Empathy Toward Strangers Triggers Oxytocin Release and Subsequent Generosity

Abstract

Empathy is related to a variety of prosocial behaviors, but the brain mechanisms producing the experience of empathy have not been fully characterized. This study investigated whether the experience of empathy raises oxytocin levels and affects subsequent generosity toward strangers. Short video clips of an emotional scene and an unemotional scene were used as stimuli. Participants rated the emotions they experienced and then played a $40 ultimatum game to gauge their generosity. We found that empathy was associated with a 47% increase in oxytocin from baseline. We also found the empathy-oxytocin response was stronger in women than in men. Higher levels of empathy were also associated with more generous monetary offers toward strangers in the ultimatum game. Our findings provide the first evidence that oxytocin is a physiologic signature for empathy and that empathy mediates generosity.

Full text

VALUES, EMPATHY, AND FAIRNESS ACROSS SOCIAL BARRIERS
Empathy toward Strangers Triggers Oxytocin
Release and Subsequent Generosity
Jorge A. Barrazaaand Paul J. Zakb
aSchool of Organizational and Behavioral Studies and Center for Neuroeconomics
Studies, Claremont Graduate University, Claremont, California 91711, USA
bDepartment of Economics and Center for Neuroeconomics Studies, Claremont Graduate
University, Claremont, California 91711, USA
Empathy is related to a variety of prosocial behaviors,but the brain mechanisms produc-
ing the experience of empathy have not been fully characterized. This study investigated
whether the experience of empathy raises oxytocin levels and affects subsequent gen-
erosity toward strangers. Short video clips of an emotional scene and an unemotional
scene were used as stimuli. Participants rated the emotions they experienced and then
played a $40 ultimatum game to gauge their generosity. We found that empathy was
associated with a 47% increase in oxytocin from baseline. We also found the empathy–
oxytocin response was stronger in women than in men. Higher levels of empathy were
also associated with more generous monetary offers toward strangers in the ultimatum
game. Our findings provide the first evidence that oxytocin is a physiologic signature
for empathy and that empathy mediates generosity.
Key words: oxytocin; empathy; distress; gender; hormones; emotion induction
Introduction
Humans are often aroused by the distress
of others. Empathy allows us to perceive an-
other’s affective state and motivates action if the
other is perceived to be in an aversive state.1–3
The enduring interest in empathy across disci-
plines (as illustrated by this issue) is caused, in
part, by its relationship to moral behaviors, as
argued by Aristotle in The Nicomachean Ethics4
and Adam Smith in The Theory of Moral Senti-
ments5as well as other scholars. Altruism can
be considered morally virtuous and has been
associated with empathy.2,6The experience of
empathy has been shown to motivate prosocial
behaviors, such as volunteering and donations
to charities.7,8
Although much is known about the behav-
ioral outcomes when people are empathic, the
Address for correspondence: Paul J. Zak, Department of Economics
and Center for Neuroeconomics Studies, Claremont Graduate University,
160 East 10th Street, Claremont, CA 91711-6165. paul.zak@cgu.edu
physiologic mechanisms of empathy are not
well understood.7,9Specifically, little is known
about how observing the aversive states of oth-
ers translates into the subjective experience of
empathy. In addition, the physiologic substrates
causing individuals to experience different em-
pathic states, such as empathic concern or per-
sonal distress when observing aversive states in
others, is unknown. We propose that the neu-
rohormone oxytocin (OT) may be part of the
brain architecture that produces experienced
empathy. OT is associated with attachment be-
haviors in mammals, and we identified OT as
a likely mechanism that causes human beings
to respond to the affective states of others. OT
is a neuroactive hormone that is directly syn-
thesized in the hypothalamus and projects to
brain areas that are associated with emotions
and social behaviors (e.g., amygdala and cingu-
late cortex).10 In socially monogamous mam-
mals, OT mediates prosocial behaviors, such as
partner preference, social recognition, parental
care, and social approach.11–14
Values, Empathy, and Fairness across Social Barriers: Ann. N.Y. Acad. Sci. 1167: 182–189 (2009).
doi: 10.1111/j.1749-6632.2009.04504.x c
!2009 New York Academy of Sciences.
