Christians
¼6.89, 95%CI ¼[6.13, 7.65]), F(1, 118) ¼
.01, p>.250,Zr
2
< .01 (see Figure 5).
Next, we tested whether Christians underperformed on sci-
entific reasoning tasks relative to non-Christians regardless of
the difficulty of the items (which would suggest that Christians
disidentify with science-relevant domains) or whether they
underperformed just on difficult items (which would suggest
that Christians experience anxiety in science-relevant
domains). A 2 (Christian Identity) 2 (Threat Condition)
2 (Easy vs. Difficult Items) mixed-model ANCOVA, with
repeated measures on the last factor, revealed only a main
effect of item difficulty (i.e., participants solved more easy than
difficult items correctly), F(1, 118) ¼14.01, p< .001, Zr
2
¼
.11, in addition to the effects described earlier. The three-way
interaction was not significant (p> .250). Thus, Christians’
underperformance on ‘‘scientific’’ tasks extends to both easy
and difficult items, suggesting that Christians disengage from
any task described as assessing scientific reasoning and not just
from tasks that are particularly difficult.
3
Study 5
If stereotypes compromise Christians’ scientific abilities in
everyday college contexts, then mere contextual cues relevant
to science should also trigger underperformance (Cheryan
et al., 2009; Murphy et al., 2007). Moreover, because the impact
of negative stereotypes is strongest among highly identified
group members (whose group membership is important to their
self-concept; Schmader, 2002), religiosity should moderate the
effects of such cues. We tested these hypotheses by having par-
ticipants complete a task in either a divinity school, which Chris-
tians should perceive as compatible with their (religious)
identity, or a physical sciences building, which Christians
should perceive as less compatible with their identity.
Method
Participants
One hundred and seven psychology students (43 men,
64 women; M
age
¼20.88, SD ¼3.74) participated in exchange
for credit. Prescreening revealed 39 participants as Christian
and 64 as non-Christian. Four participants did not specify their
religious affiliation and were dropped from analyses. Addition-
ally, 11 participants were omitted due to suspicion and two sta-
tistical outliers were excluded because their Cook’s Dscores
(.15 and .16) were more than 5 SD above the sample mean. The
final sample thus consisted of 90 individuals.
Procedure and Materials
One week prior to the study, participants completed an online
demographic survey, which included five religiosity questions
(e.g., ‘‘What is the general importance of God in your life?’’)
administered on 11-point scales (1 ¼not at all,11¼extremely;
Preston & Epley, 2009). The day before the study, participants
were e-mailed instructions to take the study in either the divi-
nity school (low-threat context; n¼47) or the physical
sciences building (high-threat context; n¼43). To increase
awareness of the context, upon arrival to their assigned build-
ing, all participants read an adapted mission statement of the
relevant department. Although the mission statement for the
divinity school stated that its faculty and students believed reli-
gion was reconcilable and compatible with other disciplines in
the humanities and sciences, the mission statement for the
physical sciences division did not mention any nonscience
disciplines.
Next, participants completed a measure described as a logi-
cal reasoning test, which consisted of 10 questions from the for-
mer Graduate Record Examination (GRE) analytical section
(e.g., ‘‘David ranks 7th from the top and 28th from the bottom
in a class. How many students are there in the class?’’ (a) 36,
(b) 35, (c) 34, (d) cannot be determined, (e) none of the above;
M¼5.31, SD ¼1.53).
Results and Discussion
We predicted that Christians’ religiosity would be negatively
correlated with the number of questions solved correctly in the
high-threat condition but not in the low-threat condition. In
other words, we expected religiosity to moderate the effects
of context (physical sciences building vs. divinity school) on
Christians’ performance. We expected no such effect for
non-Christians.
Because Christians (M¼7.32, SD ¼2.73 on an 11-point
scale) and non-Christians (M¼2.74, SD ¼1.96) differed
Figure 5. Remote associates test items solved as a function of threat
condition and Christian identity, Study 4.
6Social Psychological and Personality Science
by guest on November 11, 2015spp.sagepub.comDownloaded from
significantly in religiosity, t(88) ¼10.40, p< .001, we ana-
lyzed the data as two separate Context (0 ¼physical sciences,
1¼divinity school)Religiosity (mean centered) Interactions
for Christians and non-Christians, using multiple regression
(Aiken & West, 1991).
