Cushman,?? F.,?? Gray,?? K.,?? Gaffey,?? A.?? &?? Mendes,?? W.B.?? (in?? press)?? Simulating?? Murder:?? The?? aversion?? to?? harmful?? action.?? ?? Emotion?? ?? 1??
Simulating murder: The aversion to harmful action
Fiery Cushman1, Kurt Gray2, Allison Gaffey3 & Wendy Berry Mendes4
1Harvard University, 2The University of Maryland, 3University of Notre Dame 4University of California, San
Diverse lines of evidence point to a basic human aversion to physically harming others. First, we demonstrate that
unwillingness to endorse harm in a moral dilemma is predicted by individual differences in aversive reactivity, as
indexed by peripheral vasoconstriction. Next, we tested the specific factors that elicit the aversive response to harm.
Participants performed actions such as discharging a fake gun into the face of the experimenter, fully informed that
the actions were pretend and harmless. These simulated harmful actions increased peripheral vasoconstriction
significantly more than did witnessing pretend harmful actions or to performing metabolically-matched non-harmful
actions. This suggests that the aversion to harmful actions extends beyond empathic concern for victim harm.
Together, these studies demonstrate a link between the body and moral decision making processes.
People are averse to performing harmful actions
and often consider it morally wrong to harm a person
even when it would save many more lives (Mikhail,
2000; Petrinovich, O'Neill, & Jorgensen, 1993). Even
front-line soldiers trained and motivated to kill often
deliberately miss visible enemy targets (Grossman,
1995). This aversion to harm is essential to ordinary
human functioning, as evidenced by the antisocial
behavior of psychopaths, who are argued to lack it
(Blair, 1995). The aversion to harming others is so
basic to our moral sense that it is easy to miss an
important question: What is its psychological basis?
Our first experiment examines the link between
physiological responses and answers on a classic moral
dilemma: whether it is allowable to kill someone in
order to save many lives. We measured physiological
reactivity, linked to general aversive states, during a
non-moral task, and then examined whether it
predicted advocating the death of one person to save
others. We use autonomic changes, specifically
changes in total peripheral resistance (TPR), which are
associated with negative stress responses (Gregg,
James, Matyas, & Thorsteinsson, 1999; Mendes,
Blascovich, Hunter, Lickel, & Jost, 2007). Past
research indicates that the aversion to harm in such
moral dilemmas involves an affective component (e.g.
Greene, Sommerville, Nystrom, Darley, & Cohen,
2001, Koenigs et al., 2007; Mendez et al., 2005,
Moretto, Làdavas, Mattioli, & di Pellegrino, 2010), but
does not establish a link between the specific aversion
to harm in moral dilemmas and general aversive
reactivity in non-moral situations.
Our second experiment builds on this finding,
asking why people find the performance of harmful
actions aversive. First, aversion may stem from
empathic concern for the welfare of the victim
(Crockett et al., 2010; Hoffman, 2000; Mehrabian &
Epstein, 1972; Pizarro, 2000). For instance, we might
be averse to punching another because considering the
victim’s pain causes us psychological distress (Batson
et al., 2003; Gray, Gray & Wegner, 2007; Singer, et al.,
2004). Importantly, victim distress is not an intrinsic
property of an action itself, but rather of its expected
outcome. We call this the “outcome aversion” model:
people are averse to harmful acts because of empathic
concern for victim distress.
In addition, an aversive response might be
triggered by the basic perceptual and motoric
properties of an action, even without considering its
outcome. Blair (1995) suggests a mechanism by which
harmful actions themselves can become aversive:
When the unconditioned aversive stimulus of victim
distress (e.g. crying) is repeatedly paired with a
particular action (e.g. pushing or hitting a person),
those actions acquire a conditioned aversive response.
On this “action aversion” model, empathy is critical to
the acquisition of the aversive response to harmful
actions, but the conditioned response may
subsequently be evoked by intrinsic properties of the
Study 2 tests for “action aversion” by examining
participants’ physiological responses while either
performing or witnessing harmful actions (e.g.,
stabbing an experimenter with a rubber knife,
shooting him with a disabled handgun, etc.), or
performing similar but harmless actions (e.g., slicing a
pretend loaf of bread with a knife). Our use of
simulated actions follows past research demonstrating
that pretend stimuli can be sufficient to elicit strong
psychological responses (Rozin, Millman, &
Cushman,?? F.,?? Gray,?? K.,?? Gaffey,?? A.?? &?? Mendes,?? W.B.?? (in?? press)?? Simulating?? Murder:?? The?? aversion?? to?? harmful?? action.?? ?? Emotion?? ?? 2??
