10.1177/0146167204263780PERSONALITY AND SOCIAL PSYCHOLOGY BULLETINSchubert / BODILY FEEDBACK OF POWER
The Power In Your Hand:
Gender Differences In Bodily Feedback
From Making a Fist
Thomas W. Schubert
University of Jena
Men and women differ in the meaning they attribute to physical
coercion and bodily force. Men associate bodily force with gain
ing power, whereas women associate bodily force with expressing
loss of power. It is hypothesized that because of these associations,
performing bodily forceful behavior feeds back on appraisals of
one’s power and that bodily feedback effects will mirror the gen-
der differences in associations. Supporting these hypotheses, it
was found that unobtrusively inducing behavior related to
bodily force (making a fist) activated the concept of power in a
Stroop task for both genders but that it increased hope for power
and positive judgments of an assertively acting target for men,
whereas it decreased hope for power and led to negative judg-
ments of an assertively acting target for women.
Keywords: embodiment; bodily feedback; power; aggression
Imagine the following three persons: a boy watching
his favorite (male) soccer player shooting a penalty, the
soccer player after he scores the goal, and a politician
when he explains to his opponents that he will reach his
goals. What behavior will these three men probably
share? Probably, all three of them will clench their hands
to fists, embodying their will to have power. However,
women or girls would less likely show this behavior in the
same situations. In the present research, this difference
and the consequences of this behavior on appraisals of
one’s power are investigated.
Making a fist indicates the potential to use bodily
force. Demonstrating bodily force in this way implies the
potential willingness to perform the implied action. As
exemplified above, such behaviors are shown in negotia
tions of power and status, are associated with assertive
ness (Gitin, 1970), and may lead to others’ submission
(Tiedens & Fragale, 2003). But the impact of perform
ing such behavior may go further—it also may influence
the person performing it. Following literature on bodily
feedback (for a review, see Neumann, Förster, & Strack,
2003), the current research investigates the possibility
that there is a bodily feedback of performing gestures of
bodily force on power-related thinking. The first ques
tion is: Does making a fist influence how powerful a
person thinks he or she is?
It is important, however, to acknowledge gender dif-
ferences in the realm of bodily force or physical aggres-
sion, which persist even in modern societies. Men and
women differ in how they understand and use bodily
means for coercive actions. Campbell and her col-
leagues (Campbell & Muncer, 1994; Campbell, Muncer,
& Coyle, 1992; see Geen, 1998) argued that if men hit
someone, it is likely that they want the other to comply.
In contrast, if women hit someone, it is likely that they
feel powerless and want to express emotional distress.
Consequently, the second question addressed by the cur
rent research is: If there is a bodily feedback of power
from bodily force, does this feedback mirror the gender
difference in the meaning and use of physical coercion?
GENDER DIFFERENCES IN PHYSICAL COERCION
Gender is an important moderator for power and co
ercion processes. On average, women describe them
Author’s Note: This research was supported by a stipend from the
International Graduate College on Conflict and Cooperation. My grat
itude goes to Jan Crusius, Colin Leach, Amélie Mummendey, Kai
Sassenberg, and Beate Seibt for comments on earlier drafts; Kai
Epstude for his literature recommendations; and to Sonja Bräutigam,
Jan Crusius, Janine Dieckmann, and Matthias Jakob for their help in
the data collection. Correspondence concerning this article can be
addressed to Thomas W. Schubert, Universität Jena, Lehrstuhl
Sozialpsychologie, Humboldtstr. 26, 07743 Jena, Germany; e-mail:
PSPB, Vol. 30 No. 6, June 2004 757-769
© 2004 by the Society for Personality and Social Psychology, Inc.
selves as less dominant and assertive than do men
(Feingold, 1994) and differ from men in how they exert
power (McClelland, 1975). Gender differences are espe
cially large in bodily coercion, which can be seen as a
form of social influence with the goal to control others
(Tedeschi & Felson, 1994). Because women are, on aver
age, less strong than men, their chances to gain influ
ence by applying bodily force are less good. Cultural
norms of Western societies still discourage boys and men
less than girls and women from using bodily force as a
means of social influence. Consequently, frequency and
meaning of physical aggression differ between the
The relation between gender and frequency of coer
cive behavior is moderated by response mode (White,
2001). Boys are more aggressive than girls in the bodily
domain (but not in the relational domain; Crick &
Grotepeter, 1995). Men are more aggressive than
women when the response mode is bodily rather than
verbal (Bettencourt & Miller, 1996). Men commit more
violent crimes than women, especially toward strangers
(Felson, 2002). Thus, it seems that “men are especially
likely to be more aggressive than women when the ag-
gression is physical and assaultive” (Geen, 1998, p. 331).