4
Confirming our hypothesis, a two-way threat Context
Religiosity multiple regression was significant for Christians
(b¼.72, SE ¼0.17), t(28) ¼4.34, p< .001, but not for non-
Christians (p> .250; see Figure 6). Among Christians, religios-
ity correlated negatively with performance in the high-threat
context (b¼.63, SE ¼0.14, t(28) ¼4.45, p< .001) and
was uncorrelated with performance in the low-threat context,
b¼.09, SE ¼0.09, t(28) ¼1.03, p> .250.
General Discussion
These studies demonstrate that increasing the salience of perva-
sive antiscience stereotypes about Christians (Study 1) can lead
Christians to underperform on scientific tasks (Studies 3–5)
and disidentify with science (Study 2), especially for the highly
religious (Study 5). Although the differences between
Christians and non-Christians disappear when the stereotypes
are explicitly removed, the overall effects of these stereotypes
are pernicious. As with other groups, Christians may face a per-
petuating cycle whereby they underperform due to the existing
stereotypes, thereby confirming those original stereotypes.
Through what mechanism(s) do the negative effects of these
stereotypes emerge? One possibility is that Christians’ anxiety
about confirming the stereotypes undermines their perfor-
mance and engagement (Steele & Aronson, 1995). Alterna-
tively, Christians may disidentify with fields perceived not to
‘match’’ their religious identity (i.e., science), either because
they believe that others stereotype them as not belonging in sci-
ence (Steele, Spencer, & Aronson, 2002; Stephens et al., 2012)
or because they themselves stereotype their (religious) values
as incompatible with science (see Nosek, Banaji, & Greenwald,
2002).
Study 4, in which Christians underperformed on both diffi-
cult and easy (presumably less anxiety inducing) scientific rea-
soning items, seems inconsistent with the anxiety explanation,
although future research could test the role of anxiety more
directly (e.g., by measuring emotions or working memory prior
to the task). Regarding the distinction between other stereotyp-
ing and self-stereotyping, Study 1 demonstrated that Christians
did not personally endorse antiscience stereotypes about their
group, despite their awareness of the existence of such stereo-
types. We thus doubt that self-stereotyping is the only reason
behind Christians’ underperformance. However, perhaps being
reminded of others’ negative stereotypes leads Christians to
eventually internalize perceptions of themselves and their
group as unscientific.
Unlike women and ethnic minorities, on whom much
research about negative intellectual stereotypes has focused,
American Christians are a dominant majority group (77%of
the population; Gallup, 2012) and not one generally perceived
as disadvantaged. Yet context matters, and in scientific
domains, different proportions and a potentially very different
climate exist. Christian underrepresentation in science may be
caused by self-selection (choosing not to enter science-related
fields) as well as underperformance (not succeeding in said
fields), both of which are exacerbated by negative stereotypes.
That Christians constitute such a large proportion of the U.S.
population means that factors discouraging their participation
in scientific disciplines may vastly impact the potential flow
of students into science-related careers. Diagnosing such bar-
riers is a critical step in developing strategies for encouraging
more people to pursue their scientific ambitions. Previous
research has found that the pernicious effects of negative
stereotypes on academic performance and interest can be miti-
gated by affirming personal values (Cook, Purdie-Vaughns,
Garcia, & Cohen, 2012), increasing feelings of belongingness
(Walton & Cohen, 2011), or exposing individuals to counter-
stereotypic role models (Marx & Goff, 2005). Future research
should tailor such strategies to increase Christians’ participa-
tion in science. If effective, these strategies would be vital tools
for increasing scientific involvement and literacy in American
society.
Figure 6. Correct solutions as a function of Christian identity, threat
context, and religiosity (+1SD), Study 5.
Rios et al. 7
by guest on November 11, 2015spp.sagepub.comDownloaded from
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to
the research, authorship, and/or publication of this article.
Funding
This research was supported in part by the John Templeton Foundation
(Award #13397 and #40515). Any opinions, findings, conclusions, or
recommendations expressed in this publication are those of the authors
and do not necessarily reflect the views of the John Templeton
Foundation.
