ArticlePDF Available

Empathy, Target Distress, and Neurohormone Genes Interact to Predict Aggression for Others-Even Without Provocation

Authors:

Abstract and Figures

Can empathy for others motivate aggression on their behalf? This research examined potential predictors of empathy-linked aggression including the emotional state of empathy, an empathy target's distress state, and the function of the social anxiety-modulating neuropeptides oxytocin and vasopressin. In Study 1 (N = 69), self-reported empathy combined with threat to a close other and individual differences in genes for the vasopressin receptor (AVPR1a rs3) and oxytocin receptor (OXTR rs53576) to predict self-reported aggression against a person who threatened a close other. In Study 2 (N = 162), induced empathy for a person combined with OXTR variation or with that person's distress and AVPR1a variation led to increased amount of hot sauce assigned to that person's competitor. Empathy uniquely predicts aggression and may do so by way of aspects of the human caregiving system in the form of oxytocin and vasopressin.
Content may be subject to copyright.
http://psp.sagepub.com/
Bulletin
Personality and Social Psychology
http://psp.sagepub.com/content/40/11/1406
The online version of this article can be found at:
DOI: 10.1177/0146167214549320
2014 40: 1406Pers Soc Psychol Bull
Anneke E. K. Buffone and Michael J. Poulin
Without Provocation EvenEmpathy, Target Distress, and Neurohormone Genes Interact to Predict Aggression for Others
Published by:
http://www.sagepublications.com
On behalf of:
Society for Personality and Social Psychology
can be found at:Personality and Social Psychology BulletinAdditional services and information for
http://psp.sagepub.com/cgi/alertsEmail Alerts:
http://psp.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
What is This?
- Oct 6, 2014Version of Record >>
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
Personality and Social
Psychology Bulletin
2014, Vol. 40(11) 1406 –1422
© 2014 by the Society for Personality
and Social Psychology, Inc
Reprints and permissions:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/0146167214549320
pspb.sagepub.com
Article
On June 9, 2012, guided by his little daughter’s screams, a
Texas father rushed to a gruesome scene—a farmhand sexu-
ally assaulting his 5-year-old daughter. Within seconds, the
father beat the assaulter to death. The father was not natu-
rally an aggressive man; in fact, nearly immediately, upon
realizing that the man seemed critically injured, the father
called 911 on behalf of the man he had beaten, hoping to save
his life. The father can even be heard crying on the tape of
the 911 call. In court, the father was found not guilty of
homicide as attorneys and prosecutors found that the father
was remorseful, did not mean to kill his daughter’s attacker,
and even attempted to save his life (Valencia, Savidge, &
Maddox, 2012). What might have led this apparently normal,
even sensitive man to brutally beat his child’s attacker to
death? What may have led him to set aside his aversion to
such violence? For that matter, why do people in general
aggress on behalf of others, whether parents moved to defend
their children, policemen and soldiers fighting on behalf of
comrades, or ordinary people taking stands for colleagues
and loved ones?
To a large degree, asking why individuals aggress on
behalf of others is merely a more specific way of asking why
people do anything on behalf of others—that is, why people
engage in any form of prosocial behavior. In the present arti-
cle, we examine the possibility that the caregiving behavioral
system (Bell & Richard, 2000; Bowlby, 1969; Brown &
Brown, 2006; Preston, 2013; Shaver, Mikulincer, &
Shemesh-Iron, 2009) may facilitate aggression even as it can
elicit gentle, nurturing behaviors. We also present evidence
that certain features of the caregiving system, including the
experience of empathy and the function of the caregiving-
linked hormones vasopressin and oxytocin, can predict
aggression, thereby helping to explain why aggression on
behalf of others may occur under certain circumstances.
The Caregiving System
Several researchers have argued that prosocial behavior is
regulated in part by a caregiving behavioral system that
evolved alongside the well-researched attachment system
(Bell & Richard, 2000; Bowlby, 1969; Brown & Brown, 2006;
Preston, 2013; Shaver et al., 2009). The primary feature of this
system is a motivation to reduce the suffering and/or increase
the well-being of a valued individual perceived as being in
549320PSPXXX10.1177/0146167214549320Personality and Social Psychology BulletinBuffone and Poulin
research-article2014
1University at Buffalo, NY, USA
Corresponding Author:
Michael J. Poulin, Department of Psychology, University at Buffalo, Park
Hall, Room 207, Buffalo, NY 14260-411, USA.
Email: mjpoulin@buffalo.edu
Empathy, Target Distress, and
Neurohormone Genes Interact to Predict
Aggression for Others–Even Without
Provocation
Anneke E. K. Buffone1 and Michael J. Poulin1
Abstract
Can empathy for others motivate aggression on their behalf? This research examined potential predictors of empathy-linked
aggression including the emotional state of empathy, an empathy target’s distress state, and the function of the social anxiety-
modulating neuropeptides oxytocin and vasopressin. In Study 1 (N = 69), self-reported empathy combined with threat to
a close other and individual differences in genes for the vasopressin receptor (AVPR1a rs3) and oxytocin receptor (OXTR
rs53576) to predict self-reported aggression against a person who threatened a close other. In Study 2 (N = 162), induced
empathy for a person combined with OXTR variation or with that person’s distress and AVPR1a variation led to increased
amount of hot sauce assigned to that person’s competitor. Empathy uniquely predicts aggression and may do so by way of
aspects of the human caregiving system in the form of oxytocin and vasopressin.
Keywords
prosocial behavior, aggressive behavior, caregiving, oxytocin, vasopressin, empathy
Received November 26, 2013; revision accepted July 21, 2014
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
Buffone and Poulin 1407
need, perhaps especially when suffering is perceived as
undeserved (Goetz, Keltner, & Simon-Thomas, 2010). Many
accounts of the functioning of the caregiving system focus
on the role of the emotions people tend to have when viewing
another person in a state of distress or vulnerability. These
emotions include feelings of warmth, tenderness, and sym-
pathy—collectively referred to variously as empathy, com-
passion, or empathic concern (e.g., Batson, 1991; Goetz et
al., 2010; Mikulincer, Shaver, Bar-On, & Ein-Dor, 2010;
Sober & Wilson, 1998). Experimental manipulations and
assessments of these emotions (henceforth “empathy”) have
shown them to predict a wide array of prosocial behaviors,
including laboratory assessments of willingness to help oth-
ers (Batson, 1991), real-world volunteering (Gillath et al.,
2005; Stuermer, Snyder, & Omoto, 2005), and responsive-
ness in close relationships (Collins & Ford, 2010).
Although it is still unknown precisely how empathy moti-
vates helping behavior, biological models of the caregiving
system have focused on the role of the hormones and neu-
rotransmitters oxytocin and vasopressin (Feldman, 2012;
Goetz et al., 2010; Kemp & Guastella, 2011; Poulin, Holman,
& Buffone, 2012; Preston, 2013). Although these neurohor-
mones have many divergent effects—for example, vasopres-
sin is frequently anxiogenic whereas oxytocin can be
anxiolytic—in non-human mammals, these neurohormones
both regulate pair-bonding and the provision of parental care
(Carter, 1998), and are thus may facilitate caregiving and
empathy more broadly.
Oxytocin and Caregiving
In humans, the vast majority of the work on these neurohor-
mones has focused on oxytocin, with a literature emerging to
support the view that oxytocin is instrumental in parental
bonding with and caring for children (e.g., Bakermans-
Kranenburg & van Ijzendoorn, 2008; see Feldman, 2012, for
a review). In addition, there is some evidence, though not
conclusive, that oxytocin facilitates empathy and prosocial
behavior (for reviews, see Gonzalez-Liencres, Shamay-
Tsoory, & Brüne, 2013; Norman, Hawkley, Cole, Berntson,
& Cacioppo, 2012). Intranasal administration of oxytocin
may increase emotional empathy (Hurlemann et al., 2010),
and in several studies, intranasal administration of oxytocin
has led to increased generosity in economic games
(Baumgartner, Heinrichs, Vonlanthen, Fischbacher, & Fehr,
2008; Kosfeld, Heinrichs, Zak, Fischbacher, & Fehr, 2005;
Mikolajczak et al., 2010), though other studies have failed to
find this link (e.g., Singer et al., 2008). In addition, there is
some evidence that plasma oxytocin increases following an
empathy induction, possibly indicating that oxytocin also
helps account for the effects of empathy on subsequent
behavior (Barraza & Zak, 2009).
Research on the oxytocin receptor gene (OXTR) also pro-
vides tentative evidence for a role of oxytocin in empathy.
Individuals homozygous for the G allele of OXTR variant
rs53576, who are hypothesized to be more sensitive to the
effects of oxytocin, exhibit less stress and greater empathic
accuracy after being directed to attend to others’ feelings
(Rodrigues, Saslow, Garcia, John, & Keltner, 2009) and
engage in more charitable behavior under conditions of per-
ceived social threat (Poulin et al., 2012) than do others. In
contrast, however, OXTR variants have not been shown to
predict social integration in a large community sample
(Chang et al., 2014). It is important to note that the role of
OXTR variant rs53576 may differ by ethnicity, with some
evidence indicating that it predicts social behavior in
European American (White) individuals but not for members
of other ethnic groups (Kim et al., 2010; Poulin et al., 2012).
Vasopressin and Caregiving
Very little if any research has examined vasopressin admin-
istration and caregiving, and the links between vasopressin,
human caregiving, and empathy are unclear (Gonzalez-
Liencres et al., 2013). However, a small amount of research
suggests that vasopressin 1a receptor gene (AVPR1a) varia-
tion may influence social behavior, including generosity. For
example, those with greater numbers of “long” versus “short”
alleles of the 1a vasopressin receptor (AVPR1a) variant rs3,
who are believed to be more sensitive to the effects of vaso-
pressin, exhibit more laboratory generosity toward strangers
(Knafo et al., 2008) and more stable pair bonds (Walum et
al., 2008) than do others. Similarly, those with greater num-
bers of “long” versus “short” alleles of AVPR1a variant rs1
demonstrate more commitment to their civic duties under
conditions of perceived social threat (Poulin et al., 2012)
than do others.
Caregiving and Aggression
Given that empathy motivates a wide range of prosocial
behaviors, it is not difficult to imagine that it could motivate
helping another person by way of aggressing against a third
party. However, research has not directly tested this hypoth-
esis. Some research has examined links between empathy
and anti-social behaviors such as punishment or hostility
(Condon & DeSteno, 2011; Keller & Pfattheicher, 2013), but
these studies did not examine the context of aggression
directed toward helping a third party. Moreover, these stud-
ies have yielded inconsistent findings, with Condon and
DeSteno (2011) showing that experimentally induced empa-
thy inhibits aggression in the form of punishing a wrongdoer,
whereas Keller and Pfattheicher (2013) found that compas-
sion (closely related to empathic concern) predicts increased
hostility, albeit solely among prevention-focused individu-
als. Other prior research indicates that empathy induces par-
tiality toward a empathy target that results in disadvantaging
a third party (Batson, Klein, Highberger, & Shaw, 1995).
However, this outcome differs from aggression in that, while
partiality does result in harm to a third party, it does not
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
1408 Personality and Social Psychology Bulletin 40(11)
reflect a “proximate . . . intent to cause harm” (Anderson &
Huesmann, 2003, p. 298, emphasis added) to the third party.
Finally, a separate body of research suggests that aggression
for the sake of another results from experiencing anger on
the other’s behalf (Hoffman, 1989; Neumann, 2000; Staub,
1978; Stotland, 1969; Vitaglione & Barnett, 2003; Weiner,
1995) but has not examined the possibility that empathy
plays a direct role in eliciting aggression. In sum, no research
has yet provided evidence for a direct link between empathy
and aggression, especially in the absence of provocation or a
desire to punish a wrongdoer.
Although social-psychological literature has not tested a
direct role for empathy in helping-linked aggression, decades
of research on caregiving behavior in non-human animals
have found aggressive acts on behalf of mates and offspring
to be a fundamental part of animal caregiving (Gammie &
Nelson, 2000; Lonstein & Gammie, 2002; Wolff, 1985).
These types of aggression are facilitated, just as are gentler
aspects of pair-bonding and parenting, by vasopressin
(Bosch, Pfortsch, Beiderbeck, Landgraf, & Neumann, 2010;
Carter, 1998) and oxytocin (Bosch, Meddle, Beiderbeck,
Douglas, & Neumann, 2005; Campbell, 2010). Moreover,
there is preliminary evidence that these neurohormones,
especially oxytocin, facilitate aggressive caregiving behav-
iors in humans.
