ArticlePDF Available

Abstract and Figures

The last decades of research provided overwhelming evidence that compassion fosters a vast range of behaviors toward reducing suffering of others. In this regard, compassion has been described as a prosocial tendency par excellence, fostering helping behavior across a variety of social situations. With the present contribution, we apply a differentiated perspective on compassion. Building on just deserts theory, we argue that when other individuals suffer from unjust actions, compassion for the suffering individuals can foster harmful tendencies towards those who caused the suffering (i.e., third-party punishment). In Studies 1a to 1f, we examined a rich variety of situations in which unjust suffering occurs (i.e., terrorist attacks, sexual assaults, rape of children, and war) and documented a positive relation between compassion for suffering victims and punishment inclinations towards those who caused the suffering. Applying an experimental approach using various paradigms in Studies 2 to 6, compassion was shown to increase third-party punishment. Additional analyses revealed that (a) this increase occurs because compassion intensified moral outrage, which in turn predicted third-party punishment (Studies 2 to 6), and (b) compassion only fosters third-party punishment when suffering was caused by high (vs. low) unjust acts (Study 5). Overall, the present research reveals compassion in a different light in that harmful consequences of compassion are considered. Implications are discussed from a perspective of basic research on compassion and third-party punishment as well as from a societal perspective.
Content may be subject to copyright.
Journal of Personality and Social
Psychology
Compassion Magnifies Third-Party Punishment
Stefan Pfattheicher, Claudia Sassenrath, and Johannes Keller
Online First Publication, April 4, 2019. http://dx.doi.org/10.1037/pspi0000165
CITATION
Pfattheicher, S., Sassenrath, C., & Keller, J. (2019, April 4). Compassion Magnifies Third-Party
Punishment. Journal of Personality and Social Psychology. Advance online publication.
http://dx.doi.org/10.1037/pspi0000165
Compassion Magnifies Third-Party Punishment
Stefan Pfattheicher
Aarhus University and Ulm University
Claudia Sassenrath and Johannes Keller
Ulm University
The last decades of research have provided overwhelming evidence that compassion fosters a vast range
of behaviors toward reducing suffering of others. In this regard, compassion has been described as a
prosocial tendency par excellence, fostering helping behavior across a variety of social situations. With
the present contribution, we apply a differentiated perspective on compassion. Building on just deserts
theory, we argue that when other individuals suffer from unjust actions, compassion for the suffering
individuals can foster harmful tendencies toward those who caused the suffering (i.e., third-party
punishment). In Studies 1a to 1f, we examined a rich variety of situations in which unjust suffering occurs
(i.e., terrorist attacks, sexual assaults, rape of children, and war) and documented a positive relation
between compassion for suffering victims and punishment inclinations toward those who caused the
suffering. Applying an experimental approach using various paradigms in Studies 2 through 6, compas-
sion was shown to increase third-party punishment. Additional analyses revealed that (a) this increase
occurs because compassion intensified moral outrage, which in turn predicted third-party punishment
(Studies 2 to 6), and (b) compassion only fosters third-party punishment when suffering was caused by
high (vs. low) unjust acts (Study 5). Overall, the present research discusses compassion in a different light
in that harmful consequences of compassion are considered. Implications are discussed from a perspec-
tive of basic research on compassion and third-party punishment as well as from a societal perspective.
Keywords: compassion, empathy, harm, injustice, third-party punishment
Remember the Paris attacks of November 2015, when several
mass shootings and a suicide bombing took place in the heart of
the city. Many were hurt and killed, and substantial suffering
resulted. Strong reactions were elicited based on this suffering,
ranging from compassion for the injured to retaliatory tendencies
toward those who caused the harm. In this contribution, the main
idea put forward is that compassion for suffering individuals can
motivate harmful tendencies toward those responsible for the
suffering. Specifically, we argue that when suffering is caused by
profound unjust behavior, compassion fosters punishment toward
the perpetrator(s).
The present work has two central aims. The first is to apply a
differentiated perspective on compassion. Compassion has been
conceptualized as a prosocial tendency par excellence, fostering
helping behavior across a variety of social situations (Dalai Lama
& Ekman, 2008; Goetz, Keltner, & Simon-Thomas, 2010). In this
contribution, it is argued and shown that in situations where
individuals do not have the option to help a victim, compassion can
foster harmful consequences for other individuals (i.e., third-party
punishment of suffering-causing individuals). As such, the present
work reveals compassion in a different light. This perspective from
basic research is complemented by a societal perspective. Specif-
ically, the second aim is to understand and to document the
interpersonal consequences that result from the omnipresent con-
frontation with suffering. Indeed, people today are frequently
exposed to suffering, notably when terrorist attacks happen, when
viewing media reports about the war in Syria, or when perpetrators
violate fundamental moral standards and cause suffering (e.g., the
New Year’s Eve sexual assaults in Cologne, Germany in 2015). In
this regard, we tested the idea of compassion fostering third-party
punishment in a variety of contexts of high societal relevance (i.e.,
terrorist attacks, sexual assaults, rape of children, and war). In the
following, the current state of research on compassion is outlined.
Subsequently, the exploratory framework is presented about why
compassion fosters third-party punishment.
The Essence of Compassion
The core theme of compassion refers to being moved by anoth-
er’s suffering and possessing concerned feelings that are elicited in
response to this suffering (Haidt, 2003; Lazarus, 1991; for a
compelling review on compassion, see Goetz et al., 2010). As a
consequence of compassion, actions are elicited that aim to reduce
the negative state of the suffering individual, especially when the
suffering other deserves help and one has the resources to help
(Goetz et al., 2010; Haidt, 2003; Ministero, Poulin, Buffone, &
DeLury, 2018). In this regard, it is shown that compassion leads to
prosocial actions such as helping and supporting children in need,
Stefan Pfattheicher, Department of Psychology and Behavioral Sciences,
Aarhus University, and Department of Social Psychology, Ulm University;
Claudia Sassenrath and Johannes Keller, Department of Social Psychology,
Ulm University.
This research was supported by a grant from the Baden-Württemberg
Foundation to Stefan Pfattheicher.
Correspondence concerning this article should be addressed to Stefan
Pfattheicher, Department of Psychology and Behavioral Sciences, Aarhus
University, 8000 Aarhus C, Denmark. E-mail: sp@psy.au.dk
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Journal of Personality and Social Psychology:
Interpersonal Relations and Group Processes
© 2019 American Psychological Association 2019, Vol. 1, No. 999, 000
0022-3514/19/$12.00 http://dx.doi.org/10.1037/pspi0000165
1
aiding people who suffer from a catastrophe or a personal loss,
assisting homeless or poor people, and being supportive toward a
romantic partner (Ashar et al., 2016; Batson et al., 1989, 1997;
Cialdini, Brown, Lewis, Luce, & Neuberg, 1997; Crocker, Canev-
ello, & Lewis, 2017; Eisenberg, McCreath, & Ahn, 1988; Eisen-
berg et al., 1994).
1
One main motivational tendency emerging from compassion is
the goal of relieving pain from suffering others (Goetz et al.,
2010). Building on the notion that relieving pain is one central
element of morality (Haidt, 2003; Schein & Gray, 2018), compas-
sion is considered a moral emotion (Cameron & Payne, 2012;
Haidt, 2003; Nussbaum, 1996; Zaki, 2018). Beyond that, compas-
sion entails a specific sensitivity regarding negative social infor-
mation, for instance, regarding suffering caused by unjust behavior
(Goetz et al., 2010; Keller & Pfattheicher, 2013). When compas-
sionate individuals observe unjust suffering, they “moralize” the
event and then possess an intensified moral concern in this regard
(Bekkers & Ottoni-Wilhelm, 2016; Horberg, Oveis, & Keltner,
2011). As a consequence of their moral concern, compassionate
individuals’ resulting actions aim to reestablish justice and moral-
ity, for instance by means of helping a suffering victim (Goetz et
al., 2010).
In the following, we outline the rationale about why compassion
fosters third-party punishment. Third-party punishment in this
work is conceptualized broadly as any behavior that includes
negative sanctions and consequences or the support of negative
sanctions and consequences targeted at a third party who has
caused harm. These negative sanctions and consequences include
a broad range of tendencies such as physical or psychological harm
or inflicting financial costs for the third party. Examples of third-
party punishment are support for retribution of terror attacks,
support for the death penalty of rapists, and third-party punishment
of unfair individuals in economic games. As such, the theoretical
and empirical basis of the present work is not restricted to a single
and specific manifestation of third-party punishment but rather
makes claims regarding a greater range of third-party punishment
tendencies.
Why Compassion Fosters Third-Party Punishment:
Just Deserts Theory
We build on a prominent theory in the domain of criminology
and morality, that is, just deserts theory, to argue why compas-
sionate individuals are motivated to punish those who caused
suffering. Just deserts theory deals with the basic question of why
people punish and why they want violators of morality to be
harmed. Its basic foundation goes back to Immanuel Kant (1785/
1998), who argued that humans act as rational agents; conse-
quently, individuals must consider the consequences of their ac-
tions, and therefore would deserve punishment in case their acts
violated justice (cf. Barton, 2004; Darley, Carlsmith, & Robinson,
2000). The essence of just deserts theory is that harm targeted at an
offender should be proportionate to the strength of the moral
violation (Von Hirsch, 1976); its justification lies in the reestab-
lishment of justice and in righting morally wrong past actions.
Empirically, Carlsmith, Darley, and Robinson (2002) showed that
individuals follow the just deserts approach when they observe a
moral violation. They documented that judging an action as mor-
ally wrong led to moral outrage, that is, negative affect (anger,
contempt) provoked by the perception of a moral violation (Batson
et al., 2007; Hoffman, 1990; Montada & Schneider, 1989; Tetlock,
2002). Moral outrage, in turn, led to the inclination to harm the
offender (Carlsmith et al., 2002; Eriksson, Strimling, Andersson,
& Lindholm, 2017).
It is important to note that just deserts theory includes the idea
that contextual factors pronounce or attenuate individuals’ percep-
tion of a moral violation and consequently their moral outrage and
punishment inclinations (Carlsmith et al., 2002; Robinson & Dar-
ley, 1995). This implies that the more an action is judged as
morally wrong (because of certain contextual factors), the more
moral outrage emerges, leading to stronger tendencies to punish
the perpetrator (Fetchenhauer & Huang, 2004). Carlsmith and
colleagues (2002) found empirical support for this idea. In a series
of studies, they showed that higher seriousness of a morally wrong
action elicits more moral outrage, and as a response, stronger
motivation and support to punish the perpetrator (see also Fiske &
Tetlock, 1997; Tetlock, Kristel, Elson, Green, & Lerner, 2000).
We argue that compassion reflects such a contextual factor that
intensifies individuals’ moral outrage and punishment inclinations.
Compassion makes one sensitive toward the cared-for (Bekkers &
Ottoni-Wilhelm, 2016; Goetz et al., 2010; Haidt, 2003). In such a
state, unjust harm toward the cared-for is then perceived as more
profound, intensifying moral outrage (see Thulin & Bicchieri,
2016, and Rothschild & Keefer, 2018, for initial empirical evi-
dence). Moral outrage, in turn, should lead to the tendency to
punish the perpetrator to reestablish justice. In other words, be-
cause one cares for and feels compassion for a suffering person,
unjust harm done to this person is perceived as more morally
profound, resulting in intensified moral outrage and magnified
third-party punishment. We therefore expect that compassion for
individuals suffering from unjust behavior magnifies third-party
punishment because compassion intensifies moral outrage, which
in turn increases punishment inclination toward those who caused
unjust suffering. To summarize, we argue that compassion reflects
an intensifier of individuals’ perception of a moral violation and
the downstream consequences (moral outrage and third-party pun-
ishment); therefore, we expect that, given that suffering is caused
by unjust behavior, compassion intensifies moral outrage about
unjust suffering, which in turn increases third-party punishment
toward those responsible for the suffering.
In the studies of the present contribution, we first tested whether
compassion is positively correlated to third-party punishment
across a variety of situations in which suffering was caused by
unjust behavior (Studies 1a–1f). In Studies 2 through 6 we applied
an experimental approach and tested whether compassion in-
creases third-party punishment. We further showed that (a) this
increase occurs because compassion intensifies moral outrage,
which in turn predicted third-party punishment (Studies 2– 6), and
1
It is noteworthy that other researchers use “empathic concern” instead
of compassion to describe concerned feelings for suffering others. In fact,
empathic concern is typically used interchangeably with compassion (cf.
Batson, 2009; Condon & DeSteno, 2017; Goetz et al., 2010; Singer &
Klimecki, 2014; Singer & Steinbeis, 2009; Stellar & Keltner, 2014; Stellar
et al., 2015). As the term ‘empathy’ creates misunderstanding (Batson,
2009; DeSteno, 2015), the term compassion is used to make explicit that
the present research focuses on concerned feelings for suffering others, not
for others in general.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
2PFATTHEICHER, SASSENRATH, AND KELLER
(b) compassion fosters third-party punishment only when the ac-
tion that caused the suffering was unjust, not when the action was
not unjust (Study 5).
The Value of the Present Research
The present work makes important contributions to the litera-
ture, both from the perspective of research on compassion as well
as from the perspective of research on third-party punishment:
First, the present contribution shapes our basic understanding of
compassion in that harmful consequences of compassion are con-
sidered. So far, research on compassion has focused clearly on its
ability to reduce suffering (Goetz et al., 2010). By showing that
compassion actually motivates tendencies that entail harmful con-
sequences for other individuals (i.e., third-party punishment), we
point to a so far overlooked consequence of compassion.
In this regard, the present research is in line with previous work
from Buffone and Poulin (2014) and Keller and Pfattheicher
(2013) in showing negative interpersonal tendencies resulting from
compassion. Buffone and Poulin (2014) document that compas-
sion for a distressed person increases harming behavior toward a
third individual who faced an upcoming competition with the
distressed person. In this way, compassionate individuals helped
the suffering (distressed) other by means of handicapping the
nonsuffering individual. Thus, compassion leads to harmful be-
havior that directly helps a suffering person. Hence, the work by
Buffone and Poulin (2014) fits well in the overall picture that the
main tendency that results from compassion is to help a suffering
other. The present work goes one crucial step further, arguing and
showing that in situations in which it is difficult for compassionate
individuals to help a victim, they are motivated to harm the
suffering-causing individual. As such, we show that compassion is
not only about helping a victim. It is more than that; it is also about
punishing a perpetrator for causing harm to a person someone
cares for. Of course, punishment of a perpetrator can carry an
indirect helping function; past research has shown that (third-
party) punishment of unfair individuals reduces future unfair be-
havior to the same and other victims (Fehr & Fischbacher, 2004;
Rand & Nowak, 2013). Our main point here is that compassion not
only fosters actions that improve the state of a suffering victim; it
can also foster actions that impair another individual’s state.
Further relating the present contribution to other relevant work,
Keller and Pfattheicher (2013) showed that compassion for a
suffering individual who was harmed by others leads to a hostile
worldview that humans are bad in general given that individuals
were prevention-focused. As an extension of Keller and Pfat-
theicher (2013), the present work shows that compassion can foster
negative consequences that really affect other individuals, not only
by increasing a general negative belief about humans.
We can extend existing literature on compassion in an additional
meaningful way. So far, research on compassion has focused
mainly on compassion emerging from suffering that is not caused
by third parties (cf. Goetz et al., 2010). We document that harmful
tendencies (i.e., punishment) result from compassion when suffer-
ing is caused by third parties, given that the suffering-causing
actions are unjust. In the few existing studies that include consid-
eration of suffering caused by third parties, it is argued that anger
emerges when seeing that a person for whom one cares is unde-
servedly harmed (termed “empathic anger”; Batson et al., 2007;
Hoffman, 2000; Vitaglione & Barnett, 2003). This anger fosters
the motivation to punish the one who is responsible for the ob-
served suffering (Batson, 2009). This perspective fits the line of
argument of the present work. We argue that being in a compas-
sionate state (what Batson et al., 2007, call “care about the welfare
of a suffering other”) intensifies the perception of injustice (i.e.,
intensified moral outrage). We suggest that moral outrage and
empathic anger on the basis of undeserved (i.e., unjust) harm are
closely related, as moral outrage includes anger (besides contempt;
Tetlock, 2002; Tetlock et al., 2000). In this regard, we aim to
extend research on empathic anger as our research suggests why
empathic anger emerges (because of compassion for someone who
suffers from injustice).