Nemeroff, 1986). Action aversion predicts a robust
aversive response to pretend actions with motoric and
perceptual properties of actual harmful behaviors even
though the “perpetrator” knows that no harm will
occur, whereas outcome aversion does not.
Additionally, action aversion predicts a greater aversive
response to performing harm than witnessing it
(because only the former involves an action), whereas
outcome aversion predicts an equal aversive response
in both cases (because they yield the same outcome).
In summary, in Study 1 we test the relationship
between the moral judgment of harmful actions and
general TPR reactivity to a non-moral task. In Study 2
we test whether simulated harms specifically trigger
TPR reactivity. Additionally, we compare reactivity
for performing versus witnessing simulated harm,
testing whether reactivity depends on the anticipation
of a harmful outcome versus the performance of a
Study 1 examines the relationship between threat
reactivity and responses to a classic moral dilemma.
We tracked changes in TPR during a stressful
arithmetic task and predicted that individuals
exhibiting greater TPR reactivity would be less willing
to endorse harming one person in order to save the
lives of several others.
We recruited 108 healthy participants (81 female)
aged 19-40 years (median 24). After obtaining consent,
an experimenter applied sensors that measured
impedance cardiography (HIC 2500, Chapel Hill,
NC)), electrocardiography (Biopac ECG module,
Goleta, CA), and blood pressure responses (Colin
Prodigy II, San Antonio, TX). Impedance
cardiographic and electrocardiography signals were
sampled at 1000Hz and integrated with a Biopac
MP150. Post-acquisition waveforms were scored using
Mindware software (IMP 3.0) by trained research
assistants (see Mendes, 2009). We estimated TPR
using the standard formula:
TPR = (Mean arterial pressure/Cardiac output) x 80.
After baseline participants met a new experimenter
who asked them to count backwards quickly in steps
of 7 from a four digit number. Mental arithmetic is a
common laboratory stress task that can evoke
increases in sympathetic nervous system responding.
We used TPR change during the first minute of the
stress task as our indication of threat reactivity,
subtracting the last minute of the baseline period from
the first minute of the stress task.
Participants were recruited as part of a larger study
on physiological changes associated with emotion and
body manipulations, analyses of which are beyond the
scope of the present study. Here we report
physiological responses that occur prior to our
emotion manipulations. The body position
manipulation (leaning forward versus leaning back)
was introduced before the stress task, but had no
effect on the early physiological responses we analyze
Participants were then provided a packet of
questionnaires that included a moral dilemma. It
asked the participant to imagine being on a lifeboat
that would sink – killing all onboard –unless someone
is thrown off. One person on the lifeboat is “leaning
over the side.” Participants were asked, “Is it morally
acceptable for you to push this person overboard in
order to save the lives of the remaining passengers”,
indicating yes/no. Then they were asked, “Please
indicate how morally acceptable it would be for you to
throw this person overboard in order to save the lives
of the remaining passengers” on a seven point scale (1
= “completely unacceptable”, 7= “completely
Results and Discussion
Measurement of TPR requires several high-quality,
artifact-free physiological signals. Twenty participants
were excluded for low-quality impedance or
electrocardiograph waveform, twenty-three because
blood pressure measurements were not obtained
during the first minute of the task, three because TPR
reactivity scores differed by more than two standard
deviations from the mean, and another nine dropped
out of the study prior to the assessment of the moral
dilemmas. Responses to moral dilemmas did not
significantly differ between those with usable TPR
data to those without, t (106) = 0.60, ns.
As predicted, increased TPR reactivity2 was
reliably associated with lesser endorsement of pushing
a person overboard in our moral judgment task r = -
.31, N = 51, p < .05, although not when rated
dichotomously, t(40) < .01, ns. This result was robust
after controlling for the experimental manipulations of
posture, affect, their interaction, gender, and age, r = -
.31, N = 51, p < .05.