For example, intranasal oxytocin leads to protective
maternal behavior against “enthusiastic strangers” (Mah,
Bakermans-Kranenburg, Van IJzendoorn, & Smith, 2014). In
addition, breast-feeding human mothers have higher levels
of circulating oxytocin and increased aggressive tendencies
toward potential intruders compared with formula-feeding
mothers (Hahn-Holbrook, Holbrook, & Haselton, 2011;
Hahn-Holbrook, Holt-Lunstad, Holbrook, Coyne, & Lawson,
2011). In addition, De Dreu and colleagues have found that
oxytocin administration facilitates ingroup favoritism that
implicitly disfavors highly threatening outgroups (De Dreu,
2012; De Dreu et al., 2010; Shalvi & De Dreu, 2014).
Relatively little research has examined links between vaso-
pressin administration and aggression on another’s behalf,
but at least one study found that vasopressin administration
led to greater negative affect toward unfamiliar faces among
males, possibly consistent with more aggressive tendencies
toward strangers (Thompson, George, Walton, Orr, &
Benson, 2006).
The Present Research
In sum, research on empathy, non-human caregiving, and
neurohormonal mechanisms for prosocial behavior suggests
that situations that evoke empathy might facilitate aggres-
sion on behalf of a vulnerable other. The present research
sought to test whether assessed or elicited empathy would
lead to situation-specific aggression on behalf of another per-
son, and to explore the potential role of oxytocin and vaso-
pressin in explaining any associations between empathy and
aggression. Two studies addressed these goals. The first
study used a correlational design exploring the links between
self-reported empathy and real-world aggression on behalf
of another. The second study used an experimental design to
manipulate empathy and assess aggressive behavior on
behalf of a distressed other person in a controlled setting. In
both studies, the roles of vasopressin and oxytocin in empa-
thy-linked aggression were examined indirectly by way of
individual differences in receptor genes for these neurohor-
mones. Specifically, we examined three genetic variants
thought to increase the effects of oxytocin and vasopressin
and previously shown to predict prosocial behavior: the
AVPR1a variants rs1 and rs3 (long vs. short genotypes) and
the OXTR variant rs53576 (GG genotype vs. other).
In addition to empathy and vasopressin/oxytocin receptor
genes, both studies also examined the role of the empathy
target’s distress in predicting aggression. We expected that
empathy would only lead to aggression on behalf of clearly
distressed targets for two reasons. First, the caregiving sys-
tem is specifically responsive to distress cues from vulnera-
ble others (Goetz et al., 2010; Preston, 2013). Second, the
effects of empathy on aggression should be specific to
aggression designed to benefit the empathy target. If the
empathy target is not in need of assistance—that is, dis-
tressed—aggression should not occur. Examining the effects
of empathy in the context of both distressed and non-dis-
tressed empathy targets allowed us to test these predictions.
Our primary hypothesis was that manipulated or assessed
empathy would predict aggression on behalf of a distressed,
but not a non-distressed, empathy target. In addition, our ten-
tative prediction was that empathy’s role in predicting
aggression for distressed empathy targets would be greater
for longer versus shorter genotypes of AVPR1a rs1 and rs3 as
well as for GG versus other genotypes of OXTR rs53576.
Study 1
Method
Participants. Undergraduates (N = 69; 35 women, 31 men, 3
declined to answer; age: M = 19.69 years) in the introductory
psychology subject pool at the State University of New York
(SUNY) Buffalo were recruited for this study. Participants’
ethnic composition was 1 African or African American, 12
Asian or Asian American, 51 European or European Ameri-
can, 6 Hispanic, 1 Arab or Arab American, and 1 Native
American, with 3 individuals being of mixed race/ethnicity.
Participating young adults were compensated with research
credit for their course research requirement.
Procedure. Students signed up online for a two-part study
with the first part consisting of saliva donation in the labora-
tory and the second part consisting of an online survey. In the
first session, all participants completed informed consent
materials and then donated their saliva sample for DNA
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
Buffone and Poulin 1409
extraction via Oragene DNA collection kits (http://www.
dnagenotek.com). In the survey, participants were asked to
write and answer questions about a time in the past 12 months
when they witnessed a close other (friend, family member,
significant other, child) being hurt physically or emotionally
by a third party other than themselves. Participants were
instructed to think of only one such event and to refer to this
one event for all subsequent questions. All participants were
able to recall such an event. Participants were then fully
debriefed and thanked for their participation.
Measures
Empathy. Participants rated to what degree they felt emo-
tions associated with empathy, specifically empathic concern,
when witnessing the conflict (Batson et al., 1997; Batson,
Shaw, & Oleson, 1992). Specifically, participants rated to
what degree they felt “compassionate,” “softhearted,” “sym-
pathetic,” “moved,” “tender,” or “warm” toward their close
other at the time of the close other’s conflict (1 = “not at all”;
7 = “extremely”), with the mean of these items used as an
index of empathy (α = .88).
Aggression. Participants indicated whether they con-
fronted the other party physically, verbally, or by other
means in the conflict via three dichotomous items: “When
this conflict was going on, did you confront the other per-
son (i.e., the person who was in conflict with the person you
cared about) using physical force?” and “When this conflict
was going on, did you confront the other person (i.e., the
person who was in conflict with the person you cared about)
verbally?” as well as “When this conflict was going on, did
you confront the other person (i.e., the person who was in
conflict with the person you cared about) by other means?”
A composite score was created reflecting the total number
of physical and verbal confrontation, and confrontation by
other means (range = 0-3). A score of 0 reflected that the
individual did not intervene.
Close other’s distress. Participants rated perceived distress
of the close other by indicating to what degree they felt their
close other was emotionally harmed by the conflict with the
item, “To what extent do you think this conflict was emotion-
ally harmful to the person you cared about?” (1 = “not at all”;
7 = “extremely”), which served as the measure of how much
they perceived their close other as in distress.
AVPR1a and OXTR variables. DNA from participants’
saliva samples was examined for two variants in the gene
for the 1a vasopressin receptor (AVPR1a) and one variant
in the gene for the oxytocin receptor (OXTR). Details of
the genetic analysis are available via supplementary mate-
rials (http://pspb.sagepub.com/supplemental). The AVPR1a
polymorphisms investigated (rs1 and rs3) were regions of
DNA in which a section of nucleobases repeats multiple
times, with the number of repetitions varying among
individuals. Following prior research (Knafo et al., 2008;
Poulin et al., 2012), the number of repetitions for each poly-
morphism was dichotomized at the median to categorize
variants as “long” or “short.” Because all individuals have
a version, or allele, from each parent, individual genotypes
can be long/long, short/long, and short/short. This informa-
tion was used to create an ordinal variable representing the
number of “long” alleles (0-2) each individual possessed,
consistent with analyses in previous research (Poulin et al.,
2012).
The OXTR variant (rs53576) is a single-nucleotide poly-
morphism (SNP) consisting of a locus in the genetic code at
which either the nucleobase adenine (A) or guanine (G) may
be present. Because every individual receives one allele from
each parent, this results in three possible genotypes: AA, AG,
or GG. After these genotypes were identified (see supple-
mentary materials), a dichotomous variable was created for
our analyses representing OXTR rs53576 GG status (AA/AG
= 1, GG = 0). This procedure of contrasting GG with AA/AG
status parallels what has been done in previous research
(Kogan et al., 2011; Rodrigues et al., 2009).
Control variables. In addition to the primary variables
of interest, we also assessed several variables for the pur-
pose of ruling out alternate explanations for our findings.
We assessed how much subjects personally felt emotion-
ally harmed by the close other’s conflict situation, using an
item similar to that used to assess the close other’s degree
of distress. In addition, the kind of relationship to the close
other was assessed (e.g., same-sex friend, parent, romantic
partner) as well as the sex and age of the close other. We
also assessed trait aggression with the Buss–Perry Aggres-
sion Questionnaire (Buss & Perry, 1992) and impulsiveness
with the 16-item version of the Barratt Impulsiveness Scale
(Patton, Stanford, & Barratt, 1995). In addition, to measure
the degree of participants’ perceived overlap between their
own self-concept and that of the close other, participants
completed the Inclusion of the Other in the Self (IOS) scale
(Aron, Aron, & Smollan, 1992). The IOS scale consists of a
set of seven numbered pictures depicting increasing literal
overlap between two circles, one labeled “self” and the other
labeled “other” (1 = no overlap between self and other; 7 =
nearly complete overlap).
Results
Descriptive statistics
Confrontation. A total of 26 participants reported confront-
ing their close other’s perpetrator in one or more of the three
possible ways: physical (n = 3), verbal (n = 14), by other
means (n = 4), both physical and verbal (n = 4), or both ver-
bal and other (n = 1).1 There were no bivariate associations
of the composite variable representing confrontation with
other key study variables (see Table 1). Confrontation also
was uncorrelated with gender (coded female = 0, male = 1),
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
1410 Personality and Social Psychology Bulletin 40(11)
indicating that men and women did not tend to recall using
different amounts of aggression, rs(26) < −.13, ps = .29.
AVPR1a and OXTR polymorphisms. The genotyping of
AVPR1a rs1 revealed a range of repeat values from 300 to
325 repeats. A median split between 305 and 309 was per-
formed to distinguish short from long alleles. Genotyping of
AVPR1a rs3 yielded a range of repeat values from 319 to
348 repeats. A median split between 332 and 334 repeats was
used to discern between short and long alleles. Prior research
suggests that the function of OXTR rs53576 may be specific
to Caucasian populations (Kim et al., 2010; Poulin et al.,
2012). For this reason, only European American participants
(n = 51) were included in analyses that included this variant
(see Table 2).
All polymorphisms examined were in Hardy–Weinberg
equilibrium, and there were no significant gender differences
in genotype. The oxytocin and vasopressin polymorphisms
in this study were uncorrelated, rs(69) < −.07, ps > .39. The
vasopressin polymorphisms were marginally correlated in
this study, r(69) = .22, p = .07.
Associations between empathy and aggression. We predicted
that empathy for a distressed close other in need of assistance
(with distress measured as perceptions of perceived emo-
tional harm to the close other) would be associated with
aggression against the perpetrator that endangered the close
other. To test this prediction, a standardized regression with
aggression toward the perpetrator (physical, verbal, or by
other means) as the dependent variable and three different
independent variables—empathy, perceived distress of the
close other, and their interaction—was conducted (see Table 3;
Model I). Results revealed a significant two-way interaction
between distress of the close other and empathy, B = 0.43,
95% confidence interval (CI) = [−0.28, −0.19], β = .43, p <
.001. Simple slopes analyses with recentered interaction
terms (distress recoded so that 0 fell at 1 SD above or below
the mean) and all variables standardized revealed that when
the close other was perceived as highly distressed, empathy
marginally predicted higher likelihood of aggression on
behalf of the close other, B = 0.06, 95% CI = [−0.01, 0.14],
β = .28, p = .09; but when the close other was perceived as
low in distress, empathy significantly predicted lesser likeli-
hood to aggress against the close other’s perpetrator, B =
−0.12, 95% CI = [−0.18, −0.06], β = −.54, p < .001.
Associations of AVPR1a and OXTR genes with empathy and
distress. We also predicted that the two-way interaction
Table 1. Study 1: Descriptive Statistics and Correlations for Key Study Variables.
Variable N M (SD) Rangea
Correlations
1 2 3 4 5 6 7
1. Confrontation 69 1.15 (0.21) 0-3 1.00
2. Trait aggression 69 3.38 (1.29) 1-7 .06 1.00
3. Closeness to perpetrator 69 2.75 (2.02) 1-7 .06 .06 1.00
4. Closeness to victim 69 5.42 (1.53) 1-7 .05 −.29* .19 1.00
5. Impulsivity 69 3.49 (0.84) 1-7 .04 .38** .03 −.26* 1.00
6. Threat to self 69 3.93 (2.02) 1-7 −.01 .10 .25* .29* −.20* 1.00
7. Threat to close other 69 5.43 (1.66) 1-7 −.18 .13 .04 .01 −.15 .65*** 1.00
8. Empathy 69 3.60 (1.44) 1-7 .02 .03 .05 −.12 .20 .23
aMean of verbal confrontation and physical confrontation, as well as confrontation by other means.
*p < .05. **p < .01. ***p < .001.
Table 2. Study 1: Distribution of AVPR1a and OXTR Genotypes.
