Personality and Individual Differences 131 (2018) 222–231

Contents lists available at ScienceDirect

Personality and Individual Differences

journal homepage: www.elsevier.com/locate/paid

Religion and moral self-image: The contributions of prosocial behavior, T

socially desirable responding, and personality
⁎
Sarah J. Ward , Laura A. King
University of Missouri, Columbia, United States

A R T I C LE I N FO A B S T R A C T

Keywords: Often, the high moral self-image held by religious people is viewed with skepticism. Three studies examined the
Religion contributions of socially desirable responding (SDR), personality traits, prosocial behavior, and individual dif-
Morality ferences in prosocial tendencies to the association between religiosity and moral self-image. In Studies 1 and 2
Moral self-image (N's = 346, 507), personality traits (agreeableness, conscientiousness) and individual differences in empathy/
Prosociality
prosociality were the strongest explanatory variables for religiosity's association with moral self-image measures;
SDR and prosocial behavior contributed more weakly to this association. In Study 3 (N = 180), the effect of a
bogus pipeline manipulation on moral self-image was moderated by religiosity. Among the highly religious,
moral self-image remained high even in the bogus pipeline condition. These studies show that the association
between religiosity and moral self-image is most strongly explained by personality traits and individual differ-
ences in prosociality/empathy, rather than a desirability response bias.

1. Introduction 2013; Saroglou et al., 2005).

People generally desire to perceive themselves as moral and strive
towards embodying moral traits (e.g., Aquino & Reed II, 2002; Merritt,
Effron, & Monin, 2010). Although this motivation is true of many
people, it is particularly heightened among the religious: Several stu-
dies demonstrate religiosity is positively associated with self-reports of
moral self-image and moral values (e.g., Furrow, King, & White, 2004;
Johnston, Sherman, & Grusec, 2013; Putnam, Campbell, & Garrett,
2010; Saroglou, Pichon, Trompette, Verschueren, & Dernelle, 2005;
Vitell et al., 2009; Ward & King, under review). The high morality
espoused by religious people has been critiqued extensively and viewed
with skepticism, as religion does not generally promote moral behavior
in controlled experiments or naturalistic settings (e.g., Batson, 1976;
Batson et al., 1989; Ji, Pendergraft, & Perry, 2006; McKay &
Whitehouse, 2015) and instead only seems related to more circum-
scribed instances of prosociality (e.g., Norenzayan & Shariff, 2008;
Saroglou et al., 2005). Why is religion persistently related to viewing
oneself as highly moral, despite equivocal evidence for a more general
link between religion and actual moral behaviors? Some scholars have
posited that the association between religiosity and moral self-image
may be an artifact of socially desirable responding (SDR; Shariff, 2015;
McKay & Whitehouse, 2015); others have proposed it may be more
genuine due to religious people's heightened prosocial behavior and
shared overlap with personality traits linked to morality (e.g., Saroglou,
To date, there have not been critical empirical investigations into
the association between religiosity and moral self-image. Much of the
extant literature has probed how religion relates to a diverse range of
moral values or prosocial personality traits without focusing on ex-
planations for these associations. Here we focus broadly on moral self-
image, rather than specific values or traits, seeking to understand why
religious people perceive themselves as highly moral. The present stu-
dies investigated how SDR, prosocial behavior, personality traits, and
individual differences in prosocial tendencies account for the relation-
ship between religiosity and moral self-image. Before presenting the
studies, we briefly review the mechanisms that may explain why re-
ligiosity is associated with viewing oneself as moral.
1.1. Socially desirable responding
Certainly, skepticism of the superior morality espoused by religious
people has been longstanding within psychology. Batson and colleagues
(Batson, 1976; Batson & Gray, 1981; Batson et al., 1989) posited that
religion relates to an egoistic motivation for morality, driven more by
reputational concerns rather than altruism. Correspondingly, the most
common explanation proffered to explain (and critique) the validity of
the association between religiosity and self-reported moral character-
istics is SDR (e.g., Galen, 2012; McKay & Whitehouse, 2015; Shariff,
2015). SDR is often conceptualized as involving two facets: impression

⁎
Corresponding author at: 210 McAlester Hall, Department of Psychological Sciences, University of Missouri, Columbia, MO 65211, United States.
E-mail address: sjhtg6@mail.missouri.edu (S.J. Ward).

https://doi.org/10.1016/j.paid.2018.04.028
Received 2 February 2018; Received in revised form 18 April 2018; Accepted 19 April 2018
0191-8869/ © 2018 Elsevier Ltd. All rights reserved.
S.J. Ward, L.A. King
management and self-deception (Paulhus & Reid, 1991). Contemporary
research and theory suggest that impression management and self-de-
ception should not be characterized as response biases but instead be
viewed as capturing substantive individual differences and motivations.
For instance, impression management has been argued to reflect
individual differences in interpersonal self-control (Uziel, 2010), emo-
tional stability, and conscientiousness (Ones, Viswesvaran, & Reiss,
1996). As opposed to reflecting dishonesty, impression management is
positively associated with moral virtues, including honesty-humility, as
well as to lower frequencies of immoral behavior in a low-demand
context (de Vries, Zettler, & Hilbig, 2014; Zettler, Hilbig, Moshagen, &
de Vries, 2015). Consequently, though traditionally thought of as
merely a form of desirability bias, impression management may instead
reflect genuine honesty and a moral orientation (e.g., Zettler et al.,
2015). Similarly, self-deception, once theorized as representing an un-
conscious tendency to maintain an unjustifiably good (though honestly
held) impression of oneself (Paulhus & Reid, 1991), is now thought to
be employed strategically to boost one's interpersonal appeal, status,
and resources (e.g., Koban & Ohler, 2016; Paulhus & John, 1998; von
Hippel & Trivers, 2011). If self-reports of SDR capture genuine in-
dividual differences in valuing morality and in self-control (e.g.,
Paulhus & Trapnell, 2008; Uziel, 2010; Zettler et al., 2015), then con-
trolling for these variables is unlikely to elucidate the real link between
religiosity and moral self-image. As such, alternative operationaliza-
tions of response biases may be needed to probe this link effectively.
Religiosity is moderately correlated with impression management
and more weakly correlated with self-deception (r's = .31 and .12, re-
spectively, in a meta-analysis; Sedikides & Gebauer, 2010). Corre-
spondingly, religious people report being better than other people on
both moral and non-moral traits (e.g., Gebauer, Sedikides, & Schrade,
2017; Ludeke & Carey, 2015; Rowatt, Ottenbreit, Nesselroade, &
Cunningham, 2002). Although religiosity is related to self-enhancing
tendencies, previous literature has demonstrated that the link between
religiosity and prosocial values is not fully explained by SDR. Con-
trolling for SDR, intrinsic religiosity predicts self-reported altruism and
empathy (Saroglou et al., 2005), volunteerism (Hansen, Vandenberg, &
Patterson, 1995), and adherence to biblical commands (Rowatt et al.,
2002). Still, the extant research on this topic is limited. First, these
studies have not measured moral self-image or moral identity, so it is
unclear how SDR accounts for religiosity's role in promoting views of
oneself as moral. Additionally, past studies used a limited set of ques-
tionnaires pertaining to morality, personality, and SDR. It may be
possible that another factor, or combination of factors, explains the
association between religiosity and moral self-image, but previous
studies may have failed to identify this because they used a limited set
of measures. Also, these studies sometimes failed to report the religious
affiliation of their participants (e.g., Hansen et al., 1995; Saroglou et al.,
2005) or utilized religiously homogenous participants (e.g., Rowatt
et al., 2002), potentially leading to a restricted range on the relation-
ships observed and limiting the generalizability of these results. Finally,
SDR may affect not only self-reports of moral self-image but also reports
of personality characteristics and prosocial behaviors. As such, it is
critical to measure these variables together to tease out how much their
relationships are accounted for by shared variance with SDR.
Even if SDR helps to account for the association between religiosity
and moral self-image, there may be other factors involved as well.
Having a propensity towards personality traits associated with proso-
cial tendencies and engagement in prosocial behavior may lead re-
ligious people to perceive themselves as highly moral, possibilities to
which we now turn.
1.2. Religiosity and personality
Personality traits and individual differences in prosociality may
explain the link between religiosity and moral self-image (e.g.,
Saroglou, 2013; Saroglou et al., 2005). Agreeableness, honesty-
223
Personality and Individual Differences 131 (2018) 222–231
humility, and conscientiousness are linked to both prosocial behavior
and religiosity (e.g., Graziano, Habashi, Sheese, & Tobin, 2007;
Habashi, Graziano, & Hoover, 2016; Hilbig, Glöckner, & Zettler, 2014;
Saroglou, 2009). Prosocial personality encompasses not only these
broad traits but also more specific other-oriented values and tendencies.
Religiosity is associated with empathy, a prosocial trait, as well as with
prosocial values including benevolence and compassion (e.g., Saroglou,
2013; Saroglou, Delpierre, & Dernelle, 2004). These associations
emerge in diverse religious groups (Christians, Buddhists, Jews, and
Muslims) and hold when controlling for self-reports thought to tap SDR
(e.g. Saroglou et al., 2004; Saroglou et al., 2005). Thus, there is evi-
dence that religious people genuinely hold prosocial values, which may
cause them to perceive themselves as being highly moral.
1.3. Religiosity and prosocial behavior
Consider as well that the individual differences noted above (i.e.,
empathy and agreeableness) predict engaging in prosocial behavior
(Ozer & Benet-Martinez, 2006). As such, religious individuals may view
themselves as highly moral because they can recall themselves enga-
ging in prosocial behavior (Bem, 1967). Across a wide array of religious
groups and continents (Pelham & Crabtree, 2008), religiosity is corre-
lated with self-reported prosocial behavior, including volunteerism and
charitable donations (e.g., Monsma, 2007; Putnam et al., 2010;
Saroglou, 2013). Some controlled and naturalistic experiments have
also shown a positive association between religiosity and prosocial
behavior (e.g., Anderson & Mellor, 2009; Everett, Haque, & Rand, 2016;
Saroglou et al., 2005), though results are inconsistent (e.g., Galen,
2012; Kramer & Shariff, 2018). It is important to note that religious
prosociality is typically extended to ingroup members (e.g., Saroglou
et al., 2005), rather than broadly, suggesting it is contextualized. Still,
religious people's high valuation of their morality may arise from their
tendency to behave prosocially. In sum, broad and specific individual
differences and behavioral reports may account for a substantial por-
tion of the shared variance between religiosity and moral self-image. If
supported, this contention would suggest that the link between re-
ligiosity and moral self-image is neither wholly veridical nor completely
artifactual. Rather, it is at least partially a product of dispositional
tendencies that both facilitate actual behavior and color self-percep-
tions.
1.4. Overview and predictions
Two correlational studies and an experiment tested key potential
contributors to the relationship between religiosity and moral self-
image. Studies 2 and 3 tested the contributions of self-reported per-
sonality traits (e.g., agreeableness, conscientiousness), individual dif-
ferences in prosocial tendencies, prosocial behavior, and measures of
SDR to the association between religiosity and moral self-image. We
predicted that the association between religiosity and moral self-image
would be explained by individual differences in prosocial tendencies,
personality traits (i.e., agreeableness, honesty-humility, and con-
scientiousness), and SDR. In order to address the inherent ambiguity in
self-reports of SDR, Study 3 employed a bogus pipeline manipulation,
providing a strong test of the role of impression management in this
relationship. We predicted that the bogus pipeline would attenuate the
association between religiosity and moral self-image.
2. Study 1
2.1. Overview and predictions
Study 1 included several possible explanatory variables for the link
between religiosity and moral self-image. We measured moral self-
image using an established measure of moral identity (Aquino & Reed
II, 2002), along with a set of items developed to tap various aspects of
S.J. Ward, L.A. King Personality and Individual Differences 131 (2018) 222–231

