You are on page 1of 14

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/282235187

God Is Watching You . . . But Also Watching Over You: The Influence of
Benevolent God Representations on Secular Volunteerism Among Christians.

Article  in  Psychology of Religion and Spirituality · January 2015


DOI: 10.1037/rel0000040

CITATIONS READS
16 636

3 authors:

Kathryn A Johnson Adam Cohen


Arizona State University Arizona State University
42 PUBLICATIONS   569 CITATIONS    102 PUBLICATIONS   3,825 CITATIONS   

SEE PROFILE SEE PROFILE

Morris Alan Okun


Arizona State University
160 PUBLICATIONS   5,796 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Personality and Political Attitudes View project

Future self-continuity and persistence of college students View project

All content following this page was uploaded by Morris Alan Okun on 22 October 2015.

The user has requested enhancement of the downloaded file.


Psychology of Religion and Spirituality
God Is Watching You . . . But Also Watching Over You:
The Influence of Benevolent God Representations on
Secular Volunteerism Among Christians
Kathryn A. Johnson, Adam B. Cohen, and Morris A. Okun
Online First Publication, July 27, 2015. http://dx.doi.org/10.1037/rel0000040

CITATION
Johnson, K. A., Cohen, A. B., & Okun, M. A. (2015, July 27). God Is Watching You . . . But Also
Watching Over You: The Influence of Benevolent God Representations on Secular
Volunteerism Among Christians. Psychology of Religion and Spirituality. Advance online
publication. http://dx.doi.org/10.1037/rel0000040
Psychology of Religion and Spirituality © 2015 American Psychological Association
2015, Vol. 7, No. 4, 000 1941-1022/15/$12.00 http://dx.doi.org/10.1037/rel0000040

God Is Watching You . . . But Also Watching Over You: The Influence of
Benevolent God Representations on Secular Volunteerism
Among Christians
Kathryn A. Johnson, Adam B. Cohen, and Morris A. Okun
Arizona State University

One prominent theory is that prosociality is promoted by the belief in an authoritarian God. Building
upon this theory, we developed a theoretical model in which beliefs about the self and the world and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

volunteer motives account for differential effects of benevolent and authoritarian God representations on
This document is copyrighted by the American Psychological Association or one of its allied publishers.

secular volunteerism (benefiting those outside the family or religious group). This model was tested with
undergraduate (Study 1) and community samples (Study 2) of Christian theists. In support of our model,
representations of a benevolent God were positively associated with a benevolent self-identity and a
moral obligation, with a significant total positive indirect effect on secular volunteerism via internal
motivations. In contrast, representations of an authoritarian God were associated with a low benevolent
self-identity and a significant total negative indirect effect on secular volunteerism. The effects of God
representations on volunteerism via religious obligation and external motivation (eternal rewards) were
inconsistent across samples.

Keywords: God representation, supernatural punisher, intrinsic motivation, volunteerism

The vicissitudes of life dictate that all people will, from time to through a combination of mechanisms, including the fear of divine
time, need assistance from others. Consequently, helping others is punishment, the promise of eternal rewards, and by capitalizing on
crucial to human flourishing. Moreover, helping others benefits concerns involving social monitoring (Norenzayan, 2013; Roes &
both the recipient and the benefactor (Okun, Yeung, & Brown, Raymond, 2003). If people share the belief that God is watching in
2013). It is not surprising, then, that helping is a universally order to reward or punish, they will be more likely to help not only
recognized virtue and all religious traditions stress the importance kith and kin, but also the strangers among them. Belief in a
of helping others (Dovidio, Piliavin, Schroeder, & Penner, 2006). punishing, authoritarian God has been shown to have certain
Indeed, religious people are often the most likely to engage in societal benefits such as inhibiting cheating (Shariff & Noren-
helping behaviors such as volunteerism (Jackson, Bachmeier, zayan, 2011), and there is evidence that groups that focus on the
Wood, & Craft, 1995; Ruiter & DeGraaf, 2006). Yet the psycho- authoritarian nature of God are less subject to in-group fragmen-
logical mechanisms that link religion and helping have not been tation—particularly in harsh environments (Snarey, 1996).
well understood, particularly when prosociality is for the benefit of On the contrary, beliefs in an authoritarian God can also have
strangers. negative outcomes. When authoritarian God representations are
the focus, people become more aggressive and willing to punish
God Representations and Prosociality (Bushman, Ridge, Das, Key, & Busath, 2007), more likely to
characterize the misfortune of others as divine punishment (Aten et
A prominent theory regarding the link between religion and
al., 2008; Froese & Bader, 2010), and are sometimes actually less
prosociality is that belief in a moralizing, punishing, authoritarian
willing to help others—particularly those outside their own group,
God fosters extrinsically motivated cooperation and prosociality
in a form of parochial prosociality (Johnson, Li, Cohen, & Okun,
2013; Norenzayan et al., in press; Ritter & Preston, 2013). Au-
thoritarian God representations have also been positively associ-
ated with the value of power but negatively correlated with the
Kathryn A. Johnson, Adam B. Cohen, and Morris A. Okun, Department value of benevolence (Johnson, Okun, & Cohen, 2014). The in-
of Psychology, Arizona State University. terpretation of findings regarding the effects of belief in an au-
The authors thank the John Templeton Foundation and the Notre Dame thoritarian God on prosocial behavior is further clouded inasmuch
Science of Generosity for supporting this research. The data in Study 1 was as nearly all of the studies using primes investigating religiously
used in Kathryn A. Johnson’s doctoral dissertation and the results have linked prosociality have primed religion in general, rather than
been published online–Johnson, K. A. (2013). From beliefs to virtuous
focusing on specific beliefs about the nature of God (Galen, 2012;
behaviors: The influence of God-concepts on intentions to volunteer (Order
No. AAI3518607). Retrieved from Dissertation Abstracts International:
Preston, Ritter, & Hernandez, 2010). Thus, it is unclear whether an
Section B: The Sciences and Engineering. authoritarian God, per se, promotes prosociality.
Correspondence concerning this article should be addressed to Kathryn Recent research (Johnson, Li, et al., 2013) suggests that proso-
A. Johnson, Department of Psychology, Arizona State University, P.O. ciality is also likely to be associated with the belief that God is
Box 871104, Tempe, AZ 85287-1104. E-mail: kathryn.a.johnson@asu.edu gracious, loving, kind, and merciful—a God who watches over you

1
2 JOHNSON, COHEN, AND OKUN

like a compassionate parent figure (Benson & Spilka, 1973) or directly benefit their offspring (e.g., little league coaching; Gar-
who models compassionate behavior by healing the sick or send- land, Myers, & Wolfer, 2008) or their religious group (Johnson,
ing rain in times of drought. A benevolent God is thought to be Cohen, & Okun, 2013). However, it is often dissimilar others and
aware of one’s needs, attentive, protective, loving, and forgiving, those outside one’s religious or social group that need help—the
among other positive attributes (Froese & Bader, 2010). Repre- poor, the immigrant, prisoner, orphan, and a host of others. There-
sentations of God as benevolent are associated with positive im- fore, we specifically focused on nonkin, nonkith, “secular” vol-
ages of the self (Benson & Spilka, 1973; Roberts, 1989), the value unteering—planned acts of helping directed toward those outside
of benevolence (Johnson et al., 2014), the belief that misfortune one’s own family or religious group.
occurs due to natural causes rather than as divine punishment
(Froese & Bader, 2010), and with relatively higher rates of vol-
Theoretical Model
unteerism (Johnson et al., 2013).
Yet no theoretical model has emerged to explain how belief in We developed a theoretical model in which representations of
a benevolent God is linked to prosociality. Building upon previous God as benevolent and authoritarian were theorized to be differ-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

research discussed ahead, we sought to elucidate the relation entially associated with beliefs about the self and the world: a
This document is copyrighted by the American Psychological Association or one of its allied publishers.