AVPR1a polymorphisms OXTR polymorphism
Genotype
rs1 variant rs3 variant
Genotype
rs53576 variant
Frequency Frequency Frequency
Short/short 20 (28.99%) 33 (47.83%) AA 1 (1.96%)
Short/long 31 (44.93%) 23 (33.33%) AG 15 (29.40%)
Long/long 18 (26.09%) 13 (18.84%) GG 35 (68.63%)
Total 69 69 Total 51a
aNumbers refer to European American Participants only. In addition, it was not possible to extract sufficient DNA from one participant’s saliva sample to
perform the analysis.
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
Buffone and Poulin 1411
between empathy (empathic concern) and distress would be
stronger for individuals with long versus short alleles of
AVPR1a rs1 and r3 and also that the empathy–distress inter-
action would be stronger for individuals with the GG geno-
type of OXTR rs53576 than for individuals with the AA or
AG genotypes. Separate models tested the inclusion of each
of these three variants as a moderator, with each model
including empathy, distress, the genetic variant, all three
two-way product–term interactions among those variables,
and the three-way product–term interaction. Results of these
regression models can be seen in Table 3.
Examination of the hypothesized three-way interactions
for individuals (N = 69) with the vasopressin receptor geno-
type AVPR1a variants rs1 and rs3 (Vasopressin receptor vari-
ant × Empathy × Distress) revealed no significant interaction
for rs1 (see Table 3, Model II), but a significant three-way
interaction for rs3’s number (0-2) of long alleles, B = 0.14,
95% CI = [0.004, 0.27], β = .26, p = .04; see Table 3, Model
III.2 This three-way interaction for rs3 was probed in three
different regression models in which each of the three differ-
ent genotypes in turn was examined as the reference category
(i.e., 0) of a dummy-coded variable. These analyses
produced a significant distress by empathy interaction for
those with the long/long genotype, B = 0.82, 95% CI = [0.44,
1.21], β = .81, p < .001, or the short/long genotype B = 0.53,
95% CI = [0.30, 0.76], β = .52, p < .001, but not for those
with the short/short genotype, B = 0.24, 95% CI = [−0.07,
0.55], β = .23, p = .13. These patterns can be viewed in
Figures 1a, 1b, and 1c. Simple slope tests revealed that for
long/long individuals, empathy significantly predicted
higher aggression on behalf of a close other perceived as
highly distressed, B = 1.27, 95% CI = [0.52, 2.04], β = 1.22,
p = .001, and less aggression on behalf of a close other who
was low in distress, B = −0.38, 95% CI = [−0.86, 0.10], β =
−.36, p = .12. Similarly, for those with the short/long geno-
type, empathy predicted significantly higher aggression for a
close other high in distress, B = 0.56, 95% CI = [0.17, 0.95],
β = .54, p < .01, but less aggression on behalf of a close other
low in distress, B = −0.50, 95% CI = [−0.80, −0.20], β =
−.48, p < .01. By contrast, among individuals of the short/
short genotype, empathy predicted less aggression on behalf
of a close other either high in distress, B = −0.15, 95% CI =
[−0.60, 0.30], β = −.14, p = .51, or low in distress, B = −0.62,
95% CI = [−1.05, −0.19], β = −.59, p < .01.
Table 3. Study 1: Regression Models Predicting Confrontation of Close Other’s Perpetrator.
Model IaModel IIbModel IIIcModel IVd
Variable B (95% CI) βB (95% CI) βB (95% CI) βB (95% CI) β
Distress 0.04 [−0.19, 0.26] .04 0.02 [−0.14, 0.17] .03 0.003 [−0.14, 0.15] .01 −0.22 [−0.50, 0.06] −.22
Empathy −0.13 [−0.37, 0.10] −.13 −0.12 [−0.29, 0.05] −.18 −0.04 [−0.21, 0.12] −.06 0.05 [−0.29, 0.40] .05
AVPR1a rs1 −0.01 [−0.18, 0.15] −.02
AVPR1a rs3 −0.09 [−0.27, 0.09] −.14
OXTR rs53576 −0.18 [−0.79, 0.43] −.08
Distress ×
Empathy
0.43 [0.21, 0.66] .43*** 0.27 [0.12, 0.42] .43** 0.31 [0.17, 0.46] .49*** 0.62 [0.34, 0.90] .63***
rs1 × Distress −0.03 [−0.17, 0.11] −.05
rs1 × Empathy 0.03 [−0.12, 0.18] .05
rs1 × Distress
× Empathy
0.04 [−0.10, 0.18] .07
rs3 × Distress −0.04 [−0.18, 0.09] −.08
rs3 × Empathy 0.21 [0.04, 0.39] .35*
rs53576 ×
Distress
0.60 [0.01, 1.19] .32*
rs53576 ×
Empathy
−0.24 [0.87, 0.40] −.13
rs3 × Distress
× Empathy
0.14 [0.004, 0.27] .26*
rs53576 ×
Distress ×
Empathy
−0.93 [−1.71, −0.16] −.39*
Note. All variables are standardized in these models. Distress and Empathy were recentered for the purpose of calculating interactions in all subsequent
models.
an = 69, adjusted R2 = .17, p < .01; F(3, 65) = 5.77.
bn = 65, adjusted R2 = .15, p = .02; F(7, 57) = 2.64.
cn = 65, adjusted R2 =.25, p = .001; F(7, 57) = 4.05.
dn = 51, adjusted R2 = .35, p < .01; F(7, 43) = 4.80.
*p < .05. **p < .01. ***p < .001.
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
1412 Personality and Social Psychology Bulletin 40(11)
In addition to the three-way interaction, there was a sepa-
rate significant two-way interaction between rs3 genotype
and empathy, B = 0.21, 95% CI = [0.04, 0.39], β = .35, p =
.02. Simple slope tests indicated that for individuals with the
short/short genotype, empathy predicted decreased aggres-
sion, B = −0.26, 95% CI = [−0.47, −0.04], β = −.38, p = .02,
while there was no such association among those with the
short/long genotype, B = −0.09, 95% CI = [−0.37, 0.19], β =
−.14, p = .50, and a non-significant trend for empathy to pre-
dict greater aggression among those with the long/long geno-
type, B = 0.41, 95% CI = [−0.10, 0.92], β = .60, p = .11.
The proposed three-way interaction between the dichoto-
mous version of the oxytocin receptor gene (AA and AG = 0,
GG = 1), distress, and empathy (OXTR × Distress × Empathy)
was significant, B = −0.93, 95% CI = [−1.71, −0.16], β =
−.39, p = .02 (see Table 3, Model IV). Probing of the interac-
tion was done by examining the predicted two-way interac-
tion of empathy with distress with all variables standardized
and the oxytocin variable coded in separate analyses with
either AA/AG or GG set equal to 0. Results demonstrated
that the effect of the distress by empathy interaction was sig-
nificant for individuals with the GG genotype of OXTR, B =
0.62, 95% CI = [0.34, 0.90], β = .63, p < .001, but that there
was no such significant interaction for those with the AA or
AG genotypes, B = − 0.41, 95% CI = [−1.05, 0.22], β =
−.42, p = .20. This pattern is illustrated in Figures 2a and 2b.
Simple slopes analyses revealed that for those with the GG
genotype of OXTR, empathy predicted higher likelihood of
aggression on behalf of a close other high in distress, B =
0.68, 95% CI = [0.15, 1.20], β = .63, p = 0.01, but lower
aggression on behalf of a close other low in distress, B =
−0.57, 95% CI = [−0.91, 0.23], β = −.53, p < .01.
In addition to the three-way interaction, there was a sepa-
rate significant two-way interaction between rs53576 geno-
type and the close other’s distress, B = 0.60, 95% CI = [0.01,
1.19], β = .32, p = .045. Simple slope tests indicated that for
individuals with the AA/AG genotype, there was a non-sig-
nificant trend for distress to predict increased aggression, B =
0.22, 95% CI = [−0.11, 0.55], β = .36, p = .18, while among
those with the GG genotype, there was a non-significant
2
Aggressive Behaviors (AVPR1a rs3 LL)
0
0.5
1
1.5
2
Low EmpathyHigh Empathy
Low Distress
High Distress
0
0.5
1
1.5
Low EmpathyHigh Empathy
Low Distress
High Distress
Aggressive Behaviors (AVPR1a rs3 SL)
(a) (b)
0
0.5
1
1.5
2
Low EmpathyHigh Empathy
Low Distress
High Distress
Aggressive Behaviors (AVPR1a rs3 SS)
(c)
Figure 1. Study 1: Graphs of empathy by distress predicting confrontation of the close other’s perpetrator separately for individuals
with the (a) long/long genotype versus (b, c) any short genotype of AVPR1a rs3.
Note. Low and high values of distress are one SD below the mean and one SD above the mean, respectively. Confrontation is an index of physical, verbal,
and other confrontation (range = 0-3).
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
Buffone and Poulin 1413
trend for distress to predict decreased aggression, B = −0.13,
95% CI = [−0.31, 0.05], β = −.21, p = .14.
Gender effects. Given that there are frequently gender dif-
ferences in patterns of aggression (for a review, see Archer,
2009) and in the effects of oxytocin and vasopressin (e.g.,
Carter, 1998; Hoge et al., 2014), we examined whether our
effects would differ by gender. Looking at the analyses for
men and women separately, we found that the two-way
interaction between distress of the close other and empathy
appeared to be greater for women, B = 0.46, 95% CI = [0.05,
0.87], β = .38, p = .03, than for men, B = 0.17, 95% CI =
[−0.19, 0.52], β = .19, p = .34. The effects of empathy did not
appear to be confounded by gender differences: The Empa-
thy × Distress interaction remained significant when gender
and two-way interactions between gender and distress and
gender and empathy were added into the model, B = 0.33,
95% CI = [0.07, 0.62], t(65) = 2.55, p = .01.
The three-way interaction with AVPR1a rs3 appeared to
be greater among females, B = 0.52, 95% CI = [−0.12, 1.16],
β =.43, p = .11, than among males, B = 0.07, 95% CI =
[−0.52, 0.67], β =.09, p = .80. Similarly, the three-way inter-
action with OXTR rs53576 appeared to be greater among
females, B = −1.00, 95% CI = [−2.26, 0.26], β = −.39, p =
.11, than among males, B = −0.39, 95% CI = [−2.09, 1.30],
β = −.21, p = .63. The role of genotype did not appear to be
confounded by gender differences: There was no Gender ×
Empathy × Distress interaction (p = .41).
Follow-up analyses. A series of follow-up analyses tested
alternate explanations for our findings. First, we tested
whether trait aggression and impulsivity—two factors asso-
ciated with aggressive acts—or self–other overlap—a poten-
tial correlate of empathy—would account for the observed
associations between empathy and aggression. None of these
variables was a significant predictor of aggression, and also
did not moderate our findings as no two- or three-way inter-
actions with these variables were significant.
Next, we examined the possibility that high levels of both
empathy and aggression were merely due to high levels of
threat by entering participants’ own feelings of being threat-
ened as a covariate. All models were unaffected when this
variable was entered, and it also did not moderate any of our
results. Last, gender, kind of relationship, or age of the close
other experiencing the conflict that participants reported on
similarly did not affect any of the results when entered into
any of the regression models as covariates. In addition, these
variables did not moderate our findings as no two- or three-
way interactions with these variables were significant.
Discussion
Consistent with our hypotheses about the caregiving system
motivating aggression, it was empathy, and not trait aggres-
sion or perceptions of emotional threat toward the self, that
predicted aggression in our study. This pattern of findings
suggests that the reported aggression was situation-specific
and directed toward helping a relationship partner and not a
function of hostility (cf. Keller & Pfattheicher, 2013). As
predicted, empathy predicted aggression on behalf of a close
other if the close other was perceived to be distressed—and
thus a reasonable target of empathy—but not if the close
other was not perceived to be distressed.
In addition, empathy predicted less aggression on behalf
of a distressed close other in need among individuals with
the short/short version of the AVPR1a rs3 genetic variant, but
predicted higher aggression on behalf of a distressed, vulner-
able close other among those with at least one long version of
this variant. Similarly, the combination of target distress and
empathy significantly predicted increased empathy-linked
aggression for individuals with the GG genotype of OXTR
0
0.5
1
1.5
2
Low EmpathyHigh Empathy
Low Distress
High Distress
Aggressive Behaviors (OXTRGG)
0
0.5
1
1.5
2
Low EmpathyHigh Empathy
Low Distress
High Distress
Aggressive Actions (OXTR AA & AG)
(a)
(b)
Figure 2. Study 1: Graphs of empathy by distress predicting
confrontation of the close other’s perpetrator separately for
European American individuals with the (a) GG or (b) AA/AG
genotypes of OXTR rs53576.