The observed correlation between moral
judgment and TPR reactivity is consistent with our
Cushman,?? F.,?? Gray,?? K.,?? Gaffey,?? A.?? &?? Mendes,?? W.B.?? (in?? press)?? Simulating?? Murder:?? The?? aversion?? to?? harmful?? action.?? ?? Emotion?? ?? 3??
prediction that unwillingness to endorse harmful
action is linked with threat reactivity. This suggests
that the aversion to harmful actions may be
instantiated physiologically. However, it leaves open
the basis of this response: Does it depend upon
empathy for an actual victim, or also upon the
perceptual and motoric properties of the action itself?
whether physiological aversion can be triggered by
only the motor or perceptual properties of harmful
action. Participants were asked to perform five
simulated harmful actions, to witness another person
perform them, or to perform five simulated non-
harmful actions. We tested two predictions of the
action aversion hypothesis: simulated harmful actions
would elicit aversive reactivity despite the absence of
any harmful outcome, and this response would be
greater when performing the action than when
witnessing the action.
We recruited 108 participants (69 female) aged 18-
35 years (median 20). Participants initially consented to
a study of “pretend actions” omitting any mention of
harm. Participants completed twenty items from the
Positive and Negative Affect Schedule (PANAS;
Watson, Clark, & Tellegen, 1988) and then we applied
sensors from two different ambulatory devices that
allowed free movement during testing: VU-AMS
impedance/ECG (Amsterdam, The Netherlands), and
SpaceLab blood pressure (ABP 90207, Issaquah, WA).
The experimenter measured baseline physiological
responses during 5min rest, and then brought
participants into a room where she described the full
experimental procedure, obtained additional informed
consent, and initiated the pretend actions. The
experimenter emphasized that the participant was free
to omit any actions.
A male research assistant played the role of victim
in the “perform harm” and “witness harm”
conditions. Participants in the “perform harm”
condition were asked to perform five actions in a fixed
order: (1) smashing the victim’s shin with a hammer –
a PVC pipe was worn under a fake pant leg, (2)
smashing the victim’s hand with a rock – a rubber
hand was placed at the cuff of the shirt and the actual
hand was obscured from sight, (3) discharging a
handgun into the victim’s face – a weighty metal
replica, (4) drawing a knife across the victim’s throat –
Study 2 tested the action aversion hypothesis –
a rubber knife, and (5) smacking a baby against the
table – we used a realistic looking baby doll (see Figure
1). No verbal communication occurred between the
participant and the victim, and the victim avoided eye
contact except during the action itself. The victim
grimaced slightly during each action, but exhibited no
The experimenter initiated each action by
describing it to the subject and emphasizing that it was
pretend and no harm would occur. The participant
was instructed to contemplate performing the action
for one minute while holding the relevant implement
(e.g. gun), during which the experimenter exited the
room. The experimenter returned and said, “It is time
to perform the action.” After the subject performed
the action (or chose not to) the experimenter closed a
curtain between the participant and the victim,
instructed the participant to sit quietly for one minute,
and exited. This sequence was repeated for each
The “witness harm” condition proceeded
identically except that the participant was introduced
to two additional experimenters who assumed the
roles of perpetrator and victim. The participant heard
an identical description of the event to be performed,
and then contemplated watching that event for one
minute. The experimenter returned to the room and
asked the perpetrator to “harm” the victim, which the
perpetrator did directly in front of the subject with
neutral affect. The perpetrator and the victim were
then masked by a curtain during a one minute post-
action period. The “no harm” condition also
proceeded identically except that there were no
additional experimenters, and the participant was
asked to perform five metabolically-controlled pretend
actions: (1) hammering an imaginary nail on a block of
wood, (2) using a rock to smash a (rubber) nut, (3)
using a spray bottle to mist an imaginary plant, (4)
using a rubber knife to cut a (cardboard) loaf of bread,
and (5) smacking a hand broom against a table to
shake out dust.
After the pretend actions, participants
returned to the original room and sensors were
removed. Participants completed several
questionnaires including the 20-item PANAS and 5
hypothetical moral dilemmas drawn from previous
research (Greene et al., 2001). These moral dilemmas
asked participants whether they would perform a
harmful action in order to save many lives (e.g.
whether to smother one’s own crying baby in order to
successfully hide the whole family from enemy
Cushman,?? F.,?? Gray,?? K.,?? Gaffey,?? A.?? &?? Mendes,?? W.B.?? (in?? press)?? Simulating?? Murder:?? The?? aversion?? to?? harmful?? action.?? ?? Emotion?? ?? 4??
During the experiment impedance and
electrocardiograph were monitored continuously, but
BP measurements were manually initiated by the
experimenter with a key press. We took BP readings at
the first, third and fifth actions, and the timing of
these (before versus after the action) was varied
between participants. Thus, for each subject we
calculated three TPR reactivity scores.