The present work also advances understanding of the emergence
of third-party punishment. Past research shows that the moral
emotion of anger drives third-party punishment of individuals that
intentionally behaved unfairly toward an interaction partner (Ne-
lissen & Zeelenberg, 2009; see also Jordan, McAuliffe, & Rand,
2016). Approaches that explain the emergence of third-party pun-
ishment assume that unfair behavior reflects the violation of a
fairness norm; this perception of a norm violation elicits anger
which in turn motivates punishment of the unfair third party (Falk,
Fehr, & Fischbacher, 2003; Fehr & Fischbacher, 2004; Singer &
Steinbeis, 2009). We rely on this research, arguing that moral
outrage (which includes anger about unfair behavior) motivates
third-party punishment. Moreover, we provide an additional piece
to the puzzle on how third-party punishment emerges. Here, we
build on just deserts theory (as outlined in detail above). Just
deserts theory includes the idea that contextual factors pronounce
or attenuate individuals’ perception of a moral violation and con-
sequently their moral outrage and punishment inclinations (Carl-
smith et al., 2002; Robinson & Darley, 1995). We argue that
compassion reflects such a contextual factor in that it intensifies
the perception of unfairness (i.e., intensifies moral outrage), lead-
ing to increased third-party punishment. In sum, with the empirical
research program reported in the following, we aim to increase
knowledge and understanding about so far overlooked conse-
quences of compassion and to point to a crucial factor in the
emergence of third-party punishment.
Studies 1a–1f
The aim of the first set of studies was to put the idea of
compassion-based third-party punishment to a correlational test.
Moreover, we aimed at documenting a positive relation of com-
passion and third-party punishment in many different situations of
high societal relevance. To do so, we assessed compassion for
suffering individuals from various terrorist attacks (Paris 2015;
Istanbul 2015; Brussels 2016) in Studies 1a–1c; compassion for
suffering individuals from the war in Syria in Study 1d; compas-
sion for women suffering from sexual assaults perpetrated by
groups of young foreign men on New Year’s Eve, 2015 in Co-
logne, Germany, in Study 1e; and compassion for three children
suffering from rape by a man in Study 1f. In each of the studies,
we also assessed third-party punishment inclinations toward those
responsible for the suffering, that is, how much participants agree
to inflict harm on the suffering causing individuals (e.g., support of
retributive actions for terror attacks and support for death penalty
for the rapist).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
3
COMPASSION MAGNIFIES THIRD-PARTY PUNISHMENT
Method
Research ethics statement. The studies reported in the pres-
ent contribution were conducted in full accordance with the Ethical
Guidelines of the German Association of Psychologists (DGPs)
and the American Psychological Association (APA). Institutional
review boards or committees are not mandatory at Danish and
German universities. Data, all items, and dropout rates of all
studies presented in this contribution are available on the Open
Science Framework (see https://osf.io/fqnjm/).
Procedure. Each study was conducted online. Studies 1a–1d
were conducted via Amazon Mechanical Turk (cf. Buhrmester,
Kwang, & Gosling, 2011); Study 1e was distributed via a mailing
list of German individuals who are regularly informed about new
psychological online studies; and Study 1f was conducted via an
online platform providing access to German individuals who are
willing to participate in studies in exchange for a small amount of
money. All studies followed the same procedure. First, participants
read an introductory statement that reminded them of an event that
included suffering. For instance, the introductory statement in the
context of the Paris attacks read, “The following statements refer
to the recent terrorist attacks in Paris, France.”
After the introductory sentence(s), participants completed three
items assessing compassion toward the suffering individuals (the
items are reported below). Then they responded to one filler item
(e.g., “How I dress is important to me.”). The filler item was not
significantly related to either compassion or third-party punish-
ment in any of the studies (with a single exception of minor
empirical relevance in Study 1f). After the filler item, participants
completed an attention check item (i.e., “This is an attention check
item. Please answer: I completely agree.”). Finally, they responded
to three items assessing third-party punishment toward those re-
sponsible for the suffering. Cronbach’s s, mean values, and
standard deviations of all measures are provided in Table 1.
Compassion. To assess compassion, three items were formu-
lated based on the empathic concern scale by Davis (1983). The
compassion items in the context of the Paris attacks read, “I feel
compassion for the people suffering from the attacks in Paris”; “I
am quite touched by the people suffering from the attacks in
Paris”; and “I have concerned feelings for the people suffering
from the attacks in Paris.” Responses were given on a 7-point
Likert scale ranging from strongly disagree to strongly agree in
Studies 1a–1e and on a 5-point Likert scale in Study 1f. The items
assessing compassion in the other contexts (terrorist attacks in
Instanbul and Brussels, the war in Syria, the sexual assaults in
Cologne, and sexual assaults of children) were formulated exactly
as the items of the Paris attacks study, but the context differed
(e.g., “I feel compassion for people suffering in Syria” or “I feel
compassion for the children.”).
Third-party punishment. To assess third-party punishment,
three items were formulated in each study. The items in the context
of the Paris attacks read, “I support attacks on those who caused
the suffering of the attacks in Paris”; “Those responsible for the
attacks in Paris should be attacked”; and “If I could, I would attack
those who caused the suffering of the attacks in Paris.” The items
assessing third-party punishment in the other contexts were for-
mulated exactly as the items of the Paris attacks study, adapted to
the respective context (e.g., “If I could decide I would attack those
who cause suffering in Syria,” or “The rapist deserves the death
penalty”).
Participants. Using G
Power (Faul, Erdfelder, Buchner, &
Lang, 2009), a power analysis was conducted for a two-tailed test
to detect at least a small to medium effect (r.25); power was set
to .80 (Cohen, 1992). This power analysis revealed a required
sample size of N120 to detect a significant effect (alpha level
of .05) given there is a true effect. However, because we did not
know the “true effect,” we oversampled the studies when possible
(in terms of financial opportunities and accessibility to partici-
pants). Only in Study 1e, in which sample size (N72) was
determined by a restricted pool of participants, did we fall below
the threshold set by the power analysis. See Table 1 for how many
participants completed each study as well as for means and stan-
dard deviations regarding age and percentage of sexes. Sample
sizes in Table 1 refer to how many participants completed all items
and responded correctly to the attention check item.
Results and Discussion
An overview concerning the assessed variables is given in Table
1. It should be noted that the observed means in reported compas-
sion are relatively high, suggesting that the applied contexts in
which suffering occurred elicited compassion to a substantial
extent. Yet, as indicated by the relatively large standard deviations,
some individuals in the studies did not report strong compassionate
tendencies. Regarding third-party punishment, means differed;
they were higher when terrorist attacks happened compared with
when persons suffered from sexual assaults. The large standard
Table 1
Overview Concerning the Characteristics and Variables of Studies 1a–1f as well as Correlations Between Compassion and
Third-Party Punishment
Study Context
Sample size Age Sex Compassion
Third-party
punishment Correlation
NMean (SD) % women Mean (SD)Mean (SD)rpvalue
Study 1a Paris attacks 188 35.6 (10.9) 47.3% .73 6.31 (0.95) .90 4.99 (1.78) .22 .002
Study 1b Istanbul attacks 203 36.1 (12.2) 51.2% .86 6.07 (1.12) .87 4.73 (1.84) .23 .001
Study 1c Brussels attacks 238 36.4 (12.1) 48.7% .90 6.13 (1.16) .92 4.65 (1.94) .24 .001
Study 1d War in Syria 170 36.1 (12.0) 57.1% .94 5.40 (1.70) .91 4.61 (1.54) .20 .010
Study 1e Sexual assaults Cologne 72 27.4 (11.1) 72.9% .90 5.13 (1.35) .87 2.03 (1.37) .34 .003
Study 1f Rape of children 198 37.2 (11.6) 39.9% .81 4.77 (0.44) .97 2.96 (1.56) .18 .014
Note. Responses were given on a 7-point Likert scale in Studies 1a–1e and on a 5-point Likert scale in Study 1f.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
4PFATTHEICHER, SASSENRATH, AND KELLER
deviations in third-party punishment indicate that participants dif-
fer to a large degree in their punishment inclinations.
Most important for the present contribution are the correlations
between compassion and third-party punishment. In all six studies,
across a variety of contexts of high societal relevance, a positive
relation was found, indicating that higher compassion is accom-
panied by higher third-party punishment inclinations toward those
who caused the suffering. All relations remained robust in terms of
effect size and significance levels when controlling for sex and
age. The lowest correlation was found in the context of the death
penalty for a man who raped three children, potentially due to the
low variance on the compassion measure. The highest correlation
was found in the smallest sample; accordingly, the confidence
interval of this correlation is larger than in the other samples.
Overall, an internal meta-analysis including all samples, using
fixed effects in which the mean effect size was weighted by sample
size (cf. Goh, Hall, & Rosenthal, 2016), revealed a positive rela-
tion of small to medium size between compassion and third-party
punishment (Mr.22, SE 0.03, Z7.39, p.001, two-
tailed).
We acknowledge that these studies were of correlational nature,
thus we do not know whether compassion fosters third-party
punishment (in a causal sense). In addition, in these studies, we
have not tested whether other emotions (e.g., anger) might be
responsible for the observed correlations. Yet if we had included
other emotions, the studies would still be of correlational nature.
That is, if we really want to show that compassion is driving the
effects instead of other emotions, we need experimental studies in
which only compassion is manipulated. This was done in the
experimental studies that follow (Studies 2 to 6).
Studies 2a and 2b
Method
Procedure. Study 2a was conducted in the laboratory at a
German University, and Study 2b was conducted via Amazon
Mechanical Turk (cf. Buhrmester et al., 2011). In both studies,
compassion was manipulated first, then a manipulation check was
implemented to assess state compassion and other emotions. After
that, participants learned that suffering was caused by unjust
behavior. This aspect is important, because in this way we manip-
ulated compassion prior to imparting (and thus independent from)
the information about unjust behavior and injustice-related emo-
tions. Finally, individuals in both studies report on their third-party
punishment inclinations.
Compassion manipulation. Participants read a short text
about George (70 years of age) who reports:
Last week I broke my shoulder. As you can imagine, at my age you
don’t recover quickly, it takes longer than usual. Every time I move
my body it hurts...although I frequently take aspirins. I do go for
short walks though. But walking is difficult. After that I’m so ex-
hausted that I have to lie down.
Before reading the text, a standard procedure was used to
manipulate compassion (Batson et al., 1997; Batson, Chang, Orr,
& Rowland, 2002). Participants were randomly assigned either to
a high- or a low-compassion condition. In the high-compassion
condition, participants read these compassion-promoting instruc-
tions for how they should approach the text before reading the text
(cf. Batson et al., 1997, Study 3; Batson et al., 2002): “While
reading, imagine how the described person feels about what has
happened and how it has affected his life. Try to feel the full
impact of what he has been through and how he feels as a result.”
In the low-compassion condition, participants read, “While read-
ing, take an objective perspective toward what is described. Try
not to get caught up in how the described person feels; just remain
objective and detached.” Thus, we had a low versus high-
compassion condition in the studies.
Manipulation check. State compassion (Study 2a ␣⫽.83,
Study 2b ␣⫽.91) was assessed with the items “moved,” “sym-
pathetic,” “concerned,” and “compassionate” (cf. Batson et al.,
1997, 2002). Please note that we refer to this index as “state
compassion” in the following text. In addition, the six basic
emotions of anger, sadness, fear, surprise, happiness, and disgust
(Ekman, 1992) were assessed along with the moral emotion of
contempt (Haidt, 2003), as it has been argued that moral outrage is
reflected by the combination of anger and contempt (Tetlock,
2002; Tetlock et al., 2000).
2
Responses were given on a 7-point
Likert scale ranging from not at all to extremely. Anger and
contempt were highly correlated (Study 2a r.49, p.001,
Study 2b r.61, p.001) and averaged to compute an index of
state outrage. We refer to this index as “state outrage” in the
following text.
Third-party punishment. Before individuals reported on
their punishment inclinations, they learned that George was inten-
tionally harmed, as intentional harm is one main elicitor of injus-
tice perception (e.g., Mikula, 1993). Specifically, participants read:
“George continues: Maybe you are interested in how I broke my
shoulder. It was an accident. A young man deliberately bumped
into me. Then I fell. This is how it happened.” After reading this
text, third-party punishment inclinations toward the young man
were assessed. The items (Study 2a ␣⫽.73, Study 2b ␣⫽.93)
read, “The young man deserves hard treatment”; “I want to teach
the young man a lesson”; and “I want the young man to be severely
punished.”
Moral outrage. We have argued that compassion for unjust
suffering motivates third-party punishment because compassion
intensifies moral outrage, which in turn fosters third-party punish-
ment. In Study 2b (but not in Study 2a) we assessed moral outrage
right before punishment. The items (␣⫽.90) read, “I am angry at
the young man”; “I am outraged by the young man’s behavior”;
and “The behavior of the young man is absolutely immoral.”
Please note that we refer to this index as “moral outrage” (as
opposed to the index of “state outrage” described above) in the
following text.
Participants. Using G
Power (Faul et al., 2009), a power
analysis was conducted for a two-tailed test to detect at least a
medium effect (d.50); power was set to .80 (Cohen, 1992). This
power analysis revealed a required sample size of N128 to
detect a significant effect (alpha level of .05) given there is a true
2
We acknowledge that there is disagreement about whether or not to
include disgust in a measure of moral outrage (for a discussion see Gervais
& Fessler, 2017). Our a priori approach was to follow Tetlock and
colleagues (2000; Tetlock, 2002), arguing that moral outrage is reflected by
the combination of anger and contempt. We note, however, that the results
do not change when disgust is included in the moral outrage index.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
5
COMPASSION MAGNIFIES THIRD-PARTY PUNISHMENT
effect. In Study 2a, we obtained 105 German participants (53.3%
women; M
age
21.0, SD
age
3.4). The set threshold determined
by power analysis was not reached because in the laboratory the
time the study could be conducted was restricted. In Study 2b, 171
U.S.-American individuals completed the study (64.3% women;
M
age
39.1, SD
age
13.4). Dropouts were not significantly
uneven by condition in any of the experimental studies reported in
the present contribution (cf. Zhou & Fishbach, 2016).
Results and Discussion
Manipulation check. First, we examined in Study 2a whether
the manipulation increased (only) state compassion. This was the
case. Analysis revealed a significant difference, t(103) 3.05, p
.01; Cohen’s d0.59 in state compassion between the low-
compassion condition (M3.74, SD 1.39) and the high-
compassion condition (M4.52, SD 1.24). Next, there was no
significant difference in state outrage, t(103) 0.34, p.74;
Cohen’s d0.07 between the low-compassion condition (M
1.59, SD 0.99) and the high-compassion condition (M1.52,
SD 1.05). The low means in state outrage suggest that state
outrage did not play a role in this manipulation. The two conditions
also did not differ regarding the other assessed emotions (all ts
|1|), except for sadness which was (nonsignificantly) increased in
the high-compassion condition, t(103) 1.46, p.15; Cohen’s
d0.28. The same pattern in terms of significance levels, effect
sizes, and direction of effects could be found in Study 2b. State
compassion was significantly increased in the high-compassion
condition, t(169) 3.22, p.01; Cohen’s d0.50 but state
outrage was not, t(169) 1.49, p.14; Cohen’s d0.23. These
results, in both studies, indicate a successful and specific induction
of compassion in both studies, with sadness also increased. We
note that one can argue that sadness must be included in the state
compassion index (e.g., Batson et al., 1989; Sassenrath, Pfat-
theicher, & Keller, 2017).
3
Third-party punishment. In Study 2a, analysis revealed a
significant difference, t(103) 2.10, p.04; Cohen’s d0.41,
in third-party punishment between the low-compassion condition
(M3.66, SD 1.48) and the high-compassion condition (M
4.24, SD 1.35). This finding was replicated in Study 2b; analysis
revealed a significant difference, t(169) 2.74, p.01; Cohen’s
d0.40 in third-party punishment between the low-compassion
condition (M3.24, SD 2.02) and the high-compassion con-
dition (M4.02, SD 1.68).
Mediation analysis in Study 2b. For testing mediation, we
used the PROCESS macro provided by Hayes (2013) and ran
Model 4. First, moral outrage (which was assessed after partici-
pants learned about unjust suffering) was regressed on the exper-
imental conditions. This analysis revealed higher moral outrage in
the high-versus the low-compassion condition (B1.04, SE
0.28, t3.69, p.001), which was in turn positively related to
punishment (B0.84, SE 0.04, t19.73,p.001) when the
experimental conditions were included in the regression model
(B⫽⫺0.09, SE 0.16, t⫽⫺.59, p.55). The indirect effect
was also significant, that is, the bootstrapped 95% confidence
interval excluded zero [0.41; 1.34] (Hayes, 2013). In other words,
moral outrage mediated the effect of experimentally induced com-
passion on third-party punishment in that compassion increased
moral outrage, which in turn predicted third-party punishment.