182

Barraza & Zak: Empathy and Oxytocin Release
183
Recent studies in humans have revealed that
OT promotes prosocial behaviors, including
trust, reciprocity, and generosity measured us-
ing monetary transfers to strangers.15–19 Specif-
ically, OT levels measured in plasma were 41%
higher in subjects after a monetary transfer de-
noting trust was received compared to those
who received a randomly chosen transfer of
the same average amount. In these studies,
OT levels were positively associated with in-
creased monetary reciprocity toward the per-
son who initiated trust.17,18 We discovered
recently that endogenous OT release and self-
sacrificial reciprocity can be magnified by ex-
posing participants to touch prior to making
decisions. Fifteen minutes of moderate pressure
massage increased the change in OT after be-
ing trusted by 16% and increased reciprocity
by 243% compared to controls who rested for
15 min.20
Exogenous OT infusion studies in humans
have demonstrated its causal effect on proso-
cial behaviors. Intranasal infusion of 24 IU of
OT increased monetary transfers to a stranger
(denoting trust) by 17%.15 Further, in a mone-
tary transfer task that involves making an offer
to share a fixed sum of money, known as the
ultimatum game (UG), 40 IU of intranasal OT
increased the generosity of offers by 80% over
placebo.19 These studies show that OT is asso-
ciated with prosocial behaviors but leave open
the question of whether OT is associated with
empathy.
The Current Study
Herein we report a direct test of whether
OT is a proximate mechanism modulating
the subjective experience of empathy. We hy-
pothesized that OT would spike after ex-
posure to an emotional stimulus and would
be associated with the experience of two
empathic states: personal distress and em-
pathic concern. We also tested if elevated OT
would elicit a prosocial behavior—generosity
toward a stranger. Two behavioral tasks were
used to test the empathy–prosociality associ-
ation: offers in the UG and monetary do-
nations to charity. We used the UG, as re-
ported in Zak and colleagues,19 as it requires
perspective taking by participants—a cogni-
tive exercise that has been shown to provoke
empathy.6,21
Materials and Methods
Participants and Procedure
One hundred and forty-five college students
(52% female students, mean age 20.8 years,
SD =3.3) from the University of California,
Los Angeles (UCLA) participated in this study.
Participants were randomly assigned to one of
three groups: emotional video and UG (EU,
n=61), control video and UG (CU, n=56), or
emotional video only (E, n=24). Three partic-
ipants (one from each condition) were excluded
from analyses because of OT levels outside of
the acceptable assay range (>2500 pg/mL) at
baseline, which is 5 SD above the mean.
Participants were recruited by email and
earned $10 for agreeing to be part of the ex-
periment. Total earnings were based on the
decision task as discussed below. After con-
sent, participants were led to a private room
for their first blood draw by a licensed phle-
botomist. Participants were then seated at par-
titioned computer stations and asked to fill out a
survey. Once finished, participants viewed one
of two brief videos and were asked to rate the
degree to which particular emotions were felt.
Participants then played a single round of the
UG sequentially for money (except for the E
group). Survey, video, and UG instructions and
decisions were made via computer. No inter-
personal communication was permitted. Im-
mediately after the decision in the UG, a second
blood draw was performed for those in the EU
and CU groups. The E group had their second
blood draw after viewing the video. After the
second blood draw, participants were privately
informed of their study earnings and presented

184
Annals of the New York Academy of Sciences
with the option to donate to a charity. When
all tasks were completed, participants were pri-
vately paid by a lab administrator who was not
associated with the study. The protocol was ap-
proved by the Institutional Review Boards of
UCLA and Claremont Graduate University.
The study was double blind and no deception
of any kind was used.