between benevolent God representations and prosociality. Com- benevolent self-identity, moral obligations to help, and religious
plementing the notion that authoritarian God representations are obligations to help1 (Figure 1). Beliefs about the self and the world
theorized to facilitate prosociality via the expectation of external were then posited to predict volunteer motives. These motives, in
rewards and punishments (i.e., the approval of important others turn, were predicted to be associated with secular volunteerism.
such as God and religious group members; afterlife rewards), we In the present research, our focus was on the frequency with
posited that benevolent God representations facilitate prosocial which participants had volunteered for the benefit of those outside
behaviors via internal motivations activated by, for example, a their family or religious group during the previous year. However,
benevolent self-identity or an internalized moral obligation to help there are many situational factors that may prevent (or compel)
others. individuals from engaging in a planned behavior such as volun-
Using undergraduate (Study 1) and community (Study 2) sam- teering (e.g., lack of opportunity, disability or illness, other time
ples of Christian theists, we tested a mediational model that incor- commitments, etc.). Thus, following Johnson, Li, et al. (2013), we
porates both authoritarian and benevolent God representations. We also assessed intentions to engage in specific acts of secular
hypothesized that these two God representations would be differ- volunteerism in the future.
entially associated with a set of beliefs about the self and the world
which, in turn, would differentially motivate prosociality.
Benevolent and Authoritarian God Representations
Beliefs about the nature and attributes of God are multidimen-
Assessing Prosociality
sional (Kunkel, Cook, Meshel, Daughtry, & Hauenstein, 1999) and
Prosociality is broadly defined as acts of helping intended to vary among individuals even in the same religious group (Froese
benefit others. Helping behaviors can be classified in at least three & Bader, 2010). For example, people may think of God as in-
ways: (a) planned versus spontaneous helping, (b) indirect versus volved versus uninvolved (Froese & Bader, 2010), mystical versus
direct helping (Pearce & Amato, 1980), and (c) according to the anthropomorphic (Kunkel et al., 1999), or near versus distant
recipient of the helping behavior (i.e., ingroup or outgroup). (Spilka, Armatas, & Nussbaum, 1964). In the present research, our
Whereas spontaneous helping involves unplanned acts such as focus was on the benevolent and authoritarian dimensions in
helping a stranger change a flat tire, planned helping involves thinking about God which have been identified in previous re-
deliberate efforts to assist others (e.g., delivering weekly meals to search (Benson & Spilka, 1973; Johnson et al., 2014).
a frail grandparent). Indirect helping includes generous acts such Recent research has focused on views of a punishing, authori-
as giving one’s possessions to charity, whereas direct helping tarian God as being associated with depression, neuroticism, or
entails giving one’s time to help those in need. We chose to focus spiritual struggle (Rosmarin, Krumrei, & Andersson, 2009; Wood
on volunteerism, a relatively high-cost type of helping behavior. et al., 2010). However, other researchers have shown that author-
This is important because, although researchers have demonstrated itarian God representations are normative for many religious ad-
an association between public forms of religiosity (e.g., church herents (Froese & Bader, 2010) and, depending upon the measures
attendance) and volunteering, there has been insufficient research used, are generally uncorrelated with neuroticism or insecure at-
investigating how religious cognition might impact volunteerism. tachment styles (Johnson et al., 2014). It may be that benevolent
and authoritarian God representations simply address different
Secular Volunteerism
1
Earlier versions of the model included variables for Personal Respon-
Organizational volunteering has been defined as unpaid,
sibility (Penner, Fritzsche, Craiger, & Freifeld, 1995) and Belief in a Just
planned, direct acts of helping, within an institutional framework, World (Furnham, 2003) in the category “Beliefs about the Self and the
that provide benefits to others (Piliavin & Siegl, 2007). The World.” However, across seven different samples (including samples of
recipients of help are likely to be relatives (Bryan, Hammer, & Mormons, Muslims, Jews, and individuals who reported being Spiritual-
Fisher, 2000), people who are like the helper (Dovidio et al., but-not-Religious), the indirect paths from God representation to Secular
Volunteerism via Personal Responsibility or Belief in a Just World failed
2006), or people who share group memberships with the helper to reach significance in all but two samples. Therefore, in an effort to
(Flippen, Hornstein, Siegal, & Weitzman, 1996). Much of the provide a more parsimonious model, we have omitted these variables from
literature on volunteerism has included activities by helpers that the final model.
GOD REPRESENTATIONS AND VOLUNTEERISM 3
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Figure 1. Standardized path coefficients in the revised theoretical model in Study 1. N ⫽ 426. Dotted lines
indicate nonsignificant pathways. The correlations between variables within each class of variables (i.e., God
representations, Beliefs about the Self and the World, Volunteer Motives) and the nonsignificant direct paths
from Benevolent God and Authoritarian God to Secular Volunteerism were included in the model analyses, but
not shown here for clarity. ⴱⴱⴱ p ⱕ .001.

psychological needs and predict different social attitudes and be- tives: no motivation (amotivation), adopting a value as one’s own
haviors. Whereas authoritarian God representations may success- (identification), a sense that the value is emanating from the self
fully inhibit antisocial behavior, benevolent God representations (integration), feelings of guilt or self-esteem (introjection), com-
may be more likely to elicit prosocial intentions and behaviors. pliance with external regulations (external), and, finally, a purely
Nevertheless, we also tested the aspect of the supernatural punisher intrinsic motivation of personal interest or enjoyment. Identified,
hypothesis which posits that belief in an authoritarian God leads to integrated, and intrinsic (pleasure) motivations are considered to
a religious obligation to help and, in turn, external motivations be autonomous and internal, meaning that the desire to act is
(i.e., concern for one’s reputation before God and the religious perceived as a feeling of volition without external rewards or
group), indirectly leading to secular volunteering. punishments. SDT holds that people who are self-motivated rather
than externally controlled (by coercion or reward) are more en-
Effects of Benevolent God Representations via thusiastic, persistent, creative, and satisfied with life. Grano, Lu-
Benevolent Self-Identity cidi, Zelli, and Violani (2008) found that, as the value of helping
others becomes an integrated aspect of the self, positive attitudes
Church attendance is often positively correlated with volunteer- about helping increase, although intrinsic motivation (i.e., volun-
ism and some suggest this may be because individuals involved in teering simply for pleasure) was not a significant predictor of
organized religion receive more invitations to help from religious volunteer intentions.
leaders and peers—many of whom serve as role models of com- We predicted that benevolent God representations would be
passionate helping. One can think of other examples of benevolent positively associated with a benevolent self-identity which, in turn,
volunteer role models such as Mother Theresa or Gandhi. For would be strongly and positively associated with internal (identi-
Christians, the quintessential role model of compassionate helping fication and integration) motives to help others (H1), and nega-
may be God—who also modeled benevolence in the person of tively associated with amotivation (disinclination) (H2). Thus,
Jesus Christ. Thus, Christians may aspire to be like Jesus who benevolent God representations were hypothesized to have two
cared for the sick (Mark 5:19); fed the hungry (Matthew 9:36); and positive, indirect effects on secular volunteerism via benevolent
forgave all, even in death (Luke 23:34). Indeed, self-identifying as self-identity.
benevolent is a robust predictor of volunteerism (Finkelstein,
2009), and is underscored by research showing that a benevolent Effects of Benevolent God Representations via
self-identity predicts behavioral intentions over and above previ-
Moral Obligation
ous behavior, positive attitudes toward volunteering, or subjective
norms (Terry, Hogg, & White, 1999). Even people who do not self-identify as benevolent may still
The desire to express one’s benevolent self-identity can be value benevolence and have internalized moral intuitions about
classified as an internal motivation in terms of self-determination helping (White & Plous, 1995). When people observe others in
theory (SDT; Ryan & Deci, 2000). SDT posits six types of mo- need, internalized altruistic social norms activate a perceived
4 JOHNSON, COHEN, AND OKUN

moral obligation to help (Schwartz & Howard, 1980); and the Effects of Authoritarian God Representations
belief that one has a moral obligation to help others is often cited
as a primary reason for volunteering (Snyder & Omoto, 2000). A perceived religious obligation and extrinsic motivations (i.e.,
When reminded of their moral identity, people are also more likely fear of punishment, the expectation of reward, and reputational
to expand the sense of “we” and be willing to care for those outside concerns) are central to the supernatural watcher, or punishing
their own immediate family or social group (Reed & Aquino, God, hypothesis (Norenzayan, 2013). Therefore, we expected au-
2003). thoritarian God representations to be positively associated with
We reasoned that a perceived moral obligation may be espe- external motivations to help (H7), and also with feelings of guilt
cially keen for people who feel they have been the recipient of (introjected motivation) if one does not help (H8). In short, con-
divine help in times of trouble or blessings in times of need. sistent with the supernatural punisher hypothesis, we also expected
Religious people (here we focus on Christians) may feel obliged authoritarian God representations to have two positive, indirect
to extend those same benevolent acts toward others. Previous effects on secular volunteerism via religious obligations.
research indicates that most Christians represent God as benev-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