Additional analyses. In additional analyses, we predicted
third-party punishment with state compassion and state outrage (to
recapitulate, these were assessed right after the manipulation of
compassion but before participants learned about unjust suffering)
together in a simultaneous regression model. In this way, we
further tested whether state compassion is driving the effect of the
applied manipulation on third-party punishment and not directly
induced state outrage. Multiple regression analyses revealed that
state compassion was significantly predictive of third-party pun-
ishment (Study 2a: B0.37, SE 0.10, t3.78, p.001; Study
2b: B0.40, SE 0.09, t4.60, p.001) and relevant in terms
of effect size (Study 2a R
2
.13; Study 2b R
2
.12). In
contrast, state outrage assessed prior to the intentional harm infor-
mation, expectedly, was not significantly related in Study 2a (B
0.04, SE 0.13, t0.33, p.74) although statistically signif-
icant in Study 2b (B0.26, SE 0.13, t2.10, p.04).
However, in both studies, the effect size of state outrage was
negligible (Study 2a R
2
.00; Study 2b R
2
.02), and the
significant effect of Study 2b was not replicated in the following
studies.
Discussion. In Studies 2a and 2b, we induced compassion
without any reference to injustice or a moral violation by a third
party. It is shown that the manipulation induced compassion and
not moral outrage. We then showed that compassion intensified
moral outrage in response to the subsequently presented informa-
tion on intentional harm, which in turn predicted third-party pun-
ishment. Overall, Studies 2a and 2b provide first evidence for the
idea that compassion increases third-party punishment in an online
study and a study conducted in the laboratory, using participants
from the U.S. and Germany.
Additional Studies
We ran additional studies which are, because of redundancy and
methodological shortcomings, not reported in detail in this paper;
detailed information as well as data can be found on the OSF (see
https://osf.io/fqnjm). This package includes one online study in
which we induced compassion for a homeless person who suffered
from unjust action by her husband and a lab study in which we
induced compassion for a 12-year old student who was beaten up
by an older schoolmate. We found significantly increased third-
party punishment inclinations in the high- compared with the
low-compassion condition. However, the two studies suffer from a
main shortcoming. Specifically, we have manipulated compassion
for another individual who suffered from injustice by a third
person; this situation confounded compassion (elicited by suffer-
ing) and moral outrage (elicited by injustice). As such, we could
have directly induced moral outrage in addition to compassion.
The shortcomings are not present in Studies 2a and 2b reported
above and Studies 4 to 5 reported below; in these studies, we
induce compassion without any reference to injustice or a moral
violation by a third party.
3
We note that there were no problems in the present set of studies with
multicollinearity in terms of the assessed emotions when looking at the
correlations and the Variance Inflation Factor (VIF), expect for sadness,
which can arguably be included in the state compassion index (Batson,
2011; Olderbak et al., 2014; Sassenrath et al., 2017).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
6PFATTHEICHER, SASSENRATH, AND KELLER
We also note that we ran an additional study in which we first
manipulated compassion for a woman suffering from AIDS.
Thereafter, participants learned that AIDS was unjustly caused by
a man (the scenario was adapted from Batson et al., 1997). In this
study, we did not find a significant effect of the compassion
manipulation on third-party punishment of the man, t(139) 1.27,
p.21, d.22. There can be many reasons for a nonsignificant
effect (e.g., Simonsohn, 2015): in the present case, it is possible
that the scenario was not realistic, that (a proportion of) partici-
pants blamed the victim (e.g., Lerner, 1980) rather than felt com-
passion, or that there is no true effect. We included this study in the
internal meta-analysis reported below.
In the next studies (Studies 3a–3c), we aimed (a) to conceptually
replicate the basic effect of compassion on third-party punishment,
(b) to replicate the found mediation in Study 2b, and (c) to apply
a paradigm to obtain a behavioral measure of third-party punish-
ment. To do so, we used the classic third-party punishment para-
digm (Fehr & Fischbacher, 2004).
Studies 3a–3c
Method
Procedure. Studies 3a to 3c were conducted online via Am-
azon Mechanical Turk (cf. Buhrmester et al., 2011). Participants
first read the instructions of the third-party punishment paradigm
(Fehr & Fischbacher, 2004). Compassion was then manipulated. In
Studies 3a and 3b we measured state compassion after the manip-
ulation (as a manipulation check); in Study 3c, no manipulation
check was used to get a dependent variable unbiased by the
measurement of the manipulation check. After that, participants
had the opportunity to engage in third-party punishment. In Study
3c, we further had a measure of moral outrage (the mediator) after
the dependent variable; moral outrage was measured after the
dependent variable to get an unbiased and valid measure of third-
party punishment (Spencer, Zanna, & Fong, 2005). Please note
that in Studies 2, 5, and 6, the mediator was measured in-between
the independent and the dependent variable.
Third-party punishment paradigm. Participants were first
told that two other participants (called Player A and Player B) have
worked together with equal effort to fulfill a task. In this task, they
had together earned $10. Next, participants learned that Player A
was given full control of the distribution of these $10. They were
told that Person A took $7 while giving Person B $3, thus harming
Person B financially. Player B had no choice but to accept the
decision of Player A. As such, participants observed an unjust
distribution of money (Eriksson et al., 2017; Nelissen & Zeelen-
berg, 2009).
4
To assess punishment behavior, participants were given $10 as
endowment and provided with the possibility to punish Person A
by reducing Person A’s income out of their own endowment of
$10. Specifically, each dollar invested to punish Person A de-
creased Person A’s income by $1 (e.g., when a participant invested
$3, the participant’s income was decreased by these $3; the income
of Person A also decreased by $3). In this way, we received a
behavioral measure of third-party punishment, ranging from 0
(when null dollars were invested) to 7 (when $7 were invested).
At the beginning of the instructions, participants were informed
that one of 20 participants in the game will be paid out, as regularly
done in the field of experimental economics (i.e., random-lottery
incentive scheme; see, e.g., Bardslay et al., 2010). This makes it
possible to play the game with high and relevant stakes while
assuring that participants’ choice had, to the given probability, real
financial consequences for participants. Payment of one of 20
participants was made after data collection was completed via the
Amazon Mechanical Turk system.
Compassion. After participants learned about the unfair dis-
tribution of money, but before participants were allowed to punish,
compassion was manipulated in line with Study 2, adapted to the
present study’s context (see also Batson et al., 1997; Batson et al.,
2002). Participants were either randomly assigned to a high- or a
low-compassion condition; specifically, participants in the high-
compassion condition read, “Before you can make your decision,
imagine how Player B feels about the decision by Player A and
how it has affected Player B. Try to feel the full impact of what the
decision could mean to Player B and how s/he feels as a result.”
Participants in the low-compassion condition read, “Before you
make your decision, please take an objective perspective toward
the decision by Player A. Try not to get caught up in how Player
B feels; please remain objective and detached.” Then participants
were asked to write down their thoughts and feelings. After this
instruction, participants decided how much to reduce Player A’s
income.
Manipulation check. After the manipulation of compassion
in Studies 3a and 3b, but before participants could punish, they
respond to three items assessing their current compassionate feel-
ings for Player B. Participants were asked to indicate, on a 5-point
Likert scale, the degree to which they are currently feeling each
emotional state toward Player B. The items read (Study 3a: ␣⫽
.80; Study 3b: ␣⫽.82), “compassionate,” “moved,” “touched.”
5
Moral outrage. In Study 3c, moral outrage was measured
(after third-party punishment). The items used in Study 2b were
adapted to the present study’s context. The items read, “I think that
the decision by Player A was immoral”; “I think the split by Player
A was immoral”; and “I think that Player A’s behavior was
immoral.” Responses were given on a 6-point Likert scale ranging
from strongly disagree to strongly agree. In addition, we assessed
individuals’ emotional response toward the (unjust) distribution
decision of Person A; specifically, individuals had to indicate their
emotional state when they thought about the distribution decision
of Person A. We assessed the six basic emotions of anger, sadness,
fear, surprise, happiness, and disgust (Ekman, 1992) along with the
moral emotion of contempt (Haidt, 2003). Responses were given
on a 5-point Likert scale ranging from not at all to extremely.
Anger and contempt were highly correlated, r.48, p.001 and
averaged to compute an index of anger and contempt. This index
and the first three items were highly correlated, r.57, p.001,
4
Actually, Player A and Player B were fictitious. Participants were
debriefed in this regard at the end of the study.
5
In these studies, we have not assessed other emotions. Please note that
in Study 4, we used the same game theoretical paradigm as in Studies
3a–3c (the third-party punishment paradigm); in Study 4, we found that
other emotions (except for sadness) were not affected.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
7
COMPASSION MAGNIFIES THIRD-PARTY PUNISHMENT
z-standardized because of the different Likert scale ranges,
6
and
averaged to compute an index of moral outrage. The results re-
ported below were exactly the same when we used structural
equation modeling (SEM) and modeled a latent variable affecting
the first three items and the index of anger and contempt, and then
used this latent variable as the mediator.
Participants. Power analysis was equivalent to Study 2, so we
aimed to obtain at least N128 participants. We sampled N133
participants in Study 3a (41.4% women; M
age
33.7, SD
age
11.3),
N126 participants in Study 3b (50.0% women; M
age
34.3,
SD
age
10.2) who completed all measures, and, to adequately power
path analysis, N229 participants in Study 3c (43.7% women;
M
age
33.9, SD
age
10.8).
Results and Discussion
Manipulation check. Analysis in Study 3a of state compas-
sion toward Player B revealed higher state compassion in the
high-compassion condition (M2.77, SD 0.88) compared with
the low-compassion condition (M2.41, SD 1.04). This
difference was significant, t(131) 2.15, p.03; Cohen’s d
0.37. The same pattern emerged in Study 3b: State compassion
was higher in the high-compassion condition (M2.73, SD
1.10) compared with the low-compassion condition (M2.22,
SD 1.01). This difference was significant, t(124) 2.74, p
.01; Cohen’s d0.40. State compassion was further related to
third-party punishment (Study 3a r.33, p.001; Study 3b r
.43, p.001).
Main results. Next, it was tested whether the experimental
manipulation of compassion increased third-party punishment.
This was not the case in Study 3a, t(131) 0.03, p.97; Cohen’s
d0.01. In contrast, analyses revealed a significant difference in
Study 3b, t(124) 2.95, p.01; Cohen’s d0.47 between the
low-compassion condition (M2.16, SD 1.46) and the high-
compassion condition (M3.06, SD 1.93), as well as in Study
3c, t(227) 2.56, p.01; Cohen’s d0.43 between the
low-compassion condition (M1.84, SD 1.45) and the high-
compassion condition (M2.39, SD 1.78). Given the incon-
sistent results across the three studies, we ran an internal meta-
analysis using fixed effects in which the mean effect size was
weighted by sample size (cf. Goh et al., 2016). The effect was
significant, Md.29, Z3.22, p.01, two-tailed, such that
overall, compassion increased third-party punishment. Here, a
small-to-medium effect was found.
Mediation analysis. Next, mediation analysis was executed in
Study 3c (Hayes, 2013). Analysis revealed that moral outrage was
higher in the high- compared to the low-compassion condition
(B0.39, SE 0.11, t3.38, p.001); this in turn was
positively related to third-party punishment (B0.87, SE 0.11,
t8.02, p.001) when the experimental condition was con-
trolled for (B0.21, SE 0.19, t1.08, p.28). This indirect
effect was significant; the bootstrapped 95% confidence interval
(based on 5000 resamples) excluded zero [0.15; 0.59].
We also tested, controlling for moral outrage, whether the other
assessed emotional reactions toward the distribution decisions (i.e.,
sadness, fear, surprise, happiness, and disgust) functioned as me-
diators. This was not the case: In fact, none of the indirect effects
turned out to be significant.
In sum, moral outrage mediated the effect of compassion on
third-party punishment, thus replicating Study 2b. Moreover, we
applied a causal analytic test to the relation of compassion and
third-party punishment, including punishment behavior,inour
analyses. However, we want to acknowledge a shortcoming of
Studies 3a–3c: As we have manipulated compassion for another
individual who suffered from injustice, we could have directly
induced moral outrage in addition to compassion. Therefore, in the
next study, we manipulated compassion first and did so indepen-
dently from the injustice information.
Study 4
Method
Procedure. The overall procedure followed Studies 3a–3c.
Study 4 was conducted in the laboratory at a German University.
Participants first read the instructions of the classic third-party
punishment paradigm. Compassion for Player B was then manip-
ulated without giving any information about the unfair distribution
of money. Only after having successfully induced compassion
(which is confirmed by an independent manipulation check study,
see below) did participants learn that the individual for whom they
feel compassion (Player B) received only $3 from Player A. That
is, in this study, we manipulated compassion independently from
injustice. Finally, participants could engage in third-party punish-
ment toward Player A.
Compassion. We adapted the procedure of Batson, Batson, et
al. (1995) and told our participants that participants in the role of
Player A and Player B had to write something interesting that
happened to them recently. All participants were then told that they
will read what Player B has written. The text of Player B read (cf.
Batson, Batson, et al., 1995):
I’m supposed to write about something interesting that’s happened to
me lately. Well, I don’t know if this will be interesting to anybody
else, but the only thing I can think of is that since two days I’m not
together with my boyfriend any more. We’ve been going together
since our junior year in high school and it’s been great. Now we split
up and I’ve been real down. It’s all I think about. My friends all tell
me that I’ll meet other guys and they say all I need is for something
good to happen to cheer me up. I guess they’re right, but so far that
hasn’t happened.
Before reading the text, as in the previous studies, participants
read compassion promoting instructions for how they should ap-
proach the text (Batson, Batson, et al., 1995; Batson, Chang, Orr,
& Rowland, 2002; Batson, Polycarpou, et al., 1997).
In an independent online manipulation check study to test this
manipulation (N127), we measured state compassion (␣⫽.89)
as in Study 3a and 3b using the items “moved,” “sympathetic,”
“concerned,” and “compassionate.”
7
In addition, the six basic
6
The different Likert scale ranges (5 versus 6) applied in this study were
a mistake.
7
The manipulation check was not included in the main study (Study 4),
because we also wanted to report on a study where there is no measurement
(manipulation check and mediator) in between the manipulation of the inde-
pendent variable and the assessment of the dependent variable, because as-
sessments itself can alter the state participants are in (e.g., Kühnen, 2010).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
8PFATTHEICHER, SASSENRATH, AND KELLER
emotions along with contempt were measured. Analysis revealed
that the manipulation increased state compassion, t(125) 3.71,
p.001; Cohen’s d0.70, but did not significantly affect state
moral outrage, t(125) 0.48, p.64; Cohen’s d0.09, or any
other emotion (all ts|1.20|), except for sadness, t(125) 2.98,
p.01; Cohen’s d0.53. Overall, these results indicate a
successful induction of compassion with this procedure.
Third-party punishment. After the manipulation of compas-
sion, participants were told that Person A took $7 while giving
Person B $3. Participants could then punish Person B, following
the exact procedure of Studies 3a–3c.
Participants. Power analysis was equivalent to the previous
studies, so we aimed to obtain at least N128 participants. We
sampled N140 participants (55.7% women; M
age
21.5,
SD
age
2.9).
Results and Discussion
It was tested whether compassion increased third-party punish-
ment. This was the case; analysis revealed a significant difference,
t(138) 2.56, p.01; Cohen’s d0.43 between the low-
compassion condition (M2.46, SD 1.55) and the high-
compassion condition (M3.17, SD 1.75).
Study 4 mainly addressed an alternative explanation of the
findings of Studies 3a to 3c, specifically, that we could have
directly induced moral outrage regarding the unfair distribution of
money, in addition to compassion. The procedure used in Study 4
manipulated compassion independently from the unfair distribu-
tion of money; as such, moral outrage was not directly induced, as
indicated by the independent manipulation check study. We can
thus rule out the mentioned alternative explanation.
Study 5
In the studies reported so far, suffering was caused by unjust
behavior of the harmdoer(s). We have argued that compassion
intensifies the perception of unjust behavior and consequently
moral outrage and punishment inclinations. This reasoning would
be supported if compassion intensified third-party punishment
only in case harm was caused by unjust behavior, but not in case
harm was not caused by unjust behavior, because then compassion
is unlikely to intensify injustice perception as there is no injustice.
Consequently, in this case, compassion should not lead to moral
outrage, and consequently should not foster third-party punish-
ment. Therefore, there should be no effect of compassion on
third-party punishment given harm was caused by low unjust
actions (as the cascade of mediation is interrupted; for the general
logic of this approach, see Jacoby & Sassenberg, 2011; Pirlott &
MacKinnon, 2016; Spencer et al., 2005). Accordingly, in this
study, we manipulated the level of injustice of the harm-causing
action, orthogonally to the manipulation of compassion. We ex-
pected that compassion intensifies third-party punishment only
when harm was caused by unjust (vs. low unjust) behavior (i.e., an
interaction effect involving the manipulation of compassion and
the manipulation of injustice).