Table 1 traits (Aquino & Reed II, 2002). Participants rated their agreement with
Demographic information, studies 1 through 3. 10 statements about the importance of possessing moral traits (e.g.,
Study 1 Study 2 Study 3 honest, caring, fair). The scale has two subscales: internalization (e.g., “I
strongly desire to have these characteristics”) and symbolization (e.g., “I
Gender often wear clothes that identity me as having these characteristics”),
Men (n) 141 195 79
reflecting desires to possess or symbolically represent the traits.
Women (n) 172 296 95
Unreported (n) 0 1 1
Age M(SD) 18.99(0.91) 37.66(12.72) 18.64(1.31) 2.2.2.2. Moral self-image. To provide a more comprehensive assessment
Ethnicity
of participants' views of their morality, we also included items tapping
White/Caucasian 84.7% 77.6% 82% moral strivings and moral failings. Four moral strivings items assessed how
Black/African-American 7% 9.4% 4.5% much people strive to behave morally (e.g. “I am very committed to
Asian 6.1% 6.1% 6.7% moral values”). Six moral failings items assessed the extent to which
Hispanic/Latino 1.3% 4.9% 2.8%
participants acted immorally (e.g., “I sometimes act immorally.”).
Other 0.9% 2% 3.9%
Principal components analyses showed that the moral strivings items
Religious affiliation
loaded on one factor (eigenvalue = 2.95; 73.86% variance explained);
Christian 36.1% 29.9% 35%
Catholic 28.1% 11% 24.4%
2 factor solutions did not produce interpretable results. The moral
Atheist (non-believing) 8.3% 18.3% 11.1% failings items also loaded on one factor (eigenvalue = 3.40; 56.68%
Agnostics 16.6% 21% 19.4% variance explained) in a principal components analysis. See
Jewish 2.6% 1.6% 2.8% Supplementary Materials for all items and validation procedures.
Buddhist 1.6% 2.2% 1.7%
Protestant 2.2% 9.4% 1.7%
Muslim 1% 1% 0.6% 2.2.2.3. Prosocial scale. Participants rated the 16-item Prosocialness
Other 3.5% 4.9% 3.4% Scale for Adults (Caprara, Steca, Zelli, & Capanna, 2005), which
Total sample size 313 492 175 assessed people's likelihood of engaging in prosocial behavior and
feeling empathy (e.g., “I try to help others,” “I try to console those who
Note. Demographics provided are for the final samples used for analyses (i.e.,
are sad”).
not including participants removed for expressing suspicion in Study 3 or for
those removed for failing attention checks in Studies 1 and 2).
2.2.2.4. Empathy. The 7-item empathy subscale of the Interpersonal
Reactivity Index (Davis, 1983) assessed the extent to which people
moral self-views. In addition to SDR, we assessed individual differences
experience feelings of sympathy and concern for others (e.g., “I often
in agreeableness, conscientiousness, prosocial tendencies, and prosocial
have tender, concerned feelings for people less fortunate than me”).
behaviors. We expected that the association between religiosity and
moral self-image would be explained by individual differences in pro-
2.2.2.5. Prosocial behavior. Participants rated 11-items from the Self-
social tendencies and personality traits and by impression management.
Report Altruism Scale (Rushton, Chrisjohn, & Fekken, 1981), which
Measuring all these variables within the same study provided an op-
assessed how often people engaged in various prosocial behaviors (e.g.,
portunity to explore their joint influence on moral self-image.
“I have given money to a charity,” “I have allowed someone to go ahead
of me in line”). We added two additional items (i.e., 13 items total) to
2.2. Method
reflect the types of prosocial behavior college students were likely to
encounter (e.g., “I have stood up for a person that was being bullied,” “I
2.2.1. Participants & procedure
have turned in a wallet (or other lost item) that I found.”).
346 undergraduates completed this online study in partial fulfill-
ment of a research participation requirement (see Table 1). Sample size
2.2.2.6. Personality traits. Twelve agreeableness (e.g., “I believe that
was determined by the number of participants available during a se-
others have good intentions”) and five conscientiousness (e.g., “I make
mester. To ensure adequate representation of nonbelievers, students
plans and stick to them”) items were drawn from the International
who indicated they were atheist or agnostic on a premeasure at the
Personality Item Pool (Goldberg et al., 2006; see Supplementary
beginning of the semester were emailed an advertisement about this
Materials) to assess these constructs. We were not able to use full
study. Advertisements did not mention the purpose of the study nor
scales of these measures due to space considerations. More items were
why they qualified for it, and these participants did not receive any
included to assess agreeableness (vs. conscientiousness) due to its
additional incentives (beyond research participation credits) for parti-
greater relevance to morality variables.
cipation.1
Unless noted, items were rated on 1–7 scales with higher scores
2.2.2.7. Desirability bias. Participants completed a shortened 20-item
indicating higher agreement (or higher likelihood or frequency). Moral
version of the Balanced Inventory of Desirable Responding (BIDR;
self-image measures were completed prior to reports of prosocial be-
Paulhus & Reid, 1991; see Supplementary Materials for items) which
havior, to reduce contamination of ratings by behavioral reports.
included two subscales: impression management, the tendency to
Measures were embedded within filler measures of personality and
describe oneself favorably to others (10 items; e.g., “I never take
well-being. See Table 2 for descriptive statistics.2,3
things that don't belong to me”), and self-deceptive enhancement, the
tendency to maintain overly positive (and undeserved) self-views (10
2.2.2. Measures
items; e.g., “It would be hard for me to break any of my bad habits,”
2.2.2.1. Importance of moral identity scale. The Moral Identity Scale
recoded). Scores were computed by counting the number of extreme
measures the degree to which people's self-concepts emphasize moral
responses (6's and 7's after reverse coding appropriate items). To
provide an additional assessment of self-deception, participants also
1
These same procedures were followed in Study 3 to increase the recruitment of completed the Self-Deception Questionnaire (SDQ; Gur & Sackeim,
nonbelievers, which make up a small percentage of students at this university. 1979). The SDQ includes statements thought to be universally true but
2
Participants also completed additional measures of personality and prosociality for
ego-threatening (e.g., “Were your parents ever mean to you?”). Five
exploratory purposes in Studies 1 and 2.
3
Studies 1 and 2 included three attention check items asking participants to select a
items were dropped from the original scale (e.g., “Have you ever
particular response. Participants who failed two or more were removed from the data set wanted to rape or be raped by someone?”) in order to make the scale
(Study 1: n = 33 excluded; final N = 313; Study 2: n = 15 excluded; final N = 492). more appropriate for a contemporary sample. To simplify analyses, an