Note. Low and high values of distress are one SD below the mean and one
SD above the mean, respectively. Confrontation is an index of physical,
verbal, and other confrontation (range = 0-3).
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
1414 Personality and Social Psychology Bulletin 40(11)
rs53576, but reduced aggression for those with the AG or AA
genotypes. Together, these results suggest that empathy may
have predicted aggression due to its effects on vasopressin
and oxytocin.
Study 1 provided evidence of empathy-linked aggression
in the real world, which was consistent with our predictions.
However, several features of this study limit our ability to
draw firm conclusions about this phenomenon. First, the cor-
relational design of Study 1 prevents drawing of conclusions
about causality. In addition, Study 1’s reliance on retrospec-
tive self-reports makes the accuracy and predictive validity
of such data unclear. Moreover, given that the conflicts
reported in Study 1 were those of a close other, we cannot
say whether empathy could facilitate aggression beyond the
realm of pre-existing close relationships. Finally, we cannot
rule out the possibility that empathic anger explains the phe-
nomenon of empathy-linked aggression. Together, these
limitations suggested the need for additional research. Study
2 used an experimental design with assessed laboratory
aggression to rule out potential biases of self-report and to
establish causality. The study’s core design was a scenario in
which the third party was a stranger, as was the empathy tar-
get, but was not a wrongdoer, reducing the likelihood of
empathic anger and allowing a test of whether empathy-
linked aggression generalizes beyond close relationships.
Study 2
Method
Participants. Introductory psychology subject pool partici-
pants at SUNY Buffalo (N = 162; gender: 105 women, 57
men; age: M = 19.18 years) were recruited for this study. The
ethnic composition of the sample was 25 African or African
American, 17 Asian or Asian American, 112 European or
European American, 6 Hispanic, 1 Arab American, and 1
Native American. All participants were compensated with
course credit.
Procedure and measures. Participants were told that they
would assist with two studies: first, a study on individual dif-
ferences in social behavior, for which they would provide a
saliva sample for DNA extraction; and second, a study on
whether different personality constructs would moderate the
usually hindering effects of pain on cognitive performance.
For the second study, participants were informed that their
role would be to rate the personality of one of a pair of other
participants, supposedly seated in one of our laboratory’s
additional lab rooms, based on an essay that the individual
had written. These participants, who were actually fictional,
were described as strangers to each other. Actual participants
were further told that one of the pretend participants in this
pair would be exposed to a painful but harmless stimulus, hot
sauce (a non-reactive measure of aggression; Lieberman,
Solomon, Greenberg, & McGregor, 1999), and then compete
in a math test. To motivate the (fictional) competitors, so par-
ticipants were further led to believe, the winner would win
US$20 and the loser would win nothing. To make it clear to
participants that the hot sauce consumed would inflict pain
on the participant, participants were told that the study inves-
tigated “the effects of physical pain on performance,” that
the hot sauce was used as the “painful stimulus,” and that we
were “using hot sauce as the source of participants’ pain.”
Empathy manipulation. After donating the saliva sample,
participants completed a short demographic questionnaire
and some filler questionnaires. Immediately following the
completion of these items, participants were instructed to
read a short essay by one of the (fictional) math-test par-
ticipants (referred to hereafter as the “empathy target”).
The instructions the actual participants received were mod-
eled after Batson and colleagues (Batson et al., 1997) and
comprised the empathy manipulation, with one randomly
assigned set of participants being asked to pay attention to
how the person feels rather than all of the information pre-
sented (the high empathy condition), whereas the other set of
participants were asked to remain objective and pay attention
only to the facts presented (the low empathy condition).
Distress manipulation. The content of the empathy target’s
essay comprised the distress manipulation. All participants
read this essay, which described a series of the empathy tar-
get’s financial setbacks, including the need to pay for course
registration and to replace a totaled car. However, one (ran-
domly assigned) half of the participants read a version of the
essay that concluded with the empathy target expressing high
distress over their financial need (i.e., “I’ve never been this
low on funds and it really scares me, to be honest. What if
I need to pay for something else I didn’t expect?”), whereas
the remainder of the participants read a version in which the
empathy target expressed low distress over their financial
need (“I’ve never been this low on funds, but it doesn’t really
bother me. I’m pretty sure things will get better soon, plus at
least I have a new car”).
Hot sauce assignment (aggression). Next, participants
were given the opportunity to aggress against the empathy
target’s math-test competitor by assigning hot sauce to the
competitor. To prevent hypothesis guessing, a further part of
the presented cover story led participants to believe that the
researchers were interested in examining math performance
across different amounts of a painful stimulus, and for that
reason, the amount of hot sauce the (fictional) math-test par-
ticipants would receive was variable. Participants were told
that to “assure random assignment” to hot sauce amounts, the
researchers were asking the participants to assign hot sauce to
the competitor in the dyad about whom they had not received
any prior information. Participants were further told that the
study participants were supposedly participating in measured
the effects of pain on performance, that hot sauce served as
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
Buffone and Poulin 1415
the painful stimulus, and that we expected the stimulus to
be sufficiently painful to hinder performance. Following a
procedure similar to that of McDermott, Tingley, Cowden,
Frazzetto, and Johnson (2009), participants were asked to
select an amount of hot sauce for the empathy target’s com-
petitor on a 6-point scale ranging from no hot sauce to three
teaspoons (one tablespoon) in half-teaspoon increments.
Suspicion probe. Finally, participants were carefully
probed for suspicion via two questions: “Could you please
tell me in your own words what this study was about?” and
“Did anything about this study seem odd or out of place to
you?” Participants were counted as suspicious if they did not
believe that there was another set of participants or if they
stated that the study measured aggressing on behalf of the
empathy target. Participants were then thoroughly debriefed,
thanked for their participation, and dismissed.
AVPR1a and OXTR variables. DNA from saliva samples
was analyzed for genetic data using procedures identical to
those used in Study 1 for AVPR1a rs1 and rs3 and for OXTR
rs53576. Also as in Study 1, ordinal variables were created
for AVPR1a rs1 and rs3 representing the number of alleles
classified as “long” (0-2) according to a median split. How-
ever, examining these data revealed that relatively few indi-
viduals were of the long/long genotype of rs3 (n = 15), so
individuals with any long allele (long/long and short/long)
were grouped together and a dichotomous variable con-
trasted these individuals with those with only short alleles.
In addition, a dichotomous variable was created for OXTR
rs53576 representing GG status (AA/AG = 1, GG = 0).
Additional measures. Participants rated the empathy target
on likeability and a variety of personality items to keep con-
sistent with the cover story.
Results
Descriptive statistics
Suspicion. A total of 12 participants were suspicious.
Excluding these participants from the analyses did not alter
our results for any model except the three-way interaction
with oxytocin. Here, the three-way interaction only reached
marginal significance after excluding suspicious participants
(p = .06).
Hot sauce assignment. The mean of the hot sauce assigned
was 4.17 on the 6-point scale, corresponding to approxi-
mately 1.5 teaspoons of hot sauce assigned on average.
AVPR1a and OXTR polymorphisms. Genotyping of
AVPR1a rs1 resulted in a range of repeats from 297 to 337
repeats. A median split at 317 repeats was used to distin-
guish long from short alleles, and 134 individuals provided
a saliva sample large enough for DNA extraction. The
genotyping of AVPR1a rs3 yielded a range of repeat values
from 326 to 352 repeats. A median split between 334 and 336
repeats was used to discern between short and long alleles.
Analyses of the OXTR SNP (rs53576) revealed that within the
group of European Americans (n = 112), 104 individuals pro-
vided enough saliva for their genotype to be identified (see Table
4). All polymorphisms examined were in Hardy–Weinberg
equilibrium, and there were no significant gender differences
in genotype. In this study, all of the oxytocin and vasopressin
polymorphisms were uncorrelated, rs(162) < .10, ps > .27.
Associations between empathy and aggression. A two-way
ANOVA tested the hypothesis that empathy and distress
would interact such that participants induced to feel empathy
would assign more hot sauce to the competitor when the
empathy target was described as being highly distressed over
his or her financial need but not when the empathy target was
described as not being very distressed over his or her finan-
cial need.
The Distress × Empathy interaction was significant, F(3,
158) = 14.62, p < .001 (see Table 5; Model I). Probing via t
tests showed that empathy led to aggression when the empa-
thy target self-described as being highly distressed t(161) =
3.57, p < .001, such that those in the empathy condition allo-
cated significantly more hot sauce to the competitor (M =
4.70, SD = 1.67) than did those in the objective condition
(M = 3.27, SD = 1.81); mean difference between groups:
Table 4. Study 2: Distribution of AVPR1a and OXTR Genotypes.
AVPR1a polymorphisms OXTR polymorphism
Genotype
rs1 variant rs3 variant
Genotype
rs53576 variant
Frequency Frequency Frequency
Short/short 14 (12.84%) 35 (26.12%) AA 9 (8.65%)
Short/long 76 (69.72%) 84 (62.69%) AG 47 (45.19%)
Long/long 19 (11.19%) 15 (11.12%) GG 48 (46.15%)
Total 109 134 Total 104a
aNumbers reflect European American individuals only.
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
1416 Personality and Social Psychology Bulletin 40(11)
1.41, 95% CI = [0.63, 2.20]. However, when the target self-
described as not being too distressed over their financial situ-
ation, empathy (M = 3.91, SD = 1.74) led to marginally less
aggression, t(161) = −1. 83, p = .07, compared with remain-
ing objective (M = 4.64, SD = 1.91); mean difference between
groups: −0.74, 95% CI = [−1.53, 0.05]. Together, these
results support our prediction that the interactive effect of
target distress and observer empathy leads to aggression.
Interactions of AVPR1a and OXTR genes, empathy, and dis-
tress. As in Study 1, we predicted that the two-way interac-
tion between empathy and target distress would be stronger
for individuals with more long versus short alleles of AVPR1a
rs1 and r3, and that the empathy by target distress interaction
would be stronger for individuals with the GG genotype of
OXTR rs53576 than for individuals with the AA or AG geno-
types. To test these predictions, each polymorphism was
entered into an ANOVA predicting amount of hot sauce
assigned along with empathy, distress, all three two-way
product–term interactions, and the three-way (Genotype ×
Empathy × Distress) product–term interaction. Thus, three
three-way ANOVAs (one for each polymorphism) were
tested.
Examination of the hypothesized interactions for AVPR1a
rs1 and rs3 revealed again no significant hypothesized inter-
action for rs1. However, there was a significant two-way
interaction between rs1 genotype and target distress, F(1,
108) = 3.71, p = .03 (see Table 5, Model II). Simple contrast
tests indicated that for individuals with the short/short or
short/long genotypes, high target distress led to lower levels
of aggression (M = 3.60, SD = 1.62) compared with low dis-
tress (M = 4.51, SD = 1.79), t(108) = −2.16, p = .03; mean
difference between groups: −0.77, 95% CI = [−1.49, −0.06].
However, for individuals with the long/long genotype, high
target distress caused greater aggression (M = 4.30, SD =
2.31) compared with low distress (M = 2.67, SD = 1.41),
t(108) = 2.01, p = .047; mean difference between groups:
1.71, 95% CI = [0.02, 3.39].
In addition, the hypothesized Genotype × Empathy ×
Distress three-way interaction emerged for rs3, F(7, 126) =
4.86, p = .03 (see Table 5, Model III). Simple contrast analy-
ses revealed a significant Empathy × Distress interaction for
individuals with any long version of the gene, F(1, 126) =
16.86, p < .001, but no significant interaction for those with
the short version of the gene—that is, those of the short/short
genotype: F(1, 126) = 0.02, p = .90. This pattern can be
viewed in Figures 3a and 3b. Specifically, among those with
any long alleles, when the target self-described as highly dis-
tressed, empathy led to greater (M = 4.59, SD = 1.57) hot
sauce assignment on behalf of the target, t(133) = 4.21, p <
.001, than did remaining objective (M = 3.13, SD = 1.77);
mean difference between groups: 2.04, 95% CI = [1.08,
3.00]. By contrast, when the target self-described as low on
distress, empathy (M = 2.67, SD = 1.15) marginally led to
less aggression against the competitor, t(133) = −1.64, p =
.10, compared with remaining objective (M = 5.50, SD =
1.87); mean difference between groups: −0.85, 95% CI =
[−1.86, 0.15].
Table 5. Study 2: ANOVA Models for Hot Sauce Assignment Predicted by Empathy, Distress, and Genotype.