Campbell and her associates (Campbell et al., 1992;
Campbell & Muncer, 1994; Muncer & Campbell, 2000)
found that men conceptualize their own use of bodily
force in a more instrumental manner than do women:
Men tend to understand bodily force more than women
as a means to reach goals by controlling others or as a
useful way to gain power. Women hold a more expressive
conceptualization of their own use of bodily force than
do men. For them, using bodily force is a sign of losing
(or lost) power and of expressing this loss toward others.
Women understand bodily force more than men as emo
tional behavior and react with guilt and repression.
Women, Campbell et al. argue, associate their own use of
bodily force with abandonment rather than being in
control or gaining power.
Consequently, men and women differ concerning
associations between bodily force and experiences of
power. Men are culturally less discouraged to use bodily
force, which will frequently be associated with success
and power gain, and they understand their own bodily
force as a means to gain control. Therefore, men are
more likely to develop an association between bodily
force and being in control, or having power. Women are
culturally discouraged from using bodily force, it will fre
quently be associated with power loss, and they under
stand their use of it as loss of control. Therefore, women
are more likely to develop an association between bodily
force and powerlessness, or losing power.
To answer the question of how behavior related to
bodily force can influence a person who is performing it,
the relation between behavior and cognition has to be
analyzed. Cognition is in many ways embodied, that is,
bodily behavior and cognition are closely intercon
nected and interdependent (Barsalou, 1999; Barsalou,
Niedenthal, Barbey, & Ruppert, in press; Glenberg,
1997). Because of this interdependence, bodily behavior
can influence cognitive processing. Evidence for such
influences exists at several levels, such as effects of be
havior on language comprehension, on basic motiva
tional processes, and on the experience of feelings.
At a very basic level, comprehension of phrases can be
facilitated or inhibited by compatible or incompatible
movements, respectively (Glenberg & Kaschak, 2003;
Klatzky, Pellegrino, McCloskey, & Doherty, 1998). The
same applies for basic motivational processes such as
approach and avoidance. Some motor programs such
as flexing or extending the arm are so closely associ
ated with mental processes of approach and avoidance
that their performance influences evaluative processes
(Cacioppo, Priester, & Berntson, 1993; Chen & Bargh,
1999; Friedman & Förster, 2000; Neumann & Strack,
2000a). On an even higher level, feelings can be facili-
tated or inhibited by compatible or incompatible facial
expressions, even in the absence of self-perception pro-
cesses and inferences (Neumann & Strack, 2000b).
Strack, Martin, and Stepper (1988) showed that feelings
of amusement were facilitated by compatible facial
behavior (unobtrusively induced smiling) but inhibited
by incompatible facial behavior. Other feelings can be
influenced in a similar way. In a study by Strack and
Neumann (2000), participants who were led to furrow a
brow judged persons as less famous because the muscle
contracting facilitated the feeling of mental effort, on
which the judgment was then based.
It is likely that such effects rest to a large extent on a
facilitation of cognitive processes by compatible bodily
actions. Förster and Strack (1996) and Förster and Step
per (2000) provided evidence that performing bodily
behaviors facilitated cognitive processing of compatible
stimuli by analyzing the effects on mental capacity. They
found that the compatibility between motor behavior
and cognition facilitates processing, thereby freeing
mental capacity, whereas incompatibility of motor
behavior and cognition leads to mental load.
In sum, bodily actions facilitate compatible mental
processes (for alternative explanations, see Adelmann &
Zajonc, 1989; Laird, 1974; Zajonc & Markus, 1984). To
learn more about the influence of bodily force on power,
it is therefore necessary to look more closely on mental
representations of power.
758 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
MENTAL REPRESENTATIONS OF POWER
AND THEIR RELATION TO BODILY FORCE
Even if power is typically defined in an objective man
ner, as the actual control a person has over another per
son’s outcomes, for most effects of power it will be deci
sive whether a person mentally represents himself or
herself as powerful or powerless (Keltner, Gruenfeld, &
Anderson, 2003). Representing oneself as powerful re
sults from an appraisal of a situation as an opportunity to
have impact on others. Conversely, representing oneself
as powerless results from seeing the risk of being domi
nated by others. In the following, the term appraisal will
therefore be used to refer to such representations. Ob
jective power will have some effects even when it is not
accompanied by appraising oneself as powerful, but
most theorists agree that power appraisals are often the
true source of psychological effects of objective power
(Haidt & Rodin, 1999; Skinner, 1996). This becomes
especially clear when objective power and the subjective
appraisal diverge. If nominally powerful persons per
ceive themselves as powerless, they are likely to act de
fensively and show intrusive concerns with domination
when they are challenged (Bugental & Lewis, 1999).