olent (Johnson et al., 2014), value benevolence (Schwartz & Study 1: Test of the Theoretical Model in an
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Huismans, 1995), and often strive to express their benevolent Undergraduate Sample
values via helping (Okun, O’Rourke, Keller, Johnson, & End-
Because many Christians believe that Jesus was God incarnate,
ers, 2014).
Christians are more likely to personify God compared with mem-
In terms of SDT, people with a perceived moral obligation to
bers of other monotheistic religious groups. Therefore, we chose to
help can be said to have mixed motives. On the one hand, they are
assess the beliefs and motives of Christians in testing the effects of
internally motivated by their values and moral intuitions to help.
personified God representations on volunteerism proposed in our
Yet internalized values and moral intuitions are often derived from
theoretical model. In Study 1, we tested the theoretical model in a
and reinforced by society and the expectations of important others.
sample of Christian undergraduate college students.
Morally obliged people may also feel proud if they do, and yet
guilty if they do not, volunteer (i.e., have introjected motivations).
We expected that benevolent God representations would be Method
positively associated with a perceived moral obligation to help Participants. Participants were 454 Christian (164 Catholics
others which, in turn, would be positively associated with (a) and 262 non-Catholic Christians) undergraduates attending a large
internal motivation to express benevolent values and moral intu- public university who were enrolled in either introduction to
itions about helping (H3); and (b) introjected motivation to help psychology or online sociology courses. Sociology students re-
(H4). Thus, benevolent God representations were hypothesized to ceived extra credit and psychology students received partial credit
have two positive indirect effects on secular volunteerism via toward fulfilling a course requirement. There were 162 males (63
moral obligations. were Catholic), 263 females (101 Catholic) and one with missing
data on gender. Participants were Euro American (66%), Hispanic
Effects of Benevolent God Representation via (20%), or other (14%) ethnicity, and were distributed across three
Religious Obligation socioeconomic status (SES) classes. Most reported being “middle
class” (49%), 22% were “lower or working class,” and 29% were
Moral obligations and religious obligations can be differentiated “upper middle or upper class.”
(Morewedge & Clear, 2008). Whereas moral obligations arise Screening criteria. Belief in God is a central assumption in
from internalized values and moral intuitions, religious obligations the proposed theoretical model of religious volunteerism. There-
are construed here as being external to the self. For example, God fore, Christian participants who responded with a rating of less
commands people to help those less fortunate (e.g., Leviticus 23). than 3, “God might exist,” on a five-option multiple choice ques-
We predicted that Christians would acknowledge a religious obli- tion at the beginning of the survey, were excluded from the study
gation to help regardless of their God representation. (n ⫽ 21).
Religious obligations are, in turn, associated with a system of To guard against careless participation, we included a distractor
(often eternal) rewards and punishments. The promise of such page explaining that people often hurry through surveys and ask-
rewards can also motivate prosociality (Harrell, 2012). In an SDT ing participants not to provide a 1 to 7 Likert response to the
framework, religious obligation was expected to be most closely statement, “Here is the question that you should not answer.”
related to external motivation—volunteering in order to earn rec- Participants who responded with any rating were deemed to have
ognition, to obtain rewards, and to avoid punishment or criticism not followed instructions and were dropped (n ⫽ 7).
from important others (including God). As discussed above, how-
ever, external volunteer motivations are typically less potent than
Procedure
internal ones (Finkelstein, 2009; Stukas, Snyder, & Clary, 1999).
We expected that benevolent God representations would be The online survey consisted of three sections: Beliefs about the
positively associated with a perceived religious obligation to help Self and the World; Volunteering (i.e., Volunteer Motivations,
others which, in turn, would be positively associated with external Intentions to Volunteer in the future, and previous Secular Volun-
motivations to help (H5), and also with feelings of guilt (intro- teer experience for the benefit of those outside one’s family and
jected motivation) if one does not help (H6). Thus, we expected religious group); and Religious Beliefs. The three sections were
benevolent God representations to have two positive indirect ef- presented as though they were unrelated studies as described
fects on secular volunteerism via religious obligations. ahead. In order to guard against order effects, there were two
GOD REPRESENTATIONS AND VOLUNTEERISM 5

versions of the survey: (a) Self & World, Volunteering, and and “. . . because I want to avoid being criticized by others in my
Religion, and (b) Volunteering, Self & World, and Religion. Each religious group.” The remaining subscales, Introjected (␣ ⫽ .79)
section also included distractor items which were not included in and Amotivation (␣ ⫽ .86), were not modified. Because volun-
the model or analyzed here. teering strictly for personal pleasure does not predict volunteerism
independently of the other volunteer motivations (Grano et al.,
2008), this scale was excluded from the theoretical model.
Measures
Secular volunteerism. The final variable in the path model
Ratings for each of the following scales were 1 (strongly dis- was the frequency with which people had volunteered for the
agree) to 7 (strongly agree). benefit of those outside their family or religious group during the
God representations. Indicators of belief in a Benevolent past year (SecVolBHVR; Secular Volunteer Behavior). Partici-
God (B-God; God is helping, forgiving, merciful, gracious, gen- pants read a statement that defined organizational volunteering as
erous, compassionate, caring, accepting; ␣ ⫽ .90) and an Author- a service performed without compensation for an organization or
itarian God (A-God; God is commanding, strict, controlling, re- agency. Examples were provided of several different types of
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

stricting, stern, judging, angry, wrathful, punishing; ␣ ⫽ .86) were organizations such as churches/religious groups, social service
This document is copyrighted by the American Psychological Association or one of its allied publishers.

derived from previous research (e.g., Benson & Spilka, 1973; agencies, schools, hospitals, Red Cross, political parties, and en-
Johnson et al., 2014; Kunkel et al., 1999). Participants were asked vironmental conservation groups. Participants responded to the
to rate the extent to which they agreed that each of the randomly question, “During the PAST YEAR, how often did you perform
ordered adjectives were descriptive of God. volunteer services for an organization that benefited people outside
your family or religious group?” Participants rated their frequency
of volunteering on a 7-point scale, with response options of
Beliefs About the Self and the World
“never” (coded 1), “a few days per year” (coded 2), “1 day per
Benevolent self. A benevolent self-identity (BenSelf) was as- month” (coded 3), “2–3 days per month” (coded 4), “once a week”
sessed using a new measure adapted from Aquino and Reed’s (coded 5), “2–3 days per week” (coded 6), and “almost daily”
(2002) measure of moral identity (␣ ⫽ .79). Participants were (coded 7). The mean rating of frequency of volunteering was 2.34
asked to imagine a person who was “caring, compassionate, ac- (SD ⫽ 1.35).
cepting of others, generous, and helpful”—adjectives characteriz- Intentions to volunteer. Whereas our secular volunteerism
ing a benevolent person and derived from value scales (Schwartz, measure assessed volunteering during the past year, a secondary
1992). Following Aquino and Reed, participants were then asked goal was to assess intentions to engage in secular volunteerism in
to rate the extent to which they agreed with items such as, “I am the future. Therefore, following Johnson, Li, et al. (2013), we also
the sort of person who has these characteristics.” assessed intentions to volunteer for specific secular causes. To
Moral obligation. In light of the absence of a published scale sample across individual proclivities, we asked about a variety of
to assess perceived moral obligation to volunteer, we created a volunteer possibilities. Participants were asked how likely they
four-item scale (␣ ⫽ .77): “People have a moral obligation to would be to volunteer over the next 12 months by: “Helping
volunteer to help others,” “I personally feel I have a moral obli- underprivileged youths learn to read,” “Distributing reading ma-
gation to volunteer to help others,” “If a stranger needs help, a terials to hospice patients,” “Helping distribute food at a local food
person who is able to provide it has a moral obligation to do so,” bank,” “Helping build housing for poor families,” and “Delivering
and “Helping others is an important moral activity.” food to families whose husbands are in prison” (SecVolINT,
Religious obligation. Five items were created to assess the Secular Volunteer Intentions; ␣ ⫽ .92).
perceived religious obligation to help others (␣ ⫽ .81): “God
expects people to obey the commandments,” “I have a religious
Results
obligation to help others,” “God commands people to help one
another” “God expects people to care for the needs of others” and Descriptive statistics for model variables. Given the com-
“The scriptures command people to help others.” plexity of the model, scale scores from observed data were com-
Volunteer motivations. The items used to assess volunteer puted for each of the measured God representations, beliefs about
motivations were adapted from the Volunteer Motivations Scale self and the world, volunteer motivations, and the two secular
(Grano et al., 2008) which consists of six subscales: Integrated volunteerism variables, rather than taking a latent variable ap-
(identity as a volunteer), Identified (valuing volunteerism), Intro- proach. This was done by taking the average of all the items for
jected (feeling guilty for not volunteering), External (volunteering each scale. The means, standard deviations, and correlations
at the behest of others), Amotivation (not volunteering), and Per- among all variables in the model are shown in Table 1. As shown,
sonal Pleasure (volunteering for personal pleasure). A-God and B-God were not directly correlated with SecVolBHVR
Because Integrated and Identified motives have been shown to (M ⫽ 2.34, SD ⫽ 1.35). However, B-God was positively corre-
be highly correlated in previous research (Grano et al., 2008) and lated with SecVolINT (M ⫽ 4.66, SD ⫽ 1.34), r ⫽ .17, p ⬍ .001.
were correlated .64 in the present study, we combined the items on In terms of motivational predictors, SecVolBHVR and SecVolINT
these two scales (Internal Motivation; (␣ ⫽ .87). In so doing, we were most strongly correlated with Internal motivations, r ⫽ .34
were able to test a more parsimonious model. The External scale and .47, p ⬍ .001. This is consistent with previous research
(Grano et al., 2008) was modified to focus on extrinsic motivations applying self-determination theory to volunteer motivations
with respect to religion (␣ ⫽ .72): “I volunteer because God (Grano et al., 2008).
rewards people who help others,” “. . . because I will earn rewards Path analysis of indirect effects in the theoretical model.
in the afterlife,” “. . . because God is pleased when I volunteer,” Using MPlus 7.2 (Muthén, & Muthén, 1998 –2012), with robust
6 JOHNSON, COHEN, AND OKUN