Model IaModel IIbModel IIIcModel IVd
Variable F F F F
Empathy 3.35 1.18 2.71 2.73
Distress 12.03*** 0.02 14.84*** 1.31
AVPR1a rs1 2.12
AVPR1a rs3 0.05
OXTR rs53576 2.41
Empathy × Distress 14.62*** 3.82 16.86*** 0.05
rs1 × Empathy 0.57
rs1 × Distress 3.71*
rs1Empathy × Distress 1.67
rs3 × Empathy 0.04
rs3 × Distress 3.08
rs3 × Empathy × Distress 4.86*
rs53576 × Empathy 8.75**
rs53576 × Distress 0.75
rs53576 × Empathy ×
Distress
5.10*
an = 162; F(3, 158) = 5.69, p = .001.
bn = 109; F(11, 97) = 2.35, p = .01.
cn = 134; F(7, 126) = 3.34, p < .01.
dn = 104; F(7, 96) = 3.15, p < .01.
*p < .05. **p < .01. ***p < .001.
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
Buffone and Poulin 1417
As in Study 1, the proposed three-way OXTR × Distress ×
Empathy interaction reached significance for the European
Americans whose saliva contained enough DNA for analysis
(n = 104), F(7, 96) = 5.10, p = .03 (see Table 5, Model IV).
Simple contrast tests revealed that the only difference between
OXTR rs53576 GG individuals and AA/AG individuals
occurred in the low distress condition. Specifically, in the low
distress condition among GG individuals, empathy (M = 5.10;
SD = 1.73) trended to cause greater aggression, t(103) = 1.65,
p = .10, compared with remaining objective (M = 3.83, SD =
1.99); mean difference between groups: 1.27, 95% CI =
[−0.26, 2.79]. However, in the low distress condition among
AA/AG individuals, empathy (M = 3.09, SD = 1.64) caused
less aggression, t(103) = −2.58, p = .01, compared with
remaining objective (M = 4.88, SD = 1. 90); mean difference
between groups: −1.79, 95% CI = [−3.17, −0.42].
In addition to the three-way interaction, there was a sepa-
rate significant two-way interaction between rs53576 geno-
type and empathy, F(1, 103) = 8.75, p = .004. Simple contrast
tests indicated that for individuals with the AA/AG geno-
type, empathy did not cause any differences in aggression (M
= 3.90, SD = 1.70) compared with remaining objective (M =
3.91, SD = 2.09), t(103) = −0.04, p = .97; mean difference
between groups: −0.02, 95% CI = [−1.00, 0.96]. However,
for individuals with the GG genotype, empathy did cause
increased aggression (M = 4.62, SD = 1.86) compared with
remaining objective (M = 3.47, SD = 1.81), t(103) = 2.52, p
= .01; mean difference between groups: 1.39, 95% CI =
[0.29, 2.48]. In sum, while the three-way interaction of
empathy, need, and OXTR rs53576 genotype did not match
predictions, empathy did cause greater aggression for indi-
viduals with the GG genotype versus the AA/AG genotype,
regardless of need condition.
Gender effects. The Distress × Empathy interaction
appeared to be greater among women, F(3, 101) = 15.04, p <
.00, than among men, F(3, 53) = 1.25, p = .27. The effects of
empathy did not appear to be confounded by gender differ-
ences: The Empathy × Distress interaction remained signifi-
cant when gender and two-way interactions between gender
and distress and gender and empathy were added into the
model, t(161) = 3.75, p < .001.
The three-way interaction with OXTR rs53576 appeared
to be present in the subgroup of females, t(67) = −2.24,
95% CI = [−6.88, −0.39], p = .03, but not in the subgroup of
males, t(35) = −0.74, 95% CI = [−9.50, 4.47], p = .47. The
three-way interaction with AVPR1a rs3, by contrast,
appeared to be non-existent among females, t(85) = −0.67,
95% CI = [−2.42, 1.20], p = .51, but present among males,
t(47) = −2.94, 95% CI = [−5.38, 1.00], p < .01. The role of
genotype did not appear to be confounded by gender differ-
ences: There was no Gender × Empathy × Distress interac-
tion (p = .19).
Discussion
To build on the correlational, retrospective data of Study 1,
Study 2 used an experimental design with laboratory assess-
ment of aggressive behaviors. Study 2 again supported our
hypotheses about empathy motivating aggression. We
showed that an empathy manipulation increased aggres-
sion—amounts of hot sauce assigned—against the empathy
target’s competitor, but only when the empathy target was
described as distressed. The results from Study 2 are also
noteworthy in that they demonstrate that empathy-linked
aggression can occur for a stranger, not just a loved one,
and that provocation by the target—or indeed, any target
0
1
2
3
4
5
6
Low EmpathyHigh Empathy
Low Distress High Distress
Hot Sauce Assigned to Competitor B (AVPR1a rs3 SL & LL)
0
1
2
3
4
5
6
Low EmpathyHigh Empathy
Low Distress High Distress
Hot Sauce Assigned to Competitor B (AVPR1a rs3 SS)
(a) (b)
Figure 3. Study 2: Graphs of empathy by distress predicting assignment of hot sauce (original range = 0-6) separately for individuals
with (a) any long genotype versus those with the (b) short/short genotype for AVPR1a rs3.
Note. Scores reflect estimated marginal means.
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
1418 Personality and Social Psychology Bulletin 40(11)
characteristic—is unlikely to be the sole mechanism for
empathy-linked aggression. In this study, the competitor was
an innocent (bogus) participant about whom the test partici-
pant did not have any information, let alone any cause for
provocation.
In the current study, we again found evidence for the role
of vasopressin and oxytocin in the phenomenon of empathy-
linked aggression. The distress by empathy interaction was
significant specifically for individuals with long alleles of
AVPR1a rs3. This pattern in our data is consistent with the
possibility that vasopressin facilitates empathic responses,
including aggression, to individuals in need. In addition,
although we did not find the expected three-way interaction
for OXTR rs53576, we did find that the effects of empathy,
regardless of distress condition, were greater among OXTR
rs53576 GG individuals than among AA/AG individuals.
Although we can only speculate as to the reason why distress
functioned differently in this study versus Study 1, one pos-
sibility is that the range of distress in real-life situations such
as those examined in Study 1 provides more variability to
detect differential effects of OXTR, a possibility worth exam-
ining in future research.
In sum, Study 2 provided further documentation, in the
lab, of the role of empathy as a cause of aggression inde-
pendent of provocation by the aggression target. In addi-
tion, this study provided further evidence that oxytocin and
vasopressin may function as biological mechanisms for this
effect.
General Discussion
The present research found that assessed or elicited empathy
predicted aggression to benefit a distressed empathy target,
and that the effect of empathy may be partially explained by
the empathy-linked neurohormones vasopressin and oxyto-
cin. Specifically, Study 1 showed that when participants
reported feeling higher levels of empathy during a past epi-
sode when a close other was threatened and felt distressed,
participants were more likely to report having aggressed on
the close other’s behalf. Study 2 conceptually replicated this
pattern of results: Participants induced to be empathetic in a
lab setting assigned higher levels of a painful stimulus—hot
sauce—to a distressed empathy target’s competitor, com-
pared with participants instructed to remain objective. In
addition, in both studies, empathy predicted aggression more
strongly for OXTR rs53576 GG and AVPR1a rs3 long-allele
individuals than for others. Together, this pattern of findings
provides evidence that activation of the caregiving system
prompts aggression toward those in conflict or competition
with caregiving or empathy targets, even independent of tra-
ditional predictors of aggression such as threat to self, self–
other overlap, trait aggression and impulsivity, and in the
absence of wrongdoing or provocation from the target of
aggression. The findings of the present work have several
implications for understanding empathy and empathy-linked
aggression.
The Caregiving System and Aggression
Although the idea that people aggress on behalf of others is
not new at all (Batson et al., 1992; Hoffman, 1987, 1989;
Lickel, Miller, Stenstrom, Denson, & Schmader, 2006;
Neumann, 2000; Staub, 1978; Stotland, 1969; Vitaglione &
Barnett, 2003; Weiner, 1995), we believe the idea that empa-
thy can drive aggression absent provocation or injustice to be
quite novel. We predicted that this would be the case based on
the notion that empathy derives from the behavioral caregiv-
ing system (Goetz et al., 2010; Preston, 2013). This system is
most frequently described as being focused on nurturing and
soothing valued others, but it extends beyond these functions
and emotions to include the function of protecting vulnerable
others (Bell & Richard, 2000). Among non-human animals,
aggression on behalf of offspring (i.e., maternal and paternal
aggression) has long been studied as an integral part of the
caregiving system (Gammie & Nelson, 2000; Lonstein &
Gammie, 2002; Wolff, 1985). In humans, however, caregiv-
ing has been studied primarily in terms of prosocial tending to
offspring or other valued individuals (Bell & Richard, 2000;
Bowlby, 1969; Brown & Brown, 2006; Mikulincer, Shaver,
Gillath, & Nitzberg, 2005). Our findings affirm that, for
humans as well, caregiving and aggression are linked.
We had specifically predicted that the effects of empathy
on aggression would be accounted for by the actions of the
neurohormones vasopressin and oxytocin. Our finding that
variation in vasopressin and oxytocin receptor genes moder-
ate empathy’s effects supports this prediction. More broadly,
a role of these neurohormones in empathy’s effects is consis-
tent with the possibility that empathy facilitates a broad array
of behaviors—whether kind or aggressive—geared toward
benefiting vulnerable others (Goetz et al., 2010; Preston,
2013). These findings, combined with data from the present
studies and the known associations of vasopressin and oxy-
tocin with aggression (e.g., Bosch et al., 2005; Bosch et al.,
2010; Campbell, 2010; Carter, 1998; De Dreu, 2012; Hahn-
Holbrook, Holbrook, & Haselton, 2011; Hahn-Holbrook,
Holt-Lunstad, et al., 2011; Mah et al., 2014; Thompson et al.,
2006), make the effects of these neurohormones promising
mechanisms for empathy-linked aggression.
Implications for Real-World Aggression
Prior research only found empathy to predict aggression
toward an anger-arousing wrongdoer (Hoffman, 1989;
Neumann, 2000; Staub, 1978; Stotland, 1969; Weiner, 1995),
such as a drunk driver who injured an empathy target
(Vitaglione & Barnett, 2003). Although the situations
reported in Study 1 may have been likely to lead to anger or
provocation, in Study 2, the competitor who received hot
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
Buffone and Poulin 1419
sauce had neither indirectly nor directly caused the empathy
targets’ distress but nonetheless became the victim of aggres-
sion. If empathy can facilitate aggression independent of
provocation or wrongdoing, it may do so in a broader variety
of situations and toward a broader variety of targets than has
previously been considered. That is, empathy could facilitate
aggression any time such aggression is instrumental to help-
ing or benefiting the empathy target, independent of whether
or not such acts of aggression are just or morally sound. For
example, empathy could lead an individual to blame an inno-
cent person for a crime or misdeed to protect a friend or child
from punishment. This idea, although novel, is not entirely
without precedent: Batson and others have shown that empa-
thy can promote unjust—though not aggressive—outcomes
(Batson et al., 1995; Batson et al., 1992; Hoffman, 1987;
Lickel et al., 2006).
In the real world, situations in which empathy works
hand-in-hand with—and augments the effects of—anger in
response to provocation are likely more common than empa-
thy-linked aggression that is not fueled by angry affect.
Batson and colleagues have previously suggested that people
may sometimes be more motivated to aggress on behalf of
others than on behalf of themselves (Batson et al., 1992).
This effect may in fact be a result of empathy and may mani-
fest in real-world phenomena ranging from playground
fights to large-scale aggression. Concern over the distress
and suffering of loved ones and other ingroup members may
augment the desire to aggress seen in terrorists, gang mem-
bers, and others (cf. Cikara, Bruneau, & Saxe, 2011).
Limitations and Future Directions
The current work has limitations that point the way to future
research. First, and most notably, the functioning of the
AVPR1a and OXTR variants examined is still incompletely
understood. Although both are believed to affect sensitivity
to their respective neurohormones—vasopressin and oxyto-
cin (e.g., Knafo et al., 2008; Rodrigues et al., 2009)—the
mechanisms for these effects and the range of implications of
this increased sensitivity are unknown.
Second, in both studies, aggression was the only potential
behavioral outcome assessed. It is thus not known whether or
how empathy-linked aggression is associated with other
forms of helping, including directly providing assistance to
an empathy target. For example, it is possible that empathy-
linked aggression would be less likely if aggressing against a
third party were one of many different options to help. Future
research should examine this and other possible boundary
conditions of empathy-linked aggression.