Anderson and Berdahl (2002) recently provided evi-
dence that manipulations of objective power were medi-
ated by subjective appraisal of power; in the same man-
ner, effects of trait personality dominance also were
mediated by subjective appraisal of power. Thus, the
appraisal of one’s power is not only an epiphenomenon
of objective power, but it plays a mediating role for some
effects of objective power and it can sometimes even
override objective power.
Combining the insight that power is mentally repre
sented with the findings on bodily feedback and mean
ings of bodily force, it can be hypothesized that behav
iors related to bodily force will have effects on power
appraisals. More specifically, I predict that behavior that
is compatible with powerfulness will facilitate appraisals
of situations as affording influence on others. Likewise,
behavior that is compatible with powerlessness will fa
cilitate appraisals of situations as threatening domi
nance by others. As seen above, bodily force is compati
ble with powerfulness for men but with powerlessness
for women. Consequently, it can be expected that bodily
force will feed back powerfulness to men but powerless
ness to women. Of importance, women are expected to
associate bodily force with powerlessness, not because
they are more frequently forced in this way but because
their own application of bodily force is associated with
power loss. Finally, it is important to note that these ef
fects of bodily feedback from bodily force are expected
to occur unconsciously, without inferences or self-
OVERVIEW OF PRESENT RESEARCH
In the following three studies, these ideas will be
tested by looking at how judgments of situations and oth
ers are influenced by bodily feedback from bodily force.
The hypothesis is that bodily feedback from bodily force
leads men to have more hope for power, and judge oth
ers accordingly, whereas it leads women to have less hope
for power and judge others from the perspective of a vic
tim. But before these hypotheses are tested in Studies 2
and 3, Study 1 establishes that the concept of power is
activated by performing a behavior associated with
bodily force. As a manipulation of bodily feedback, all
three studies use making a fist in the experimental con
ditions. Making a fist is the abbreviation of hitting some
one or something; it implies the ability and deter
mination that one will use bodily force (Gitin, 1970).
Bodily force also may be associated with emotions and
mood. It might be that performing such behavior in
duces a negative mood in women, or a positive mood in
men, or increases arousal. If this was the case, one could
argue that these variables mediate the effects of the ma
nipulation. Therefore, mood and arousal are assessed in
all three studies to check whether making a fist has any
effect on these variables.
STUDY 1: WHEN YOUR BODY TELLS YOU
THAT POWER IS AT STAKE
I argued that bodily feedback from bodily force will be
moderated by gender. However, this implies that bodily
force will activate the concept of power itself for both
genders. Study 1 tested whether making a fist increases
accessibility of power-related words independent of gen
der. To rule out that the effects hinge on inferential pro
cesses and self-perception, making a fist was induced
such that participants were not aware of its relation to
bodily force (Strack et al., 1988) and participants were
thoroughly questioned for awareness of this association.
In addition, it was tested whether bodily force also acti
vates the concept of aggression itself. If making a fist
would activate aggression, one could argue that any
effects on power would be mediated by this effect.
At the same time, as participants performed either be
havior related to bodily force or neutral behavior, the ac
cessibility of words associated with power, with aggres
sion, and of neutral words was assessed in a Stroop task.
The Stroop task measures how long it takes participants
to name the color of the presented word. It is assumed
that the higher the accessibility of a word, the more it in
terferes with color-naming, resulting in longer reaction
times (MacLeod, 1991). The study had a 2 (hand posi
Schubert / BODILY FEEDBACK OF POWER 759
tion: fist vs. neutral) × 2 (gender) × 2 (word type: power
vs. aggression) factorial design with the first two factors
being varied between and the third factor within partici
pants. Reaction times to neutral words were later used as
a covariate to control for interindividual differences in
Fourteen women and 11 men took part in the study
(ages 20 to 27, M = 22.4, SD = 2.1). When asked for goals
of the study and possible influences of their hand posi
tion on reaction times, none of them guessed the true
purpose of the experiment.