Table 1
Means, Standard Deviations, and Zero-Order Correlations Between the Manifest Variables in the Theoretical Model, Study 1

B-God A-God Ben Self Moral Oblig Relig Oblig Amotive Internal Introject External Vol Intent

Mean 6.21 3.28 5.98 4.70 5.26 1.84 4.94 3.64 3.22 4.66
SD .89 1.23 .92 1.10 1.32 1.17 1.10 1.42 1.31 1.34
A-God ⫺.00
BenSelf .34ⴱⴱ ⫺.22ⴱⴱ
MoralOblig .23ⴱⴱ .06 .31ⴱⴱ
ReligOblig .38ⴱⴱ .36ⴱⴱ .13ⴱⴱ .31ⴱⴱ
Amotive ⫺.26ⴱⴱ .23ⴱⴱ ⫺.48ⴱⴱ ⫺.20ⴱⴱ ⫺.05
Internal .20ⴱⴱ .02 .34ⴱⴱ .42ⴱⴱ .21ⴱⴱ ⫺.23ⴱⴱ
Introjected .02 .05 .11ⴱ .40ⴱⴱ .20ⴱⴱ .09 .47ⴱⴱ
External .13ⴱⴱ .28ⴱⴱ ⫺.00 .21ⴱⴱ .39ⴱⴱ .25ⴱⴱ .25ⴱⴱ .40ⴱⴱ
Vol Intent .17ⴱⴱ ⫺.00 .25ⴱⴱ .30ⴱⴱ .22ⴱⴱ ⫺.22ⴱⴱ .47ⴱⴱ .27ⴱⴱ .20ⴱⴱ
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Vol Secular .02 .03 .01 .11ⴱⴱ .06 ⫺.05 .34ⴱⴱ .16ⴱⴱ .04 .15ⴱⴱ
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Note. N ⫽ 425.

p ⬍ .05. ⴱⴱ p ⬍ .01.

maximum likelihood (MLR) estimation, we tested a path model Indirect effects of B-God on secular volunteer behavior.
specifying the predicted paths among the manifest variables in There were multiple indirect effects that help to understand why
accord with our theoretical model All variables within each class representations of God as benevolent may facilitate volunteering
of variables (i.e., God representations, beliefs, and motivations) for the benefit of those outside the family or religious group (Table
were allowed to correlate. For example, Amotivation, Internal, 2). As expected, B-God was positively associated with BenSelf
Introjected, and External motives were allowed to correlate with which, in turn, was positively associated with Internal Motivation
each other in the volunteer motives class. The direct paths from (H1), which was the strongest motivational predictor of SecVolB-
God representations (i.e., B-God and A-God) to SecVolBHVR HVR, as can be seen in Figure 1. Also as expected, B-God was
were also included in the model. positively associated with Moral Obligation which, in turn, was
We evaluated the model fit by examining the chi-square positively associated with Internal Motivation (H3), a positive
goodness-of-fit test, root mean square error of approximation predictor of SecVolBHVR. The positive effect of B-God on
(RMSEA), comparative fit index (CFI), and standardized root- SecVolBHVR via Amotivation (H2) was not significant.
mean-square residual (SRMR). An RMSEA less than .05, CFI Also contrary to our predictions (H4), although Moral Obliga-
greater than .95, and SRMR less than .05 indicate good model fit tion was associated with Introjected Motivation (i.e., guilt), Intro-
(Kline, 2005). We found that the fit of the theoretical model to the jected Motivation was not a significant predictor of SecVolBHVR.
data was adequate but not good according to conventional stan- A simple explanation is that the reduction of negative emotional
dards, ␹2(19) ⫽ 78, p ⬍ .05, RMSEA ⫽ .09, CFI ⫽ .92, SRMR ⫽
states such as guilt is insufficient to motivate people to help
.06.
unrelated others. Moreover, as can be seen in Table 1, Introjected
Post hoc model modifications are typically required to develop
Motivation was also correlated with both Internal Motivation, r ⫽
better fitting path models (Byrne, 2012). The goal is to specify the
.47, p ⬍ .001 and External Motivation, r ⫽ .40, p ⬍ .001, but less
most parsimonious model with both good theoretical justification
highly correlated with SecVolBHVR, r ⫽ .16, p ⫽ .001. It may be
for each estimated parameter and with good fit statistics. To
that feelings of guilt for not volunteering come into play only via
determine which paths to add, we consulted the modification
the other motivational pathways.
indices provided in the MPlus output. Modification indices are
estimates of the incremental reduction in chi-square that could be B-God was also positively associated with Religious Obligation;
achieved by the addition of specified pathways. However, paths however, contrary to our predictions (H6), Religious Obligation
suggested by the modification indices were added only if doing so was not a significant predictor of Introjected Motivation, suggest-
was reasonable within the theoretical framework as discussed ing that guilt for not helping may be activated by the failure to
above. follow internalized values and moral intuitions (i.e., Moral Obli-
The modification indices suggested that adding a negative path gation) rather than religious imperatives. Instead, and as expected,
from A-God to BenSelf would improve the model fit. This is Religious Obligation was associated with External Motivation
consistent with studies showing that self-identity often reflects (H5). External motivation (religious rewards for helping) was not
representations of God (Roberts, 1989). In other words, if people a significant predictor of SecVolBHVR.
represent God as punishing, wrathful, and stern, they may be less Indirect effects of A-God on secular volunteer behavior. In
likely to see themselves as compassionate and caring. The incre- contrast to the punishing God hypothesis, the indirect effects of
mental improvement in chi-square for the added pathway was A-God on SecVolBHVR were negative, with a significant total of
significant, p ⬍ .001. The respecified model provided a good fit the standardized indirect effects ⫽ ⫺2.89, p ⫽ .004 (see Table 2).
for the data, MLR ␹2(18) ⫽ 52, p ⬍ .05, RMSEA ⫽ .07, CFI ⫽ As expected, A-God was positively associated with Religious
.96, SRMR ⫽ .04. The standardized coefficients for the significant Obligation which, in turn, was positively associated with External
pathways predicting SecVolBHVR in the final model are depicted Motivations (H7). However, as discussed above, External Motiva-
in Figure 1. tions were not a positive predictor of SecVolBHVR. Furthermore,
GOD REPRESENTATIONS AND VOLUNTEERISM 7

Table 2
Standardized Indirect Effects Relative to Standard Errors From Benevolent (B-God) and
Authoritarian (A-God) to Secular Volunteer Behavior and Intentions in Studies 1 and 2

Volunteer behavior Volunteer intentions


(SecVolBHVR) IE/SE (SecVolINT) IE/SE
Model pathway Study 1 Study 2 Study 1 Study 2

Benevolent God
B-God ¡ BenSelf ¡ Amotivation ⫺.64 ⫺1.42 2.85ⴱⴱ 1.34
B-God ¡ BenSelf ¡ Internal 3.67ⴱⴱ 6.18ⴱⴱ 3.54ⴱⴱ 5.98ⴱⴱ
B-God ¡ Moral Oblig ¡ Internal 3.71ⴱⴱ 6.10ⴱⴱ 3.47ⴱⴱ 5.85ⴱⴱ
B-God ¡ Moral Oblig ¡ Introjected .36 ⫺.37 1.13 ⫺.00
B-God ¡ ReligOblig ¡ Introjected .35 ⫺.37 .83 ⫺.00
B-God ¡ ReligOblig ¡External ⫺1.34 ⫺2.92ⴱⴱ 2.08ⴱ ⫺.55
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Total indirect effects of benevolent God 2.64ⴱⴱ 5.38ⴱⴱ 6.18ⴱⴱ 7.51ⴱⴱ


This document is copyrighted by the American Psychological Association or one of its allied publishers.