Method
Procedure. Study 5 was conducted online via Amazon Me-
chanical Turk (cf. Buhrmester et al., 2011). Compassion was
manipulated first; the level of injustice (low vs. high) was manip-
ulated orthogonally thereafter. We then assessed moral outrage,
finally, third-party punishment was measured. Overall, a 2 (low vs.
high compassion) by 2 (low vs. high injustice) between-factorial
design was applied. The applied scenario was identical to Studies
2a and 2b (70-year-old George suffered a broken shoulder). Re-
sponses in this study were given on a 5-point Likert scale ranging
from strongly disagree to strongly agree.
Compassion. Compassion for George was manipulated as in
Studies 2a and 2b. This manipulation specifically increased com-
passion and no other emotional states, as indicated by the success-
ful manipulation checks in Studies 2a and 2b.
Injustice. The level of injustice involved in the harm that
occurred was manipulated after the compassion manipulation (on
a separate page). Injustice was manipulated by varying whether or
not there was intent to cause suffering. Past research shows con-
vincingly that individuals consider harm causing action to be far
more unjust when the action was done intentionally, versus when
the very same action was done unintentionally (e.g., Cushman,
2008; Hechler & Kessler, 2018; Malle & Knobe, 1997). Specifi-
cally, participants were given additional information. In the low-
injustice condition, participants read that “Georg continues: Maybe
you are interested in how I broke my shoulder. It was an accident.
A young man unintentionally bumped into me. Then I fell. It
wasn’t the man’s fault. This is how it happened.” In the high-
injustice condition, participants read that “Georg continues: Maybe
you are interested how I broke my shoulder. It was an accident. A
young man deliberately bumped into me. Then I fell. This is how
it happened.” The manipulation of injustice was pretested in an
online study (N51).
8
Results revealed that participants consider
the behavior of Person A to be more unjust in the high-injustice
condition (M6.03, SD 1.55) than in the low-injustice con-
dition (M2.57, SD 1.96). This difference was significant:
t(49) 6.98, p.001, d1.96.
Moral outrage. We assessed moral outrage right before pun-
ishment using the same items as in Study 2b (e.g., “I am angry at
the young man”). Moral outrage was reliably measured; Cron-
bach’s alpha was .96.
Third-party punishment. The same items as in Studies 2a
and 2b were used (e.g., “I want the young man to be severely
punished”). Third-party punishment was reliably measured; Cron-
bach’s alpha was .95.
Participants. In this study, we wanted to obtain very high
power (P.99) to detect at least an effect of small to medium size
(f.20). Using G
Power (Faul et al., 2009), the power analysis
revealed a required sample size of N462 to detect a significant
interaction effect (alpha level of .05) given there is a true effect.
Financial opportunities when the study was run made it possible to
overpower the study and therefore to reduce the likelihood of Type
II error, and to detect even small effects. We obtained complete
data from 973 U.S.-American individuals (62.0% women; M
age
36.7, SD
age
12.1).
8
Injustice was measured using three items (␣⫽.99). The items read,
“The behavior of the young man was immoral,” “The behavior of the
young man was unjust,” and “The behavior of the young man was morally
wrong.”
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
9
COMPASSION MAGNIFIES THIRD-PARTY PUNISHMENT
Results and Discussion
Main results. Means of the conditions are presented in Figure
1. Analyses of variance yielded a significant main effect of com-
passion, F(1, 969) 7.16, p.01, p
2.01, and a significant
main effect of injustice, F(1, 969) 452.63, p.001, p
2.32.
The main effects were qualified by a significant two-way interac-
tion, F(1, 969) 9.32, p.01, p
2.01. Decomposing the
interaction revealed that when injustice was high, participants in
the high-compassion condition reported significantly more third-
party punishment (M2.83, SD 1.23) compared with the
low-compassion condition (M2.48, SD 1.16), F(1, 969)
16.37, p.001, p
2.02. In contrast, such an effect of compas-
sion on third-party punishment was not present when injustice was
low: Participants in the high-compassion condition did not report
significantly more third-party punishment inclinations (M1.32,
SD 0.64) compared with the low-compassion condition (M
1.35, SD 0.70), F(1, 969) 0.07, p.79, p
2.00).
Mediation analysis. For testing the mediation, we used the
PROCESS macro provided by Hayes (2013) and we ran Model 8.
This model tests the assumptions that compassion intensifies moral
outrage, which in turn predicts third-party punishment, but that this
cascade only takes place in the high-injustice condition, and not in
the low-injustice condition. Accordingly, we tested for a condi-
tional indirect effect (conditioned by the injustice manipulation).
Because of the complexity of the following statistical models,
we predominantly report the relevant findings of the model in plain
text. Yet, all statistical parameters are provided on the OSF (https://
osf.io/fqnjm). It was first tested whether the significant positive
relation of compassion and third-party punishment given high
injustice is explained by moral outrage. This was indeed the case:
Compassion intensified moral outrage, which in turn was posi-
tively related to third-party punishment. The indirect effect was
also significant (the 95% bootstrapped CI excluded zero [0.16;
0.50]). This was already shown in Study 2b (using the same
scenario), so Study 5 directly replicates this finding. In contrast,
moral outrage did not function as significant mediator in the
low-injustice condition (the 95% bootstrapped CI included zero
[0.09; 0.11]).
Next, the full conditional indirect effect model was tested.
Analyses revealed a significant interaction of compassion and
injustice when moral outrage was set as the dependent variable.
Specifically, compassion significantly intensified moral outrage
for individuals in the high-injustice condition but not in the low-
injustice condition. Moral outrage, in turn, predicted third-party
punishment when compassion, injustice, as well as their interac-
tion, were controlled for. Overall, there was a significant index for
a conditional indirect effect (the 95% bootstrapped CI excluded
zero [0.13; 0.52]).
Discussion. Overall, we found empirical support for our main
assumption that compassion magnifies third-party punishment:
The highly unjust action by the young man increased third-party
punishment (vs. a low unjust action); this increase was magnified
by compassion (see Figure 1). The latter increase could be ex-
plained by higher moral outrage in the high- compared to the
low-compassion condition; moral outrage was, in turn, positively
related to third-party punishment. These findings replicate Studies
2b and 3c. Overall, Study 5 showed that compassion intensifies
moral outrage and thus serves as a magnifier of third-party pun-
ishment given high injustice. It should be noted that very low
levels of third-party punishment emerged in the low injustice
conditions. We argue that the low means had to emerge, because
low (or no) injustice should lead to very low levels of third-party
punishment inclinations. It is actually an interesting question for
future research to examine at what level of injustice does compas-
sion start to intensify moral outrage and third-party punishment.
Studies 6a and 6b
In all studies so far, we made use of a standard approach to
manipulate compassion (i.e., the “Batson-manipulation”; e.g., Bat-
son, 2009; Batson, Batson, et al., 1995; Batson, Polycarpou, et al.,
1997). This manipulation includes the presentation of a sympa-
thetic suffering target, and in the high-compassion condition par-
ticipants are asked to imagine how the described person feels and
are encouraged to feel for the suffering person; in the control
condition, however, participants are asked to remain objective and
detached. Recently, this manipulation was critically discussed:
McAuliffe, Forster, Philippe, and McCullough (2018) showed that
participants who received no instructions how they should ap-
proach a suffering target reported similar levels of compassion
compared with a standard high-compassion condition (i.e., taking
the perspective and feel for the suffering target), and higher levels
of compassion than a standard low-compassion (i.e., instructions to
stay objective and detached). McAuliffe and colleagues (2018)
conclude that when confronted with a sympathetic suffering target,
individuals respond spontaneously with compassion, and that en-
couraging to feel for the suffering target (using the “Batson-
manipulation”) adds only little additional compassion. Moreover,
one can argue that by encouraging individuals to remain objective
and detached, the low-compassion condition involves downregu-
lation of compassion (and emotions in general); thus an effect
might not be driven by compassion but instead by downregulation
of emotions. To address these issues, in the next two studies, we
manipulated compassion in a different way and used different
control conditions.
Method
Procedure. The procedure and material followed Study 5.
Only the manipulation of compassion differed, and only the high-
injustice condition was applied in these studies. We had a com-
Figure 1. Third-party punishment tendencies as a function of injustice
and compassion (Study 5).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
10 PFATTHEICHER, SASSENRATH, AND KELLER
passion condition and a (two) control condition(s) in Study 6a (6b).
In both studies, we first manipulated compassion and assessed the
manipulation check. We then provided information that George
suffered from unjust behavior by a young man, assessed moral
outrage, and finally third-party punishment inclinations toward the
young man who bumped into the old man (George).
As an additional measure, we exploratively assessed identifica-
tion with the suffering target (using the item “I identified with the
old man”) as an alternative/additional explanation why compas-
sion should foster third-party punishment. The idea is that com-
passion for a suffering target might increase identification with the
harmed (Cassell, 2009), which in turn drives punishment. Because
we did not find any significant effects involving identification
(e.g., identification was not increased in the compassion condi-
tions), we did not consider identification when further reporting
the results.
Compassion manipulation. To manipulate compassion, we
relied on work of Klimecki, Leiberg, Ricard, and Singer (2014), as
well as from Stellar, Cohen, Oveis, and Keltner (2015) and Oveis,
Horberg, and Keltner (2010). To induce compassion, these authors
used sets of videos or pictures showing suffering people, but did
not offer any explicit instructions of how to approach the stimuli
(as in the “Batson-manipulation”). This way of inducing compas-
sion is in line with McAuliffe et al. (2018), arguing that individuals
respond spontaneously with compassion for a sympathetic suffer-
ing target. In our two studies, we made use of this idea.
In Study 6a, adapted from Klimecki et al. (2014) and Stellar et
al. (2015), participants in the compassion condition read, “In the
following, you will watch a video about George. George is 61
years old and homeless. The video takes two and a half minutes.”
Participants then watched a video of a suffering homeless old man
(George) who talks about his difficult life on the street. Partici-
pants in the control group read, “Later in the study, you will get to
know George. George is 61 years old and homeless. Before you
judge a situation involving George, we ask you to watch a video
about how to build a fence. The video takes two and a half
minutes.” Participants then watched a video about how to build a
fence (cf. Stellar et al., 2015, Study 3). Videos in both conditions
lasted 140 seconds.
In Study 6b, we adapted the procedure by Stellar et al. (2015,
Study 2) and Oveis et al. (2010). Participants in the compassion
condition read, “In the following, you will see 5 pictures of
George. George is 61 years old. Every picture is presented for 5
seconds. The next picture will always appear automatically.” Par-
ticipants then saw five pictures of the suffering old homeless man
of Study 6a. Participants in the first control group read the same
text and then saw five different pictures of a nonsuffering man of
roughly the same age sitting on a chair. Participants in the second
control group read, “In the following, you will get to know George.
George is 61 years old. Later in the study, you will judge a
situation involving George.” In this condition, participants re-
ceived no pictures. As such, in both Studies 6a and 6b, we had no
explicit instructions to downregulate emotions in the control
group.
Manipulation check, moral outrage, and third-party punish-
ment. The same items as in Study 2b and Study 5 were used. All
scales showed satisfying reliability (s.85). The manipulation
check items were assessed right after watching the videos (Study
6a) or pictures (Study 6b). Then participants were given informa-
tion that George suffered from unjust behavior. Participants read,
“George reports about a recent accident: ‘Last week I broke my
shoulder. A young man deliberately bumped into me. Then I fell.
This is how it happened.’” Finally, moral outrage and third-party
punishment inclinations were assessed.
Participants. Power analysis was equivalent to previous stud-
ies (Studies 2– 4), so we aimed to obtain at least N64 partici-
pants per cell. In Study 6a (two conditions) we sampled N220
participants (52.3% women; M
age
36.8, SD
age
12.2). In Study
6b (three conditions) we sampled N332 participants (53.6%
women; M
age
36.5, SD
age
12.6).
Results and Discussion
Manipulation check. In Study 6a, analysis revealed a signif-
icant difference, t(218) 17.76, p.001; Cohen’s d2.39 in
state compassion between the compassion condition (M5.61,
SD 1.19) and control condition (M2.45, SD 1.44). The
same pattern emerged in Study 6b, F(2, 329) 84.60, p.001;
state compassion was significantly higher, t(218) 10.90, p
.001; Cohen’s d1.47 in the compassion condition (M5.19,
SD 1.29) than in the first control condition (M3.21, SD
1.40), and it was significantly higher, t(219) 11.96, p.01;
Cohen’s d1.61 than in the second control condition (M2.90,
SD 1.54). As such, in both studies, we successfully induced
compassion using a different procedure than in previous studies.
Next, in Study 6a, although state outrage was on a low level (the
means were far below the midpoint on the 1 to 7-point scale, p
.001), there was a significant difference in state outrage, t(218)
8.05, p.001; Cohen’s d1.08 between the compassion
condition (M2.83, SD 1.71) and the control condition (M
1.33, SD 0.94). The same pattern emerged in Study 6b, F(2,
329) 41.00, p.001; state outrage was significantly higher,
t(218) 7.70, p.001; Cohen’s d1.04 in the compassion
condition (M2.68, SD 1.49) than in the first control condition
(M1.39, SD 0.94), and it was significantly higher, t(219)
6.94, p.001; Cohen’s d0.93 than in the second control
condition (M1.48, SD 1.04).
9
In addition, happiness was
reduced (Study 6a Cohen’s d⫽⫺1.26, Study 6b d⫽⫺1.11) and
disgust was increased (Study 6a Cohen’s d0.57, Study 6b d
0.61). We address below the problem that not only compassion
was altered by running a multiple regression analysis including
state compassion, state outrage, and the other emotions as predic-
tor to see which emotion drives the effects (see section “Additional
Analyses” below).
Third-party punishment. In Study 6a, analysis revealed a
significant difference, t(218) 3.59, p.001; Cohen’s d0.48
in third-party punishment between the compassion condition (M
5.09, SD 1.45) and the control condition (M4.37, SD 1.53).
This finding was replicated in Study 6b, F(2, 329) 8.43, p
.001; third-party punishment was significantly higher, t(218)
2.93, p.01; Cohen’s d0.39 in the compassion condition (M
9
We also tested in independent studies the material we relied on in our
studies, i.e., Klimecki et al. (2014); Oveis et al. (2010), and Stellar et al.
(2015), and found the same pattern; that is, compassion was increased, but
so was state outrage on a low level. This is not surprising, because the
stimuli include suffering people and the displayed suffering can be attrib-
uted to injustice and thus increase state outrage.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
11
COMPASSION MAGNIFIES THIRD-PARTY PUNISHMENT
4.68, SD 1.63) than in the first control condition (M4.02,
SD 1.73), and it was significantly higher, t(219) 3.99, p
.001; Cohen’s d0.54 than in the second control condition (M
3.75, SD 1.82).
Additional analyses. Because both state compassion and state
outrage were increased by the experimental manipulation, we
predicted third-party punishment in a multiple regression analysis
to test which emotion is crucial and responsible for the effect of the
experimental manipulation on third-party punishment. Accord-
ingly, we simultaneously entered state compassion and state out-
rage and the other emotions in a multiple regression model. In both
studies, whereas state compassion was a significant and relevant
(in terms of effect size) predictor of third-party punishment
(R
2
s.10), state outrage and the other emotions were not a
significant predictor and not relevant in terms of effect size
(R
2
s.015). We conclude that although the manipulation in-
creased state outrage on a low level, in addition to compassion (as
in other research, see footnote 9), our analyses show that state
compassion but not state outrage is crucial, and it is responsible for
the effect of the experimental manipulation on third-party punish-
ment.
Mediation analysis. To reduce redundancy, the statistical pa-
rameters are reported in detail on the OSF. In line with the
previous studies, the analyses showed that the experimental ma-
nipulation significantly increased moral outrage. Again, state com-
passion and not state outrage was responsible for the effect of the
manipulation; only state compassion was a significant and relevant
predictor of moral outrage in terms of effect size (R
2
s.10),
whereas state outrage was not a significant predictor and not
relevant in terms of effect size (R
2
s.013). Moral outrage in
turn predicted third-party punishment. The bootstrapped 95% con-
fidence intervals in both studies excluded zero (lower 95% CIs
.16).
Discussion. In Studies 6a and 6b we replicated the findings of
the previous studies using a different manipulation of compassion.