Ultimatum Game
Participants played a single round of the UG
to assess generosity toward a stranger.19,22 In
the UG, participants were randomly put into
dyads and in each dyad were randomly as-
signed to the role of decision maker 1 (DM1) or
decision maker 2 (DM2). Both DMs received
extensive and identical instructions for the UG,
including examples. In our version of the UG,
DM1s were endowed with $40 and were asked
to choose an offer of a split of this money to the
DM2 in his or her dyad. DM2s had no endow-
ment. If DM2 accepted the offer from DM1,
both DMs were subsequently paid the money
according to the accepted division. However,
both DMs earned nothing if DM2 rejected the
offer from DM1. The UG task is designed to
have participants consider how the DM2 in the
dyad would react to an offer (perspective tak-
ing) because DM2s can reject offers. A rejection
of the offer from DM1 in the UG allows DM2
to punish DM1 for stinginess but at a cost of the
loss of the money offered. Although most UG
experiments are played with $10,23 a $40 en-
dowment was used in this experiment in order
to compensate participants for two blood draws
as well as to explore whether a parametric re-
lationship existed between DM2 offers and the
change in OT, as has been found for DM2s in a
related monetary decision task called the trust
game.18,20
Blood Draw
After consent, all participants had 20 mL of
blood drawn by a licensed phlebotomist from
an antecubital vein. Two, 8-mL, EDTA, whole-
blood tubes and one serum-separator tube were
drawn while maintaining a sterile field and us-
ing a Vacutainer c
!(BD, Franklin Lakes, NJ,
USA). Those in the EU and CU groups had
a 20-mL second blood draw immediately fol-
lowing their decision in the UG. Participants
were prompted to make their decisions serially
so that the decision and blood draw were tem-
porally close, typically occurring within 2 min
after the decision, as in Zak and colleagues.18
Participants in the E group received a second
20-mL blood draw following viewing and rat-
ing of the emotional video. Blood tubes were
immediately placed on ice after being drawn.
The tubes were then placed in a refrigerated
centrifuge and spun at 1500 rpm for 12 min at
4◦C. Plasma and serum were removed from
the tubes and placed into 2-mL microtubes
with screw caps. These tubes were immedi-
ately placed on dry ice and then transferred
to a −70◦C freezer until analysis.
Assays
Five hormones were assayed using either
radioimmunoassay (RIA) or enzyme-linked
immunosorbent assays (ELISA). Andrenocorti-
cotropin hormone (ACTH) (plasma-RIA) sam-
ples were assayed using a kit produced by
DiaSorin, Inc. (Stillwater, MN, USA), corti-
sol (serum-RIA) samples were assayed using
a Diagnostic Systems Laboratories (Webster,
TX, USA) kit, and progesterone (serum-RIA)
and estradiol (serum-RIA) were assayed with
kits from Siemens Healthcare Diagnostics Inc.
(Los Angeles, CA, USA). OT was assayed us-
ing a competitive ELISA assay from Assay De-
signs, Inc. (Ann Arbor, MI, USA). The inter-
and intra-assay coefficients of variations for OT
were 7.8% at 484.68 pg/mL and 10.2% at
494.63 pg/mL (10 replicates), respectively. All
tests were performed by the Endocrine Core
Laboratory of the Yerkes National Primate Re-
search Center at Emory University, Atlanta,
GA.

Barraza & Zak: Empathy and Oxytocin Release
185
Surveys
Participants filled out several survey instru-
ments to examine the effects of personality
factors on OT release and behavior. Instru-
ments included the Interpersonal Reactivity In-
dex,24 the Affect Intensity Measure,25 the Big
Five Inventory,26 Cognitive Hardiness,27 along
with basic demographic questions taken from
Zak and colleagues.18
Video
All participants, using headphones, privately
viewed a 2-min long video in their partitioned
computer stations. Participants in the EU and
E conditions watched a video in which a father
explains his current experiences with his 2-year-
old son who has terminal brain cancer. The
video includes scenes of the child in the hospital
and with his father who narrates the video clip.
Participants in the CU condition watched a
clip similar in length with images of the father
and child. However, the narration was of the
father describing a day at the zoo and has no
mention of the child’s illness or any expression
of concern for the child.