Authoritarian God
A-God ¡ BenSelf ¡ Amotivation .63 1.35 ⫺2.78ⴱⴱ ⫺1.34
A-God ¡ BenSelf ¡ Internal ⫺3.65ⴱⴱ ⫺3.93ⴱⴱ ⫺3.42ⴱⴱ ⫺3.82ⴱⴱ
A-God ¡ ReligOblig ¡ Introjected .35 ⫺.37 .81 ⫺.00
A-God ¡ ReligOblig ¡ External ⫺1.32 ⫺2.74ⴱⴱ 2.08ⴱ ⫺.55
Total indirect effects of authoritarian God ⫺2.89ⴱⴱ ⫺4.82ⴱⴱ ⫺1.87† ⫺3.82ⴱⴱ
ⴱ ⴱⴱ

p ⫽ .06. p ⬍ .05. p ⬍ .01.

Religious Obligation was not a significant predictor of Study 2: Test of the Theoretical Model in a
SecVolBHVR via Introjected (H8). Community Sample
Although not hypothesized, there was a significant negative
indirect effect of A-God on secular volunteerism via BenSelf and In concert, the significant indirect pathways observed in Study
Internal Motivations. This is in contrast to the positive effects of 1 highlight the ways in which representations of God as benevolent
B-God via BenSelf. may foster secular volunteerism in a Christian population. In Study
Indirect effects of God representations on secular volunteer 2, we sought to generalize our finding by testing our respecified
intentions. Although our primary focus was on the indirect model in a community sample of Christian adults (Figure 2).
effects of God representations on past secular volunteer behavior,
we also tested the indirect effects of God representations on future Participants and Procedure
secular volunteer intentions. The model fit the data well, MLR
Participants were 804 workers on the Mechanical Turk website
␹2 (18) ⫽ 52, p ⬍ .001, RMSEA ⫽ .07, CFI ⫽ .96, SRMR ⫽ .04.
(301 male, 503 female) who were paid $2 after completion of the
As shown in Table 2, there was a significant positive total indirect Benevolence Survey (see Study 1). Belief in God was assessed
effect of B-God and a negative total indirect effect of A-God on using a 5-option multiple choice question in a prescreening survey,
SecVolINT. and participants were disqualified if belief in the existence of God
was rated less than 3, “God might exist.” As in Study 1, the data
Discussion for Christians only (Catholics, n ⫽ 230, and non-Catholic Chris-
tians, n ⫽ 574) are reported here. Participants were mainly Euro
A number of tentative conclusions can be derived from the final American (72%), Black or African American (12%), Hispanic
model. The results of Study 1 using a sample of Christian under- (7%), Asian or Asian American (5%), or other (4%). In terms of
graduates indicate that belief in a benevolent God—rather than SES, 46% reported being “middle class,” 42% “working or lower
belief in a punishing God—is more strongly and positively asso- class,” 12% “upper middle or upper class,” with four participants
ciated with both previous secular volunteer experience and with not responding. With respect to age, 42% of the participants were
intentions to volunteer for specific secular causes in the future. 19 –29 years old; 27% from 30 to 39 years of age; 23% from 40 to
This can be explained in that a benevolent self-identity and a 49 years of age; and 8% over age 50.
perceived moral obligation to help others were both positively
associated with representations of God as benevolent, but nega-
Results and Discussion
tively associated or uncorrelated with authoritarian God represen-
tations. Further, a benevolent self-identity and the belief that one As in Study 1, scores for both B-God and A-God were uncor-
has a moral (but not necessarily a religious) obligation to help were related with SecVolBHVR. However, B-God was positively cor-
positive predictors of internal motivations to volunteer. Consistent related with (SecVolINT), r ⫽ .23, p ⬍ .001. Scores for A-God
with previous research (Gagné, 2003; Grano et al., 2008), internal were negatively correlated with SecVolINT, r ⫽ ⫺.13, p ⫽ .015.
motivation was a stronger predictor of secular volunteerism than Although not presented here, the means, standard deviations, and
was external motivation. Introjected motivation, feeling guilty correlations between the variables were identical in direction and
about not helping, was not a significant predictor of secular vol- similar in magnitude when compared with those in Study 1 (see
unteerism after controlling for internal and external motives. Table 1).
8 JOHNSON, COHEN, AND OKUN
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Figure 2. Path model from Study 1 with standardized path coefficients from Study 2. N ⫽ 804. Dotted lines
indicate nonsignificant pathways. The correlations between variables within each class of variables (i.e., God
representations, Beliefs about the Self and the World, Volunteer Motives) and the nonsignificant direct paths
from Benevolent God and Authoritarian God to Secular Volunteerism were included in the model analyses, but
not shown here for clarity. ⴱⴱ p ⱕ .01. ⴱⴱⴱ p ⱕ .001.

In specifying the model for Study 2, the same 10 variables from negative total indirect effect of A-God on SecVolINT (see
Study 1 were used. The respecified model from Study 1 provided Table 2).
a good fit to the data, ␹2 (18) ⫽ 103, p ⬍ .05, RMSEA ⫽ .08, Discussion. The results of Study 2 indicate that belief in a
CFI ⫽ .95, SRMR ⫽ .04. benevolent God—rather than an authoritarian God—is likely
Indirect effects of B-God on secular volunteer behavior. to be indirectly associated with both previous secular volunteer
As in Study 1, there were several significant indirect pathways behavior and the intention to volunteer in the future for the
from B-God to SecVolBHVR (see Table 2). The pathway from benefit of those outside the family or religious group. Benev-
B-God ¡Benevolent Self ¡ Internal Motivation ¡ SecVolB- olent God representation was a positive predictor of a benevo-
HVR was positive and significant; the path from B-God ¡ lent self-identity and a perceived moral obligation to help,
Moral Obligation ¡ Internal Motivation ¡ SecVolBHVR was whereas Authoritarian God representation was a negative pre-
also positive and significant. Although external motivation and dictor of a benevolent self-identity and uncorrelated with a
secular volunteerism were uncorrelated in terms of the bivariate moral obligation to help. This is important because a benevolent
correlation, the path from B-God ¡ Religious Obligation ¡ self-identity and moral obligation were significant positive pre-
External Motivation ¡ SecVolBHVR was negative and signif-
dictors of internal motivations, the strongest positive predictor
icant, suggesting that perceived religious obligations may have
of secular volunteerism.
a statistical suppressor effect via external motivations and may
Both benevolent and authoritarian God representations were
undermine secular volunteerism (see also Geiser, Okun, &
positively correlated with a perceived religious obligation to help
Grano, 2014).
others. However, the model results pertaining to religious obliga-
Indirect effects of A-God on secular volunteer behavior.
As in Study 1, the total indirect effects of A-God on SecVolB- tion were more difficult to interpret. As in Study 1, religious
HVR were significant and negative. This was explained by the obligation was a negative predictor of previous secular volunteer
negative path from A-God ¡ Benevolent Self ¡ Internal behavior via external motivation in the full model. This suggests
Motivation ¡ SecVolBHVR (see Table 2). Also, because Ex- that religious obligation may actually suppress internal motiva-
ternal Motivation was a negative predictor of SecVolBHVR in tions to volunteer for the benefit of those outside the family or
the full model, the path from A-God to SecVolBHVR via religious group (see also Geiser et al., 2014). These results are also
Religious Obligation was significant and negative. Other indi- in accord with previous research showing that external motives
rect paths from A-God to SecVolBHVR were not significant. (i.e., external rewards and punishments) are often less effective in
Indirect effects of God representations on secular volunteer eliciting volunteerism (Finkelstein, 2009; Gagné, 2003; Grano et
intentions. In terms of secular volunteer intentions, the model al., 2008; Stukas et al., 1999). We note, however, that the negative
provided a good fit to the data, ␹2 (18) ⫽ 101, p ⬍ .001, indirect effects of religious obligation on previous secular volun-
RMSEA ⫽ .08, CFI ⫽ .95, SRMR ⫽ .04. There was a signif- teer behavior were not significant with respect to secular volunteer
icant positive total indirect effect of B-God and a significant intentions.
GOD REPRESENTATIONS AND VOLUNTEERISM 9