This manipulation builds on the idea that individuals respond
spontaneously with compassion for a sympathetic suffering target
(McAuliffe et al., 2018). Although the manipulation also increased
state outrage, we showed statistically that only compassion was
crucial and responsible for increased third-party punishment.
Please note that in the previous studies reported in the contribution,
we had a more specific manipulation of compassion that did not
increase state outrage (the “Batson-manipulation”). This manipu-
lation, however, included an explicit instruction to downregulate
emotions in the control group. This shortcoming was addressed in
Studies 6a and 6b. We conclude that there is overall strong
evidence that compassion increases third-party punishment. This is
also supported by the internal meta-analysis reported in the fol-
lowing.
Internal Meta-Analysis
We ran an internal meta-analysis including the experimental
studies (Studies 2 to 6) and the three additional studies reported on
the OSF (see also section “Additional Studies” in this paper). In
total, this resulted in 13 main effects between an experimental
(compassion) condition and a control condition included in the
analysis. Again, we used fixed effects in which the mean effect
size was weighted by sample size (Goh et al., 2016). The overall
effect was significant, Md0.38, SE 0.04, Z9.17, p.001,
two-tailed. The lower 95% CI of the effect was 0.30, and the upper
95% CI was 0.46. As such, the present set of studies suggests that
compassion increases third-party punishment. Here, a small-to-
medium effect was found.
General Discussion
People are frequently confronted with suffering, as is evident in
the media’s extensive coverage of terrorist attacks, the suffering of
war, violent acts such as homicides, and brutal beatings of inno-
cent persons. Given the omnipresent confrontation with suffering,
we aimed to provide a better understanding of resulting interper-
sonal consequences. In Studies 1a–1f, we have shown that com-
passion for individuals suffering from unjust actions was corre-
lated with punishment inclinations toward those responsible for the
suffering. This relation could be documented across a variety of
contexts of high societal relevance (i.e., terrorist attacks, sexual
assaults, rape of children, and war). In the consecutive studies
(Studies 2– 6), we applied an experimental approach and showed
that compassion magnifies punishment inclinations toward those
responsible for unjust suffering. Additional analyses aiming to
explain this relation revealed that compassion intensified moral
outrage about the unjust suffering, moral outrage, in turn, was
related to third-party punishment.
In sum, we have shown across six studies in the laboratory and
online, as well as across different countries, that compassion for
people suffering from unjust actions magnifies third-party punish-
ment toward those who have caused unjust suffering. These results
of the present set of studies have numerous implications for
research on compassion and third-party punishment, as well as for
society. These implications are discussed in the following sections.
Implications for Research on Compassion
Emotions prepare for action (Scherer, 2005). With respect to
compassion, the last decades of research on this moral emotion
have shown that compassion leads to a gamut of pro-social behav-
iors, ranging from caring for children and babies in need to helping
people who contracted a disease; compassion even prepares indi-
viduals to engage in proenvironmental behavior (Goetz et al.,
2010; Pfattheicher, Sassenrath, & Schindler, 2016). However, pos-
sible antisocial or harm-causing tendencies of compassion were
almost completely overlooked. The few exceptions include re-
search by Batson, Batson, et al. (1995), showing that compassion
increases allocation of resources to a suffering individual, thus
exploiting, however, the resources of the collective. Lupoli, Jam-
pol, and Oveis (2017) document that compassion increases lying
behavior to protect a suffering person. Buffone and Poulin (2014)
show that compassion for a distressed person increases harming
behavior toward a third individual who faces an upcoming com-
petition with the distressed person. These findings indicate that
compassion can have negative consequences for others. Still, they
fit the overall picture that the main emerging tendency of com-
passion is to help a suffering other (Goetz et al., 2010); in each
study, the person for whom compassion is felt benefits from
compassionate individuals’ actions. In the present work, we ex-
amined what happens when compassionate individuals cannot
directly help a victim (i.e., improving the state of a suffering
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
12 PFATTHEICHER, SASSENRATH, AND KELLER
individual by his or her own actions) but when the possibility to
harm the perpetrator is given (i.e., impair the state of another
individual). In such situations, compassion increases harm, that is,
punishment. Of course, we must acknowledge that our findings
remain restricted to situations in which it is difficult for individuals
to help. However, direct helping is often difficult. Think, for
example, about when terrorist attacks happen, when viewing me-
dia reports about the war in Syria, or when perpetrators commit
sexual assaults; it is difficult to help the victims directly, but one
can generate retaliatory tendencies and support for punishment of
the suffering-causing third parties. Indeed, our findings apply to
these specific situations, but this fact does not make the findings
less impactful.
Our findings can additionally be discussed in reference to work
by Batson, Klein, Highberger, and Shaw (1995). These authors
tested consequences of compassion when altruism and justice are
at odds, specifically whether compassion and resulting helping for
one individual can increase injustice to other individuals. Empir-
ically it is shown that with low feelings of compassion, individuals
acted in line with the justice principle: The vast majority of
participants did not reassign a terribly ill child on a waiting list to
immediate help at the expense of others who were more needy.
However, when inducing compassion for the child, a majority of
participants violated the justice principle and helped the child at
the expense of others. Batson, Klein, et al. (1995) created a
situation in which altruism and justice are at odds and show that
compassion motivates people to help—at the costs of violating the
justice principle. Batson, Klein, et al. (1995, p. 1053) states that “if
empathy can be evoked for the victims of injustice, then these two
motives can be made to work together rather than at odds.” This is
exactly what we show in the present contribution: Compassion for
a victim suffering from injustice intensifies the perception of
injustice and its downstream consequences. Compassion thus re-
flects, as Batson, Klein, et al. (1995, p.1053) put it, the “emotional
fire” that intensifies emotional reactions to cared-for suffering
others.
We furthermore want to discuss the implications of the present
work with respect to incidental and integral compassion. Condon
and DeSteno (2011) have shown that incidental compassion, elic-
ited by an uninvolved sad third person who suffered because she
found out her brother had cancer, reduces punishment of a third
uninvolved person who cheated on a task to win money. In the
present work, we have shown that integral compassion can in-
crease third-party punishment. Note, however, that our study and
the study by Condon and DeSteno (2011) not only differed in the
aspect of incidental versus integral compassion, but also regarding
third-party punishment. Third-party punishment in our studies was
about the perpetrator who hurt a victim the participant was feeling
compassion toward, whereas the study by Condon and DeSteno
(2011) was about third-party punishment toward an unrelated
individual who has not hurt someone but has cheated on a task.
Taken together, these findings point to the complexity of compas-
sion in the context of third-party punishment, in that harm-
reducing as well as harm-increasing tendencies can result from
compassion.
A relevant question that needs further discussion is whether (a)
compassionate individuals consider harming a perpetrator as an
ultimate goal in itself, or (b) whether harm serves as an instru-
mental means to reach a higher goal, that is, to reestablish justice.
The empirical findings of the present work speak in favor of the
latter, because compassion-based third-party punishment is ex-
plained by moral outrage which represents a driving force for
actions that reestablish justice (Carlsmith et al., 2002). Yet we
acknowledge that we did not test these two explanatory accounts
against each other. In discussing the specific motives of compas-
sionate punishers, one can also consider the possibility that indi-
viduals punish to deter future suffering; in this regard, the seminal
work by Carlsmith and colleagues (2002, 2008) shows that al-
though people typically express strong preferences for deterrence,
individuals’ actual behavior is driven exclusively by just deserts
concerns. Tan and Xiao’s (in press) recent work supports this idea
and shows that third-party punishment is driven by just deserts
concerns rather than by deterrence motives. Of course, this does
not conclusively answer the question about whether third-party
punishment based on compassion may also be driven by just
deserts concerns.
In addition we want to emphasize that additional processes can
contribute to compassion increasing third-party punishment. That
is to say, we only provide empirical support for compassion
intensifying moral outrage, which in turn motivates third-party
punishment. It could be that compassion is accompanied by being
more identified with the victim (Cassell, 2009), which increases
moral outrage about the victim being harmed (although we did not
find initial support for this idea in Studies 6a and 6b), or that
compassion is accompanied by a self-other overlap with the suf-
fering person (Batson et al., 1997), motivating third-party punish-
ment. It could also be that compassionate individuals who punish
a perpetrator want to signal to others that they do not accept unjust
behavior (Jordan, Hoffman, Bloom, & Rand, 2016). Basically, we
do not argue for a single mediator perspective; it seems reasonable
to assume that additional processes can contribute to compassion
increasing third-party punishment (cf. Bullock et al., 2010; Fiedler,
Schott, & Meiser, 2011, and Fiedler, Harris, & Schott, 2018, for
this basic idea).
The overall picture in the literature on compassion suggests that
compassion might function as a possibility to reduce escalation of
violence (Batson & Ahmad, 2001; Condon & DeSteno, 2011;
Singer & Steinbeis, 2009), but given the findings of the present
research, compassion might also add fuel to the fire (Antonetti &
Maklan, 2017; Condon & DeSteno, 2017). Specifically, it has been
shown that peer punishment (such as third-party punishment in the
present contribution) can create a downward cycle of retribution
and counterpunishment (Nikiforakis, 2008; Pinker, 2011), leading
to suboptimal intra- and interpersonal states (Dreber, Rand, Fuden-
berg, & Nowak, 2008; Pfattheicher, Böhm, & Kesberg, 2018).
From this perspective, we cloud the rosy picture that is typically
drawn of compassion as well as the notion that one can reach
global peace by practicing compassion whenever suffering occurs
(Dalai Lama & Ekman, 2008). These thoughts are in line with
recent considerations about the limits of empathy and compassion
(e.g., Bloom, 2016; Breithaupt, 2015; Decety & Cowell, 2014;
Sassenrath, Hodges, & Pfattheicher, 2016).
Understanding the Emergence of Third-Party Punishment
The present research has implications for our understanding of
how and why third-party punishment toward those who caused
unjust suffering emerges. Past research has explained the emer-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
13
COMPASSION MAGNIFIES THIRD-PARTY PUNISHMENT
gence of third-party punishment by arguing that unfair behavior
reflects the violation of a fairness norm; this norm violation elicits
anger, which in turn motivates punishment of the third party (Fehr
& Fischbacher, 2004; Nelissen & Zeelenberg, 2009; Jordan,
McAuliffe, & Rand, 2016). With our contribution, we point to a
significant factor contributing to this cascade: compassion for a
suffering victim. We show that compassion magnifies anger
(moral outrage), which in turn fosters third-party punishment. As
such, we contribute to a better understanding about the evolution
of third-party punishment, and thus add one piece of significant
knowledge to the existing literature.
We further emphasize that the present research has applied a
broad conceptualization of third-party punishment, incorporating
any behavior that includes negative sanctions and consequences or
the support of negative sanctions and consequences targeted at a
third party that has caused harm. This broad conceptualization
made it possible to include a broad range of tendencies, including
support for retribution of terror attacks, support of the death
penalty for rapists, and third-party punishment of unfair individu-
als in economic games. As such, we did not restrict our study
program to economic games (which one could argue is the current
gold standard in the investigation of third-party punishment);
rather, we aimed to test the idea of compassion motivating third-
party punishment across a vast variety of social situations and
across different moral domains (e.g., harm, fairness; Graham et al.,
2011), which reflects a strength of the present contribution.
Finally, we argue that the idea of compassion fostering third-
party punishment can also be applied to intergroup hate and
aggression, for instance in the Middle East region (Böhm, Thiel-
mann, & Hilbig, 2018; Bruneau, Cikara, & Saxe, 2017; Vanman,
2016). Today, political leaders from Israel and Palestine stress past
suffering caused by members of the other group. When people feel
compassion for suffering individuals of their own group, they
might judge this suffering as particularly unjust, eliciting moral
outrage and, consequently, increase the chances for intergroup
aggressive acts. Basically, we argue that compassion for victims of
the in-group might stimulate punishment inclinations and aggres-
sive tendencies toward out-group members who caused the suffer-
ing (cf. Gordijn, Wigboldus, & Yzerbyt, 2001). We do not, how-
ever, want to demonize compassion as it is an important emotion
that has the potential to motivate pro-social behavior (Haidt, 2003;
Goetz et al., 2010). What seems important is for whom compassion
is felt. In this regard, Mazziotta, Feuchte, Gausel, and Nadler
(2014) have shown that thinking about harm during war that was
inflicted by one’s own group (i.e., a perpetrator-focus) elicits
compassionate feelings for victims of an outgroup and increased
willingness to engage in cross-group contact, in contrast to think-
ing about harm that was inflicted on the own group (i.e., a
victim-focus; see also Roth, Shane, & Kanat-Maymon, 2017). By
combining our research and the work of Mazziotta et al. (2014),
important policy implications can be made. Compassion for vic-
tims’ suffering, caused by their own harmful actions, or by the
in-group, seems beneficial for peace (Mazziotta et al., 2014); in
contrast, as shown in the present research, compassion for victims
suffering from harmful actions caused by third persons can lead to
punishment inclinations, potentially stimulating escalation of vio-
lence rather than preventing it (Condon & DeSteno, 2017). Basi-
cally, we argue that one should be cautious about claims that
promote the practice of unconditional compassion whenever suf-
fering occurs (Dalai Lama & Ekman, 2008).
Limitations and Outlook
At this point we acknowledge limitations of the present work
and point to potential future research. First, we want to discuss
three methodological points. The first methodological point is the
discussion of possible demand characteristics. We think that two
arguments speak against the possibility that demand characteristics
are driving the effects of the present contribution. First, one could
assume that it is socially demanded that one shows helping behav-
ior toward a victim, especially when one is asked to feel compas-
sion (as the work by Batson and colleagues suggest). However, we
think that social demands are less likely present when it comes to
punishment (and thus harm) of a perpetrator. We acknowledge,
however, that we have no empirical data for this claim. Second, in
our four studies that made use of the third-party punishment
paradigm (Studies 3a–3c, 4), punishment was incentivized and
thus was associated with financial costs. Accordingly, adhering to
demand characteristics (resulting in increased punishment) is
costly, which likely reduces its occurrence. Overall, although we
cannot fully rule out that demand characteristics might have played
a role, we argue that they are too weak to explain away the basic
effect of compassion magnifying third-party punishment.
The second methodological point is that we have not, in the
present set of studies, examined compassion-based third-party
punishment in face-to-face situations. We have deliberately
avoided face-to-face interactions and used scenario versions and a
game theoretical paradigm (i.e., the third-party punishment para-
digm) because these were relatively neutral. Hence, these situa-
tions are more parsimonious regarding the information that is
salient in the situation. Accordingly, possible confounding factors
that may come along with real face-to-face interactions (e.g.,
attractiveness, gender) can be reduced. This leaves room for future
research to study the idea of compassion-based third-party pun-
ishment in face-to-face interactions, for instance in front of a court
where judges might feel compassion for raped children, fostering
their inclination to impose the death penalty on the rapist.
The third methodological point that we want to discuss is that in
Studies 3a–3c (and two studies reported above in the section
“Additional Studies”) we have manipulated compassion for an-
other individual who suffered from injustice by a third person. As
such, this situation confounded compassion (elicited by suffering)
and moral outrage (elicited by injustice). On the one hand, this
confound is, of course, a shortcoming. On the other hand, it is a
strength in the sense that the present research program has in-
cluded situations in which suffering and injustice naturally occur
together, which is often the case. Terrorist attacks provide a useful
example; these create substantial suffering that is highly unjust.
The same applies to the rapist that creates highly unjust suffering
in children, or people suffering from unjust actions in the Syrian
war. The confound is addressed in Studies 2 and 4 6. Here,
compassion was manipulated prior to injustice so that compassion
and injustice were independent. These situations might less likely
occur in real life but are, from a methodological perspective, more
valid as the confounding condition is avoided. In sum, it is a
strength of the present contribution that situations are included in
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
14 PFATTHEICHER, SASSENRATH, AND KELLER
which both suffering and injustice naturally occur together as well
as situations in which compassion precedes injustice.
We further note that we have examined effects of compassion
on third-party punishment on a short-time scale. Thus, we cannot
draw inferences about effects of constant and ongoing suffering
and whether retaliatory tendencies on the basis of compassion
become chronic (e.g., chronic hostile tendencies against those
responsible for terrorist attacks). Past research has shown that
overwhelming levels of suffering can lead to a collapse of com-
passion, especially for individuals who are skilled at emotion
regulation (Cameron & Payne, 2011). Thus, in cases where indi-
viduals constantly self-regulate compassion, we expect that retal-
iatory tendencies are less likely to become chronic. However, if
individuals do not self-regulate recurring compassion toward in-
dividuals suffering from injustice, hostile tendencies on the basis
of compassion might become chronic. Indeed, past research has
revealed positive relations between chronic compassion and
chronic hostility toward other people in general (Keller & Pfat-
theicher, 2013), yet long-term effects of the exposure to suffering
still need to be investigated.