Video Ratings
At the end of the video, participants were
asked to rate the degree to which they expe-
rienced particular emotions while viewing the
video. This list included 12 adjectives previ-
ously used to assess empathy toward others5
(e.g., sympathetic,compassion,moved, tender, warm,
soft-hearted) and personal distress (e.g., anxious,
distressed,sad, annoyed, frightened, disturbed ). Partic-
ipants rated these adjectives from 1 (did not feel
this way at all)to5(felt this way very much). Com-
posite measures were created for both empathy
(α=0.75) and distress (α=0.60).
Donation Task
After the UG and second blood draw, partic-
ipants were informed of their study earnings in
private and presented with the opportunity to
donate any amount of their study earnings to
one of two well-known charities (St. Jude Chil-
dren’s Hospital, or the American Red Cross).
The experimenters informed participants that
there was no obligation to donate and that their
decision to donate was anonymous.
Results
Response to Video
There was no change in OT in those who
viewed the emotional video (EU +E: base-
line OT =474.87 pg/mL, SD =306.75, post-
video OT =448.91 pg/mL, SD =288.72;
one-tailed paired ttest, P=0.21, n=80).
There was a significant decrease in OT in those
who viewed the control video (CU: baseline
OT =464.96 pg/mL, SD =341.90, post-video
OT =377.64 pg/mL, SD =250.95; two-tailed
paired ttest, P=0.03). However, separating
emotional video conditions we found different
results. OT significantly increased among par-
ticipants in the E condition who viewed the
emotional video but did not play the UG (base-
line OT =401.83 pg/mL, SD =230.06, post-
video OT =592.19 pg/mL, SD =225.34;
two-tailed paired ttest, P=0.004). Al-
ternatively, there was a significant decrease
in OT for those who viewed the emotional
video and played the UG (EU: baseline
OT =502.57 pg/mL, SD =328.75, post-
video OT =394.56 pg/mL, SD =293.08;
two-tailed paired ttest, P<0.001). As
Figure 1 shows, the emotional video increased
OT but not when the second blood draw fol-
lowed the UG.
Emotional Ratings
Participants in the E and EU conditions
rated the emotional video as eliciting greater
empathy than those in the CU condition (CU
M=2.88, E M=3.46, P<0.001; EU
M=3.53, P<0.001, both one-tailed ttests)

186
Annals of the New York Academy of Sciences
Figure 1. Change in oxytocin (OT) from baseline to post video across all conditions; UG,
ultimatum game. ∗denotes a significant difference at p <.05.
Figure 2. Participant video ratings of empathy and distress across all conditions.∗denotes
a significant difference at p <.05.
and distress (CU M=2.74, E M=2.93,
P=0.08; EU M=3.17, P<0.001,
both one-tailed ttests). Participants in the E
and EU condition reported statistically equal
experiences of empathy (P=0.63) and distress
(P=0.12). See Figure 2.
Using simple correlations across all con-
ditions, there was no relationship between
a change in OT and subjective empathy
(r=0.053, P=0.28) or distress (r=−0.074,
P=0.21). We discovered, however, a high
correlation between self-reported empathy and
distress (r=0.81, P<0.001). As a result, partial
correlations were examined between OT and
each emotion, controlling for the other one.
Controlling for these cross-effects, a spike in
OT was significantly associated with increased
feelings of empathy (r=0.197, P=0.01) as well
as decreased feelings of distress (r=−0.188,
P=0.02). A hierarchical regression analysis
was conducted to test if empathy and distress
predicted a change in OT when controlling
for changes in other hormones (ACTH, corti-
sol). Changes in hormones related to OT and
gender were entered in the first step, followed
by empathy and distress video ratings in the
second step of the equation. This analysis re-
vealed a significant overall regression equation:

Barraza & Zak: Empathy and Oxytocin Release
187
F(5, 118) =3.09, P<0.01. Both empathy (β=
0.294, P<0.05) and distress (β=−0.301,
P<0.05) were significant predictors of the
change in OT in different directions. This result
was maintained in separate analyses controlling
for basal levels of progesterone or estradiol in
women.