Notes
1. In Studies 2–4, we asked Christian participants whether they iden-
tified as ‘‘Catholic’’ (range from 32%in Study 4 to 55%in Study
2), ‘‘Protestant’’ (range from 26%in Study 3 to 51%in Study 4), or
(Studies 3–4) ‘‘other’’ (32%in Study 3, 17%in Study 4). Among
Christians, there were no main effects of denomination or interac-
tions with condition on science identification (Study 2) or perfor-
mance (Studies 3–4; ps > .19). However, given our low statistical
power to detect such interactions, and the absence of more fine-
grained distinctions, we do not consider these null results conclu-
sive and believe the potential effects of denomination (and related
variables such as fundamentalism) warrant future investigation.
2. In Studies 3–4, highest education (1 ¼some high school,6¼doc-
torate or professional degree) was unrelated either to Christian
identity or condition (ps > .250), and controlling for it did not affect
our results.
3. When the easy and difficult items were analyzed separately, the
Condition Christian Identity Interaction was significant for dif-
ficult items (p¼.028) and approached significance for easy items
(p¼.109). Notably, the difference between Christians’ and non-
Christians’ performance in the scientific reasoning condition was
significant for both sets of items (ps¼.012).
4. The three-way interaction between condition, religious identity
(0 ¼non-Christian,1¼Christian), and religiosity was marginal
(b¼.58, SE ¼0.32), t(82) ¼1.85, p¼.068, likely because reli-
giosity (on which Christians scored higher than non-Christians)
was mean centered across the entire sample, regardless of religious
identity.
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and
interpreting interactions. Thousand Oaks, CA: Sage.
Aronson, J., Lustina, M. J., Good, C., Keough, C., Steele, C. M., &
Brown, J. (1999). When white men can’t do math: Necessary and
sufficient factors in stereotype threat. Journal of Experimental
Social Psychology,35, 29–46.
Calhoun, C., Aronczyk, M., Mayrl, D., & VanAntwerpen, J. (2007).
The religious engagements of American undergraduates.Brook-
lyn, NY: Social Science Research Council.
Cheryan, S., Plaut, V. C., Davies, P. G., & Steele, C. M. (2009). Ambi-
ent belonging: How stereotypical cues impact gender participation
in computer science. Journal of Personality and Social Psychol-
ogy,97, 1045–1060.
Committee on Equal Opportunities in Science and Engineering.
(2014). Broadening participation in America’s science and engi-
neering work force. Retrieved from http://www.nsf.gov/od/iia/
activities/ceose/reports/Full_2011-2012_CEOSE_Report_to_
Congress_Final_03-04-2014.pdf
Cook, J. E., Purdie-Vaughns, V., Garcia, J., & Cohen, G. L. (2012).
Chronic threat and contingent belonging: Protective benefits of
values affirmation on identity development. Journal of Personality
and Social Psychology,102, 479–496.
Dar-Nimrod, I., & Heine, S. J. (2006). Exposure to scientific theories
affects women’s math performance. Science,314, 435.
Dorfman, J., Shames, V. A., & Kihlstrom, J. F. (1996). Intuition, incu-
bation, and insight: Implicit cognition in problem solving. In G.
Underwood (Ed.), Implicit cognition (pp. 257–296). Oxford, Eng-
land: Oxford University Press.
Fiske, S. T., Cuddy, A. J. C., Glick, P., & Xu, J. (2002). A model of
(often mixed) stereotype content: Competence and warmth respec-
tively follow from perceived status and competition. Journal of
Personality and Social Psychology,82, 878–902.
Gallup. (2011). More than 9 in 10 Americans continue to believe in God.
Retrieved from http://www.gallup.com/poll/147887/americans-
continue-believe-god.aspx
Gallup. (2012). In U.S., 77%identify as Christian. Retrieved from
http://www.gallup.com/poll/159548/identify-christian.aspx
Gervais, W. M., & Norenzayan, A. (2012). Analytic thinking pro-
motes religious disbelief. Science,336, 493–496.
Gross, N., & Simmons, S. (2009). The religiosity of American college
and university professors. Sociology of Religion,70, 101–129.
Inzlicht, M., & Ben-Zeev, T. (2000). A threatening intellectual envi-
ronment: Why females are susceptible to experiencing problem-
solving deficits in the presence of males. Psychological Science,
11, 365–371.