One subject elected not to perform any
actions and was excluded from all analyses. Thirty-
three participants were excluded from physiological
analyses due to equipment malfunction, experimenter
error, or biologically implausible measurements.
Individual data points greater than two standard
deviations from the group mean were excluded from
analysis. Several participants were excluded from
behavioral analysis because of incomplete responses to
post-test survey items.
An ANCOVA of post task negative affect
(controlling for pre-task scores) yielded a significantly
effect F(2, 101) = 16.90, p < .001. Simple effects
revealed increased negative affect for perform harm (p
< .001) and witness harm (p < .001) compared with no
harm, but no difference between perform and witness
conditions (p > .25).
We used the general linear model to test for
effects of condition (perform vs. witness vs. no harm)
and measurement period (pre-action contemplation vs.
post-action recovery) on TPR reactivity, treating
participant as a random effect and employing robust
standard errors. There was a significant effect of
measurement period β1= -.31, z= -3.30, p<.01: Across
all three conditions, TPR reactivity was greater pre-
action than post-action. We treated the perform harm
condition as the comparison condition, and found that
TPR reactivity was significantly greater than for
witness harm β2= -.23, z= -2.08, p<.05 and no harm
β3= -.26, z= -2.65, p<.01 (Table 1). Additional
analyses revealed no significant effect of order (earlier
versus later actions) on TPR reactivity, and no
significant interactions between condition,
measurement period, or task order.
We performed a supplementary analysis of
TPR reactivity taken from the very first pre-action
contemplation period, a point at which participants
had anticipated the task, but had not performed or
witnessed any action. TPR reactivity for the perform
condition (M=89.8) was significantly higher than for
the witness condition (M=30.7) β= -.37, t=2.36, p<.05
and the no harm condition (M=29.9) β=-.37, t=2.34,
Paralleling our analysis in Study 1, we assessed
whether greater TPR reactivity was associated with
lesser endorsement of harming one person in order to
save several others. We calculated a summary TPR
score for each participant, averaging their reactivity
measurements and adjusting post-action TPR scores
to match pre-action TPR scores according to the
relevant coefficient (β1) of the GLM presented above.
We then correlated TPR reactivity with the mean
judgment across five moral dilemmas. Collapsed
across all three conditions this correlation was
significant r= -.32, N=71 p<.01. The relationship was
larger and significant for the witness harm condition
r= -.49, 95% CI -.12- -.87, N=25, p<.05, smaller and
non-significant correlation for the perform harm
condition r= -.36, CI -.06-.78, N=23, p<.10, and
smallest and non-significant for the no harm condition
r= -.24, CI -.20-.68, N=23, p<.28. Analysis of variance
revealed no significant effect of condition on mean
moral judgment F(2, 97)=0.75, p=.47. Even though
our condition effects did not influence later moral
judgments on hypothetical scenarios, individual
differences in TPR reactivity predicted moral
judgments as in Study 1.
We investigated individuals’ aversion to harmful
actions. Study 1 demonstrated that individuals
exhibiting greater threat reactivity were less likely to
endorse harm in order to save lives. This finding
corroborates past evidence suggesting a more potent
aversive response to the idea of performing direct
harm than to allowing indirect harm to more distant
others. It also suggests that individual differences in
total peripheral resistance during non-stressful tasks
can predict this aversion to harm.
Study 2 investigated the psychological basis of
this aversive response. Performing simulated harmful
actions evoked robust TPR reactivity despite
participants’ full awareness that no actual harm would
be caused. TPR reactivity was lower among
participants asked to witness harmful actions or to
perform metabolically-controlled non-harmful actions.
Moreover, TPR reactivity differed between conditions
Cushman,?? F.,?? Gray,?? K.,?? Gaffey,?? A.?? &?? Mendes,?? W.B.?? (in?? press)?? Simulating?? Murder:?? The?? aversion?? to?? harmful?? action.?? ?? Emotion?? ?? 5??
during the very first pre-action measurement period,
prior to participants performing or witnessing any
action at all. Thus, simply contemplating performing a
simulated harmful action leads to greater
vasoconstriction than contemplating witnessing a
These findings suggest an aversion to performing
harmful actions that extends beyond the expectation
of a harmful outcome. Clearly, TPR reactivity to
pretend harmful action (such as hitting a plastic baby
doll) cannot be attributed to explicit belief that harm
will occur. Nevertheless, pretend events could trigger
the imagination of harmful outcomes. Critically,
however, outcome aversion predicts similar affective
states for witnessing and performing harm, while
action aversion predicts a unique aversive response to
performing harm, as we observed. We therefore
consider it unlikely that the TPR reactivity associated
with performing simulated harmful actions was caused
solely by consideration of a harmful outcome, such as
empathic concern for victim distress.