In addition, in Study 2, we did not have participants taste
the hot sauce themselves to verify its painful nature, and we
did not have a statement by the third party claiming that they
disliked hot foods—both common procedures in research
using hot sauce as a measure of aggression. However,
instructions to participants made it very clear that
the (pretend) other participants were participating in a study
supposedly measuring the effects of pain on performance,
that hot sauce served as the painful stimulus, and that we
expected performance to decrease in response to the inges-
tion of hot sauce. Nonetheless, future research should exam-
ine other types of aggression as potential forms of
empathy-linked aggression.
Finally, in both studies, the pattern of results suggested
that the effects of empathy and the role of oxytocin and (to a
lesser extent) vasopressin may have been greater among
females than among males, but our studies were not powered
to test these gender differences. Given the sexually dimor-
phic roles of these neurohormones in non-human animals,
researchers with interests specifically in gender differences
may benefit from studies designed specifically to test the
prediction that empathy may promote aggression differen-
tially in females versus males.
In sum, the present work represents the first evidence
that empathy predicts aggressive behavior on its own, even
independent of provocation, and in conjunction with the
empathy-linked hormones/neuropeptides oxytocin and
vasopressin. These findings shed new light on the nature of
empathy, as well as on predictors of aggression and vio-
lence. Our findings additionally reaffirm the view that the
caregiving behavioral system functions in humans to
address not only needs for nurturance and comfort, but also
needs for protection. In closing, we should note that our
findings also provide a note of caution given recent enthu-
siasm for interventions that involve administering caregiv-
ing-related neurohormones or empathy training (Bornstein,
2010). Just as the self-esteem movement was not a panacea
leading to happy, successful, and well-adapted children
(Baumeister, Campbell, Krueger, & Vohs, 2003), oxytocin
and/or empathy interventions may not stop problems such
as bullying and other forms of aggression and violence,
because aggression itself may result from empathy. Whether
empathy is gentle or fierce depends very much on for whom
the empathy is felt.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect
to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, author-
ship, and/or publication of this article.
Notes
1. Due to the small number of people reporting “other” forms of
aggression, we also conducted all Study 1 analyses with a depen-
dent variable comprised just of physical and verbal aggression.
The results were substantively identical (i.e., all significant
results and directions of association were unchanged). In addi-
tion, to address possible violations of the assumption of hetero-
skedasticity given the count nature of this variable, all analyses
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
1420 Personality and Social Psychology Bulletin 40(11)
for Study 1 were examined using Huber–White robust standard
errors (Huber, 1967; White, 1980); results were substantively
identical.
2. Although many investigations of AVPR1a rs1 and rs3 rely on
dichotomizing repeat lengths into “long” and “short,” we also
conducted all analyses reported in Studies 1 and 2 using the
within-person mean number of repeats as a continuous variable.
Results obtained using this strategy yielded the same pattern of
results as reported in the studies’ respective results sections.
Supplemental Material
The online supplemental material is available at http://pspb.
sagepub.com/supplemental.
References
Anderson, C. A., & Huesmann, L. R. (2003). Human aggression:
A social-cognitive view. In M. A. Hogg & J. Cooper (Eds.),
SAGE handbook of social psychology (pp. 296-323). Newbury
Park, CA: SAGE.
Archer, J. (2009). Does sexual selection explain human sex differ-
ences? Behavioral and Brain Sciences, 32, 249-311.
Aron, A., Aron, E. N., & Smollan, D. (1992). Inclusion of Other
in the Self scale and the structure of interpersonal closeness.
Journal of Personality and Social Psychology, 63, 596-612.
Bakermans-Kranenburg, M. J., & van Ijzendoorn, M. H. (2008).
Oxytocin receptor (OXTR) and serotonin transporter (5-htt)
genes associated with observed parenting. Social Cognitive
and Affective Neuroscience, 3, 128-134.
Barraza, J. A., & Zak, P. J. (2009). Empathy toward strangers trig-
gers oxytocin release and subsequent generosity. Annals of the
New York Academy of Sciences, 1167, 182-189.
Batson, C. D. (1991). The altruism question: Toward a social-psy-
chological answer. Hillsdale, NJ: Lawrence Erlbaum.
Batson, C. D., Klein, T. R., Highberger, L., & Shaw, L. L. (1995).
Immorality from empathy-induced altruism: When compas-
sion and justice conflict. Journal of Personality and Social
Psychology, 68, 1042-1054.
Batson, C. D., Sager, K., Garst, E., Kang, M., Rubchinsky, K., &
Dawson, K. (1997). Is empathy-induced helping due to self-
other merging? Journal of Personality and Social Psychology,
73, 495-509.
Batson, C. D., Shaw, L. L., & Oleson, K. C. (1992). Differentiating
affect, mood, and emotion: Toward functionally based con-
ceptual distinctions. In M. S. Clark (Ed.), Review of person-
ality and social psychology, Vol. 13: Emotion (pp. 294-326).
Newbury Park, CA: SAGE.
Baumgartner, T., Heinrichs, M., Vonlanthen, A., Fischbacher, U.,
& Fehr, E. (2008). Oxytocin shapes the neural circuitry of trust
and trust adaptation in humans. Neuron, 58, 639-650.
Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K.
D. (2003). Does high self-esteem cause higher performance,
interpersonal success, happiness, or healthier lifestyles?
Psychological Science in the Public Interest, 4(1), 1-44.
Bell, D. C., & Richard, A. J. (2000). Caregiving: The forgotten ele-
ment in attachment. Psychological Inquiry, 11, 69-83.
Bornstein, D. (2010, November 8). Fighting bullying with babies.
The New York Times. Retrieved from http://opinionator.blogs.
nytimes.com/2010/11/08/fighting-bullying-with-babies/
Bosch, O. J., Meddle, S. L., Beiderbeck, D. I., Douglas, A. J., &
Neumann, I. D. (2005). Brain oxytocin correlates with maternal
aggression: Link to anxiety. The Journal of Neuroscience, 25,
6807-6815.
Bosch, O. J., Pfortsch, J., Beiderbeck, D. I., Landgraf, R., &
Neumann, I. D. (2010). Maternal behaviour is associated with
vasopressin release in the medial preoptic area and bed nucleus
of the stria terminalis in the rat. Journal of Neuroendocrinology,
22, 420-429.
Bowlby, J. (1969). Attachment and loss, Vol. 1: Attachment. New
York, NY: Basic Books.
Brown, S. L., & Brown, R. M. (2006). Selective Investment Theory:
Recasting the functional significance of close relationships.
Psychological Inquiry, 17, 1-29.
Buss, A. H., & Perry, M. P. (1992). The aggression questionnaire.
Journal of Personality and Social Psychology, 63, 452-459.
Campbell, A. (2010). Oxytocin and human social behaviour.
Personality and Social Psychology Review, 14, 281-295.
Carter, C. S. (1998). Neuroendocrine perspectives on social
attachment and love. Psychoneuroendocrinology, 23, 779-
818.
Chang, S., Glymour, M., Rewak, M., Cornelis, M. C., Walter,
S., Koenen, K. C., . . .Kubzansky, L. D. (2014). Are genetic
variations in OXTR, AVPR1A, and CD38 genes important
to social integration? Results from two large U.S. Cohorts.
Psychoneuroendocrinology, 39, 257-268.
Cikara, M., Bruneau, E. G., & Saxe, R. R. (2011). Us and them:
Intergroup failures of empathy. Current Directions in
Psychological Science, 20, 149-153.
Collins, N. L., & Ford, M. B. (2010). Responding to the needs of oth-
ers: The caregiving behavioral system in intimate relationships.
Journal of Social and Personal Relationships, 27, 235-244.
Condon, P., & DeSteno, D. (2011). Compassion for one reduces
punishment for another. Journal of Experimental Social
Psychology, 47, 698-701.
De Dreu, C. K. W. (2012). Oxytocin modulates cooperation within
and competition between groups: An integrative review and
research agenda. Hormones and Behavior, 61, 419-428.
De Dreu, C. K. W., Greer, L. L., Handgraaf, M. J. J., Shalvi, S., Van
Kleef, G. A., Baas, M., . . . Feith, S. W. W. (2010). The neu-
ropeptide oxytocin regulates parochial altruism in intergroup
conflict among humans. Science, 328, 1408-1411.
Feldman, R. (2012). Oxytocin and social affiliation in humans.
Hormones and Behavior, 61, 380-391.
Gammie, S. C., & Nelson, R. J. (2000). Maternal and mating-
induced aggression is associated with an elevation of citrul-
line immunoreactivity in the paraventricular nucleus in prairie
voles. Journal of Comparative Neurology, 418, 182-192.
Gillath, O., Shaver, P. R., Mikulincer, M., Nitzberg, R. E., Erez, A.,
& van IJzendoorn, M. H. (2005). Attachment, caregiving, and
volunteering: Placing volunteerism in an attachment-theoreti-
cal framework. Personal Relationships, 12, 425-446.
Goetz, J. L., Keltner, D., & Simon-Thomas, E. (2010). Compassion:
An evolutionary analysis and empirical review. Psychological
Bulletin, 136, 351-374.
Gonzalez-Liencres, C., Shamay-Tsoory, S. G., & Brüne, M.
(2013). Towards a neuroscience of empathy: Ontogeny, phy-
logeny, brain mechanisms, context, and psychopathology.
Neuroscience & Biobehavioral Reviews, 37, 1537-1548.
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
Buffone and Poulin 1421
Hahn-Holbrook, J., Holbrook, C., & Haselton, M. G. (2011).
Parental precaution: Neurobiological means and adaptive
ends. Neuroscience & Biobehavioral Reviews, 35, 1052-
1066.
Hahn-Holbrook, J., Holt-Lunstad, J., Holbrook, C., Coyne, S.,
& Lawson, E. T. (2011). Maternal defense: Breastfeeding
increases aggression by decreasing stress. Psychological
Science, 22, 1288-1295.
Hoffman, M. L. (1987). The contribution of empathy to justice and
moral judgment. In N. Eisenberg & J. Strayer (Eds.), Empathy
and its development (pp. 47-80). New York, NY: Cambridge
University Press.
Hoffman, M. L. (1989). Empathy and prosocial activism. In N.
Eisenberg, J. Reykowski, & E. Staub (Eds.), Social and moral
values (pp. 65-86). Hillsdale, NJ: Lawrence Erlbaum.
Hoge, E. A., Anderson, E., Lawson, E. A., Bui, E., Fischer, L.
E., Khadge, S. D., . . . Simon, N. M. (2014). Gender mod-
erates the effect of oxytocin on social judgments. Human
Psychopharmacology: Clinical and Experimental, 29, 299-304.
Huber, P. J. (1967). The behavior of maximum likelihood esti-
mates under nonstandard conditions. Proceedings of the
Fifth Berkeley Symposium on Mathematical Statistics and
Probability, 1, 221-233.
Hurlemann, R., Patin, A., Onur, A. O., Cohen, M. X., Baumgartner,
T., Metzler, S., . . .Kendrick, K. M. (2010). Oxytocin enhances
amygdala-dependent socially reinforced learning and emo-
tional empathy in humans. The Journal of Neuroscience,
30(14), 4999-5007.
Keller, J., & Pfattheicher, S. (2013). The compassion–hostility
paradox: The interplay of vigilant, prevention-focused self-
regulation, compassion, and hostility. Personality and Social
Psychology Bulletin, 39, 1518-1529.
Kemp, A. H., & Guastella, A. J. (2011). The role of oxytocin in
human affect: A novel hypothesis. Current Directions in
Psychological Science, 20, 222-231.
Kim, H. S., Sherman, D. K., Sasaki, J. Y., Xu, J., Chu, T. Q., Ryu,
C., . . . Taylor, S. E. (2010). Culture, distress, and oxytocin
receptor polymorphism (OXTR) interact to influence emotional
support seeking. Proceedings of the National Academy of
Sciences, 107, 15717-15721.
Knafo, A., Israel, S., Darvasi, A., Bachner-Melman, R., Uzefovsky,
F., Cohen, L., . . .Ebstein, R. P. (2008). Individual differences
in allocation of funds in the dictator game associated with
length of the arginine vasopressin 1a receptor rs3 promoter
region and correlation between rs3 length and hippocampal
mRNA. Genes, Brain and Behavior, 7, 266-275.
Kogan, A., Saslow, L. R., Impett, E. A., Oveis, C., Keltner, D., &
Rodrigues Saturn, S. (2011). Thin-slicing study of the oxytocin
receptor (OXTR) gene and the evaluation and expression of the
prosocial disposition. Proceedings of the National Academy of
Science, 108(48), 19189-19191.