Kray, L. J., Galinsky, A. D., & Wong, E. M. (2006). Thinking within the
box: The relational processing style elicited by counterfactual mind-
sets. Journal of Personality and Social Psychology,91, 33–48.
Markman, K. D., Lindberg, M. J., Kray, L. J., & Galinsky, A. D.
(2007). Implications of counterfactual structure for creative gener-
ation and analytical problem solving. Personality and Social Psy-
chology Bulletin,33, 312–324.
Marsh, H. W., & O’Neill, R. (1984). Self-description questionnaire
III: The construct validity of multidimensional self-concept ratings
by late adolescents. Journal of Educational Measurement,21,
153–174.
Marx, D. M., & Goff, P. A. (2005). Clearing the air: The effect of
experimenter race on target’s test performance and subjective
experience. British Journal of Social Psychology,44, 644–657.
Moss-Racusin, C. A., van der Toorn, J., Dovidio, J. F., Brescoll, V. L.,
Graham, M. J., & Handelsman, J. (2012). Scientific diversity inter-
ventions. Science,343, 615–616.
Mrazek, M. D., Chin, J. M., Schmader, T., Hartson, K. A., Smallwood,
J., & Schooler, J. W. (2011). Threatened to distraction: Mind-
wandering as a consequence of stereotype threat. Journal of
Experimental Social Psychology,47, 1243–1248.
Murphy, M. C., Steele, C. M., & Gross, J. J. (2007). Signaling threat:
How situational cues affect women in math, science, and engineer-
ing settings. Psychological Science,18, 879–885.
8Social Psychological and Personality Science
by guest on November 11, 2015spp.sagepub.comDownloaded from
Nosek, B. A., Banaji, M. R., & Greenwald, A. G. (2002). Math¼male,
me¼female, therefore mathme. Journal of Personality and
Social Psychology,83, 44–59.
O’Brien, L. T., & Crandall, C. S. (2003). Stereotype threat and arou-
sal: Effects on women’s math performance. Personality and Social
Psychology Bulletin,29, 782–789.
Pennycook, G., Cheyne, J. A., Seli, P., Koehler, D. J., & Fugelsang, J.
A. (2012). Analytic cognitive style predicts religious and paranor-
mal belief. Cognition,123, 335–346.
Preston, J. L., & Epley, N. (2009). Science and God: An automatic
opposition between ultimate explanations. Journal of Experimen-
tal Social Psychology,45, 238–241.
Schmader, T. (2002). Gender identification moderates stereotype
threat effects on women’s math performance. Journal of Experi-
mental Social Psychology,38, 194–201.
Shenhav, A., Rand, D. G., & Greene, J. D. (2012). Divine intuition:
Cognitive style influences belief in God. Journal of Experimental
Psychology: General,141, 423–428.
Sherkat, D. E. (2011). Religion and scientific literacy in the United
States. Social Science Quarterly,92, 1134–1150.
Spencer, S. J., Steele, C. M., & Quinn, D. M. (1999). Stereotype threat
and women’s math performance. Journal of Experimental Social
Psychology,35, 4–28.
Steele, C. M. (1997). A threat in the air: How stereotypes shape
intellectual identity and performance. American Psychologist,
52, 613.
Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellec-
tual test performance of African Americans. Journal of Personality
and Social Psychology,69, 797–811.
Steele, C. M., Spencer, S. J., & Aronson, J. (2002). Contending with
group image: The psychology of stereotype and social identity
threat. Advances in Experimental Social Psychology,34, 379–440.
Stephens, N. M., Fryberg, S. A., Markus, H. R., Johnson, C. S., & Cov-
arrubias, R. (2012). Unseen disadvantage: how American universi-
ties’ focus on independence undermines the academic performance
of first-generation college students. Journal of Personality and
Social Psychology,102, 1178.
Stone, J., Lynch, C. I., Sjomeling, M., & Darley, J. M. (1999). Stereo-
type threat effects on Black and White athletic performance. Jour-
nal of Personality and Social Psychology,77, 1213.
Walton, G. M., & Cohen, G. L. (2011). A brief social-belonging inter-
vention improves academic and health outcomes of minority stu-
dents. Science,331, 1447–1451.
Zuckerman, M., Silberman, J., & Hall, J. A. (2013). The relation
between intelligence and religiosity: A meta-analysis and some
proposed explanations, Personality and Social Psychology Review,
17, 325–354.