General Discussion Roberts, 2008; Schloss & Murray, 2011). However, refraining
from antisocial behavior is not necessarily equivalent to engaging
We identified two pathways that account for the differential in prosocial behavior (Janoff-Bulman, Sheikh, & Hepp, 2009).
influences of benevolent and authoritarian God representations on While we do not dispute that authoritarian God representations
secular volunteerism. First, whereas benevolent God representa- discourage antisocial behavior within a group, this seems to only
tions are positively associated with a belief that the self is also partly explain why people are willing to help others— especially
benevolent, authoritarian God representations were uncorrelated when those others are outside one’s family or religious group.
(or negatively correlated) with a benevolent self-identity; and a We speculate that representations of God as authoritarian might
benevolent self-identity, in turn, was strongly associated with also be associated with the denigration of value-violating out-
internal volunteer motivations—the strongest predictor of secular groups (see also Froese & Bader, 2010), which, in turn, would
volunteerism. Second, representations of God as benevolent were most likely reduce helping (Batson, Denton, & Vollmecke, 2008).
associated with a perceived moral obligation to volunteer whereas This view is consistent with the idea that the religious prosociality
authoritarian God representations were not. Moral obligations are fostered by an authoritarian God might be parochial in nature
grounded on internalized values and moral intuitions about help-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

(Norenzayan, 2013). In other words, belief in an authoritarian God


ing. Like a benevolent self-identity, moral obligations are associ-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

may inspire helping in some contexts, but helping may be circum-


ated with internal motivations to help—a robust predictor of sec- scribed for the benefit of the ingroup (Blogowska & Saroglou,
ular volunteerism. This is consistent with previous research 2011; Preston & Ritter, 2013).
showing that people often volunteer as an expression of their In contrast to the punishing God hypothesis, we reasoned that
prosocial values (Okun et al., 2014), and helping others who are focusing on the benevolent aspects of God’s nature is more likely
unlike the self or outside one’s social group may provide an added to impel Christian theists to help others— even when the benefi-
opportunity to act on benevolent God beliefs and benevolent ciaries are those outside the religious or social group (Johnson, Li,
self-concepts (see also Stukas, Worth, Clary, & Snyder, 2009). et al., 2013, 2014). This can be explained in that benevolent God
Religious obligations appear to have a direct effect on external representations are associated with an individual’s benevolent self-
motivation but with inconsistent indirect effects on secular volun- identity and sense of moral obligation to help all others and, in
teerism. More specifically, in the community sample, the relation- turn, associated with internal motivations to engage in prosocial
ship between religious obligation and secular volunteer behavior behaviors such as volunteering for secular causes. While much
appears to be mediated by external motivation yet the indirect theory about religious beliefs and prosocial behavior has focused
effect is negative, suggesting that a perceived religious obligation on an authoritarian, punishing God, our results suggest that we also
may reduce secular volunteering. In contrast, the association be- need to consider the influence of belief in a benevolent God on
tween religious obligation and secular volunteer intentions in the prosocial behavior.
student sample appears to be mediated by external motivation but Our results also speak to an active debate as to whether or not
with a corresponding increase in secular volunteer intentions. religious prosociality actually exists at all (Galen, 2012; Preston et
Moreover, the significant negative indirect effect of religious ob- al., 2010). It may be that the mixed results in previous research are
ligation on secular volunteer behavior was reduced to nonsignifi- due, in part, to a failure to recognize individual differences in
cance with respect to future volunteer intentions. One possibility is religious beliefs. For example, experiments investigating the ef-
that as religious obligations become more salient (and, we specu- fects of religion on prosociality often utilize priming stimuli re-
late, with a corresponding increase in religious group commit- ferring to “God” or “religion” or “prayer” (Gervais & Norenzayan,
ments), Christians are less likely to have had volunteer experience 2012; Shariff & Norenzayan, 2007; cf., Pichon, Boccato, & Saro-
for the benefit of those outside the religious group even though glou, 2007). However, it is unclear which God representations or
they report being willing to do so. other components of religion these primes may actually be acti-
vating (Ritter & Preston, 2013). Because U.S. Christians are likely
Implications of This Research
to have higher scores on ratings of God as benevolent rather than
This research helps to integrate two disparate literatures—the authoritarian (Johnson et al., 2014), general religious primes may
psychology of religion and the psychology of volunteerism—to be activating benevolent God representations for many partici-
suggest that varying representations of God are associated with pants. However, depending upon the participant sample or the
different beliefs which, in turn, undergird intentions to volunteer specific religious words used, general religion primes may also
and volunteerism behavior for the benefit of those outside one’s sometimes be associated with authoritarian God representations
family or religious group. which could yield different experimental effects.
Psychology of religion. First, in terms of the psychology of Psychology of volunteerism. In the last century, part-time
religion, social scientists are becoming increasingly aware of the volunteerism has become a key avenue for extending help to those
extent to which religious and spiritual beliefs influence social in need. Although paid professionals still administer many chari-
attitudes. One prominent theory regarding the importance of reli- table organizations, it is estimated that nearly half of the adult
gious beliefs posits that belief in a high, punishing God watching population in the U.S. and Canada participate in at least some
from above fostered prosociality and, consequently, enabled the volunteer activities each year and many humanitarian efforts sim-
development of large scale societies (Norenzayan, 2013). There is ply would not exist without the support of volunteers. Moreover,
evidence that authoritarian God representations are associated with the benefits for the volunteer include increased subjective well-
strict social norms (Froese & Bader, 2010), the inhibition of being, self-esteem, educational and occupational achievement, re-
cheating (Shariff & Norenzayan, 2011), and the facilitation of duced risk of mortality, and reduced substance abuse and truancy
ingroup cohesion (Roes & Raymond, 2003) (cf., Sanderson & rates in youths (Wilson, 2000). Volunteering is a win-win situation
10 JOHNSON, COHEN, AND OKUN

because it benefits both the recipient and the volunteer (Okun et teer efforts were also outgroups in terms of social class, ethnicity,
al., 2013). Our results suggest that developing strategies focused local versus distant locality, and so forth.
on benevolence or on the benevolent nature of God may foster We also do not claim to have included all possible explanatory
helping, generally—and volunteering, specifically, thereby serving variables in our theoretical model. Indeed, people volunteer for
the betterment of the community as well as promoting the psycho- many other reasons that we have not examined here (Snyder &
logical well-being of volunteers. Omoto, 2000). In that same vein, we wish to stress that we are not
suggesting that being religious or that belief in God are prerequi-
sites for volunteering. Many nonreligious and nonbelieving indi-
Limitations and Future Directions viduals generously donate their resources and time every day for a
Path analysis is an excellent statistical tool for uncovering the host of secular causes. For example, merely priming a moral
links among a set of variables. If the specified model fits the data identity, without any explicit reference to religious beliefs, elicits
well, the significant pathways in the model begin to shed some a universalistic prosociality (Reed & Aquino, 2003). Other inter-
light on the possible causal relations among the variables. How- esting variables, such as personality, may also play an important
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

role in fostering volunteerism (Carlo, Okun, Knight, & de Guz-


ever, the cross-sectional data we have presented here can only go
This document is copyrighted by the American Psychological Association or one of its allied publishers.