Moreover, our research remains silent regarding the default
behavioral tendency that results from compassion. This thought
refers to the question of whether the first impulse of compassionate
individuals is to help a person suffering from injustice, or whether
the first impulse is to harm those responsible for the suffering.
Because compassion is elicited by the attention to a suffering
person, we assume that the first impulse is to help the person and
to only later turn attention to those who caused the suffering. These
considerations are in line with research of Weng, Fox, Hessentha-
ler, Stodola, and Davidson (2015; see also Jordan, Hoffman, et al.,
2016) who offered their participants the possibility of helping a
victim and punishing the responsible harmdoer. The authors only
found an effect of compassion training on helping behavior, not on
punishment. These findings suggest that the default tendency of
compassionate individuals is to help, yet with no possibility of
helping, as in our research, compassion can also foster third-party
punishment.
We emphasize that the present research incorporates a variety of
situations in which compassion-based third-party punishment may
emerge, yet the applicability of the basic idea to additional situa-
tions of high societal relevance seems possible. One can wonder
whether compassion for suffering victims causes judges to inflict
stronger punishment on a perpetrator. It even seems possible that
compassion for a victim might increase the chance that an accused
perpetrator receives a death sentence. This idea was put to a first
correlational test in Study 1f but has not been experimentally
investigated; thus, causal conclusions cannot be drawn. Moreover,
the idea of compassion-based third-party punishment can be ap-
plied to intergroup contexts, for instance, to the conflict between
Israel and Palestine where substantial suffering, compassion, and
retaliation occur. Basically, we argue that the present research can
reflect a solid basis for studies on compassion and third-party
punishment in a variety of social contexts.
Conclusions
In this contribution, we have shown that punishment inclinations
can be magnified by an emotion that is considered to be a prosocial
tendency par excellence, that is, compassion (Dalai Lama & Ek-
man, 2008; Goetz et al., 2010). Specifically, by showing that
compassion relates to harmful tendencies when unjustified suffer-
ing has occurred, we applied a differentiated perspective on com-
passion and its interpersonal consequences. As such, the current
research opens a new avenue of research for studying compassion,
hopefully inspiring both basic research as well as research from an
applied, societal perspective.
References
Antonetti, P., & Maklan, S. (2017). Concerned protesters: From compas-
sion to retaliation. European Journal of Marketing, 51, 983–1010.
http://dx.doi.org/10.1108/EJM-03-2016-0145
Ashar, Y. K., Andrews-Hanna, J. R., Yarkoni, T., Sills, J., Halifax, J.,
Dimidjian, S., & Wager, T. D. (2016). Effects of compassion meditation
on a psychological model of charitable donation. Emotion, 16, 691–705.
http://dx.doi.org/10.1037/emo0000119
Bardslay, N., Cubitt, R., Loomes, G., Moffatt, P., Starmer, C., & Sugden,
R. (2010). Experimental economics. Rethinking the rules. New Jersey:
Princeton University Press.
Barton, A. (2004). Just deserts theory. In M. Bosworth (Ed.), Encyclopedia
of prisons and correctional facilities (pp. 504 –507). Thousand Oaks,
CA: Sage.
Batson, C. D. (2009). These things called empathy. In J. Decety & W. Ickes
(Eds.), The social neuroscience of empathy (pp. 3–16). Cambridge, MA:
MIT Press. http://dx.doi.org/10.7551/mitpress/9780262012973.003
.0002
Batson, C. D. (2011). Altruism in humans. United Kingdom: Oxford
University Press.
Batson, C. D., & Ahmad, N. (2001). Empathy-induced altruism in a
prisoner’s dilemma II: What if the target of empathy has defected?
European Journal of Social Psychology, 31, 25–36. http://dx.doi.org/10
.1002/ejsp.26
Batson, C. D., Batson, J. G., Griffitt, C. A., Barrientos, S., Brandt, J. R.,
Sprengelmeyer, P., & Bayly, M. J. (1989). Negative-state relief and the
empathy–altruism hypothesis. Journal of Personality and Social Psy-
chology, 56, 922–933. http://dx.doi.org/10.1037/0022-3514.56.6.922
Batson, C. D., Batson, J. G., Todd, R. M., Brummett, B. H., Shaw, L. L.,
& Aldeguer, C. M. (1995). Empathy and the collective good: Caring for
one of the others in a social dilemma. Journal of Personality and Social
Psychology, 68, 619 – 631. http://dx.doi.org/10.1037/0022-3514.68.4
.619
Batson, C. D., Chang, J., Orr, R., & Rowland, J. (2002). Empathy, atti-
tudes, and action: Can feeling for a member of a stigmatized group
motivate one to help the group? Personality and Social Psychology
Bulletin, 28, 1656 –1666. http://dx.doi.org/10.1177/014616702237647
Batson, C. D., Kennedy, C. L., Nord, L. A., Stocks, E. L., Fleming,
D. Y. A., Marzette, C. M.,...Zerger, T. (2007). Anger at unfairness: Is
it moral outrage? European Journal of Social Psychology, 37, 1272–
1285. http://dx.doi.org/10.1002/ejsp.434
Batson, C. D., Klein, T. R., Highberger, L., & Shaw, L. L. (1995).
Immorality from empathy-induced altruism: When compassion and jus-
tice conflict. Journal of Personality and Social Psychology, 68, 1042–
1054. http://dx.doi.org/10.1037/0022-3514.68.6.1042
Batson, C. D., Polycarpou, M. P., Harmon-Jones, E., Imhoff, H. J., Mitch-
ener, E. C., Bednar, L. L.,...Highberger, L. (1997). Empathy and
attitudes: Can feeling for a member of a stigmatized group improve
feelings toward the group? Journal of Personality and Social Psychol-
ogy, 72, 105–118. http://dx.doi.org/10.1037/0022-3514.72.1.105
Bekkers, R., & Ottoni-Wilhelm, M. (2016). Principle of care and giving to
help people in need. European Journal of Personality, 30, 240 –257.
http://dx.doi.org/10.1002/per.2057
Bloom, P. (2016). Against empathy: The case for rational compassion.
New York, NY: HarperCollins.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
15
COMPASSION MAGNIFIES THIRD-PARTY PUNISHMENT
Böhm, R., Thielmann, I., & Hilbig, B. E. (2018). The brighter the light, the
deeper the shadow: Morality also fuels aggression, conflict, and vio-
lence. Behavioral and Brain Sciences, 41, e98.
Breithaupt, F. (2015). Why empathy is not the best basis for humanitari-
anism. Global Cooperation Research Papers, 9, 5–16.
Bruneau, E. G., Cikara, M., & Saxe, R. (2017). Parochial empathy predicts
reduced altruism and the endorsement of passive harm. Social Psycho-
logical and Personality Science, 8, 934 –942. http://dx.doi.org/10.1177/
1948550617693064
Buffone, A. E., & Poulin, M. J. (2014). Empathy, target distress, and
neurohormone genes interact to predict aggression for others-even with-
out provocation. Personality and Social Psychology Bulletin, 40, 1406 –
1422. http://dx.doi.org/10.1177/0146167214549320
Buhrmester, M., Kwang, T., & Gosling, S. D. (2011). Amazon’s Mechan-
ical Turk: A new source of inexpensive, yet high-quality, data? Per-
spectives on Psychological Science, 6, 3–5. http://dx.doi.org/10.1177/
1745691610393980
Bullock, J. G., Green, D. P., & Ha, S. E. (2010). Yes, but what’s the
mechanism? (don’t expect an easy answer). Journal of Personality and
Social Psychology, 98, 550 –558. http://dx.doi.org/10.1037/a0018933
Cameron, C. D., & Payne, B. K. (2011). Escaping affect: How motivated
emotion regulation creates insensitivity to mass suffering. Journal of
Personality and Social Psychology, 100, 1–15. http://dx.doi.org/10
.1037/a0021643
Cameron, C. D., & Payne, B. K. (2012). The cost of callousness: Regu-
lating compassion influences the moral self-concept. Psychological Sci-
ence, 23, 225–229. http://dx.doi.org/10.1177/0956797611430334
Carlsmith, K. M., & Darley, J. M. (2008). Psychological aspects of retrib-
utive justice. Advances in Experimental Social Psychology, 40, 193–236.
http://dx.doi.org/10.1016/S0065-2601(07)00004-4
Carlsmith, K. M., Darley, J. M., & Robinson, P. H. (2002). Why do we
punish? Deterrence and just deserts as motives for punishment. Journal
of Personality and Social Psychology, 83, 284 –299. http://dx.doi.org/10
.1037/0022-3514.83.2.284
Cassell, E. J. (2009). Compassion. In S. J. Lopez & C. R. Snyder (Eds.),
Oxford library of psychology. Oxford handbook of positive psychology
(pp. 393– 403). New York, NY: Oxford University Press.
Cialdini, R. B., Brown, S. L., Lewis, B. P., Luce, C., & Neuberg, S. L.
(1997). Reinterpreting the empathy-altruism relationship: When one into
one equals oneness. Journal of Personality and Social Psychology, 73,
481– 494. http://dx.doi.org/10.1037/0022-3514.73.3.481
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159.
http://dx.doi.org/10.1037/0033-2909.112.1.155
Condon, P., & DeSteno, D. (2011). Compassion for one reduces punish-
ment for another. Journal of Experimental Social Psychology, 47, 698 –
701. http://dx.doi.org/10.1016/j.jesp.2010.11.016
Condon, P., & DeSteno, D. (2017). Enhancing compassion: Social psy-
chological perspectives. In E. Seppala, J. Doty, M. Worline, D. Cam-
eron, S. Brown, & E. Simon-Thomas (Eds.), Oxford handbook of com-
passion science (pp. 287–298). United Kingdom: Oxford University
Press.
Crocker, J., Canevello, A., & Lewis, K. A. (2017). Romantic relationships
in the ecosystem: Compassionate goals, nonzero-sum beliefs, and
change in relationship quality. Journal of Personality and Social Psy-
chology, 112, 58 –75. http://dx.doi.org/10.1037/pspi0000076
Cushman, F. (2008). Crime and punishment: Distinguishing the roles of
causal and intentional analyses in moral judgment. Cognition, 108,
353–380. http://dx.doi.org/10.1016/j.cognition.2008.03.006
Dalai Lama, & Ekman, P. (2008). Emotional awareness: Overcoming the
obstacles to psychological balance and compassion. New York, NY:
Henry Holt.
Darley, J. M., Carlsmith, K. M., & Robinson, P. H. (2000). Incapacitation
and just deserts as motives for punishment. Law and Human Behavior,
24, 659 – 683. http://dx.doi.org/10.1023/A:1005552203727
Davis, M. H. (1983). Measuring individual differences in empathy: Evi-
dence for a multidimensional approach. Journal of Personality and
Social Psychology, 44, 113–126. http://dx.doi.org/10.1037/0022-3514
.44.1.113
Decety, J., & Cowell, J. M. (2014). Friends or foes: Is empathy necessary
for moral behavior? Perspectives on Psychological Science, 9, 525–537.
http://dx.doi.org/10.1177/1745691614545130
DeSteno, D. (2015). Compassion and altruism: How our minds determine
who is worthy of help. Current Opinion in Behavioral Sciences, 3,
80 – 83. http://dx.doi.org/10.1016/j.cobeha.2015.02.002
Dreber, A., Rand, D. G., Fudenberg, D., & Nowak, M. A. (2008). Winners
don’t punish. Nature, 452, 348 –351. http://dx.doi.org/10.1038/nature
06723
Eisenberg, N., Fabes, R. A., Murphy, B., Karbon, M., Maszk, P., Smith,
M.,...Suh, K. (1994). The relations of emotionality and regulation to
dispositional and situational empathy-related responding. Journal of
Personality and Social Psychology, 66, 776 –797. http://dx.doi.org/10
.1037/0022-3514.66.4.776
Eisenberg, N., McCreath, H., & Ahn, R. (1988). Vicarious emotional
responsiveness and prosocial behavior: Their interrelations in young
children. Personality and Social Psychology Bulletin, 14, 298 –311.
http://dx.doi.org/10.1177/0146167288142008
Ekman, P. (1992). An argument for basic emotions. Cognition and Emo-
tion, 6, 169 –200. http://dx.doi.org/10.1080/02699939208411068
Eriksson, K., Strimling, P., Andersson, P. A., & Lindholm, T. (2017).
Costly punishment in the ultimatum game evokes moral concern, in
particular when framed as payoff reduction. Journal of Experimental
Social Psychology, 69, 59 – 64. http://dx.doi.org/10.1016/j.jesp.2016.09
.004
Falk, A., Fehr, E., & Fischbacher, U. (2003). On the nature of fair behavior.
Economic Inquiry, 41, 20 –26. http://dx.doi.org/10.1093/ei/41.1.20
Faul, F., Erdfelder, E., Buchner, A., & Lang, A. G. (2009). Statistical
power analyses using G
Power 3.1: Tests for correlation and regression
analyses. Behavior Research Methods, 41, 1149 –1160. http://dx.doi.org/
10.3758/BRM.41.4.1149
Fehr, E., & Fischbacher, U. (2004). Third-party punishment and social
norms. Evolution and Human Behavior, 25, 63– 87. http://dx.doi.org/10
.1016/S1090-5138(04)00005-4
Fetchenhauer, D., & Huang, X. (2004). Justice sensitivity and distributive
decisions in experimental games. Personality and Individual Differences,
36, 1015–1029. http://dx.doi.org/10.1016/S0191-8869(03)00197-1
Fiedler, K., Harris, C., & Schott, M. (2018). Unwarranted inferences from
statistical mediation tests—An analysis of articles published in 2015.
Journal of Experimental Social Psychology, 75, 95–102. http://dx.doi
.org/10.1016/j.jesp.2017.11.008
Fiedler, K., Schott, M., & Meiser, T. (2011). What mediation analysis can
(not) do. Journal of Experimental Social Psychology, 47, 1231–1236.
http://dx.doi.org/10.1016/j.jesp.2011.05.007
Fiske, A. P., & Tetlock, P. E. (1997). Taboo trade-offs: Reactions to
transactions that transgress the spheres of justice. Political Psychology,
18, 255–297. http://dx.doi.org/10.1111/0162-895X.00058
Gervais, M. M., & Fessler, D. M. T. (2017). On the deep structure of social
affect: Attitudes, emotions, sentiments, and the case of “contempt.”
Behavioral and Brain Sciences, 40, e225. http://dx.doi.org/10.1017/
S0140525X16000352
Goetz, J. L., Keltner, D., & Simon-Thomas, E. (2010). Compassion: An
evolutionary analysis and empirical review. Psychological Bulletin, 136,
351–374. http://dx.doi.org/10.1037/a0018807
Goh, J. X., Hall, J. A., & Rosenthal, R. (2016). Mini meta-analysis of your own
studies: Some arguments on why and a primer on how. Social and Person-
ality Psychology Compass, 10, 535–549. http://dx.doi.org/10.1111/spc3
.12267
Gordijn, E. H., Wigboldus, D., & Yzerbyt, V. (2001). Emotional conse-
quences of categorizing victims of negative outgroup behavior as in-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
16 PFATTHEICHER, SASSENRATH, AND KELLER
group or outgroup. Group Processes & Intergroup Relations, 4, 317–
326. http://dx.doi.org/10.1177/1368430201004004002
Graham, J., Nosek, B. A., Haidt, J., Iyer, R., Koleva, S., & Ditto, P. H.
(2011). Mapping the moral domain. Journal of Personality and Social
Psychology, 101, 366 –385. http://dx.doi.org/10.1037/a0021847
Haidt, J. (2003). The moral emotions. In R. J. Davidson, K. R. Scherer, &
H. H. Goldsmith (Eds.), Handbook of affective sciences (pp. 852– 870).
United Kingdom: Oxford University Press.
Hayes, A. F. (2013). Introduction to mediation, moderation, and condi-
tional process analysis: A regression-based approach. New York, NY:
Guilford Press.
Hechler, S., & Kessler, T. (2018). On the difference between moral outrage
and empathic anger: Anger about wrongful deeds or harmful conse-
quences. Journal of Experimental Social Psychology, 76, 270 –282.
http://dx.doi.org/10.1016/j.jesp.2018.03.005
Hoffman, M. L. (1990). Empathy and justice motivation. Motivation and
Emotion, 14, 151–172. http://dx.doi.org/10.1007/BF00991641
Hoffman, M. L. (2000). Empathy and moral development: Implications for
caring and justice. United Kingdom: Cambridge University Press. http://
dx.doi.org/10.1017/CBO9780511805851
Horberg, E. J., Oveis, C., & Keltner, D. (2011). Emotions as moral
amplifiers: An appraisal tendency approach to the influences of distinct
emotions upon moral judgment. Emotion Review, 3, 237–244. http://dx
.doi.org/10.1177/1754073911402384
Jacoby, J., & Sassenberg, K. (2011). Interactions do not only tell us when,
but can also tell us how: Testing process hypotheses by interaction.