The Stroop task presented nine words related to
power (rule, win, achieve, influence, mighty, authority, power
ful, strong, influential) and nine words related to aggres
sion (attack, hate, violence, murder, brutal, hit, aggressive,
argument, fight; some of them taken from Mussweiler &
Note that the power words do not imply a
direction. Instead, they can refer both to the self (Iamin
fluential) and to others (They are influential). Thus, pro-
cessing of these words should be facilitated both when
powerfulness and powerlessness is activated. Three lists
with five, nine, and nine neutral words (e.g., antenna, bi-
cycle, read) were assembled and used for practice trials,
general speed assessment, and concept trials.
Mood and arousal were measured by self-report. To
measure current mood and arousal, participants were
asked how they currently felt on scales from 0 (very bad)
to 10 (very good) and 0 (absolutely calm)to10(very aroused;
Neumann & Strack, 2000a). Further measures dis
tinguished between tense and energetic arousal
(Schultheiss & Brunstein, 1999). Participants rated on
11-point scales from 0 (no, not at all)to10(yes, fully)
whether they felt active, energetic, and lethargic for ener
getic arousal and nervous, restless, and calm for tense
One female and one male experimenter asked stu
dents at the campus to participate in an experiment on
“Hemisphere activation” (adapted from Friedman &
Förster, 2000) in exchange for a chocolate bar. Partici
pants were run alone or two at a time. Instructions were
given on the computer. The study was described as inves
tigating how manual actions activate brain hemispheres
and how this activation affects word comprehension. Be
fore the actual manipulation, the Stroop task was ex
plained. Participants were told to press the left cursor
key for a blue word and the right cursor key for a green
word with index and middle finger of the right hand;
they practiced 10 trials. Words remained on the screen
until an answer was given, but not more than 2,000 ms.
Feedback on correct, wrong, or too slow answers was
given during practice trials only. Next, participants read
about the children’s game “rock-paper-scissors” and saw
a line drawing of each hand position. In the fist condi
tion, participants were asked to make a “rock” with their
left hand while holding the hand slightly above the table.
In the neutral gesture condition, they were asked to
make “scissors.” It was stressed repeatedly that this hand
position had to be held during the whole experiment. At
no point was the word fist used to avoid additional activa
tion of semantic knowledge structures associated with
fist. While holding the assigned gesture, participants
continued the Stroop task. First, 18 trials presented nine
neutral words once in each color (general speed assess
ment). After a short break, the power, aggression, and
remaining neutral words were presented in a random
order, each word appearing once in each color, resulting
in a total of 54 trials. Order of trials was randomized such
that each word occurred once in the first half and once
in the second half of the trials. The experiment was
programmed in DMDX (Forster & Forster, 2003).
Following the Stroop task, and while still holding the
assigned hand position, self-reports on mood and
arousal were assessed. Afterward, participants gave
demographic information and were then asked what
they thought the purpose of the study was and whether
they suspected any influence of the hand position on the
speed of their reactions. Finally, they were debriefed and
given their candy.
Initial analyses explored differences between the
“blue” and “green” answers, which were assigned to the
middle or the index finger, respectively. It was found that
combining answers from both fingers led to the same
results as analyzing only the responses given with the
index finger but that middle finger answers showed no
effects. Therefore, only index finger answers were ana
lyzed. All wrong answers (4.6%) were discarded.
To account for interindividual differences in reaction
speed, a general speed covariate was computed by aver
aging reaction times to all neutral words except practice
trials. Reaction times to the power and the aggression
words were averaged separately. First, these two scores
were treated as repeated measures in a 2 (hand position)
× 2 (gender) × 2 (word type) ANCOVA with the averaged
reaction times on neutral words as a covariate. Besides a
significant effect of the covariate, F(1, 20) = 50.66, p <
.001, a main effect of hand position emerged, F(1, 20) =
6.74, p = .017, indicating that overall reactions were
slower when participants made a fist, which was qualified
by an interaction with word type, F(1, 20) = 4.66, p = .043,
indicating that this difference was larger for power than
760 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
for aggression words (see Table 1). This interaction was
further explored in two separate 2 (hand position) × 2
(gender) ANCOVAs for power-related and aggression-
related words and the same covariate. The covariate
was significant for both power and aggression words, F(1,
20) = 58.17 and F(1, 20) = 29.33, ps < .001. As Figure 1
shows, after controlling for the speed on neutral words,
reactions to power-related words were as predicted sig-
nificantly slower when participants made a fist (M =
534.98, SE = 16.93) than when they made a neutral ges
ture (M = 439.92, SE = 19.48), F(1, 20) = 12.78, p = .002,
This effect was not moderated by gender, F <1,
and there was no gender main effect, F(1, 20) = 1.21, p =
.284. Although reactions to aggression-related words
were also slightly slower when participants made a fist
(M = 506.19, SE = 19.89) than when not (M = 465.55, SE =
22.89), this difference was not significant, F(1, 20) = 1.69,
p = .208. None of the remaining effects was significant,
F s< 1.