Author Biographies
Kimberly Rios is an Assistant Professor of Psychology at Ohio Uni-
versity. She studies individuals’ responses to self-concept and social
identity threats, and she has developed a particular interest in threats
to religious identity.
Zhen Hadassah Cheng is a PhD candidate at the University of Ore-
gon. She is interested in understanding people’s perceptions of atheists
and theists, and developing interventions to combat stereotype threat
and to increase diversity in academia.
Rebecca R. Totton is a PhD student in Social Psychology at Ohio
University. She is interested in the psychological experiences of
minority groups.
Azim F. Shariff is Assistant Professor of Psychology at the University
of Oregon. He studies the psychology of religion, among other things.
Rios et al. 9
by guest on November 11, 2015spp.sagepub.comDownloaded from
  • ... For example, Nadal and colleagues (2012) found the following religious microaggressions themes for Muslim Americans: (a) Endorsing Religious Stereotypes of Muslims as Terrorists, (b) Pathology of the Muslim Religion, (c) Assumption of Religious Homogeneity, (d) Exoticization, (e) Islamophobic and Mocking Language, and (f) Alien in Own Land. Although being stereotyped as terrorists may be unique to Muslim Americans, Jews may experience microaggressions due to being stereotyped as cheap ( Nadal et al., 2010), and Christians may experience microaggressions due to the stereotype that they are intolerant or unscientific ( Hyers & Hyers, 2008;Rios et al., 2015). Religious microaggressions differences may not only emerge as a result of belonging to different religious groups, but differences may also be pronounced when looking at members who are part of a religious majority versus those in a religious minority. ...
  • ... Preliminary research suggests religious threat can erode protective benefits of religion (e.g., buffering against stress; Ysseldyk, Matheson, & Anisman, 2011) and undermine interest and performance in intellectual domains (i.e., where negative stereotypes about religion are prevalent; Rios et al., 2015). People also may respond to religious threat by derogating out-group members and evaluating in-group members more favorably ( Hunter et al., 2004;Ysseldyk, Haslam, Matheson, & Anisman, 2012). ...
  • ... Although much research has examined the link between religion and prosociality toward people in general (Norenzayan et al., 2016;Shariff et al., 2015), fewer studies have considered how religious identity affects prosociality toward ingroup versus outgroup members (but seeEverett et al., 2016;Preston & Ritter, 2013), particularly when reputational concerns are salient. ...

Comment

May 10, 2016
Columbia University
What a fascinating concept!
Article
February 2012 · Journal of Biological Chemistry · Impact Factor: 4.57
    The therapeutic benefit of the small heat shock protein αB-crystallin (HspB5) in animal models of multiple sclerosis and ischemia is proposed to arise from its increased capacity to bind proinflammatory proteins at the elevated temperatures within inflammatory foci. By mass spectral analysis, a common set of ∼70 ligands was precipitated by HspB5 from plasma from patients with multiple... [Show full abstract]
    Article
    January 2007
      Cytokines, once thought to be specialized molecules of the immune system, are now being investigated also for their synaptic and inflammatory action on the central nervous system (CNS). These proteins and their receptors can be synthesized in the brain by glial and neuronal cells and contribute to two main types of action: modulation of neuronal excitability and local inflammatory responses.... [Show full abstract]
      Article
      May 2012 · Frontiers in Immunology
        Historically, small heat shock proteins (sHSPs) have been extensively studied in the context of being intracellular molecular chaperones. However, recent studies looking at the role of sHSPs in neurological diseases have demonstrated a near universal upregulation of certain sHSPs in damaged and diseased brains. Initially, it was thought that sHSPs are pathological in these disease states... [Show full abstract]
        Article
        June 2012 · Journal of Experimental Medicine · Impact Factor: 12.52
          Comparison of transcriptomic and proteomic data from pathologically similar multiple sclerosis (MS) lesions reveals down-regulation of CD47 at the messenger RNA level and low abundance at the protein level. Immunohistochemical studies demonstrate that CD47 is expressed in normal myelin and in foamy macrophages and reactive astrocytes within active MS lesions. We demonstrate that CD47(-/-) mice... [Show full abstract]
          Discover more