Emotional Ratings and Other Hormones
Across all conditions, emotional ratings were
not significantly correlated with any of the
hormones (ACTH, cortisol, estradiol, proges-
terone) at baseline or post video. However,
when controlling for one another, empathy and
distress were significantly correlated with the
change in cortisol (empathy r=0.134, P=
0.07; distress r=−0.203, P=0.01). Empathy
and distress ratings were marginally correlated
with post-video OT (empathy r=0.121, P=
0.09; distress r=−0.118, P=0.09) and post-
video cortisol (empathy r=0.162, P=0.04;
distress r=−0.171, P=0.03).
Generosity in UG
Of the participants playing as DM1 (n=56),
35 (62.5%) offered an equal split, 20 (26%)
made unequal offers of $10–$19, and two made
supra-equal offers of $21 and $30. There were
49 participants who participated as DM2s.
Of those, only one person in the EU and
one in the CU groups rejected DM1 offers
of $10; all other offers were accepted. DM2s
whose offers were rejected were removed from
subsequent analyses. There were no differ-
ences in the mean DM1 offers between the
EU (M=$18.18, SD =$3.19) and CU
(M=$17.62, SD =$4.54; one-tailed ttest,
P=0.29) conditions.
Consistent with our hypothesis, DM1 pro-
posals in the UG were positively correlated with
reported empathy after the video (r=0.239,
P=0.05). There was a weak relationship be-
tween DM1 offers and distress levels (r=0.171,
P=0.11). Similar to studies of OT in the trust
game,18 DM1 behavior was uncorrelated with
the change in OT (r=−0.150, P=0.14). The
amount offered to DM2s was marginally nega-
tively correlated to the change in DM2 cortisol
(r=−0.202, P=0.07) but not to change in
OT (r=−0.120, P=0.20). Controlling for
gender and changes in ACTH and cortisol, the
money offered to DM2s did not predict change
in OT (β=−0.08, P=0.54).
Charitable Donations
Forty-four participants (32%) made mone-
tary donations (M=$6.09, SD =6.31). Do-
nations were significantly correlated with the
amount sent by DM1s (r=0.356, P=0.004).
Among all participants, donations were pos-
itively related to the change in cortisol (r=
0.146, P=0.05) but were unrelated to the
change in OT (r=−0.010, P=0.45) or to
the change in ACTH (r=−0.084, P=0.18).
Donations were not associated with emotional
video ratings (empathy r=−0.088, P=0.16;
distress r=−0.080, P=0.19).
Gender and Personality
Pooling all conditions, we found that emo-
tional ratings (controlling for one another) were
more strongly associated with changes in OT
for women (empathy r=0.245, P=0.03; dis-
tress r=−0.258, P=0.02) than for men (em-
pathy r=0.158, P=0.11; distress r=−0.134,
P=0.15). Behaviorally, more women made
charitable donations than men (23% of men
versus 41% of women, χ2=4.78, P=0.03) and
gave more in donations than men (M=$2.89
versus M=$1.08; two-tailed ttest, P=0.02).
The average amount sent by DM1s was also
greater for women than men (women $18.85,
men $17.10; one-tailed ttest, P=0.05). The
change in OT was associated with increased
dispositional empathy (r=.187, p=.02) as
measured in the IRI. No other personality vari-
ables were associated with basal OT or the
change in OT.

188
Annals of the New York Academy of Sciences
Discussion
There were three main findings from this
study. First, viewing an emotional video raised
OT by an average of 47% over baseline com-
pared to those who watched an emotionally
neutral video. Second, there was a positive
relationship between the degree of empa-
thy experienced and the change in OT.
Third, an increase in experienced empathy
was associated with greater generosity in the
UG.