Our findings do not contradict the role of
empathy and victim distress on the aversion to
harmful actions. These elements may be especially
important for the performance of “real” harmful
actions with actual consequences; moreover, they may
play a key developmental acquisition of action
aversion via associative pairing (Blair, 1995). Just as
you cannot help but swoon when smelling the
perfume or cologne associated with your first love,
people cannot help but feel upset when doing actions
typically associated with victim distress. Yet, as
important as the aversion to victim distress may be,
our results suggest a dissociable aversion based on
A forceful, automatic aversive response to the
surface properties of harmful actions may explain
otherwise puzzling human behaviors. In battlefield
behavior and hypothetical moral judgment people
resist doing direct harm despite explicit knowledge
that it could save many lives. Similarly, in our study,
people experienced a strong aversive response to
performing pretend harmful actions despite the
explicit knowledge that no harm would be caused.
These cases highlight a dissociation between our
explicit knowledge of the consequences of our actions
and our automatic affective responses to actions
(Dayan & Niv, 2008; Kahneman, 2003; LeDoux,
The action aversion model also suggests a
darker side: when banal or novel actions lack motoric
and perceptual properties associated with harm, they
may fail to trigger an aversive response. Signing one’s
name to a torture order or pressing the button that
releases a bomb each have real, known consequences
for other people, but as actions they lack salient
properties reliably associated with victim distress. A
notable parallel is evident in moral judgment: People
consider it morally worse to cause harm through direct
physical engagement than at a distance (Cushman,
Young & Hauser, 2006; Greene, Cushman, Stewart,
Lowenberg, Nystrom & Cohen, 2009). We
demonstrate that TPR reactivity increases during
(pretend) harmful actions and also correlates with
judgments of moral dilemmas. Yet, while
circumstantial evidence implicates a role for action
aversion in moral judgment, further research is
As such, our study highlights the advantage of
taking lessons from hypothetical moral dilemmas and
translating them into more active behaviors. Few past
studies directly target the human aversion to harm
using an active behavioral paradigm (Martens, Kosloff,
Greenberg, Landau, & Schmader, 2007; Milgram,
1974). This is no surprise: it is hard to get one person
to harm another ethically and in a laboratory.
Moreover, past studies often targeted situational
factors that promote harm, rather than the affective
systems that discourage it. To ask why people do harm
is a critical research question; our complementary
question is why people do not. Our study suggests that
the use of simulated harmful actions is sufficient to
generate an aversive response. Surely this response is
weaker than the aversion experienced by a solider on
the battlefield, or the captain of a sinking ship;
nevertheless, the aversion to simulated harm in the
laboratory may provide insight into the psychology
underlying the aversion to actual harm in the world
We thank our undergraduate research assistants
and other members of the Health and Physiology
Laboratory and Moral Cognition Laboratory for their
assistance in collecting the data. This research was
supported by funding from NHLBI grant (HL079383)
and the Mind, Brain and Behavior Initiative at Harvard
Cushman,?? F.,?? Gray,?? K.,?? Gaffey,?? A.?? &?? Mendes,?? W.B.?? (in?? press)?? Simulating?? Murder:?? The?? aversion?? to?? harmful?? action.?? ?? Emotion?? ?? 6??
Batson, D.C., Lishner, D.A., Carpenter, A., Dulin, L.,
Harjusola-Webb, S., Stocks, E.L., Gale, S., Hassan,
O. & Sampat, B. (1993). “…As you would have
them do onto you”: Does imagining yourself in
the other’s place stimulate moral action?
Personality and Social Psychology Bulletin, 29(9), 1190-
Blair, R. J. R. (1995). A cognitive developmental
approach to morality: investigating the
psychopath. Cognition, 57, 1-29.
Cushman, F. A., Young, L., & Hauser, M. D. (2006).
The role of conscious reasoning and intuitions in
moral judgment: Testing three principles of harm.
Psychological Science, 17(12), 1082-1089.
Crockett, M., Clark, L., Hauser, M., & Robbins, T.
(2010). Serotonin selectively influences moral
judgment and behavior through effects on harm
aversion. Proceedings of the National Academy of
Sciences, 107(40), 17433.