man, 2005). Still, research shows that, across multiple cultures,


so far in explaining how beliefs about God might influence an
religious people are more likely to volunteer (Ruiter & DeGraaf,
individual’s thoughts, feelings, and behavior. We are aware of only
2006); and, in the present research, we demonstrated that repre-
a handful of published experiments that have activated specific
sentations of God as authoritarian and benevolent are differentially
God representations, as opposed to ideas about religion or about
associated with a set of beliefs about the self and the world which,
God in general (Bushman et al., 2007; Johnson et al., 2013;
in turn, differentially motivate religious prosociality—at least in
Roberts, 1989). It is even less clear how and to what extent one’s
terms of secular volunteerism.
own identity, moral intuitions, values, or motivations might shape
representations of God—perhaps in a bidirectional manner or as a
Conclusion
reinforcing cycle (see Epley, Converse, Delbosc, Monteleone, &
Cacioppo, 2009; Ross, Lelkes, & Russell, 2012). Longitudinal and We show here that, for Christians, representations of God as
experimental research is warranted to substantiate the causal path- benevolent are likely to indirectly lead to volunteerism for the
ways specified in our theoretical model. benefit of those outside the family or religious group. The explan-
There are also other God representations that were not examined atory model investigated here can potentially be extended to other
in the present research. For example, we suggested that Jesus may religious and secularist groups as well, whose members can also
serve as a benevolent role model for Christians; however, this was choose to focus on the good, the benevolent, and not the judgmen-
not directly tested in our model. Representations of God are tal, nature of their own human role models, saints, and deities. We
complex, and Christians may also think of God as a cosmic force suggest this may impel people to positively change their self-
rather than as a personal being; as a spirit within rather than as a concept and sense of moral obligation, cultivate helping motiva-
distant, sovereign ruler; and as engaged versus disinterested in tions, and thereby promote prosociality and universal compassion
human affairs. These and other God representations may influence over and above selfish concerns.
the values, intentions, and behaviors of Christians and other reli-
gious individuals in ways that we have not considered in the References
present research. It is also likely that people have more than one Aquino, K., & Reed, A., II. (2002). The self-importance of moral identity.
representation of God (Davis, Moriarty, & Mauch, 2013) and these Journal of Personality and Social Psychology, 83, 1423–1440. http://dx
may be activated in different contexts. .doi.org/10.1037/0022-3514.83.6.1423
Additionally, although it is useful and informative to understand Aten, J. D., Moore, M., Denney, R. M., Bayne, T., Stagg, A., Owens,
the beliefs and volunteer behaviors of Christians, representations S., . . . Jones C. (2008). God images following hurricane Katrina in south
of God as benevolent and authoritarian exist in nearly all theistic Mississippi: An exploratory study. Journal of Psychology and Theology,
religions. In Islam, for example, Allah is represented as both 36, 249 –257.
Batson, C. D., Denton, D. M., & Vollmecke, J. T. (2008). Quest religion,
punitive and merciful, and the same in Judaism; Hindu deities may
anti-fundamentalism, and limited versus universal compassion. Journal
be malevolent or benevolent. Further, some religious groups may
for the Scientific Study of Religion, 47, 135–145. http://dx.doi.org/
be more likely to cultivate a reverential fear of God and have a 10.1111/j.1468-5906.2008.00397.x
more powerful sense of religious obligation relative to Christians Benson, P., & Spilka, B. (1973). God image as a function of self-esteem
(e.g., Muslims; Warner, Kilinc, Hale, Cohen, & Johnson, 2015). In and locus of control. Journal for the Scientific Study of Religion, 12,
that case, the positive indirect effect of authoritarian God repre- 297–310. http://dx.doi.org/10.2307/1384430
sentations may be enhanced. Blogowska, J., & Saroglou, V. (2011). Religious fundamentalism and
We also chose to focus on a relatively high-cost prosocial limited prosociality as a function of the target. Journal for the Scientific
behavior among Christians: secular volunteerism. However, indi- Study of Religion, 50, 44 – 60. http://dx.doi.org/10.1111/j.1468-5906
.2010.01551.x
viduals in other religious groups may be more inclined to engage
Bryan, A. D., Hammer, J. C., & Fisher, J. D. (2000). Whose hands reach
in other kinds of helping behaviors rather than volunteerism (e.g.,
out to the homeless? Patterns of helping among high and low commu-
charitable giving). Benevolent and authoritarian God representa- nally oriented individuals. Journal of Applied Social Psychology, 30,
tions (or any number of other religious cognitions) may influence 887–905. http://dx.doi.org/10.1111/j.1559-1816.2000.tb02501.x
these other types of helping in different ways. Further, we did not Bushman, B. J., Ridge, R. D., Das, E., Key, C. W., & Busath, G. L. (2007).
assess whether the nonfamily, nonreligious beneficiaries of volun- When god sanctions killing: Effect of scriptural violence on aggression.
GOD REPRESENTATIONS AND VOLUNTEERISM 11

Psychological Science, 18, 204 –207. http://dx.doi.org/10.1111/j.1467- sonality and Social Psychology, 96, 521–537. http://dx.doi.org/10.1037/
9280.2007.01873.x a0013779
Byrne, B. M. (2012). Structural equation modeling with MPlus: Basic Johnson, K. A., Li, Y. J., Cohen, A. B., & Okun, M. A. (2013). Friends in
concepts, applications, and programming. New York, NY: Routledge. high places: The influence of authoritarian and benevolent God-concepts
Carlo, G., Okun, M. A., Knight, G. P., & de Guzman, M. R. T. (2005). The on social attitudes and behaviors. Psychology of Religion and Spiritu-
interplay of traits and motives on volunteering: Agreeableness, extra- ality, 5, 15–22. http://dx.doi.org/10.1037/a0030138
version, and prosocial value motivation. Personality and Individual Johnson, K. A., Cohen, A. B., & Okun, M. A. (2013). Intrinsic religiosity
Differences, 38, 1293–1305. http://dx.doi.org/10.1016/j.paid.2004.08 and volunteering during emerging adulthood: A comparison of Mor-
.012 mons with Catholics and non-Catholic Christians. Journal for the Sci-
Davis, E. B., Moriarty, G. L., & Mauch, J. C. (2013). God images and God entific Study of Religion, 52, 842– 851. http://dx.doi.org/10.1111/jssr
concepts: Definitions, development, and dynamics. Psychology of Reli- .12068
gion and Spirituality, 5, 51– 60. http://dx.doi.org/10.1037/a0029289 Johnson, K. A., Okun, M. A., & Cohen, A. B. (2014). The mind of the
Dovidio, J. F., Piliavin, J. A., Schroeder, D. A., & Penner, L. A. (2006). Lord: The development of a scale to assess benevolent and authoritarian
The social psychology of prosocial behavior. New York, NY: Psychol- God-concepts. Psychology of Religion and Spirituality, Advance online
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

ogy Press. publication. http://dx.doi.org/10.1037/rel0000011


This document is copyrighted by the American Psychological Association or one of its allied publishers.

Epley, N., Converse, B. A., Delbosc, A., Monteleone, G. A., & Cacioppo, Kline, R. B. (2005). Principles and practice of structural equation mod-
J. T. (2009). Believers’ estimates of God’s beliefs are more egocentric eling (2nd ed.). New York, NY: Guilford Press.
than estimates of other people’s beliefs. PNAS: Proceedings of the Kunkel, M. A., Cook, S., Meshel, D. S., Daughtry, D., & Hauenstein, A.
National Academy of Sciences of the United States of America, 106, (1999). God images: A concept map. Journal for the Scientific Study of
21533–21538. http://dx.doi.org/10.1073/pnas.0908374106 Religion, 38, 193–202. http://dx.doi.org/10.2307/1387789
Finkelstein, M. A. (2009). Intrinsic vs. extrinsic motivational orientations Morewedge, C. K., & Clear, M. E. (2008). Anthropomorphic God concepts
and the volunteer process. Personality and Individual Differences, 46, engender moral judgment. Social Cognition, 26, 182–189. http://dx.doi
653– 658. http://dx.doi.org/10.1016/j.paid.2009.01.010 .org/10.1521/soco.2008.26.2.182
Flippen, A. R., Hornstein, H. A., Siegal, W. E., & Weitzman, E. A. (1996). Muthén, L. K., & Muthén, B. O. (1998 –2012). Mplus user’s guide (7th
A comparison of similarity and interdependence as triggers for in-group ed.). Los Angeles, CA: Author.
formation. Personality and Social Psychology Bulletin, 22, 882– 893. Norenzayan, A. (2013). Big gods: How religion transformed cooperation
http://dx.doi.org/10.1177/0146167296229003 and conflict. Princeton, NJ: Princeton University Press.
Froese, P., & Bader, C. (2010). America’s four Gods: What we say about Norenzayan, A., Shariff, A. F., Willard, A. K., Slingerland, E., Gervais,
God - and what that says about us. New York, NY: Oxford University W. M., McNamara, R. A., & Henrich, J. (in press). The cultural evolu-
Press. tion of prosocial religions. Behavioral and Brain Sciences.
Furnham, A. (2003). Belief in a just world: Research progress over the past Okun, M., O’Rourke, H., Keller, B., Johnson, K., & Enders, C. (2014).
decade. Personality and Individual Differences, 34, 795– 817. http://dx Value-expressive volunteer motivation and volunteering by older adults:
.doi.org/10.1016/S0191-8869(02)00072-7 Relationships with religiosity and spirituality. Journal of Gerontology,
Gagné, M. (2003). The role of autonomy support and autonomy orientation Series B: Psych Sciences and Social Sciences. Advance online publica-
in prosocial behavior engagement. Motivation and Emotion, 27, 199 – tion. http://dx.doi.org/10.1093/geronb/gbu029
223. http://dx.doi.org/10.1023/A:1025007614869 Okun, M. A., Yeung, E. W.-H., & Brown, S. (2013). Volunteering by older
Galen, L. W. (2012). Does religious belief promote prosociality? A critical adults and risk of mortality: A meta-analysis. Psychology and Aging, 28,
examination. Psychological Bulletin, 138, 876 –906. http://dx.doi.org/ 564 –577. http://dx.doi.org/10.1037/a0031519
10.1037/a0028251 Pearce, P., & Amato, P. R. (1980). A taxonomy of helping: A multidimen-
Garland, D. R., Myers, D. M., & Wolfer, T. A. (2008). Social work with sional scaling analysis. Social Psychology Quarterly, 43, 363–371.
religious volunteers: Activating and sustaining community involvement. http://dx.doi.org/10.2307/3033956
Social Work, 53, 255–265. http://dx.doi.org/10.1093/sw/53.3.255 Penner, L. A., Fritzsche, B. A., Craiger, J. P., & Freifeld, T. R. (1995).
Geiser, C., Okun, M. A., & Grano, C. (2014). Who is motivated to Measuring the prosocial personality. In J. Butcher & C. D. Spielberger
volunteer? A latent profile analysis linking volunteer motivation to (Eds.), Advances in personality assessment (Vol. 10, pp. 147–163).
frequency of volunteering. Psychological Test and Assessment Model- Hillsdale, NJ: Erlbaum.
ing, 56, 3–24. Pichon, I., Boccato, G., & Saroglou, V. (2007). Nonconscious influences of
Gervais, W. M., & Norenzayan, A. (2012). Like a camera in the sky? religion on prosociality. European Journal of Social Psychology, 37,
Thinking about God increases public self-awareness and socially desir- 1032–1045. http://dx.doi.org/10.1002/ejsp.416
able responding. Journal of Experimental Social Psychology, 48, 298 – Piliavin, J. A., & Siegl, E. (2007). Health benefits of volunteering in the
302. http://dx.doi.org/10.1016/j.jesp.2011.09.006 Wisconsin longitudinal study. Journal of Health and Social Behavior,
Grano, C., Lucidi, F., Zelli, A., & Violani, C. (2008). Motives and deter- 48, 450 – 464. http://dx.doi.org/10.1177/002214650704800408
minants of volunteering in older adults: An integrated model. The Preston, J. L., & Ritter, R. S. (2013). Different effects of religion and God
International Journal of Aging & Human Development, 67, 305–326. on prosociality with the ingroup and outgroup. Personality and Social
http://dx.doi.org/10.2190/AG.67.4.b Psychology Bulletin, 39, 1471–1483. http://dx.doi.org/10.1177/
Harrell, A. (2012). Do religious cognitions promote prosociality? Ratio- 0146167213499937
nality and Society, 24, 463– 482. http://dx.doi.org/10.1177/ Preston, J. L., Ritter, R. S., & Hernandez, J. I. (2010). Principles of
1043463112463930 religious prosociality: A review and reformulation. Social and Person-
Jackson, E. F., Bachmeier, M. D., Wood, J. R., & Craft, E. A. (1995). ality Psychology Compass, 4, 574 –590. http://dx.doi.org/10.1111/j
Volunteering and charitable giving: Do religious and associational ties .1751-9004.2010.00286.x
promote helping behavior? Nonprofit and Voluntary Sector Quarterly, Reed, A., II, & Aquino, K. F. (2003). Moral identity and the expanding
24, 59 –78. http://dx.doi.org/10.1177/089976409502400108 circle of moral regard toward out-groups. Journal of Personality and
Janoff-Bulman, R., Sheikh, S., & Hepp, S. (2009). Proscriptive versus Social Psychology, 84, 1270 –1286. http://dx.doi.org/10.1037/0022-3514
prescriptive morality: Two faces of moral regulation. Journal of Per- .84.6.1270
12 JOHNSON, COHEN, AND OKUN