224
S.J. Ward, L.A. King Personality and Individual Differences 131 (2018) 222–231

Table 2
Descriptive statistics and correlations for all measures in study 1.
REL MSI MIS MF PSB EMP AG CON IM SD

Religiosity .92
Moral self-image .35⁎⁎ .68
Moral identity–symbolization .33⁎⁎ .49⁎⁎ .80
Moral failings −.04 −.25⁎⁎ −.24⁎⁎ .84
Prosocial behavior .20⁎⁎ .32⁎⁎ .36⁎⁎ −.21⁎⁎ .78
Empathy/prosociality .23⁎⁎ .57⁎⁎ .44⁎⁎ −.20⁎⁎ .53⁎⁎ .95
Agreeableness .19⁎⁎ .62⁎⁎ .46⁎⁎ −.12 .42⁎⁎ .76⁎⁎ .89
Conscientiousness .18⁎⁎ .28⁎⁎ .26⁎⁎ −.32⁎⁎ .19⁎⁎ .12⁎ .24⁎⁎ .74
Impression management .21⁎⁎ .35⁎⁎ .34⁎⁎ −.45⁎⁎ .20⁎⁎ .33⁎⁎ .36⁎⁎ .28⁎⁎ .72
Self-deception .06 .26⁎⁎ .18⁎⁎ −.38⁎⁎ .22⁎⁎ .24⁎⁎ .36⁎⁎ .37⁎⁎ .48⁎⁎ .67
Partials-moral self-image – – .22⁎⁎ .35⁎⁎ .31⁎⁎ .27⁎⁎ .30⁎⁎ .32⁎⁎ .30⁎⁎ .34⁎⁎
M(SD) 3.72(1.73) 5.58(0.90) 4.72(0.98) 4.39(1.07) 3.24(0.53) 5.08(0.96) 5.40(0.81) 4.54(1.01) 2.76(2.25) –

Note. N = 313. Coefficients on the diagonals are alpha reliabilities. Empathy/prosociality is a composite of these traits; moral self-image is an aggregate of moral
strivings and moral identity-internalization. Partial correlations between religiosity and moral self-image, when controlling for other variables are shown on the
bottom row of this table.
⁎⁎
p < .001.
⁎
p < .05.

aggregate self-deception scale was formed using standardized SDQ,
α = .80, and BIDR-self-deception, α = .67, scores.
2.2.2.8. Religiosity. Participants also completed the Revised Intrinsic/
Extrinsic Religiosity scale (Gorsuch & McPherson, 1989).4 Eight items
assessed intrinsic religiosity (e.g., “My whole approach to life is based
on my religion”). We used the five positively worded items from this
scale here and in subsequent studies (dropping 3 reverse-coded items;
see Supplementary Materials for items), consistent with recent
psychometric evidence demonstrating the reverse-coded items are
distinct and can invalidate the intrinsic religiosity scale (Cohen et al.,
2017). This 5-item intrinsic religiosity subscale (termed “religiosity”)
was used in all analyses. We chose to focus on intrinsic religiosity,
rather than other operationalizations of religiosity, because it reflects a
global motivation towards religion and is applicable to the broad range
of religious groups in our sample. In addition, the measure of intrinsic
religiosity we use from the Revised Intrinsic/Extrinsic Religiosity scale
is one of the most widely utilized measures of religious motivation
available (e.g., Cohen et al., 2017; Hill, 2005).
2.3. Results
The Prosocialness Scale for Adults and empathy scales were strongly
related, r = .72, p < .001, and combined into a composite (labeled
“empathy/prosociality”). The moral identity-internalization and moral
strivings scales were strongly correlated, r = .52; p < .001, and were
combined into a composite, referred to as moral self-image in all sub-
sequent analyses (see Supplementary Materials for analyses with se-
parate constructs). As Table 2 shows, religiosity was positively related
to moral self-image, prosocial behavior, and empathy. Religiosity was
unrelated to moral failings. In addition, religiosity was related to im-
pression management (but not self-deception), agreeableness, and
conscientiousness—all of which were, in turn, associated with moral
identity. Thus, these data provided an opportunity to test impression
management, self-reported prosocial behavior, empathy, and person-
ality traits as explanations of the link between religiosity and moral
self-image.5 The bottom row of Table 2 shows partial correlations
4
In Studies 1 through 3, we also included the extrinsic religiosity subscale and a 1-item
measure God belief. In all studies, intrinsic religiosity and God belief were strongly cor-
related, r's ≥ .71, p's < .001. We used intrinsic religiosity rather than God belief in
analyses as it is a more widely used measure.
5
It is notable that religiosity was positively correlated with self-reported prosocial
behavior. When controlling for impression management and self-deception, prosocial
behavior was still significantly related to religiosity, partial r = .18, p = .002, and to
moral self-image and moral identity, partial r's = .23, .19, p's < .001.
between religiosity and moral self-image controlling for each possible
explanatory variable. No single variable substantially accounted for the
association between religiosity and moral self-image.
Regression equations, shown in Table 3, examined whether con-
trolling for a combination of these factors (and gender) accounted for the
relationship of religiosity to moral self-image. Because SDR is linked to
not only religiosity and moral self-image but also to other personality
traits and prosocial behavior, it made sense to enter multiple predictors
simultaneously in this model. To avoid multicollinearity in regression
estimates, we aggregated prosocial personality variables. A principal
component analysis showed that the agreeableness, empathy, and pro-
sociality all loaded onto one factor (eigenvalue = 2.46; 82.11% var-
iance). For the aggregate prosocial personality variable, α = .89; item-
total correlations ranged from .74 for empathy to .82 for prosociality.
Entered alone, religiosity predicted moral self-image, ΔR2 = .12,
β = .35, p < .001. As shown in Table 3, controlling for gender, per-
sonality variables, prosocial behavior, and SDR (impression manage-
ment, self-deception) in separate steps of a hierarchical regression
model, religiosity remained a significant predictor of moral self-image.
Comparing the contributions of the putative explanatory variables
showed that, when entered in models with religiosity, the aggregate
prosocial personality variable had a stronger influence on moral self-
image, z's = 5.26, 5.59, p's < .001, than impression management or
prosocial behavior, which did not differ from each other, z = .39,
p = .69.
2.4. Brief discussion
Religiosity was positively associated with moral self-image, as well
as with variables that could explain this association, including prosocial
behavior, empathy, and impression management. Religiosity's associa-
tion with moral self-image remained robust when controlling for
gender, personality variables, prosocial behavior, and SDR. Prosocial
personality variables were the strongest explanatory variables for this
association.
Although religiosity was positively related to various measures of
moral self-image and prosociality, it was unrelated to moral failings:
Religious people do not view themselves as less immoral than others.
This finding provides some evidence that religious people do not report
themselves favorably on all questions pertaining to morality. The re-
sults of Study 1 demonstrate prosocial behavior only minimally influ-
enced this association, though the measures of prosocial behavior and
moral identity used here were limited. Study 2 sought to corroborate
these associations using broader measures of prosocial behavior and
moral self-image.

225
S.J. Ward, L.A. King Personality and Individual Differences 131 (2018) 222–231

Table 3
Regression models predicting aggregate moral self-image from Study 1 Measures.
Full model Prosocial behavior Personality SDR

⁎
Demographics ΔR = .02
2

Sex .12⁎
Prosocial behavior ΔR2 = .09⁎⁎ ΔR2 = .10⁎⁎
Prosocial behavior −.03 .26⁎⁎
Personality traits ΔR2 = .31⁎⁎ ΔR2 = .42⁎⁎
Empathy/prosociality/agreeableness .57⁎⁎ .55⁎⁎
Conscientiousness .14⁎ .15⁎
Socially desirable responding (SDR) ΔR2 = .02⁎ ΔR2 = .14⁎⁎
Impression management .12⁎ .23⁎⁎
Self-deception −.03 .13⁎
Final step-religiosity ΔR2 = .04⁎⁎ ΔR2 = .08⁎⁎ ΔR2 = .04⁎⁎ ΔR2 = .08⁎⁎
Religiosity .20⁎⁎ .30⁎⁎ .19⁎⁎ .29⁎⁎

Note. N=312. All values presented are standardized beta weights. Empathy/prosociality/agreeableness is a composite of these traits. Columns represent separate
regression models. The full model controlled for sex, prosocial behavior, personality traits, and socially desirable responding on separate steps of the model. The three
models shown to the right of the full model control for each set of variables in separate models. Entered alone, religiosity predicted moral self-image, ΔR2 = .12,
β = .35, p < .001.
⁎⁎
p < .001.
⁎
p < .05.