Past research has purported that emotional
videos may induce OT release.28 We provide
the first direct evidence for this claim, and
we have demonstrated both a statistically and
quantitatively significant increase in OT after
an emotional stimulus. Even more compelling,
we discovered a positive parametric relation-
ship between the experience of empathy and
the change in OT. The relationship between
empathy and the change in OT was especially
strong for women. Moreover, we found that the
empathic concern subscale of the IRI, a mea-
sure of dispositional empathy (e.g., sympathy,
compassion), to be the only personality vari-
able to predict a spike in OT. The lack of a
relationship between the change in OT when
the emotional video was followed by the UG is
likely a result of the time lag between the video
and the second blood draw, which was required
for instructions and UG decisions. The half-life
of OT is very short, with estimates of between
1–2 min.29
We also reported that the experience of em-
pathy positively influenced prosociality. Partic-
ipants who were empathically engaged by the
video they viewed made more generous offers
in the UG. Those who made more generous of-
fers also donated more money to charity, with
this effect associated with physiologic distress
(a positive change in cortisol). Donations were
highest among women in the sample. At the
same time, the change in OT was strongest
among women. Post-hoc analyses found that
these gender differences were not driven by the
upregulation of OT by estrogen.30
We also found an interesting counteracting
effect of distress on OT release. Empathy
and distress were highly related in our sample
and they appear to work against each other
at a physiologic level. Psychologists have also
distinguished between empathy and distress as
motivators to help others.6,21 Batson’s
empathy–altruism hypothesis6,31 posits that
these affective states lead to divergent moti-
vations to help others. Those who experience
distress are motivated to reduce their own
aversive state, while those who experience em-
pathy are focused on relieving the aversive state
of another.32 Our physiologic data support
the separation of these two effects in relation
to OT. Interestingly, empathy and distress
levels were also associated with changes in
cortisol. In animal studies, cortisol suppresses
OT release.33 In human studies the findings
are less clear; OT administration suppresses
cortisol induced by social stress,34,35 but
cortisol administration increases plasma OT
levels.36,37 Our study showed that cortisol was
elevated in people who reported experiencing
empathy while it declined in those reporting
distress.
This study indicates that OT is a physio-
logic signature for empathy and modulates two
types of prosocial behaviors: generosity in the
UG and charitable donations. These findings
identify a proximate mechanism that explains
why humans help each other—even at a cost to
themselves.
Conflicts of Interest
The authors declare no conflicts of interest.
References
1. Dovidio, J.F. et al . 1991. The arousal: Cost-reward
model and the process of intervention. In Review of
Personality and Social Psychology: Prosocial Behavior, Vol.
2. M.S. Clark, Ed.: 86–118. Sage. Newbury Park,
CA.
2. de Waal, F.B.M. 2008. Putting the altruism back into
altruism: The evolution of empathy. Annu. Rev. Psychol.
59: 279–300.

Barraza & Zak: Empathy and Oxytocin Release
189
3. Preston, S.D. & F.B.M. de Waal. 2002. Empathy: Its
ultimate and proximate bases. Behav. Brain Sci. 25:
1–72.
4. Aristotle. 2000. T he Nicomachean Ethics. Cambridge
University Press. Cambridge, UK.
5. Smith, A. 2006 [1759]. A Theory of Moral Sentiments.
Dover Publications. Mineola, NY.
6. Batson, C.D. 1991. The Altruism Question: Toward a
Social-Psychological Answer. Lawrence Erlbaum. Hills-
dale, NY.
7. Dovidio, J.F. et al . 2006. The Social Psychology of Prosocial
Behavior. Lawrence Erlbaum. Hillsdale, NY.
8. Brooks, A.C. 2006. Who Really Cares. Basic Books.
New York, NY.
9. Eisenberg, N. & P.A. Miller. 1987. The relation of
empathy to prosocial and related behaviors. Psychol.
Bull. 101: 91–119.
10. Sofroniew, M.V. 1983. Vasopressin and oxytocin in
the mammalian brain and spinal cord. Trends. Neurosci.
6: 467–472.
11. Insel, T.R. 1997. A neurobiological basis of social
attachment. Am. J. Psychiatry 154: 726–735.
12. Insel, T.R. & L.J. Young. 2001. The neurobiology of
attachment. Nat. Rev. Neurosci. 2: 129–136.
13. Carter, C. 2003. Developmental consequences of
oxytocin. Physiol. Behav. 79: 383–397.
14. Keverne, E.B. & Curley, J.P. 2004. Vasopressin, oxy-
tocin and social behaviour. Curr. Opin. Neurobiol. 14:
777–783.