Kosfeld, M., Heinrichs, M., Zak, P. J., Fischbacher, U., & Fehr,
E. (2005). Oxytocin increases trust in humans. Nature, 435,
673-676.
Lickel, B., Miller, N., Stenstrom, D. M., Denson, T. F., & Schmader,
T. (2006). Vicarious retribution: The role of collective blame
in intergroup aggression. Personality and Social Psychology
Review, 10, 372-390.
Lieberman, J. D., Solomon, S., Greenberg, J., & McGregor, H. A.
(1999). A hot new way to measure aggression: Hot sauce allo-
cation. Aggressive Behavior, 25, 331-348.
Lonstein, J. S., & Gammie, S. C. (2002). Sensory, hormonal, and
neural control of maternal aggression in laboratory rodents.
Neuroscience Biobehavioral Review, 26, 869-888.
Mah, B. L., Bakermans-Kranenburg, M. J., Van Ijzendoorn, M. H.,
& Smith, R. (2014). Oxytocin promotes protective behavior in
depressed mothers: A pilot study with the enthusiastic stranger
paradigm. Depression and Anxiety.
McDermott, R., Tingley, D., Cowden, J., Frazzetto, G., &
Johnson, D. D (2009). Monoamine oxidase gene (MAOA)
predicts behavioral aggression following provocation.
Proceedings of the National Academy of Science, 106(7),
2118-2123.
Mikolajczak, M., Gross, J. J., Lane, A., Corneille, O., De Timary,
P., & Luminet, O. (2010). Oxytocin makes people trusting, not
gullible. Psychological Science, 21, 1072-1074.
Mikulincer, M., Shaver, P. R., Bar-On, N., & Ein-Dor, T. (2010).
The pushes and pulls of close relationships: Attachment inse-
curities and relational ambivalence. Journal of Personality and
Social Psychology, 98, 450-468.
Mikulincer, M., Shaver, P. R., Gillath, O., & Nitzberg, R. A.
(2005). Attachment, caregiving, and altruism: Boosting attach-
ment security increases compassion and helping. Journal of
Personality and Social Psychology, 89, 817-839.
Neumann, R. (2000). The causal influences of attributions on emo-
tions: A procedural priming approach. Psychological Science,
11, 179-182.
Norman, G. J., Hawkley, L. C., Cole, S. W., Berntson, G. G., &
Cacioppo, J. T. (2012). Social neuroscience: The social brain,
oxytocin, and health. Social Neuroscience, 7, 18-29.
Patton, J. H., Stanford, M. S., & Barratt, E. S. (1995). Factor struc-
ture of the Barratt Impulsiveness Scale. Journal of Clinical
Psychology, 51, 768-774.
Poulin, M. J., Holman, E. A., & Buffone, A. E. K. (2012). The
neurogenetics of nice: Receptor genes for oxytocin and vaso-
pressin with threat to predict prosocial behavior. Psychological
Science, 23, 446-452.
Preston, S. D. (2013). The origins of altruism in offspring care.
Psychological Bulletin, 139, 1305-1341.
Rodrigues, S. M., Saslow, L. R., Garcia, N., John, O. P., & Keltner,
D. (2009). Oxytocin receptor genetic variation relates to empa-
thy and stress reactivity in humans. Proceedings of the National
Academy of Sciences, 106, 21437-21441.
Shalvi, S., & De Dreu, C. K. W. (2014). Oxytocin promotes group
serving dishonesty. Proceedings of the National Academy of
Science, 111, 5503-5507.
Shaver, P. R., Mikulincer, M., & Shemesh-Iron, M. (2009). A
behavioral-systems perspective on prosocial behavior. In M.
Mikulincer, P. R. Shaver, M. Mikulincer, & P. R. Shaver
(Eds.), Prosocial motives, emotions, and behavior: The better
angels of our nature (pp. 73-92). Washington, DC: American
Psychological Association.
Singer, T., Snozzi, R., Bird, G., Petrovic, P., Silani, G., Heinrichs,
M., & Dolan, R. J. (2008). Effects of oxytocin and prosocial
behavior on brain responses to direct and vicariously experi-
enced pain. Emotion, 8, 781-791.
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
1422 Personality and Social Psychology Bulletin 40(11)
Sober, E., & Wilson, D. S. (1998). Unto others: The evolution and
psychology of unselfish behavior. Cambridge, MA: Harvard
University Press.
Staub, E. (1978). Positive social behavior and morality: Social and
personal influences. Oxford, UK: Academic Press.
Stotland, E. (1969). Exploratory investigations of empathy.
Advances in Experimental Social Psychology, 4, 271-313.
Stuermer, S., Snyder, M., & Omoto, A. M. (2005). Prosocial emo-
tions and helping: The moderating role of group membership.
Journal of Personality and Social Psychology, 88, 532-546.
Thompson, R. R., George, K., Walton, J. C., Orr, S. P., & Benson,
J. (2006). Sex-specific influences of vasopressin on human
social communication. Proceedings of the National Academy
of Sciences, 103, 7889-7894.
Valencia, N., Savidge, M., & Maddox, J. (2012, June 20). Grand
jury: No indictment against father who killed daughter’s
alleged abuser. CNN.com. Retrieved from http://www.cnn.
com/2012/06/19/us/texas-abuser-father/index.html
Vitaglione, G. D., & Barnett, M. A. (2003). Assessing a new
dimension of empathy: Empathic anger as a predictor of
helping and punishing desires. Motivation and Emotion, 27,
301-325.
Walum, H., Westberg, L., Henningsson, S., Neiderhiser, J. M.,
Reiss, D., Igl, W., . . . Lichtenstein, P. (2008). Genetic variation
in the vasopressin receptor 1a gene (AVPR1a) associated with
pair-bonding behavior in humans. Proceedings of the National
Academy of Sciences, 105, 14153-14156.
Weiner, B. (1995). Inferences of responsibility and social motiva-
tion. In M. P. Zanna (Ed.), Advances in experimental social
psychology (pp. 1-47). San Diego, CA: Academic Press.
White, H. (1980). A heteroskedasticity-consistent covariance
matrix estimator and a direct test for heteroskedasticity.
Econometrica, 48, 817-830.
Wolff, J. O. (1985). Maternal aggression as a deterrent to infan-
ticide in Peromyscus leucopus and P. maniculatus. Animal
Behaviour, 33, 117-123.
at University at Buffalo Libraries on October 7, 2014psp.sagepub.comDownloaded from
... The rs53576 polymorphism, which also exerted a significant effect on self-reported altruism in our sample, is the most intensively studied OXTR polymorphism, implicated in a variety of prosocial behaviors, including altruistic giving (Poulin et al., 2012), attachment (Costa et al., 2009;Kim et al., 2010) and romantic love (Acevedo et al., 2019), as well as related emotional states, such as empathy (Buffone and Poulin, 2014;Gong et al., 2017;Huetter et al., 2016;Rodrigues et al., 2009;Smith et al., 2014;Uzefovsky et al., 2015) and trust (F. Krueger et al., 2012;Nishina et al., 2015). ...
... N = 109) = 0.002, p = .962Altruistic giving(Poulin et al., 2012); altruistic love towards a romantic partner(Acevedo et al., 2019); amygdala, hypothalamus, septum and ventral pallidum structure and function(Acevedo et al., 2019;Tost et al., 2010); antisocial behavior(Verona et al., 2018); attachment style(Costa et al., 2009;Kim et al., 2010); autism (S.Wu et al., 2005); empathy( Buffone and Poulin, 2014;Gong et al., 2017;Huetter et al., 2016;Rodrigues et al., 2009;Smith et al., 2014;Uzefovsky et al., 2015); parental sensitivity (Bakermans-Kranenburg and vanIjzendoorn, 2008;Israel et al., 2009;Sturge-Apple et al., 2012); prosociality(Kogan et al., 2011;Tost et al., 2010); trust (F.Krueger et al., 2012;Nishina et al.N = 138) = 1.34, p = .248Altruistic giving(Israel et al., 2009); antisocial behavior(Dadds et al., 2014;Malik et al., 2012;Waller et al., 2016); autism( Campbell et al., 2011;Lerer et al., 2008); parental sensitivity( Feldman et al.(Furman et al., 2011;H. ...
Article
Altruism is a type of prosocial behavior that is carried out in the absence of personal benefit or even at an expense to self. Trait altruism varies greatly across individuals, and the reasons for this variability are still not fully understood. Growing evidence suggests that altruism may be partly determined by the oxytocin receptor (OXTR) gene, which regulates the emotions underlying altruistic attitudes, such as empathy and trust. Neuroimaging and lesion studies have also implied several higher-order brain regions, including the prefrontal cortex, in altruistic behaviors. Yet the existing reports are contradictory and suggest that the top-down control exercised by the prefrontal cortex may promote both altruistic and self-interested behaviors and, thus, could obscure one’s natural proclivity towards altruism encoded by OXTR. Here, we hypothesized that extensive prefrontal damage would result in an increased influence of the OXTR genotype on one’s altruistic attitudes and actions. To test this hypothesis, we recruited 115 male combat veterans with penetrating traumatic brain injury to the prefrontal cortex and other brain regions, as well as 35 demographically matched control subjects without brain injury. Participants completed a self-report altruism questionnaire and were genotyped for four OXTR single nucleotide polymorphisms implicated in prosocial behavior, including rs53576, rs1042778, rs2254298 and rs7632287. Consistent with the previous studies, we found that individuals homozygotic for the G allele of rs53576 and rs7632287 were significantly more altruistic than carriers of at least one “vulnerable” A allele. Remarkably, in patients with prefrontal cortex damage, greater lesion extent was associated with significantly lower altruism scores in carriers of the A allele of rs7632287, but not in G-homozygotes, suggesting that significant disruption of the prefrontal cortex increased the influence of genetic polymorphisms on prosocial behavior. This study presents the first account of an interaction effect between the OXTR genotype and the location and extent of brain damage.
... Vitaglione and Barnett (2003) demonstrated that adults who reported greater empathic anger in response to witnessing individuals who have been intentionally hurt were motivated to support these individuals and to punish the person who offended. Similar findings have been shown in more general bystander research (Buffone & Poulin, 2014;Gummerum et al., 2016;van Doorn et al., 2018). In the context of bullying, defending research has shown that dispositional levels of empathic anger are associated with less passivity and more overall defending (Pozzoli et al., 2017), as well as intended prosocial and aggressive defending specifically (Steinvik et al., 2023). ...
Article
Full-text available
Bullying remains a significant social problem among youth and many school interventions focus on increasing bystander defending in order to reduce bullying. In this study, we employed a between-groups experimental design to examine the differential effects of brief empathy and compassion activation on different bystander responses to bullying, including (1) empathic distress, empathic anger, compassion, and (2) intended bystander behaviors (i.e., passive bystanding, aggressive defending, and prosocial defending). Participants were 110 adolescents (Mage = 13.99, SD = 0.88, age range = 13–16 years; 49.1% females), randomly assigned to an experimental group that involved a 10-min visualization exercise that focused on increasing empathy [EM] or compassion [CM], or to an active control condition [FI]. Following the visualization exercise, students viewed four short bullying videos, followed by completing self-report measures of empathy-related responses and intended bystander behaviors. Analysis of variance [ANOVAs] revealed that adolescents in the CM condition reported less empathic distress and empathic anger in response to the bullying videos than the EM and FI groups. Yet, there were no further differential effects between the three conditions on responses to the bullying videos, which emphasizes the need for future research to assess more comprehensive interventions for increasing adolescents’ compassion and prosocial defending.
... Gottschall argues that societies succeed or fail depending on how they manage these tensions. Buffone and Poulin (2014) provided empirical evidence to show that empathy for others motivate aggression on their behalf. Their study examined potential predictors of empathy-linked aggression and an empathy target's distress state. ...
Chapter
The decline in violence and growing awareness of human rights can be viewed as natural experiments in history answering the question if humans are moral creatures. A dual-process approach, which assumes the intuitive and reflective systems, is adopted to examine how the reflective system controls the intuitive system, which possibly produces our cruelty. Reading a story may enhance our reflective system to suppress our cruelty. However, the power of the reflective system is weak in arousing people’s action against violence. Hence, theory of mind (ToM) accompanied by emotional empathy is necessary for the suppression of cruelty. However, ToM is modular in nature and one of the subsystems of the intuitive system; hence, empathy is narrow in its focus, which possibly causes social fragmentation and political polarisation.
... According to psychological research, the more profound our affection for the victim (higher attachment), the greater our tendency to attack the perpetrator (Buffone & Poulin, 2014). ...