3. Study 2 3.2. Method

3.1. Overview and predictions 3.2.1. Participants & procedure

Study 2 aimed to conceptually replicate and extend the findings
from Study 1 using a larger, more diverse adult sample as well as new
measures of prosocial behavior, moral self-image, and personality
traits. New measures of prosocial behavior included self-reported vo-
lunteerism and charitable donations as well as a lengthier version of the
prosocial behavior measure from Study 1. In addition to the moral
identity-internalization scale from Study 1, we included a new measure
of moral trait ratings designed to capture people's perceptions of how
well they embody various moral traits. Finally, we included measures of
honesty-humility and agreeableness from the HEXACO personality in-
ventory (Lee & Ashton, 2016) in order to conceptually replicate the
pattern of findings for these measures from Study 1 using established
scales. We predicted that, consistent with Study 1, individual differ-
ences in prosocial tendencies and personality traits (i.e., agreeableness,
honesty-humility) would be the strongest explanatory variables of the
association between religiosity and moral self-image, with prosocial
behavior and SDR contributing more weakly.
507 adults on Mechanical Turk (paid $1) completed this online
study. Sample size was determined by our budget. Table 1 shows de-
mographics. Participants first completed personality and moral self-
image measures, then reports of prosocial behavior, and, finally, mea-
sures of SDR and the 5-item religiosity measure from Study 1. Table 4
shows descriptive statistics. Unless specified, items were rated on scales
ranging from 1 to 7, with higher scores indicating higher agreement.
3.2.2. Measures
Participants completed the Moral Identity-Internalization subscale
(Aquino & Reed II, 2002), empathy/prosocial personality (Caprara
et al., 2005; Davis, 1983), and prosocial behavior scales from Study 1.
Four additional items from the Self-Report Altruism Scale (Rushton
et al., 1981) were added to the prosocial behavior scale (17 items total;
see Supplementary Materials). Participants completed the religiosity
measure used in Study 1, as well as the full BIDR (Paulhus & Reid,
1991).

Table 4
Correlations among measures, study 2.
REL MSI EMP AG HH Cons PRO VOL CHA RelCh SecCh IM SD

Religiosity .94
Moral self-image .26⁎⁎ .74
Empathy/prosociality .22⁎⁎ .61⁎⁎ .84
Agreeableness .14⁎⁎ .28⁎⁎ .36⁎⁎ .88
Honesty-humility .14⁎⁎ .54⁎⁎ .44⁎⁎ .34⁎⁎ .86
Conscientiousness .15⁎ .45⁎⁎ .30⁎⁎ .29⁎⁎ .33⁎⁎ .87
Prosocial behavior .14⁎ .11⁎ .35⁎⁎ .15⁎⁎ −.00 .07 .90
Volunteerism .19⁎⁎ .13⁎ .20⁎⁎ .16⁎⁎ .05 .03 .33⁎⁎ .96
Charity: non-monetary .14⁎ .09 .15⁎ .04 .10 .05 .20⁎⁎ .23⁎⁎ –
Charity: religious .45⁎⁎ .19⁎⁎ .17⁎⁎ .15⁎ .13⁎ .16⁎ .25⁎⁎ .25⁎⁎ .15⁎ –
Charity: secular .06 .20⁎⁎ .25⁎⁎ .17⁎⁎ .18⁎⁎ .14⁎ .28⁎⁎ .34⁎⁎ .17⁎ .27⁎⁎ –
Impression management .18⁎⁎ .53⁎⁎ .41⁎⁎ .37⁎⁎ .63⁎⁎ .42⁎⁎ .03 .04 .06 .08 .15⁎ .85
Self-deception .09 .44⁎⁎ .29⁎⁎ .37⁎⁎ .39⁎⁎ .49⁎⁎ .10⁎ −.02 .01 .06 .15⁎ .60⁎⁎ .80
Partials-moral Self-image – .17⁎⁎ .24⁎⁎ .23⁎⁎ .24⁎⁎ .25⁎⁎ .26⁎⁎ .27⁎⁎ .21⁎⁎ .26⁎⁎ .20⁎⁎ .25⁎⁎
M(SD) 3.46(2.06) −0.00(0.90) 6.16(0.92) – – 5.22(1.30) 2.43(0.74) – – – – – 5.07(4.00)

Note. N=492. Coefficients on the diagonals are alpha reliabilities. Empathy/prosociality is a composite of these traits; moral self-image is an aggregate of moral traits
and moral identity-internalization. Partial correlations between religiosity and moral self-image, when controlling for other variables are shown on the bottom row of
this table.
⁎⁎
p < .001.
⁎
p < .05.

226
S.J. Ward, L.A. King Personality and Individual Differences 131 (2018) 222–231

3.2.2.1. Moral trait ratings. Participants rated moral traits from the Table 5
prototypical attributes associated with moral person-concepts in past Regression model predicting aggregate moral self-image from study 2 measures.
research (Walker & Pitts, 1998; see Supplementary Materials). Full model Prosocial Personality SDR
Presented in randomized order, 13 morally relevant traits were rated behavior
from 1 “not at all” to 9 “very well” in terms of how much they described
Demographics ΔR2 = .07⁎⁎
participants; 3 fillers (e.g., intelligent) were also included.6
Gender −.07
Age .02
3.2.2.2. Prosocial behavior & charitable donations. Participants were Prosocial behavior ΔR2 = .02 ΔR2 = .03
provided with a definition of volunteerism—involving unpaid Prosocial behavior −.03 .03
Volunteerism .009 −.01
activities done through an organization—adapted from the Current Non-monetary .01 .04
Population Survey (Bureau of Labor Statistics, 2016). Participants charity donations
reported their average volunteer hours per month during the last Religious charity .003 −.005
12 months, ranging from “less than 5 h” to “20 h or more”, M donations
Secular charity −.08 .01
(SD) = 1(1), i.e., less than 5 h; and the total number of occasions they
donations
volunteered, ranging from “1 to 5” to “15 or more”, M(SD) = 1(1), i.e., Personality traits ΔR2 = .31⁎⁎ ΔR2 = .48⁎⁎
1 to 5 occasions. These variables were skewed (> 1.18) so they were Empathy/ .41⁎⁎ .55⁎⁎
log transformed and aggregated into a composite because they were prosociality/
strongly correlated, r = .92, p < .001.7 Participants also reported the agreeableness/
honesty-humility
estimated amount their household (including family but not
Conscientiousness .09⁎ .21⁎⁎
roommates) donated to secular, M(SD) = $175.49(387.51), and Socially desirable ΔR2 = .03⁎⁎ ΔR2 = .31⁎⁎
religious organizations, M(SD) = $227.72($1165.51). To address responding (SDR)
distributional issues (skew > 6.47; kurtosis, > 60.94), these variables Impression .16⁎ .38⁎⁎
management
were log transformed.
Self-deception .08 .20⁎⁎
Final step-religiosity ΔR2 = .01⁎ ΔR2 = .07⁎⁎ ΔR2 = .01⁎⁎ ΔR2 = .03⁎⁎
3.2.2.3. Personality. Participants completed the 16-item agreeableness Religiosity .14⁎ .29⁎⁎ .12⁎⁎ .18⁎⁎
and a 12-item honesty-humility subscales (excluding greed avoidance
Note. N=492. All values presented are standardized beta weights. Empathy/
items) from the HEXACO Personality Inventory (Lee & Ashton, 2016)
prosociality/agreeableness/honesty-humility is a composite of these traits.
and 5-items from the NEO-Five Factor Inventory (Costa Jr. & McCrae,
Columns represent separate regression models. The full model controlled for
1992; see Supplementary Materials) to measure conscientiousness. As sex, prosocial behavior, personality traits, and socially desirable responding on
in Study 1, the conscientiousness measure was kept brief due to space separate steps of the model. The three models shown to the right of the full
considerations. model control for each set of variables in separate models. Entered alone, re-
ligiosity predicted the aggregate moral self-image variable, ΔR2 = .07, β = .26,
3.3. Results p < .001.
⁎⁎
p < .001.
⁎
As in Study 1, empathy and prosociality were strongly correlated, p < .05.
r = .73, p < .001, and were aggregated. The moral identity-inter-
in Table 4, no single variable fully accounted for the association be-
nalization and moral trait scales were strongly correlated, r = .60;
tween religiosity and moral traits.
p < .001, and were combined into a composite (after standardizing the
Next, we examined the association between religiosity and moral
scales), referred to as moral self-image in subsequent analyses (see
self-image when controlling for gender, personality, prosocial behavior,
Supplementary Materials for analyses with separate constructs). As
and socially desirable responding in regression analyses. To reduce
shown in Table 4, religiosity was correlated with moral self-image,
multicollinearity, we first aggregated the prosocial personality vari-
prosocial tendencies (e.g., empathy/prosociality) and general person-
ables. A principal component analysis showed that honesty-humility,
ality (agreeableness, honesty-humility, conscientiousness). Religiosity
agreeableness, empathy, and prosociality all loaded onto one factor
was also related to volunteerism and charitable donations, though it
(eigenvalue = 2.31; 57.82% variance). These variables were ag-
was unrelated to secular charity donations. Moral self-image was po-
gregated into a prosocial personality variable, α = .75; item-total cor-
sitively associated with prosocial behavior, religious/secular charitable
relations ranged from .41 for agreeableness to .66 for empathy.
donations, empathy, agreeableness, honesty-humility, impression
When entered alone, religiosity predicted the aggregate moral self-
management, and self-deception.8 As shown in the partial correlations
image variable, ΔR2 = .07, β = .26, p < .001. As shown in Table 5,
controlling for reports of prosocial behavior, personality traits, and
6
During the moral trait ratings, participants also rated words pertaining to kindness (6 socially desirable responding, religiosity remained a significant pre-
items; e.g., compassionate, helpful; see Supplementary Materials). Kindness and moral dictor of moral self-image. When comparing the contributions of these
trait words were strongly correlated, r = .73; p < .001. The kindness words were ex-
variables (when entered in models with religiosity), the aggregate
cluded from the analyses due to overlap with the moral trait words. A principal com-
ponents analysis with varimax rotation showed that morality and kindness traits con- prosocial personality variable had a stronger influence on the aggregate
formed to a two-factor solution, with 12 moral traits loading on one factor moral self-image variable than impression management or prosocial
(eigenvalue = 9.92; 55.10% variance) and the 6 kindness loading on the other (eigen- behavior, z's = 5.33, 11.48; p's < .0001.9 Impression management had
value = 1.66; 9.19% variance). Although the results were not entirely consistent with a stronger influence on the aggregate moral self-image variable than
simple structure, they generally conformed to the expected pattern. The “virtuous” item
loaded on both factors, but was retained with the moral traits as it is more conceptually
prosocial behavior, z = 7.20, p < .0001.
similar.
7
An aggregate volunteerism variable formed without log transformation exhibited
similar patterns of association with religiosity, r = .18, p < .001, and moral self-image, (footnote continued)
r = .12, p = .009. and secular/religious charity donations, partial r's > .13, p < .008. Moral identity was
8
We also sought to examine how SDR affected results. When controlling for impression no longer related to secular/religious charity donations when controlling for impression
management and self-deception, religiosity was still significantly associated with proso- management and self-deception, partial r's ± < .08, p's > .10.
9
cial behavior, volunteerism, non-monetary charity donations, and donations to religious The prosocial behavior measure (Self-Report Altruism Scale) was not strongly cor-
charities, partial r's > .12, p's < .02. Controlling for impression management and self- related enough with volunteerism or charity donations to warrant forming an aggregate
deception, moral traits were still significantly related to volunteerism, prosocial behavior, prosocial behavior variable so it was used alone in analyses.