Ritter, R. S., & Preston, J. L. (2013). Representations of religious words: Shariff, A. F., & Norenzayan, A. (2011). Mean gods make good people:
Insights for religious priming research. Journal for the Scientific Study Different views of God predict cheating behavior. International Journal
of Religion, 52, 494 –507. http://dx.doi.org/10.1111/jssr.12042 for the Psychology of Religion, 21, 85–96. http://dx.doi.org/10.1080/
Roberts, C. W. (1989). Imagining God: Who is created in whose image? 10508619.2011.556990
Review of Religious Research, 30, 375–386. http://dx.doi.org/10.2307/ Snarey, J. (1996). The natural environment’s impact upon religious ethics:
3511298 A cross-cultural study. Journal for the Scientific Study of Religion, 35,
Roes, F. L., & Raymond, M. (2003). Belief in moralizing gods. Evolution 85–96. http://dx.doi.org/10.2307/1387077
and Human Behavior, 24, 126 –135. http://dx.doi.org/10.1016/S1090- Snyder, M., & Omoto, A. N. (2000). Doing good for self and society:
5138(02)00134-4 Volunteerism and the psychology of citizen participation. In M. Van-
Rosmarin, D. H., Krumrei, E. J., & Andersson, G. (2009). Religion as a Vugt, M. Snyder, T. R. Tyler, & A. Biel (Eds.), Cooperation in modern
predictor of psychological distress in two religious communities. Cog- society: Promoting the welfare of communities, states and organizations
nitive Behaviour Therapy, 38, 54 – 64. http://dx.doi.org/10.1080/ (pp. 127–141). New York, NY: Routledge.
16506070802477222 Spilka, B., Armatas, P., & Nussbaum, J. (1964). The concept of God: A
Ross, L. D., Lelkes, Y., & Russell, A. G. (2012). How Christians reconcile factor-analytic approach. Review of Religious Research, 6, 28 –36. http://
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

their personal political views and the teachings of their faith: Projection dx.doi.org/10.2307/3510880
as a means of dissonance reduction. Proceedings of the National Acad- Stukas, A. A., Snyder, M., & Clary, E. G. (1999). The effects of “manda-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

emy of Sciences of the United States of America, 109, 3616 –3622. tory volunteerism” on intentions to volunteer. Psychological Science,
http://dx.doi.org/10.1073/pnas.1117557109 10, 59 – 64. http://dx.doi.org/10.1111/1467-9280.00107
Ruiter, S., & DeGraaf, N. D. (2006). National context, religiosity, and Stukas, A. A., Worth, K. A., Clary, E. G., & Snyder, M. (2009). The
volunteering: Results from 53 countries. American Sociological Review, matching of motivations to affordances in the volunteer environment:
71, 191–210. http://dx.doi.org/10.1177/000312240607100202 An index for assessing the impact of multiple matches on volunteer
Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the outcomes. Nonprofit and Voluntary Sector Quarterly, 38, 5–28. http://
facilitation of intrinsic motivation, social development, and well-being. dx.doi.org/10.1177/0899764008314810
American Psychologist, 55, 68 –78. http://dx.doi.org/10.1037/0003- Terry, D. J., Hogg, M. A., & White, K. M. (1999). The theory of planned
066X.55.1.68 behaviour: Self-identity, social identity and group norms. British Jour-
Sanderson, S. K., & Roberts, W. W. (2008). The evolutionary forms of the nal of Social Psychology, 38, 225–244. http://dx.doi.org/10.1348/
religious life: A cross-cultural, quantitative analysis. American Anthro- 014466699164149
pologist, 110, 454 – 466. http://dx.doi.org/10.1111/j.1548-1433.2008 Warner, C. M., Kilinc, R., Hale, C. W., Cohen, A. B., & Johnson, K. A.
.00078.x (2015). Religion and public goods provision: Experimental and inter-
Schloss, J. P., & Murray, M. J. (2011). Evolutionary accounts of belief in view evidence from Catholicism and Islam in Europe. Comparative
supernatural punishment: A critical review. Religion, Brain & Behavior, Politics, 47, 189 –209.
1, 46 –99. White, J. A., & Plous, S. (1995). Self-enhancement and social responsi-
Schwartz, S. H. (1992). Universals in the content and structure of values: bility: On caring more, but doing less, than others. Journal of Applied
Theory and empirical tests in 20 countries. In M. Zanna (Ed.), Advances Social Psychology, 25, 1297–1318. http://dx.doi.org/10.1111/j.1559-
in experimental social psychology (Vol. 25, pp. 1– 65). San Diego, CA: 1816.1995.tb02619.x
Academic Press. Wilson, J. (2000). Volunteering. Annual Review of Sociology, 26, 215–240.
Schwartz, S. H., & Howard, J. A. (1980). Explanations of the moderating http://dx.doi.org/10.1146/annurev.soc.26.1.215
effect of responsibility denial on the personal norm-behavior relation- Wood, B. T., Worthington, E. L., Jr., Exline, J. J., Yali, A. M., Aten, J. D.,
ship. Social Psychology Quarterly, 43, 441– 446. http://dx.doi.org/ & McMinn, M. R. (2010). Development, refinement, and psychometric
10.2307/3033965 properties of the attitudes toward God scale (ATGS-9). Psychology of
Schwartz, S. H., & Huismans, S. (1995). Value priorities and religiosity in Religion and Spirituality, 2, 148 –167. http://dx.doi.org/10.1037/
four Western religions. Social Psychology Quarterly, 58, 88 –107. http:// a0018753
dx.doi.org/10.2307/2787148
Shariff, A. F., & Norenzayan, A. (2007). God is watching you: Priming
God concepts increases prosocial behavior in an anonymous economic Received December 30, 2014
game. Psychological Science, 18, 803– 809. http://dx.doi.org/10.1111/j Revision received March 24, 2015
.1467-9280.2007.01983.x Accepted May 27, 2015 䡲

View publication stats

You might also like