Dayan, P., & Niv, Y. (2008). Reinforcement learning:
The good, the bad and the ugly. Current Opinion in
Neurobiology, 18, 185-196.
Gray H.M., Gray K., & Wegner D.M. (2007).
Dimensions of mind perception. Science,
Greene, J. D., Cushman, F. A., Stewart, L. E.,
Lowenberg, K., Nystrom, L. E., & Cohen, J. D.
(2009). Pushing moral buttons: The interaction
between personal force and intention in moral
judgment. Cognition, 111(3), 364-371.
Greene, J. D., Sommerville, R. B., Nystrom, L. E.,
Darley, J. M., & Cohen, J. D. (2001). An fMRI
investigation of emotional engagement in moral
judgment. Science, 293, 2105-2108.
Grossman, D. (1995). On killing. Boston: Little, Brown
Gregg, M. E., James, J. E., Matyas, T. A., &
Thorsteinsson, E. B. (1999). Hemodynamic
profile of stress-induced anticipation and recovery.
International Journal of Psychophysiology, 34,
Hoffman, M. L. (2000). Empathy and Moral Development.
Cambridge: Cambridge University Press.
Kahneman, D. (2003). Maps of bounded rationality:
Psychology for behavioral economics. American
economic review, 93(5), 1449-1475.
LeDoux, J. (1996). The Emotional Brain. New York:
Simon and Shuster.
Martens, A., Kosloff, S., Greenberg, J., Landau, M., &
Schmader, T. (2007). Killing begets killing:
Evidence from a bug-killing paradigm that initial
killing fuels subsequent killing. Personality and Social
Psychology Bulletin, 33(9), 1251.
Mehrabian, A., & Epstein, N. (1972). A measure of
emotional empathy. Journal of Personality, 40(4),
Mendes, W., Blascovich, J., Hunter, S., Lickel, B., &
Jost, J. (2007). Threatened by the unexpected:
Physiological responses during social interactions
with expectancy-violating partners. Journal of
Personality and Social Psychology, 92(4), 698.
Mikhail, J. M. (2000). Rawls' Linguistic Analogy: A study
of the 'generative grammar' model of moral theory described
by John Rawls in 'A theory of justice'. Unpublished
Doctoral Dissertation, Cornell University, Ithaca.
Milgram, S. (1974). Obedience to Authority: An
experimental view. New York: Harper & Row
Moretto, G., L‡davas, E., Mattioli, F., & di Pellegrino,
G. (2010). A psychophysiological investigation of
moral judgment after ventromedial prefrontal
damage. Journal of Cognitive Neuroscience, 22(8), 1888-
Petrinovich, L., O'Neill, P., & Jorgensen, M. J. (1993).
An empirical study of moral intuitions: towards an
evolutionary ethics. Journal of Personality and Social
Psychology, 64(3), 467-478.
Pizarro, D. A. (2000). Nothing more than feelings?
The role of emotions in moral judgment. Journal
for the Theory of Social Behavior, 30(4), 355-375.
Rozin, P., Millman, L., & Nemeroff, C. (1986).
Operation of the laws of sympathetic magic in
disgust and other domains. Journal of Personality and
Social Psychology, 50, 703-712.
Watson, D., Clark, L. A., & Tellegen, A. (1988).
Development and validation of brief measures of
positive and negative affect: the PANAS scales.
Journal of Personality and Social Psychology, 54(6),
1. We also examined the association between
responses on moral dilemmas and changes in
cardiac output. Consistent with the “threat”
profile CO decreases were associated with less
endorsement of harming others, r (N=54) =
.29, p < .04. In Study 2 we examine responses
during a task that does not meet the
requirements of a motivated performance
situation, so we focus on TPR in both studies
Cushman,?? F.,?? Gray,?? K.,?? Gaffey,?? A.?? &?? Mendes,?? W.B.?? (in?? press)?? Simulating?? Murder:?? The?? aversion?? to?? harmful?? action.?? ?? Emotion?? ?? 7??
Figure 1: Harmful and non-harmful actions used in Experiment 2.
Table 1: Mean TPR reactivity by condition and measurement period. N indicates number of observations, with up to
three observations per subject
TPR Change: M (SD, N)
Condition Pre-action Post-action
Perform 75 (73, 34) 47 (90, 28)
Witness 34 (73, 49) 10 (88, 24)
No Harm 29 (83, 48) -28 (68, 16)