227
S.J. Ward, L.A. King Personality and Individual Differences 131 (2018) 222–231

3.4. Brief discussion 4.2.1.1. Online premeasures. At least 24 h prior to the lab session,
participants completed online measures of SDR and religiosity,
Corroborating Study 1, these results demonstrated that the link embedded in filler measures of personality and well-being. We
between religiosity and moral self-image was most strongly explained included the shortened 20-item version of the BIDR from Study 1;
by prosocial personality traits, rather than prosocial behavior or SDR. impression management, M(SD) = 4.68(3.18), α = .79; self-deceptive
Measures of prosocial behavior were less strongly related to moral self- enhancement, M(SD) = 4.83(3.44), α = .62 (subscales correlated
image than prosocial and general personality traits, providing some r = .33, p < .001). Participants also completed the 5-item religiosity
rationale as to why they were less influential to religiosity's association measure from Studies 1 and 2, M(SD) = 3.57(1.74), α = .93, which was
with moral self-image. Studies 1 and 2 relied on questionnaire measures used in all analyses.
of SDR, which, as noted previously, may reflect substantive variables
rather than bias. As such, to target impression management in parti- 4.2.1.2. Lab session measures. Procedures were modeled on
cular, Study 3 employed an experimental manipulation, the bogus pi- Heintzelman et al., 2014, Study 3. Participants visited the lab
peline. This procedure provided a more stringent test of the role of SDR individually. Upon arrival, an experimenter taped sensors to
in accounting for religious people's reports of high moral-image. participants' left cheek and inner wrist. Sensors were attached to a
machine that lit up, appearing functional. Participants were assigned
randomly to one of two conditions. Bogus pipeline participants
4. Study 3 (n = 90) were told that the sensors detected muscle movement
associated with deception. They completed a “guilty knowledge test,”
4.1. Overview and predictions ostensibly to calibrate the machine. Control participants (n = 90) were
instructed that the purpose of the sensors was to assess muscle activity
Bogus pipeline (i.e., fake lie detector) manipulations attenuate so- and were shown fake feedback as an example of the machine's
cially desirable self-presentation, leading to higher reports of undesir- functionality. All participants then completed the dependent
able characteristics (Roese & Jamieson, 1993) and lower reports of measures. Descriptive statistics for all measures completed during the
desirable characteristics (Heintzelman, Trent, & King, 2015). Because lab session are shown in Table 6.
experimental manipulations of impression management—such as the 4.2.1.2.1. Personality traits. The agreeableness and
bogus pipeline—lack the ambiguity associated with self-reports, they conscientiousness items from Study 1 were rated.
provide a stronger examination of the role of impression management 4.2.1.2.2. Moral self-image. To capture a range of moral self-
in accounting for the high moral self-image among the religious. Two perceptions, we included the moral failings and moral strivings items
predictions were possible: First, if impression management is critical to from Study 1. Because people are highly inclined to view themselves as
explaining the link between religiosity and moral self-image, then this moral and self-enhance in the moral domain (e.g., Eriksson & Funcke,
relationship should diminish in the bogus pipeline condition, as we 2014; Van Lange & Sedikides, 1998), we also included items that
would expect people with high religiosity to lower their reports of reflected overly high moral self-regard and were thus likely to be more
moral self-image when they feel that deception can be detected. Al- sensitive to the manipulation. Seven inflated moral self-image items
ternatively, if religious people genuinely believe they possess the high reflected unrealistically high views of one's morality (e.g., “I always live
moral self-image they report, then such reports should remain un- up to my moral aspirations,” “I am a more moral person than my peers,”
changed by the manipulation. We predicted that condition would “I am proud of how moral I behave”). Refer to the Supplementary
moderate the association between religiosity and moral self-image, re- Materials for items and validation procedures. To disguise our focus on
ducing the association in the bogus pipeline (vs. control) condition. morality, all moral self-image items were embedded within the
personality measures. We expected that these various facets would
tap into moral self-image and moral self-enhancement.
4.2. Method Using these ad hoc moral self-image measures was necessary be-
cause no extant scales provide measures of moral self-perceptions.
4.2.1. Participants & procedure Available scales measure, instead, whether a person wants to embody a
180 undergraduates completed the study in partial fulfillment of range of moral traits (e.g., Moral Identity Scale; Aquino & Reed II,
research participation requirements. Table 1 shows demographic 2002) or is empathic/altruistic (e.g., Prosocial Personality Battery,
characteristics. Sample size was determined by the number of partici- Penner, Fritzsche, Craiger, & Freifeld, 1995; the Altruism Scale;
pants available during a semester. Rushton, Chrisjohn, & Fekken, 1981).
4.2.1.2.3. Immoral behavior. Participants rated 16 statements
Table 6 regarding the frequency (from 1 “Never” to 7 “Often”) with which
Condition effect of lab measures, study 3.
they engaged in various immoral behaviors relevant to college students
Measure Condition means (e.g., “cheat on an exam,” “lie to your parents”; see Supplementary
Materials).
α Grand Control Bogus t(173) d
Finally, participants were probed for suspicion and debriefed. Five
mean (SD) (n = 90) Pipeline
(n = 85) participants in the bogus pipeline condition expressed doubts about its
validity and were removed from analyses.10
Inflated moral .84 4.67(.95) 4.79(0.95) 4.54(0.94) 1.71⁎ 0.26
self-image
Moral strivings .85 4.98(1.12) 5.06(1.10) 4.90(1.15) 0.93 0.14
4.3. Results
Moral failings .82 4.53 4.39 4.65(1.04) −1.70⁎ −0.26
(1.00) (0.97) Examining pre-test measures showed that religiosity was correlated
Agreeableness .79 5.27 (.76) 5.41(0.73) 5.14(0.78) 2.31⁎ 0.35 moderately with impression management, r = .27, p < .001, but was
Conscientiousness .71 4.38 4.61(0.97) 4.18(1.09) 2.76⁎⁎ 0.42
(1.05)
Immoral actions .85 3.96 (.81) 3.78(0.79) 4.08(0.92) −2.30⁎ −0.35 10
Retaining suspicious participants, results for the key interaction presented below
show marginal main effects of religiosity and condition, β < ± .18, p's > .069, quali-
Note. All tests conducted were one-tailed. fied by a marginally significant religiosity X condition interaction, ΔR2 = .02, β = .19,
⁎
p < .05. p = .060. For controls, religiosity marginally predicted inflated moral self-image, β = .18,
⁎⁎
p < .01. p = .084; within the bogus pipeline, β = .45, p = .001.