15. Kosfeld, M. et al. 2005. Oxytocin increases trust in
humans. Nature 435: 673–676.
16. Zak, P.J. 2005. Trust: A temporary human attach-
ment facilitated by oxytocin. Behav. Brain Sci. 28: 368–
369.
17. Zak, P.J., R. Kurzban & W.T. Matzner. 2004. The
neurobiology of trust. Ann. N. Y. Acad. Sci. 1032: 224–
227.
18. Zak, P.J., R. Kurzban & W.T. Matzner. 2005. Oxy-
tocin is associated with human trustworthiness. Horm.
Behav. 48: 522–527.
19. Zak, P.J., A. Stanton & S. Ahmadi. 2007. Oxytocin
increases generosity in humans. PLoS ONE 2: e1128.
20. Morhenn, V.B. et al. 2008. Monetary sacrifice among
strangers is mediated by endogenous oxytocin release
after physical contact. Evol. Hum. Behav. 29: 375–
383.
21. Davis, M.H. 1996. Empathy: A Social Psychological Ap-
proach. Westview Press. Boulder, CO.
22. G¨
uth, W., R. Schmittberger & B. Schwarze. 1982.
An experimental analysis of ultimatum bargaining.
J. Econ. Behav. Organ. 3: 367–388.
23. Camerer, C. 2003. Behavioral Game Theory: Experi-
ments in Strategic Interaction. Princeton University Press.
Princeton, NJ.
24. Davis, M.H. 1983. Measuring individual differences
in empathy: Evidence for a multidimensional ap-
proach. J. Pers. Soc. Psychol. 44: 113–126.
25. Larsen R.J. & E. Diener. 1987. Affect intensity as an
individual difference characteristic: A review. J. Res.
Pers. 21: 1–39.
26. John, O.P., E.M. Donahue & R.L. Kentle. 1991. The
Big Five Inventory- Versions 4a and 54. University of Cal-
ifornia, Berkeley, Institute of Personality and Social
Research. Berkeley, CA.
27. Nowack, K.M. 1990. Initial development of an in-
ventory to assess stress and health risk. Am. J. Health
Promot. 4: 173–180.
28. Silvers, J.A. & J. Haidt. 2008. Moral elevation can
induce nursing. Emotion 8: 291–295.
29. Meyer, C. et al. 1987. Relationship between oxytocin
release and amplitude of oxytocin cell neurosecretory
bursts during suckling in the rat. J. Endocrinol. 114:
263–270.
30. McCarthy, M.M. 1995. Estrogen modulation of oxy-
tocin and its relation to behavior. In Oxytocin: Cellular
and Molecular Approaches in Medicine and Research: 235–
245.
31. Batson, C.D. & K.C. Oleson. 1991. Current status
of the empathy-altruism hypothesis. In Review of Per-
sonality and Social Psychology: Prosocial Behavior, Vol. 2.
M.S. Clark, Ed.: 62–85. Sage. Newbury Park, CA.
32. Batson, C.D. 1987. Prosocial motivation: Is it ever
truly altruistic? In Advances in Experimental Social Psy-
chology, Vol. 20. L. Berkowitz, Ed.: 65–122.
33. Carter, C.S. 1998. Neuroendocrine perspectives on
social attachment and love. Psychoneuroendocrin 23:
779–818.
34. Ditzen, B. et al. 2008. Intranasal oxytocin increases
positive communication and reduces cortisol levels
during couple conflict. Biol. Psychiatry. in press.
35. Heinrichs, M. et al. 2003. Social support and oxytocin
interact to suppress cortisol and subjective responses
to psychosocial stress. Biol. Psychiatry 54: 1389–
1398.
36. Tops, M., J.M. van Peer & J. Korf. 2007. Indi-
vidual differences in emotional expressivity predict
oxytocin responses to cortisol administration: Rel-
evance to breast cancer? Biol. Psychol. 75: 119–
123.
37. Tops, M. et al. 2007. Anxiety, cortisol, and attach-
ment predict plasma oxytocin. Psychophysiology 44:
444–449.