Article
Full-text available
In psychological literature, compassion is conceptualized as the desire to alleviate the suffering of others. It is presumed that Buddhist compassion refers to the same definition. However, from a cultural psychology perspective, it is essential to compare and contrast the definition of compassion and the content of Buddhist compassion since these two terms originate from two distinctly different cultures. As indicated by Richard Shweder, the collective human psyche exhibits "one mind, many mentalities." To bridge this research gap, we first made a "cultural interpretation" of compassion and that of Buddhist compassion to examine the discrepancy between these two. Second, we made a "scientific interpretation" of Buddhist compassion to elucidate this construct as well as related dimensions. Specifically, to make a "cultural interpretation" of compassion and that of Buddhist compassion, we adopted both word analysis and Kwang-Kuo Hwang's "cultural system approach." Word analysis showed that compassion and Buddhist compassion seem alike. However, the cultural system approach revealed that these two differ since Buddhist philosophy attempts to resolve the potential downside of compassion by proposing the "four immeasurables" (mettā, karuṇā, muditā, and upekkhā). The joy element (muditā) can help balance our depression in the face of other people's sufferings. Further, the equality element (upekkhā) can help reduce our attachment to loved ones, thereby lowering in-group favoritism and out-group dehumanization. The four immeasurables are a pathway to move from ordinary people's imperfect compassion towards flawless, boundless "great compassion" demonstrated by buddhas. Furthermore, to make a “scientific interpretation,” we analyzed the content of Buddhist compassion from a social science perspective and found that it consists of four dimensions: edification for wrongdoers, alleviation for sufferers, serenity within, and equality without (or EASE). Hence, Buddhist compassion can be defined as "helping sufferers and wrongdoers in a calm and equal manner." By elucidating the construct and the check-balance system of Buddhist compassion, this paper offers a theoretical conceptualization of Buddhist compassion for future empirical validation. Further, this paper also serves as an example of Hwang's "cultural system approach" and "theory building method" to construct a "scientific microworld."
... This might offer possible explanations for the lack of significant associations between rs53576 and generalized empathy in European cohorts; this was likely due to the inclusion of additional studies, some of which may have primarily indexed empathy via CE (e.g. an additional study using the RMET(Krueger et al., 2012)), thus reducing the observed effect in European studies. This further highlights the need to distinguish between empathy subdomains.Of the included studies in this analysis, 7 were newly added(Buffone & Poulin, 2014;Flasbeck et al., 2018;Krueger et al., 2012;Montag et al., 2017;Pearce et al., 2017;Sachdev et al., 2010;N. Wu et al., 2012) over the Gong et al. meta-analysis, and 5 were excluded (one study examining children at risk for autism (N. ...
Thesis
Full-text available
Social cognition refers to the range of skills and abilities that enable humans to detect and process information from one’s social environment, formulate a mental understanding of one’s social situation, and behave in socially appropriate ways. These include abilities such as theory of mind (ToM; also referred to as cognitive empathy), affective empathy, and social perception, as well as social behaviour. A growing body of research has sought to gain an understanding of how these phenomena manifest in the ageing process, as opposed to younger adults. The general aim of this thesis was to study the changes in social cognition with ageing, examine its relationship with other cognitive functions, and determine its association with genetic, neuroanatomical, and socioenvironmental factors. The first study explored the effect of polymorphisms of the oxytocin receptor gene on empathy using meta-analysis of existing studies including novel data from the Sydney Memory and Ageing Study (Sydney MAS). The second study developed a short-form version of the Reading the Mind in the Eyes Test (RMET), an assessment for theory of mind, via machine learning algorithms using Sydney MAS data. The third study explored social cognitive performance in Sydney MAS participants with subjective cognitive decline, mild cognitive impairment, and dementia. The fourth study identified key neuroimaging regions associated with empathy using volumetric analysis. The fifth study comprehensively indexed social cognition in nondemented community-dwelling older adults, and identified which subdomains were related to the ageing process and to other factors. This thesis found that normal ageing saw mild changes in ToM and social perception, and executive function somewhat compensated for this performance. Neurocognitive disorders were associated with far-reaching changes in these subdomains and some changes in social behavior. Empathy was related to volumes of the insula, supramarginal gyrus, and frontal lobe small vessel disease, and was not related to genetic sensitivity to oxytocin. A short-form version of the RMET was also developed. These findings improve on the understanding of social cognitive abilities in older adults, and facilitate the adoption of social cognition measures in clinical settings involving older adults.
... For instance, in order to evoke hatred toward immigrants, politicians create and highlight narratives about the victims of crimes committed by individual immigrants [178]. Likewise, Buffone and Poulin [179] showed that when participants were motivated to feel empathy for a student who was financially in need, they were more prone to administer a larger dose of hot sauce to the student's innocent competitor in a competition for a cash prize. These findings support the view that increased IFG activity is not only related to reduced creativity [180] but is also related to increased emotional empathy. ...
Article
Full-text available
Conflicts between groups are difficult to resolve, partly because humans tend to be biased in judging outgroup members. The aim of the current article is to review findings on the link between creativity and conflict-related biases and to offer a model that views creative cognition as an ability that may contribute to overcoming conflict-related biases. Our proposed model conforms to the twofold model of creativity. According to this model, creativity involves a generation phase and an evaluation phase, and these phases correspond to the neural mechanisms that underlie conflict-related biases. Specifically, we contend that the generation phase of creativity affects conflict-related biases by exerting an influence on stereotypes and prejudice, outgroup-targeted emotions, and ingroup empathy biases, all of which rely on the default mode network. Conversely, the evaluation phase of creativity, which is usually associated with activation in the executive control network and action-observation system, may be related to herding behaviors. Building on the shared mechanisms of creativity and conflicts, we propose that studies examining creativity-based interventions may be effective in promoting reconciliation.
Article
Full-text available
L’empathie joue un rôle fondamental dans la motivation à vouloir le bien-être des autres. Cependant, elle a des coûts cognitifs, émotionnels et énergétiques. Elle est limitée et fragile. Cet article intègre des connaissances en provenance de la théorie de l’évolution, des neurosciences, de la psychologie et de l’économie comportementale pour démontrer pourquoi et comment l’empathie est non-consciemment et rapidement modulée par divers signaux sociaux et facteurs situationnels. L’empathie n’est donc pas toujours une source d’information fiable dans la prise de décision morale. Ce cadre théorique des neurosciences sociales explique que la prise de décision guidée par l’empathie seule peut parfois éroder nos valeurs éthiques. Clarifier le rôle de l’empathie dans les processus de décision morale a des implications dans les domaines du droit, de la santé et des politiques publiques.
Chapter
Empathy needs to be communicated. This chapter begins by exploring how this may take place in music therapy. Communicating empathy may be transformational in itself, fostering a sense of validation for the client. Through experiencing the therapist’s empathy, clients may grow in their ability to identify and differentiate their emotions, accept and take ownership of them, and express and shape their emotions more clearly. Empathic concern may also be a starting point, prompting additional compassionate action. The empathy-altruism hypothesis is mentioned in terms of the associations between empathy and prosocial action. The chapter explores how empathy may guide a music therapist’s clinical decision-making and compassionate action. As empathy is complex, this chapter also acknowledges how empathy may lead to inaction, distancing, or even harm (through, for example, pathological altruism or advocative exploitative empathy). It closes by discussing the importance of self-empathy and self-compassion.KeywordsInsightful empathyCommunicating empathyClinical decision-makingValidationEmpathic concernCompassionate actionSelf-empathyMusic therapy
Article
Full-text available
Two experiments tested the idea that empathy-induced helping is due to self–other merging. To manipulate empathy, half of the participants in each experiment received instructions to remain objective while hearing about a young woman in need (low-empathy condition), and half received instructions to imagine her feelings (high-empathy condition). To check generality of the empathy–helping relationship, half in each empathy condition learned that the young woman was a student at their university (shared group membership), and half learned that she was a student at a rival university (unshared group membership). Self-reported empathy for and willingness to help the young woman were assessed, and 3 measures of self–other merging were taken. In each experiment, an empathy–helping relationship was found, unqualified by group membership, that could not be accounted for by any of the merging measures.
Article
Full-text available
The purpose of the present study was to revise the Barratt Impulsiveness Scale Version 10 (BIS-10), identify the factor structure of the items among normals, and compare their scores on the revised form (BIS-11) with psychiatric inpatients and prison inmates. The scale was administered to 412 college undergraduates, 248 psychiatric inpatients, and 73 male prison inmates. Exploratory principal components analysis of the items identified six primary factors and three second-order factors. The three second-order factors were labeled Attentional Impulsiveness, Motor Impulsiveness, and Nonplanning Impulsiveness. Two of the three second-order factors identified in the BIS-11 were consistent with those proposed by Barratt (1985), but no cognitive impulsiveness component was identified per se. The results of the present study suggest that the total score of the BIS-11 is an internally consistent measure of impulsiveness and has potential clinical utility for measuring impulsiveness among selected patient and inmate populations.
Article
Full-text available
Significance Very little is known about the biological foundations of immoral behavior. We report here the results of a double-blind, placebo-controlled experiment showing that the hormone oxytocin promotes group-serving dishonesty. Compared with participants receiving placebo, participants receiving oxytocin lied more to benefit their groups, did so quicker, and did so without expectation of reciprocal dishonesty from their group members. A control setting ruled out that oxytocin drives self-serving dishonesty. These findings support the functional approach to morality and reveal the underlying biological circuitries associated with group-serving dishonesty.
Article
The purpose of the present study was to revise the Barratt Impulsiveness Scale Version 10 (BIS-10), identify the factor structure of the items among normals, and compare their scores on the revised form (BIS-11) with psychiatric inpatients and prison inmates. The scale was administered to 412 college undergraduates, 248 psychiatric inpatients, and 73 male prison inmates. Exploratory principal components analysis of the items identified six primary factors and three second-order factors. The three second-order factors were labeled Attentional Impulsiveness, Motor Impulsiveness, and Nonplanning Impulsiveness. Two of the three second-order factors identified in the BIS-11 were consistent with those proposed by Barratt (1985), but no cognitive impulsiveness component was identified per se. The results of the present study suggest that the total score of the BIS-11 is an internally consistent measure of impulsiveness and has potential clinical utility for measuring impulsiveness among selected patient and inmate populations.
Article
A new questionnaire on aggression was constructed. Replicated factor analyses yielded 4 scales: Physical Aggression, Verbal Aggression, Anger, and Hostility. Correlational analysis revealed that anger is the bridge between both physical and verbal aggression and hostility. The scales showed internal consistency and stability over time. Men scored slightly higher on Verbal Aggression and Hostility and much higher on Physical Aggression. There was no sex difference for Anger. The various scales correlated differently with various personality traits. Scale scores correlated with peer nominations of the various kinds of aggression. These findings suggest the need to assess not only overall aggression but also its individual components.
Chapter
Publisher Summary It is possible for one person to experience an emotion when he or she perceives that another person is experiencing an emotion. The relationship between action and the sharing of feelings is obviously not a simple or direct one. It is possible to study so subtle and important a phenomenon as empathy in the laboratory and to examine some of the determinants of empathy. The process leading to empathy can be understood in terms of cognitive variables such as the mental set that the person has when he or she observes the other. The form or type of social relationships between one person and another influences the amount of empathy, presumably because the form of the social relationship influences the manner of perceiving the other and thinking about him or her. Individual differences in reactions to social situations, in perceiving the other, and in thinking about him or her must be considered in predicting how much empathizing will occur. These individual differences appear to be determined in part by the birth order of the person.
Article
Objective The neuropeptide oxytocin is implicated in social processing, and recent research has begun to explore how gender relates to the reported effects. This study examined the effects of oxytocin on social affective perception and learning.Methods Forty-seven male and female participants made judgments of faces during two different tasks, after being randomized to either double-blinded intranasal oxytocin or placebo. In the first task, “unseen” affective stimuli were presented in a continuous flash suppression paradigm, and participants evaluated faces paired with these stimuli on dimensions of competence, trustworthiness, and warmth. In the second task, participants learned affective associations between neutral faces and affective acts through a gossip learning procedure and later made affective ratings of the faces.ResultsIn both tasks, we found that gender moderated the effect of oxytocin, such that male participants in the oxytocin condition rated faces more negatively, compared with placebo. The opposite pattern of findings emerged for female participants: they rated faces more positively in the oxytocin condition, compared with placebo.Conclusions These findings contribute to a small but growing body of research demonstrating differential effects of oxytocin in men and women. Copyright © 2014 John Wiley & Sons, Ltd.