European Journal of Social Psychology, 41, 180 –190. http://dx.doi.org/
10.1002/ejsp.762
Jordan, J. J., Hoffman, M., Bloom, P., & Rand, D. G. (2016). Third-party
punishment as a costly signal of trustworthiness. Nature, 530, 473– 476.
http://dx.doi.org/10.1038/nature16981
Jordan, J., McAuliffe, K., & Rand, D. (2016). The effects of endowment
size and strategy method on third party punishment. Experimental Eco-
nomics, 19, 741–763. http://dx.doi.org/10.1007/s10683-015-9466-8
Kant, I. (1998). Groundwork for the metaphysic of morals. United King-
dom: Cambridge University Press. (Original work published 1785)
http://dx.doi.org/10.1017/CBO9780511809590.004
Keller, J., & Pfattheicher, S. (2013). The compassion-hostility paradox:
The interplay of vigilant, prevention-focused self-regulation, compas-
sion, and hostility. Personality and Social Psychology Bulletin, 39,
1518 –1529. http://dx.doi.org/10.1177/0146167213499024
Klimecki, O. M., Leiberg, S., Ricard, M., & Singer, T. (2014). Differential
pattern of functional brain plasticity after compassion and empathy
training. Social Cognitive and Affective Neuroscience, 9, 873– 879.
http://dx.doi.org/10.1093/scan/nst060
Kühnen, U. (2010). Manipulation checks as manipulation: Another look at
the ease-of-retrieval heuristic. Personality and Social Psychology Bul-
letin, 36, 47–58. http://dx.doi.org/10.1177/0146167209346746
Lazarus, R. (1991). Emotion and adaptation. New York, NY: Oxford
University Press.
Lerner, M. J. (1980). Belief in a just world. A fundamental delusion. New
York, NY: Plenum Press. http://dx.doi.org/10.1007/978-1-4899-0448-5
Lupoli, M. J., Jampol, L., & Oveis, C. (2017). Lying because we care:
Compassion increases prosocial lying. Journal of Experimental Psychol-
ogy: General, 146, 1026 –1042. http://dx.doi.org/10.1037/xge0000315
Malle, B. F., & Knobe, J. (1997). The folk concept of intentionality.
Journal of Experimental Social Psychology, 33, 101–121. http://dx.doi
.org/10.1006/jesp.1996.1314
Mazziotta, A., Feuchte, F., Gausel, N., & Nadler, A. (2014). Does remem-
bering past ingroup harmdoing promote postwar cross-group contact?
Insights from a field-experiment in Liberia. European Journal of Social
Psychology, 44, 43–52. http://dx.doi.org/10.1002/ejsp.1986
McAuliffe, W. H. B., Forster, D. E., Philippe, J., & McCullough, M. E.
(2018). Digital altruists: Resolving key questions about the empathy-
altruism hypothesis in an Internet sample. Emotion, 18, 493–506. http://
dx.doi.org/10.1037/emo0000375
Mikula, G. (1993). On the experience of injustice. European Review of
Social Psychology, 4, 223–244. http://dx.doi.org/10.1080/14792779
343000077
Ministero, L. M., Poulin, M. J., Buffone, A. E. K., & DeLury, S. (2018).
Empathic concern and the desire to help as separable components of
compassionate responding. Personality and Social Psychology Bulletin,
44, 475– 491. http://dx.doi.org/10.1177/0146167217741345
Montada, L., & Schneider, A. (1989). Justice and emotional reactions to
the disadvantaged. Social Justice Research, 3, 313–344. http://dx.doi
.org/10.1007/BF01048081
Nelissen, R. M., & Zeelenberg, M. (2009). When guilt evokes self-
punishment: Evidence for the existence of a Dobby Effect. Emotion, 9,
118 –122. http://dx.doi.org/10.1037/a0014540
Nikiforakis, N. (2008). Punishment and counter-punishment in public good
games: Can we really govern ourselves? Journal of Public Economics,
92, 91–112. http://dx.doi.org/10.1016/j.jpubeco.2007.04.008
Nussbaum, M. (1996). Compassion: The basic social emotion. Social
Philosophy & Policy, 13, 27–58. http://dx.doi.org/10.1017/S0265052
500001515
Olderbak, S., Sassenrath, C., Keller, J., & Wilhelm, O. (2014). An
emotion-differentiated perspective on empathy with the emotion specific
empathy questionnaire. Frontiers in Psychology, 5, 653. http://dx.doi
.org/10.3389/fpsyg.2014.00653
Oveis, C., Horberg, E. J., & Keltner, D. (2010). Compassion, pride, and
social intuitions of self-other similarity. Journal of Personality and
Social Psychology, 98, 618 – 630. http://dx.doi.org/10.1037/a0017628
Pfattheicher, S., Böhm, R., & Kesberg, R. (2018). The advantage of
democratic peer punishment in sustaining cooperation within groups.
Journal of Behavioral Decision Making, 31, 562–571. http://dx.doi.org/
10.1002/bdm.2050
Pfattheicher, S., Sassenrath, C., & Schindler, S. (2016). Feelings for the
suffering of others and the environment: Compassion fosters proenvi-
ronmental tendencies. Environment and Behavior, 48, 929 –945. http://
dx.doi.org/10.1177/0013916515574549
Pinker, S. (2011). The better angels of our nature: The decline of violence
in history and its causes. London, UK: Penguin Books.
Pirlott, A. G., & MacKinnon, D. P. (2016). Design approaches to experi-
mental mediation. Journal of Experimental Social Psychology, 66, 29 –
38. http://dx.doi.org/10.1016/j.jesp.2015.09.012
Rand, D. G., & Nowak, M. A. (2013). Human cooperation. Trends in
Cognitive Sciences, 17, 413– 425. http://dx.doi.org/10.1016/j.tics.2013
.06.003
Robinson, P. H., & Darley, J. M. (1995). Justice, liability, and blame:
Community views and the criminal law. Boulder, CO: Westview Press.
Roth, G., Shane, N., & Kanat-Maymon, Y. (2017). Empathising with the
enemy: Emotion regulation and support for humanitarian aid in violent
conflicts. Cognition and Emotion, 31, 1511–1524. http://dx.doi.org/10
.1080/02699931.2016.1237348
Rothschild, Z. K., & Keefer, L. A. (2018). Righteous or self-righteous
anger? Justice sensitivity moderates defensive outrage at a third-party
harm-doer. European Journal of Social Psychology, 48, 507–522. http://
dx.doi.org/10.1002/ejsp.2349
Sassenrath, C., Hodges, S. D., & Pfattheicher, S. (2016). It’s all about the
self: When perspective taking backfires. Current Directions in Psycho-
logical Science, 25, 405– 410. http://dx.doi.org/10.1177/096372141
6659253
Sassenrath, C., Pfattheicher, S., & Keller, J. (2017). I might ease your pain,
but only if you’re sad: The impact of the role of empathized emotion in
the empathy-helping association. Motivation and Emotion, 41, 96 –106.
http://dx.doi.org/10.1007/s11031-016-9586-2
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
17
COMPASSION MAGNIFIES THIRD-PARTY PUNISHMENT
Schein, C., & Gray, K. (2018). The theory of dyadic morality: Reinventing
moral judgment by redefining harm. Personality and Social Psychology
Review, 22, 32–70. http://dx.doi.org/10.1177/1088868317698288
Scherer, K. R. (2005). What are emotions? And how can they be mea-
sured? Social Sciences Information/Information Sur les Sciences Socia-
les, 44, 695–729. http://dx.doi.org/10.1177/0539018405058216
Simonsohn, U. (2015). Small telescopes: Detectability and the evaluation
of replication results. Psychological Science, 26, 559 –569. http://dx.doi
.org/10.1177/0956797614567341
Singer, T., & Klimecki, O. M. (2014). Empathy and compassion. Current
Biology, 24, R875–R878. http://dx.doi.org/10.1016/j.cub.2014.06.054
Singer, T., & Steinbeis, N. (2009). Differential roles of fairness- and
compassion-based motivations for cooperation, defection, and punish-
ment. Annals of the New York Academy of Sciences, 1167, 41–50.
http://dx.doi.org/10.1111/j.1749-6632.2009.04733.x
Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005). Establishing a causal
chain: Why experiments are often more effective than mediational
analyses in examining psychological processes. Journal of Personality
and Social Psychology, 89, 845– 851. http://dx.doi.org/10.1037/0022-
3514.89.6.845
Stellar, J. E., Cohen, A., Oveis, C., & Keltner, D. (2015). Affective and
physiological responses to the suffering of others: Compassion and vagal
activity. Journal of Personality and Social Psychology, 108, 572–585.
http://dx.doi.org/10.1037/pspi0000010
Stellar, J. E., & Keltner, D. (2014). Compassion. In M. Tugade, L. Shiota,
& L. Kirby (Eds.), Handbook of positive emotion (pp. 329 –341). New
York, NY: Guilford Press.
Tetlock, P. E. (2002). Social functionalist frameworks for judgment and
choice: Intuitive politicians, theologians, and prosecutors. Psychological
Review, 109, 451– 471. http://dx.doi.org/10.1037/0033-295X.109.3.451
Tetlock, P. E., Kristel, O. V., Elson, S. B., Green, M. C., & Lerner, J. S.
(2000). The psychology of the unthinkable: Taboo trade-offs, forbidden
base rates, and heretical counterfactuals. Journal of Personality and
Social Psychology, 78, 853– 870. http://dx.doi.org/10.1037/0022-3514
.78.5.853
Thulin, E. W., & Bicchieri, C. (2016). I’m so angry I could help you: Moral
outrage as a driver of victim compensation. Social Philosophy & Policy,
32, 146 –160. http://dx.doi.org/10.1017/S0265052516000145
Vanman, E. J. (2016). The role of empathy in intergroup relations. Current
Opinion in Psychology, 11, 59 – 63. http://dx.doi.org/10.1016/j.copsyc
.2016.06.007
Vitaglione, G. D., & Barnett, M. A. (2003). Assessing a new dimension of
empathy: Empathic anger as a predictor of helping and punishing de-
sires. Motivation and Emotion, 27, 301–325. http://dx.doi.org/10.1023/
A:1026231622102
Von Hirsch, A. (1976). Doing justice: The choice of punishments. New
York, NY: Hill and Wang.
Weng, H. Y., Fox, A. S., Hessenthaler, H. C., Stodola, D. E., & Davidson,
R. J. (2015). The role of compassion in altruistic helping and punishment
behavior. PLoS ONE, 10, e0143794. http://dx.doi.org/10.1371/journal
.pone.0143794
Zaki, J. (2018). Empathy is a moral force. In K. Gray & J. Graham (Eds.),
Atlas of moral psychology (pp. 49 –58). New York, NY: Guilford Press.
Zhou, H., & Fishbach, A. (2016). The pitfall of experimenting on the web:
How unattended selective attrition leads to surprising (yet false) research
conclusions. Journal of Personality and Social Psychology, 111, 493–
504. http://dx.doi.org/10.1037/pspa0000056
Received January 29, 2018
Revision received August 27, 2018
Accepted August 27, 2018
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
18 PFATTHEICHER, SASSENRATH, AND KELLER
... In the third-party punishment paradigm (see Pfattheicher, Sassenrath, & Keller, 2019, for the same procedure), participants were told that "Your job is to judge a decision made by Person A. Person A has participated in a decision-making game including money. In this game, Person A has been randomly paired with another real person (called 'Person B' in the following). ...
... In the third-party punishment paradigm, by comparison, a minority of participants did not destroy the other person's payment (n = 287, 29.2%). This is in line with previous research on the third-party punishment paradigm, showing that a majority punish unfair other individuals (e.g., Pfattheicher, Sassenrath, & Keller, 2019). Over seventy percent (or 696 participants) reduced the Sadistic affect. ...
... Moreover, while we find that effects of boredom consist even after controlling for another negative emotion (e.g., frustration), research on sadism would benefit from including other negative mood states in analysis (e.g., anger or pain). Chester and DeWall (2017; see also Chester et al., 2019), for instance, found increased sadistic tendencies (both in sadistic affect and behavior) after individuals were socially rejected or provoked, while anger plays a strong role in social rejection and provocation (e.g., Chow et al., 2008;Pfattheicher, Sassenrath, & Keller, 2019). We hope that the present contribution stimulates future research that further advances our understanding of factors promoting sadistic tendencies. ...
Article
Full-text available
What gives rise to sadism? While sadistic behavior (i.e., harming others for pleasure) is well-documented, past empirical research is nearly silent regarding the psychological factors behind it. We help close this gap by suggesting that boredom plays a crucial role in the emergence of sadistic tendencies. Across nine diverse studies, we provide correlational and experimental evidence for a link between boredom and sadism. We demonstrate that sadistic tendencies are more pronounced among people who report chronic proneness to boredom in everyday life (Studies 1A-1F, N = 1780). We then document that this relationship generalizes across a variety of important societal contexts, including online trolling; sadism in the military; sadistic behavior among parents; and sadistic fantasies (Studies 2-5, N = 1740). Finally, we manipulate boredom experimentally and show that inducing boredom increases sadistic behavior (i.e., killing worms; destroying other participants’ pay; Studies 6-9, N = 4097). However, alternatives matter: When several behavioral alternatives are available, boredom only motivates sadistic behavior among individuals with high dispositional sadism (Study 7). Conversely, when there is no alternative, boredom increases sadistic behavior across the board, even among individuals low in dispositional sadism (Studies 8 & 9). We further show that excitement and novelty seeking mediate the effects of boredom, and that boredom not only promotes sadistic (proactive) aggression, but reactive aggression as well (Study 9). Overall, the present work contributes to a better understanding of sadism and highlights the destructive potential of boredom. We discuss implications for basic research on sadism and boredom, as well as applied implications for society at large.
... Measures. We measured affective empathy for those most vulnerable to the virus with three items, adapted from Pfattheicher et al. (2019; in all samples, Cronbach's α > .81). The items read, "I am very concerned about / I feel compassion for / I am quite moved by what can happen to … those most vulnerable to coronavirus (COVID-19)." ...
... State empathy. After the conditions, participants responded to three items assessing state empathy, as in previous research (Batson et al., 1997;Pfattheicher et al., 2019). The items read, "Right now, I am … compassionate," "… moved," "… touched" (α = .94). ...
Article
Full-text available
The COVID-19 pandemic presents a major challenge to societies all over the globe. To curb the spread of the disease, two measures implemented in many countries are minimizing close contact between people (“physical distancing”) and wearing face masks. In the present research, we tested the idea that physical distancing and wearing face masks can be the result of a prosocial emotional process — empathy for those most vulnerable to the virus. In four preregistered studies (N=3,718, Western population), we show that (i) empathy indeed relates to the motivation to adhere to physical distancing and to wearing face masks, and (ii) inducing empathy for those most vulnerable to the virus promotes the motivation to adhere to these measures (whereas merely providing information about its importance is not). In sum, the present research provides a better understanding of the promoting factors underlying the willingness to follow two important measures during the COVID-19 pandemic.
... Measures. We measured affective empathy for those most vulnerable to the virus with three items, adapted from Pfattheicher et al. (2019; in all samples, Cronbach's α > .81). The items read, "I am very concerned about / I feel compassion for / I am quite moved by what can happen to … those most vulnerable to coronavirus (COVID-19)." ...
... State empathy. After the conditions, participants responded to three items assessing state empathy, as in previous research (Batson et al., 1997;Pfattheicher et al., 2019). The items read, "Right now, I am … compassionate," "… moved," "… touched" (α = .94). ...
Preprint
Full-text available
The COVID-19 pandemic presents a major challenge to societies all over the globe. To curb the spread of the disease, two measures implemented in many countries are minimizing close contact between people (“physical distancing”) and wearing face masks. In the present research, we tested the idea that physical distancing and wearing face masks can be the result of a genuine prosocial emotion—empathy for those most vulnerable to the virus. In four pre-registered studies (total N = 3,718‬, Western population), we show that (i) empathy is indeed a basic motivation for physical distancing and wearing face masks, and (ii) inducing empathy for those most vulnerable to the virus promotes the motivation to adhere to these measures (whereas providing mere information about its importance is not). In sum, the present research provides a better understanding of the promoting factors underlying the willingness to follow two important measures during the COVID-19 pandemic.