228
S.J. Ward, L.A. King
5.3
5.1
Inflated Moral self-image
4.9
4.7 Control
4.5
Lie
Detector
4.3
4.1
3.9
3.7
3.5
Low Religiosity High Religiosity
Fig. 1. Study 3: religiosity × bogus pipeline condition predicting inflated moral
self-image
Generated regression lines predicting inflated moral self-image from religiosity
for individuals ± 1 SD from the mean on religiosity, moderated by condition
(control, β = .18, p = .084; or bogus pipeline, β = .49, p < .001), Study 3.
uncorrelated with self-deception, r = −.03, p = .68. Impression man-
agement and self-deception scores did not differ by condition, t's <
.73, p's > .47. However, bogus pipeline participants scored higher on
the religiosity measure completed in the online premeasures, M
(SDs) = 3.86(1.77), than controls, 3.33(1.69), t(173) = −2.01,
p = .046, d = −0.30, an apparent failure of random assignment. To
examine predictions for religiosity, this variable was mean centered
within conditions.11
Table 6 shows that compared to controls, the bogus pipeline led to
lowered inflated moral self-image, agreeableness, and conscientious-
ness, and higher reports of unethical acts and moral failings. This is
consistent with past research showing the bogus pipeline increases re-
ports of socially undesirable characteristics (e.g., Roese & Jamieson,
1993). We expected the bogus pipeline to lower religiosity's association
with inflated moral self-image and moral strivings. When inflated moral
self-image was regressed on mean centered religiosity, dummy-coded
condition (control = 0, bogus pipeline = 1), and their interaction,
marginal main effects of condition, β = −.14, p = .057; B
(SE) = −0.26(0.13); CI = [−0.52, 0.008] and religiosity, β = .18,
p = .064; B(SE) = 0.18(0.09); CI = [−0.01, 0.36], were qualified by a
religiosity X condition interaction, ΔR2 = .02, β = .21, p = .037; B
(SE) = 0.28(0.13); CI = [0.02, 0.55]. As shown in Fig. 1, contrary to
predictions, religiosity was more strongly associated with inflated moral
self-image in the bogus pipeline, β = .49, p < .001; B
(SE) = 0.46(0.09); CI = [0.28, 0.64], compared to the control condi-
tion, β = .18, p = .084; B(SE) = .18(.10); CI = [−0.02, 0.37]. As can
be seen, this interaction reflects the tendency of people with low re-
ligiosity to reduce their inflated moral self-image in the bogus pipeline.
Controlling for impression management, β =.37 p < .001, and the
condition X impression management interaction, β = −.03, p = .75;
ΔR2 = .17 for step, this religiosity X condition interaction was mar-
ginally significant, ΔR2 = .02, β = .19, p = .057.
Religiosity did not moderate condition effects for moral strivings,
moral failings, immoral actions, agreeableness, or conscientiousness, all
interaction β's < ± .16, all p's > .11. Collapsing across conditions,
religiosity was associated with moral strivings, r = .39, p < .001,
agreeableness, r = .19, p = .012, and conscientiousness, r = .21,
p = .005. Religiosity was marginally negatively associated with im-
moral actions, r = −.15, p = .053, and unrelated to moral failings,
r = –.11, p = .14. Impression management did not interact with
11
The key interaction reported for this study remains the same if using standardized
religiosity (vs. mean centered within condition). When inflated moral self-image was
regressed on standardized religiosity, condition, and their interaction, a main effect of
condition, β = −.18, p = .012, and a marginal main effect of religiosity, β = .19,
p = .064, were qualified by the condition X religiosity interaction, β = .20, p = .047.
229
Personality and Individual Differences 131 (2018) 222–231
religiosity, condition, or religiosity X condition, to predict inflated
moral self-image or moral strivings (all effects, β's < ± –.11, p's >
.14).
4.4. Brief discussion
These results further challenge the purported role of impression
management in explaining the association between religiosity and
moral self-image: Even when attached to an ostensible lie detector,
religiosity predicted inflated moral self-image. People responded to the
bogus pipeline manipulation in an expected manner on ratings of per-
sonality traits and immoral behavior, and these effects did not depend
on participants' religiosity. Thus, it appears that religious people are not
any more or less prone to the effects of the bogus pipeline manipulation
on ratings besides inflated moral self-image. Religiosity was similarly
related to moral strivings in both the bogus pipeline and control con-
ditions.
The fact that religiosity remained associated with inflated moral
self-image and moral strivings in the bogus pipeline condition might
seem to indicate that the association is due to self-deception. However,
religiosity was unrelated to self-deception measures across Studies 1–3.
Religiosity was related to self-reported impression management in
Study 3, consistent with Studies 1 and 2 and past research (Sedikides &
Gebauer, 2010). Although the bogus pipeline is thought to attenuate
conscious attempts at impression management, there may be more
unconscious aspects of impression management that would not be af-
fected by the bogus pipeline manipulation. Impression management
measures have been viewed as representing a communal form of self-
deception concerned with projecting a moralistic bias that may operate
on an unconscious level (Paulhus & John, 1998). The potentially un-
conscious nature of this self-deceptive moralistic bias could help ex-
plain why religious people's inflated moral self-image was unchanged in
the bogus pipeline.
5. General discussion
The strong moral self-image reported by religious people has puz-
zled researchers for decades. Religion's association with morality has
been attributed to egoistic and reputational concerns, namely SDR (e.g.,
Batson, 1976; Batson et al., 1989; Shariff, 2015). Yet, personality traits
and individual differences in prosociality may also help account for the
high moral self-image reported by religious people (e.g., Saroglou,
2013; Saroglou et al., 2005). The present studies suggest that the as-
sociation between moral self-image measures and religiosity is most
strongly explained by individual differences in prosocial personality
and traits, including empathy, agreeableness, and honesty-humility.
Prosocial behavior and impression management partially accounted for
religiosity's association with moral self-image measures. These results
demonstrate that religion shares a multifaceted association with moral
self-image that is explained by substantive personality traits and in-
dividual differences in prosociality, rather than merely resulting from
self-presentation concerns. The link between religion and moral self-
image need not be viewed as reflecting a direct relationship from re-
ligion to behavior nor as an artifact of bias. Rather, it reflects a variety
of variables including individual differences that may predispose in-
dividuals to religiosity, prosocial behavior, and moral self-image.
SDR has often been regarded as explaining religious people's high
moral self-image (e.g., McKay & Whitehouse, 2014; Shariff, 2015). The
present studies do not support a strong role of impression management
in the link between religiosity and moral self-image. In particular, the
results of Study 3 would seem to speak especially strongly against this
possibility.
In Studies 1 and 2, impression management partially accounted for
the association between religiosity and moral self-image measures, but
it was less influential than prosocial personality traits. With regard to
the potential role of self-deception, results are less clear cut. Although
S.J. Ward, L.A. King
self-reports of self-deception were not related to religiosity across
Studies 1–3, the ambiguity inherent in these measures renders conclu-
sions tentative. Future research might seek to develop experimental
manipulations of self-deception in order to probe its contribution to the
link between religiosity and moral self-image.
These studies identify the critical role of prosocial personality traits
in informing religious people's moral self-image. Yet, much remains to
be known about the causality between personality, values, and religion.
Generally, personality traits, especially agreeableness and con-
scientiousness, are considered to predict religiosity rather than result
from it (e.g., Saroglou, 2009). Conscientiousness and agreeableness
may dispose individuals towards religious membership, as religion can
provide a structured outlet for pursuing prosocial goals and fostering
social connections (e.g., Saroglou & Munoz-Garcia, 2008). However,
religion may also potentially influence prosocial personality traits and
values. Religion could strengthen prosocial inclinations, like empathy,
due to its emphasis on moral values and service. Specifying the causal
associations between religion and prosocial personality tendencies re-
mains an avenue for future inquiry. In addition, in the present studies
we treated prosocial behavior as a predictor of moral self-image, yet it is
also possible the causal direction works in the opposite direction: Just
as people who behave prosocially might come to see themselves as
more moral, people who perceive themselves as moral might have a
higher propensity to behave morally to cohere with their self-percep-
tion. Examining the causal directions between moral self-image and
morally relevant behaviors is an important area for future research.
Although these results illuminate reasons for the association be-
tween religiosity and moral self-image, there are several limitations
worth acknowledging. First, none of these studies included behavioral
measures of morality or prosociality, limiting the conclusions that can
be drawn about the veridicality of the enhanced moral self-image
among religious people. Future studies should assess how religiosity
and moral self-image relate to a broader range of self-reported and
naturalistic moral behaviors to better evaluate to what extent religious
people are reporting genuine versus enhanced reports of moral self-
image and values. In Studies 1 and 2, reports of prosocial behaviors did
little to explain the association between religiosity and moral self-
image. There are several possibilities for why this may be the case.
Religious people may imbue more everyday actions with moral sig-
nificance, whereas people who are not religious may have a stricter
definition of what constitutes a moral action (e.g., McKay &
Whitehouse, 2015). Religious people may also have more biased recall
of morally relevant actions, leading them to believe they are especially
moral. An interesting area for future research is examining whether
increasing the salience of morally relevant behaviors would attenuate
the association between religiosity and moral self-image. The link be-
tween moral self-image and religiosity may be due to religious people
being less apt to incorporate actual behavior into their assessments of
morality.
Another potential limitation is that in addition to established scales,
we relied on some measures of moral self-image that have not been
validated in past research. We made this decision because the estab-
lished scales provided inadequate measures of moral self-views. In
Studies 1 and 2, the ad hoc measures of moral self-strivings/moral traits
correlated strongly with established measures of morality/prosociality,
providing evidence they all assessed similar underlying constructs.
Nevertheless, it would be valuable for future research to accrue evi-
dence for the validity of these moral strivings/moral traits scales.
Moreover, it would also be beneficial for future research to randomize
(or counterbalance) the order of scales in case the order of scale ad-
ministration used here (moral self-image measures before prosocial
behavior scales) contributed to the results.
A final limitation is that most of our religious participants were
either Christian or Catholic, and these samples were all collected in the
United States. The low sample sizes of people in some religious groups
in these studies (e.g., Muslims, Buddhists, Jews) prevented an
230
Personality and Individual Differences 131 (2018) 222–231
investigation into whether contributors to moral self-image differed
across religious groups. Moreover, religious self-enhancement is more
pronounced in religiously-predominant countries (e.g., Gebauer et al.,
2017; Sedikides & Gebauer, 2010), such as the United States. Religion
may also be perceived as more morally normative in the Unites States
than in other cultural contexts, such as Europe. The association between
religiosity and moral self-image may differ in other countries and per-
haps within some religious groups so it is important for future research
to address these possibilities before the present results can be gen-
eralized to new contexts and religious groups.
Religion has a persistent influence on moral self-image that is ex-
plained by prosocial personality traits, conscientiousness, and socially
desirable responding. Further understanding the antecedents and con-
sequences of the high moral self-image of religious people remains an
intriguing area for future inquiry.
Declaration of conflicting interests
The authors declare no potential conflicts of interest with respect to
the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, au-
thorship, and/or publication of this article.
Appendix A. Supplementary Materials
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.paid.2018.04.028.
References
Anderson, L. R., & Mellor, J. M. (2009). Religion and cooperation in a public goods ex-
periment. Economics Letters, 105, 58–60. http://dx.doi.org/10.1007/s10726-015-
9462-6.
Aquino, K., & Reed, A., II (2002). The self-importance of moral identity. Journal of
Personality and Social Psychology, 83, 1423–1440. http://dx.doi.org/10.1037/0022-
3514.83.6.1423.
Batson, C. D. (1976). Religion as prosocial: Agent or double agent? Journal for the
Scientific Study of Religion, 29–45. http://dx.doi.org/10.2307/1384312.
Batson, C. D., Oleson, K. C., Weeks, J. L., Healy, S. P., Reeves, P. J., Jennings, P., & Brown,
T. (1989). Religious prosocial motivation: Is it altruistic or egoistic? Journal of
Personality and Social Psychology, 57, 873–884. http://dx.doi.org/10.1037/0022-
3514.57.5.873.
Bem, D. J. (1967). Self-perception: An alternative interpretation of cognitive dissonance
phenomena. Psychological Review, 74, 183–200.
Bureau of Labor Statistics (2016). Volunteering in the United States-2015. News Release:
Bureau of Labor Statistics. Retrieved from http://www.bls.gov/news.release/pdf/
volun.pdf.
Caprara, G. V., Steca, P., Zelli, A., & Capanna, C. (2005). A new scale for measuring adults'
prosocialness. European Journal of Psychological Assessment, 21, 77–89. http://dx.doi.
org/10.1027/1015-5759.21.2.77.
Cohen, A. B., Mazza, G. L., Johnson, K. A., Enders, C. K., Warner, C. M., Pasek, M. H., &
Cook, J. E. (2017). Theorizing and measuring religiosity across cultures. Personality
and Social Psychology Bulletin, 43(12), 1724–1736. http://dx.doi.org/10.1177/
0146167217727732.
Costa, R. T., Jr., & McCrae, R. R. (1992). Revised NEO personality inventory and NEO five-
factor inventory manual. Odessa, FL: Psychological Assessment Resources.
Davis, M. H. (1983). Measuring individual differences in empathy: Evidence for a mul-
tidimensional approach. Journal of Personality and Social Psychology, 44, 113–126.
http://dx.doi.org/10.1037/0022-3514.44.1.113.
Eriksson, K., & Funcke, A. (2014). Humble self-enhancement: Religiosity and the better-
than-average effect. Social Psychological and Personality Science, 5(1), 76–83. http://
dx.doi.org/10.1177/1948550613484179.
Everett, J. A., Haque, O. S., & Rand, D. G. (2016). How good is the samaritan, and why?
An experimental investigation of the extent and nature of religious prosociality using
economic games. Social Psychological and Personality Science, 7, 248–255. http://dx.
doi.org/10.1177/1948550616632577.
Furrow, J. L., King, P. E., & White, K. (2004). Religion and positive youth development:
Identity, meaning, and prosocial concerns. Applied Developmental Science, 8, 17–26.
http://dx.doi.org/10.1207/S1532480XADS0801_3.
Galen, L. W. (2012). Does religious belief promote prosociality? A critical examination.
Psychological Bulletin, 138, 876–906. http://dx.doi.org/10.1037/a0028251.
Gebauer, J. E., Sedikides, C., & Schrade, A. (2017). Christian self-enhancement. Journal of
S.J. Ward, L.A. King
Personality and Social Psychology.
Goldberg, L. R., Johnson, J. A., Eber, H. W., Hogan, R., Ashton, M. C., Cloninger, C. R., &
Gough, H. G. (2006). The international personality item pool and the future of public-
domain personality measures. Journal of Research in Personality, 40, 84–96. http://dx.
doi.org/10.1016/j.jrp.2005.08.007.
Gorsuch, R. L., & McPherson, S. E. (1989). Intrinsic/extrinsic measurement: I/E-revised
and single-item scales. Journal for the Scientific Study of Religion, 28, 348–354. http://
dx.doi.org/10.2307/1386745.
Graziano, W. G., Habashi, M. M., Sheese, B. E., & Tobin, R. M. (2007). Agreeableness,
empathy, and helping: A person × situation perspective. Journal of Personality and
Social Psychology, 93, 583. http://dx.doi.org/10.1037/0022-3514.93.4.583.
Gur, R. C., & Sackeim, H. A. (1979). Self-deception: A concept in search of a phenomenon.
Journal of Personality and Social Psychology, 37, 147–169. http://dx.doi.org/10.1037/
0022-3514.37.2.147.