... State empathy. After the item measuring vaccination intention, participants responded to three items assessing state empathy, as in previous research (Batson et al., 1997;Pfattheicher et al., 2019Pfattheicher et al., , 2020 ." Labels ranged from 1 = "strongly disagree" to 7 = "strongly agree." ...
Preprint
Full-text available
Objective: An effective vaccine against COVID-19 is a desired solution to curb the spread of the disease. However, vaccine hesitancy might hinder high uptake rates and thus undermine efforts to eliminate COVID-19 once an effective vaccine is available. The present contribution addresses this issue by examining two promising ways of increasing the intention to get vaccinated against COVID-19. Methods: We conducted two pre-registered online studies (N = 2,315 participants from the UK) in which we either measured (Study 1) or manipulated (Study 2) knowledge about and beliefs in herd immunity, as well as empathy for those most vulnerable to the virus. As a dependent variable, we assessed individuals’ self-reported vaccination intention if a vaccine against COVID-19 became available. Results: We show that the motivation to get vaccinated against COVID-19 is related to and causally promoted by both mere information about herd immunity and by empathy. Thus, interventions that combine cognitive and affective information related to others’ potential suffering appear most effective in increasing the intention to get vaccinated against COVID-19. Conclusions: The present research provides a better understanding of the intention to get vaccinated against COVID-19 and highlights two evidence-based possibilities for policymakers in promoting vaccine uptake.
... However, there are significant differences in the moral emotions between bystanders and victims. For example, victims' moral aversion results from losses of one's own wellbeing, while bystanders' aversion originates from one's beliefs in human equality and empathy (Pfattheicher, Sassenrath & Keller, 2019). Therefore, we further investigated the genetic foundations of victims' moral emotions. ...
Article
Full-text available
Immoral behaviors make individuals abominate and punish transgressors. Inspired by the associations between the Val66Met polymorphism of brain-derived neurotropic factor (BDNF) gene and emotional responses following negative events, we investigated whether this polymorphism was also associated moral emotions such as punishment and forgiveness following interpersonal transgression. To do so, we categorized 340 individuals according to the BDNF Val66Met and assessed moral emotions using 12 hypothetic scenarios with different conditions of intention and interpersonal consequence. The results indicated that this polymorphism was significantly associated with moral aversion and punishment towards transgressors. Victims with the Val/ Val genotype expressed less aversion and punishment than the Met carriers, regardless of intention and interpersonal consequence. Moreover, this polymorphism was associated with forgiveness. Victims with the Val/Val genotype expressed more forgiveness than the Met carriers. Taken together, these findings highlight the importance of the BDNF Val66Met to moral emotions.
... An important component of compassion is the desire to help and reduce the suffering of others (Goetz et al., 2010;Lazarus, 1991). In fact, there is a solid body of evidence showing that compassion increases prosocial, altruistic, helping, and donation behavior (Batson et al., 1981;Betancourt, 1990;DeSteno, 2015;Eisenberg & Miller, 1987;Sassenrath et al., 2017; see also Pfattheicher et al., 2019). Such beneficial consequences of compassion have been demonstrated in various contexts, including climate change (Lu & Schuldt, 2016), proenvironmental contexts (Pfattheicher et al., 2016), and hand hygiene behavior . ...
Article
Citizens can engage in wildlife conservation by participating in citizen science (CS) projects about wildlife. Interest in wildlife and science has been found to be one motivation for CS participation. Yet, we suggest that emotional responses, such as compassion for suffering, injured, or distressed wildlife, are relevant motivations that have so far been understudied. Compassion is known to increase behavioral intentions to alleviate suffering and, thus, is likely to have beneficial consequences for wildlife CS projects. Therefore, our two studies investigated the impact of different wildlife photographs on compassion, attitudes toward CS, and intentions to engage in CS. We found that photographs of distressed raccoons and foxes increased compassion, which thereby increased attitudes toward CS projects and some intentions to participate in these projects. Thus, compassion may be a relevant factor for increasing citizens’ engagement in wildlife conservation. We discuss the implications of our findings for CS and wildlife conservation.
... Wenn Achtsamkeit das Mitgefühl steigert, sollte einem Opfer z. B. von Gewalt verstärkt geholfen, aber gemäß neueren Befunden der Täter auch stärker bestraft werden ( Pfattheicher et al. 2019). Die theoretisch weiterführende Frage ist hier also nicht, ob Achtsamkeit prosoziales Verhalten fördert, sondern was die Randbedingungen dafür sind. ...
Article
Full-text available
Mindfulness is currently a hot topic in psychological research as well as in many parts of the society. This article aims to critically evaluate the current state of research on this topic. First, several methodological limitations regarding the empirical investigation of mindfulness become apparent, restricting valid conclusions about the actual effects. In addition, a research gap on undesirable consequences is identified. Potential negative effects of mindfulness in interpersonal and moral domains, for example, are theoretically plausible and are discussed. In sum more high quality research is needed in many areas.
... Other positive states can as well. For example, recent evidence suggests that compassion felt toward victims increases TPP (Pfattheicher, Sassenrath, & Keller, 2019). These moral emotions nudge people toward acts designed to reinforce cooperation through strengthening relationship bonds, norms for moral behaviour, or signals that a person can be counted on to favour fair outcomes. ...
Article
Third-party punishment occurs when a perpetrator of a transgression is punished by another person who was not directly affected by the transgression (i.e. a third-party). Given gratitude’s demonstrated ability to enhance both cooperation and the value people place on future-rewards, its capacity to increase third-party punishment – a phenomenon theorised to increase future cooperative behaviour – was investigated. In two experiments, participants were randomly assigned to experience one of three emotional states (i.e. gratitude, happiness, or neutrality) prior to making decisions about how much of a previous financial endowment they would spend to punish a person who transgressed against another at differing degrees within the context of a dictator game. As expected, punishment expenditures decreased for all participants as a dictator’s decision became fairer. Of primary interest, however, participants who felt grateful, as compared to those who felt neutral or happy, engaged in significantly more third-party punishment across dictator splits that were not altruistic in nature.
... Within the context of the experiment, highly empathic participants might have experienced more anger towards the offender when the norm violation of the unfair offender was highlighted. Supporting this rationale, a recent study has shown that the degree of empathic concern also serves as a positive predictor of third-party punishment, via the mediating role of moral outrage (Pfattheicher, Sassenrath, & Keller, 2019). Facing both help and punishment options, those with a higher level of empathic concern might have a stronger feeling of moral outrage especially when the norm violation becomes salient. ...
Article
Full-text available
Norm violations (e.g., unfair transgressions) are often met with punishment even by people who are not directly affected. However, punishing a transgressor is not the only option for a bystander to restore justice. Empathic concerns may dictate instead to give a helping hand to a victim. Using a pre‐registered, fully incentivized eye‐tracking study (N = 47), we investigated the cognitive mechanism linking bystanders’ empathic concern and justice‐restoring intervention behaviour. The results show that not only the decision to intervene (i.e., either costly compensating the victim or punishing the transgressor) but also the attention directed towards a victim’s payoffs (i.e., measured by the proportion of fixations) during the decision‐making period systematically varied with the individual level of empathic concern. Exploring this link further, we additionally instructed participants to focus on specific components of the norm violation, namely the (un)fair conduct of the offender or the victim's feelings. Surprisingly, highly empathic bystanders were more likely to punish the offender when the norm violation was highlighted. However, we did not observe the modulation of the instructed focus on the link between gaze‐based measures and empathic concern. Overall, these results provide initial evidence about the interacting impact of empathic concern as well as the focus on specific components of the norm violation when bystanders respond to unfair transgressions.
Article
Full-text available
Moral violations seem to elicit moral outrage because of the wrongfulness of the deed. However, recent studies have questioned the existence of moral outrage, because moral violations are confounded with the harm done to victims. Such harm elicits empathic anger rather than moral outrage (Batson et al., 2007; Batson et al., 2009). Thus, moral outrage is triggered by the wrongfulness of an action (i.e., a perpetrator's intention to harm), whereas empathic anger is triggered by its harmfulness (i.e., the actual harm done). Four studies (N = 1065) in varying contexts orthogonally crossed these antecedents of anger to differentiate between moral outrage and empathic anger. The results demonstrate that anger mainly emerged from the intention to harm, rather than the actual harm done. In contrast, the actual harm elicited empathy with victims. The findings suggest that anger about moral violations emerges separately from empathic reactions, although these reactions are difficult to distinguish in most instances. Likewise, the intention to harm provoked a willingness to punish the perpetrator much more than the actual harm did. Moral violations thus elicit moral outrage independently of their harmful consequences, even though such anger may often overlap with concern for others.
Article
Full-text available
While bystanders' outrage over moral transgressions may represent a genuine desire to restore justice, such expressions can also be self-serving—alleviating guilt and bolstering one's moral status. Four studies examined whether individual differences in observer justice sensitivity (JSO) moderate the degree to which outrage at third-party harm-doing reflects concerns about one's own moral identity rather than justice per se. Among participants low (vs. high) in JSO, feelings of guilt predicted greater outrage and desire to punish a corporation's sweatshop labor practices (Studies 1 & 2). Furthermore, affirming one's personal moral identity reduced outrage and support for punishing a corporate harm-doer among those low, but not high in JSO (Studies 3 & 4). Similar moderation was absent for other forms of justice sensitivity and just world beliefs. Effects were not explained by negative affect, empathy, personal harm, or political orientation. Results suggest that JSO uniquely differentiates defensive and justice-driven moral outrage. (150/150).
Article
Full-text available
We argue that in addition to the positive effects and functionality of morality for interactions among in-group members as outlined in the target article, morality may also fuel aggression and conflict in interactions between morality-based out-groups. We summarize empirical evidence showing that negative cognitions, emotions, and behaviors are particularly likely to appear between out-groups with opposing moral convictions.
Article
Full-text available
In social dilemma situations, individuals benefit from uncooperative behavior while exploiting resources of the collective. One prominent solution to prevent uncooperative behavior and to increase cooperation is to establish a system of costly peer punishment, that is, the possibility for every individual involved in the dilemma to impose costly punishment on interaction partners. However, recent research revealed that, in contrast to a situation without punishment, peer punishment is inefficient and maladaptive in the sense that the total payoff is reduced and punishment of cooperative individuals (i.e., antisocial punishment) is possible. In the present work, we propose that a system of democratic peer punishment, that is, direct and equal participation of each individual in the punishment decision-making process with punishment only executed when a majority has voted for its execution, can address the shortcomings of a peer punishment system. Using iterated public goods games, we show higher cooperation levels, higher total payoffs and reduced executed punishment in the democratic compared to a peer punishment system. Moreover, we document that fairness perceptions, satisfaction, and interpersonal trust are increased in the democratic punishment system. Implications for how cooperation and democratic punishment systems may evolve are discussed.
Article
Full-text available
Researchers have identified the capacity to take the perspective of others as a precursor to empathy-induced altruistic motivation. Consequently, investigators frequently use so-called perspective-taking instructions to manipulate empathic concern. However, most experiments using perspective-taking instructions have had modest sample sizes, undermining confidence in the replicability of results. In addition, it is unknown whether perspective-taking instructions work because they increase empathic concern or because comparison conditions reduce empathic concern (or both). Finally, some researchers have found that egoistic factors that do not involve empathic concern, including self-oriented emotions and self-other overlap, mediate the relationship between perspective-taking instructions and helping. The present investigation was a high-powered, preregistered effort that addressed methodological shortcomings of previous experiments to clarify how and when perspective-taking manipulations affect emotional arousal and prosocial motivation in a prototypical experimental paradigm administered over the internet. Perspective-taking instructions did not clearly increase empathic concern; this null finding was not due to ceiling effects. Instructions to remain objective, on the other hand, unequivocally reduced empathic concern relative to a no-instructions control condition. Empathic concern was the most strongly felt emotion in all conditions, suggesting that distressed targets primarily elicit other-oriented concern. Empathic concern uniquely predicted the quality of social support provided to the target, which supports the empathy-altruism hypothesis and contradicts the role of self-oriented emotions and self-other overlap in explaining helping behavior. Empathy-induced altruism may be responsible for many prosocial acts that occur in everyday settings, including the increasing number of prosocial acts that occur online.
Article
Full-text available
Empathic failures are common in hostile intergroup contexts; repairing empathy is therefore a major focus of peacebuilding efforts. However, it is unclear which aspect of empathy is most relevant to intergroup conflict. Although trait empathic concern predicts prosociality in interpersonal settings, we hypothesized that the best predictor of meaningful intergroup attitudes and behaviors might not be the general capacity for empathy (i.e., trait empathy), but the difference in empathy felt for the in-group versus the out-group, or “parochial empathy.” Specifically, we predicted that out-group empathy would inhibit intergroup harm and promote intergroup helping, whereas in-group empathy would have the opposite effect. In three intergroup contexts—Americans regarding Arabs, Hungarians regarding refugees, Greeks regarding Germans—we found support for this hypothesis. In all samples, in-group and out-group empathy had independent, significant, and opposite effects on intergroup outcomes, controlling for trait empathic concern.
Article
Full-text available
The nature of harm—and therefore moral judgment—may be misunderstood. Rather than an objective matter of reason, we argue that harm should be redefined as an intuitively perceived continuum. This redefinition provides a new understanding of moral content and mechanism—the constructionist Theory of Dyadic Morality (TDM). TDM suggests that acts are condemned proportional to three elements: norm violations, negative affect, and—importantly—perceived harm. This harm is dyadic, involving an intentional agent causing damage to a vulnerable patient (A→P). TDM predicts causal links both from harm to immorality (dyadic comparison) and from immorality to harm (dyadic completion). Together, these two processes make the “dyadic loop,” explaining moral acquisition and polarization. TDM argues against intuitive harmless wrongs and modular “foundations,” but embraces moral pluralism through varieties of values and the flexibility of perceived harm. Dyadic morality impacts understandings of moral character, moral emotion, and political/cultural differences, and provides research guidelines for moral psychology.
Article
Recent attempts to improve on the quality of psychological research focus on good practices required for statistical significance testing. The scrutiny of theoretical reasoning, though superordinate, is largely neglected, as exemplified here in a common misunderstanding of mediation analysis. Although a test of a mediation model X ➔ Z ➔ Y is conditional on the premise that the model applies, alternative mediators Z′, Z″, Z‴ etc. remain untested, and other causal models could underlie the correlation between X, Y, Z, researchers infer from a single significant mediation test that they have identified the true mediator. A literature search of all mediation analyses published in 2015 in Sciencedirect shows that the vast majority of studies neither consider alternative causal models nor alternative mediator candidates. Ignoring that mediation analysis is conditional on the truth of the focal mediation model, they pretend to have demonstrated that Z mediates the influence of X on Y. Recommendations are provided for how to overcome this dissatisfying state of affairs.
Article
When do people experience versus regulate responses to compassion-evoking stimuli? We hypothesized that compassionate responding is composed of two factors—empathic concern and the desire to help—and that these would be differentially affected by perspective taking and self-affirmation. Exploratory (Study 1) and confirmatory (Study 2) factor analyses indicated that a compassion measure consisted of two factors corresponding to empathic concern and the desire to help. In Study 1 (N = 237), participants with high emotion regulation ability reported less empathic concern for multiple children than for one, but perspective taking prevented this effect. In Study 2 (N = 155), participants reported less desire to help multiple children, but only in the presence of self-affirmation. In both the studies, empathic concern predicted greater distress while the desire to help predicted greater chances of donating. Compassionate responding may consist of two separable facets that collapse under distinct conditions and that predict distinct outcomes.
Article
Prosocial lies, or lies intended to benefit others, are ubiquitous behaviors that have important social and economic consequences. Though emotions play a central role in many forms of prosocial behavior, no work has investigated how emotions influence behavior when one has the opportunity to tell a prosocial lie—a situation that presents a conflict between two prosocial ethics: lying to prevent harm to another, and honesty, which might also provide benefits to the target of the lie. Here, we examine whether the emotion of compassion influences prosocial lying, and find that compassion causally increases and positively predicts prosocial lying. In Studies 1 and 2, participants evaluated a poorly written essay and provided feedback to the essay writer. Experimentally induced compassion felt toward the essay writer (Study 1) and individual differences in trait compassion (Study 2) were positively associated with inflated feedback to the essay writer. In both of these studies, the relationship between compassion and prosocial lying was partially mediated by an enhanced importance placed on preventing emotional harm. In Study 3, we found moderation such that experimentally induced compassion increased lies that resulted in financial gains for a charity, but not lies that produced financial gains for the self. This research illuminates the emotional underpinnings of the common yet morally complex behavior of prosocial lying, and builds on work highlighting the potentially harmful effects of compassion—an emotion typically seen as socially beneficial.