Habashi, M. M., Graziano, W. G., & Hoover, A. E. (2016). Searching for the prosocial
personality: A big five approach to linking personality and prosocial behavior.
Personality and Social Psychology Bulletin, 42, 1177–1192. http://dx.doi.org/10.1177/
0146167216652859.
Hansen, D. E., Vandenberg, B., & Patterson, M. L. (1995). The effects of religious or-
ientation on spontaneous and nonspontaneous helping behaviors. Personality and
Individual Differences, 19. http://dx.doi.org/10.1016/0191-8869(95)00016-Y.
Heintzelman, S. J., Trent, J., & King, L. A. (2015). Revisiting desirable response bias in
well-being reports. The Journal of Positive Psychology, 10, 167–178. http://dx.doi.org/
10.1080/17439760.2014.927903.
Hilbig, B. E., Glöckner, A., & Zettler, I. (2014). Personality and prosocial behavior:
Linking basic traits and social value orientations. Journal of Personality and Social
Psychology, 107, 529. http://dx.doi.org/10.1037/a0036074.
Hill, P. C. (2005). Measurement in the psychology of religion and spirituality. In R. F.
Paloutzian, & C. L. Park (Eds.). Handbook of the psychology of religion and spirituality
(pp. 43–61). New York: The Guilford Press.
Ji, C. H. C., Pendergraft, L., & Perry, M. (2006). Religiosity, altruism, and altruistic hy-
pocrisy: Evidence from Protestant adolescents. Review of Religious Research, 156–178.
Johnston, M. E., Sherman, A., & Grusec, J. E. (2013). Predicting moral outrage and re-
ligiosity with an implicit measure of moral identity. Journal of Research in Personality,
47(3), 209–217. http://dx.doi.org/10.1016/j.jrp.2013.01.006.
Koban, K., & Ohler, P. (2016). Ladies, know yourselves! Gentlemen, fool yourselves!
Evolved self-promotion traits as predictors for promiscuous sexual behavior in both
sexes. Personality and Individual Differences, 92, 11–15. http://dx.doi.org/10.1016/j.
paid.2015.11.056.
Kramer, S. R., & Shariff, A. F. (2018). Religious prosociality: A meta-analysis. (Manuscript
in preparation. University of Oregon. Cited with permission).
Lee, K., & Ashton, M. C. (2016). Psychometric properties of the HEXACO-100. Assessment.
http://dx.doi.org/10.1177/1073191116659134.
Ludeke, S. G., & Carey, B. (2015). Two mechanisms of biased responding account for the
association between religiousness and misrepresentation in big five self-reports.
Journal of Research in Personality, 57, 43–47. http://dx.doi.org/10.1016/j.jrp.2015.
03.003.
McKay, R., & Whitehouse, H. (2015). Religion and morality. Psychological Bulletin, 141,
447–473. http://dx.doi.org/10.1037/a0038455.
Merritt, A. C., Effron, D. A., & Monin, B. (2010). Moral self-licensing: When being good
frees us to be bad. Social and Personality Psychology Compass, 4, 344–357. http://dx.
doi.org/10.1111/j.1751-9004.2010.00263.x.
Monsma, S. V. (2007). Religion and philanthropic giving and volunteering: Building
blocks for civic responsibility. Interdisciplinary Journal of Research on Religion, 3.
Norenzayan, A., & Shariff, A. F. (2008). The origin and evolution of religious prosociality.
Science, 322(5898), 58–62. http://dx.doi.org/10.1126/science.1158757.
Ones, D. S., Viswesvaran, C., & Reiss, A. D. (1996). Role of social desirability in per-
sonality testing for personnel selection: The red herring. Journal of Applied Psychology,
81(6), 660–679. http://dx.doi.org/10.1037/0021-9010.81.6.660.
Ozer, D., & Benet-Martinez, V. (2006). Personality and the prediction of consequential
outcomes. Annual Review of Psychology, 57, 401–421.
Paulhus, D. L., & John, O. P. (1998). Egoistic and moralistic biases in self-perception: The
interplay of self-deceptive styles with basic traits and motives. Journal of Personality,
66(6), 1025–1060.
Paulhus, D. L., & Reid, D. B. (1991). Enhancement and denial in socially desirable re-
sponding. Journal of Personality and Social Psychology, 60, 307–317. http://dx.doi.
231
Personality and Individual Differences 131 (2018) 222–231
org/10.1037/0022-3514.60.2.307.
Paulhus, D. L., & Trapnell, P. D. (2008). Self-presentation on personality scales: An
agency-communion framework. In O. P. John, R. W. Robins, & L. A. Pervin (Eds.).
Handbook of personality (pp. 493–517). New York, NY: Guilford.
Pelham, B., & Crabtree, S. (2008). Worldwide, highly religious more likely to help others.
http://www.gallup.com/poll/111013/worldwide-highly-religious-more-likely-help-
others.aspx.
Penner, L. A., Fritzsche, B. A., Craiger, J. P., & Freifeld, T. R. (1995). Measuring the
prosocial personality. Advances in Personality Assessment, 10, 147–163.
Putnam, R. D., Campbell, D. E., & Garrett, S. R. (2010). American grace: How religion
divides and unites us. New York, NY, US: Simon & Schuster.
Roese, N. J., & Jamieson, D. W. (1993). Twenty years of bogus pipeline research: A cri-
tical review and meta-analysis. Psychological Bulletin, 114, 363–375. http://dx.doi.
org/10.1037/0033-2909.114.2.363.
Rowatt, W. C., Ottenbreit, A., Nesselroade, K. R., & Cunningham, P. A. (2002). On being
holier-than-thou or humbler-than-thee: A social-psychological perspective on re-
ligiousness and humility. Journal for the Scientific Study of Religion, 41, 227–237.
http://dx.doi.org/10.1111/1468-5906.00113.
Rushton, J. P., Chrisjohn, R. D., & Fekken, G. C. (1981). The altruistic personality and the
self-report altruism scale. Personality and Individual Differences, 2, 293–302. http://dx.
doi.org/10.1016/0191-8869(81)90063-5.
Saroglou, V. (2009). Religiousness as a cultural adaptation of basic traits: A five-factor
model perspective. Personality and Social Psychology Review. http://dx.doi.org/10.
1177/1088868309352322.
Saroglou, V. (2013). Religion, spirituality, and altruism. In K. I. Pargament, J. J. Exline, &
J. W. Jones (Eds.). APA Handbook of psychology, religion and spirituality (vol. 1, pp.
439–457)Washington, DC: American Psychological Association. http://dx.doi.org/
10.1037/14045-024.
Saroglou, V., Delpierre, V., & Dernelle, R. (2004). Values and religiosity: A meta-analysis
of studies using Schwartz's model. Personality and Individual Differences, 37, 721–734.
http://dx.doi.org/10.1016/j.paid.2003.10.005.
Saroglou, V., & Munoz-Garcia, A. (2008). Individual differences in religion and spiri-
tuality: An issue of personality traits and/or values. Journal for the Scientific Study of
Religion, 47(1), 83–101. http://dx.doi.org/10.1111/j.1468-5906.2008.00393.x.
Saroglou, V., Pichon, I., Trompette, L., Verschueren, M., & Dernelle, R. (2005). Prosocial
behavior and religion: New evidence based on projective measures and peer ratings.
Journal for the Scientific Study of Religion, 44, 323–348. http://dx.doi.org/10.1111/j.
1468-5906.2005.00289.x.
Sedikides, C., & Gebauer, J. E. (2010). Religiosity as self-enhancement: A meta-analysis of
the relation between socially desirable responding and religiosity. Personality and
Social Psychology Review, 14, 17–36. http://dx.doi.org/10.1177/1088868309351002.
Shariff, A. F. (2015). Does religion increase moral behavior? Current Opinion in Psychology.
6, 108–113. http://dx.doi.org/10.1016/j.copsyc.2015.07.009.
Uziel, L. (2010). Rethinking social desirability scales from impression management to
interpersonally oriented self-control. Perspectives on Psychological Science, 5, 243–262.
http://dx.doi.org/10.1177/1745691610369465.
Van Lange, P. A., & Sedikides, C. (1998). Being more honest but not necessarily more
intelligent than others: Generality and explanations for the Muhammad Ali effect.
European Journal of Social Psychology, 28, 675–680. http://dx.doi.org/10.1002/(SICI)
1099-0992(199807/08)28:4<675::AID-EJSP883>3.0.CO;2-5.
Vitell, S., Bing, M. N., Davison, H., Ammeter, A. P., Garner, B. L., & Novicevic, M. M.
(2009). Religiosity and moral identity: The mediating role of self-control. Journal of
Business Ethics, 88(4), 601–613. http://dx.doi.org/10.1007/s10551-008-9980-0.
Von Hippel, W., & Trivers, R. (2011). The evolution and psychology of self-deception.
Behavioral and Brain Sciences, 34(1), 1–16. http://dx.doi.org/10.1017/
S0140525X10001354.
de Vries, R. E., Zettler, I., & Hilbig, B. E. (2014). Rethinking trait conceptions of social
desirability scales: Impression management as an expression of honesty-humility.
Assessment, 21(3), 286–299. http://dx.doi.org/10.1177/1073191113504619.
Walker, L. J., & Pitts, R. C. (1998). Naturalistic conceptions of moral maturity.
Developmental Psychology, 34, 403–419. http://dx.doi.org/10.1037/0012-1649.34.3.
Ward, S.J., King, L.A. (under review). Moral self-regulation, moral identity, and re-
ligiosity. University of Missouri, Columbia.
Zettler, I., Hilbig, B. E., Moshagen, M., & de Vries, R. E. (2015). Dishonest responding or
true virtue? A behavioral test of impression management. Personality and Individual
Differences, 81, 107–111. http://dx.doi.org/10.1016/j.paid.2014.10.007.