MOOD AND AROUSAL
For both tense and energetic arousal, the respective
scales were internally consistent, α = .82 and α = .92, re
spectively. Gender × Hand Position ANOVAs were con
ducted on their average scores, the general arousal mea
sure, and on self-reported mood. No effect reached
significance, all Fs < 1.6, ps > .20, η
s < .07.
To provide evidence that performing behavior of
bodily force increases the accessibility of the power con
cept, it was assessed whether making such a gesture in
creases interferences by power-related words in a Stroop
paradigm. It was found that making a forceful gesture
slows down color naming for power-related words signifi
cantly. This indicates that power-related words were
more accessible for participants who made a fist than for
participants who did not make a fist. Of importance, this
was true for both men and women, although it should be
noted that the low sample size might have limited the
chance to find gender differences. This result might be
surprising at first given that all words were related to hav
ing power, but it has to be noted that the words did not
denote who had the power, and therefore, both power
lessness and powerfulness should lead to increased
Furthermore, the results show that the gesture had no
effect on words that were related to aggression. Thus,
making a fist does not directly activate aggression itself,
Schubert / BODILY FEEDBACK OF POWER 761
TABLE 1: Reaction Times for Power- and Aggression-Related
Words in a Stroop Task, Depending on Participants’
Hand Position and Gender, Study 1
Word Type Hand Position MSEM SE
Power-related Neutral 460.84 21.50 419.01 34.49
Fist 544.91 25.07 525.06 22.62
Aggression-related Neutral 482.68 25.26 448.42 40.52
Fist 515.82 29.45 496.57 26.57
NOTE: SE instead of SD is reported because means are estimated con
trolling for reaction times to neutral words.
Reaction Time to Power Words (ms)
Reaction Time to Aggression Words (ms)
Figure 1 Mean reaction times in Stroop task (+ SE) for power-related words (left) and aggression-related words (right) as a function of
participants’ hand position and gender, Study 1.
NOTE: Means are controlled for reaction time to neutral words.
presumably because the population we drew from in this
study is usually not involved in physical violence that
would create such an association. At the same time, the
effect on power words shows that such gestures are asso
ciated with having influence in general. Furthermore,
the gesture did not affect mood or arousal, which is not
surprising given that it was not held for a long time or
accompanied by any positive or negative feedback on
performance, which would probably be necessary to
cause a change in mood or even emotion (Stepper &
STUDY 2: WHEN YOUR BODY TELLS YOU
THAT THERE IS (NO) HOPE FOR POWER
Building on the finding that a bodily force gesture
activates the power concept for both men and women,
it is now possible to test the hypothesized divergence of
feedback effects of bodily force on men and women.
Above, it was argued that bodily force would facilitate sit
uated conceptualizations of gaining power for men,
whereas it would facilitate situated conceptualizations of
losing power for women. To measure these effects, Study
2 uses a test developed for the study of need for power
(McClelland, 1975), the Multi-Motive Grid (MMG;
Sokolowski, Schmalt, Langens, & Puca, 2000). In the tra-
dition of the TAT, but with the advantage of an easier
assessment, it measures perceived applicability of motive-
congruent thoughts to pictorial stimuli. The MMG items
are pretested as indicating concerns with power, achieve-
ment, and affiliation. Because the items for achievement
also have strong associations to control, power and
achievement were combined into a single score of hope
for control, following suggestions by Woike and col
leagues (Woike, 1995; Woike, Lavezzary, & Barsky, 2001).
Making a fist again served as a manipulation of bodily
force. It is predicted that making a fist increases interpre
tations related to hope for control for men but decreases
such interpretations for women.
Participants answered the MMG while making either
a fist or a neutral manual gesture. Thus, the study had a 2
(hand position: fist vs. neutral) × 2 (gender) between-
subjects factorial design with two main dependent vari
ables: hope for control and hope for affiliation.
In total, data from 42 women and 36 men were col
lected. Four men had to be excluded because they
guessed the true purpose of the study. Mean age was 20.9
(SD = 3.9, ranging from 17 to 37).
The MMG was slightly modified, rewording some
items and excluding filler items. The shortened version
measured hope for power, achievement, and affiliation.
Fourteen drawings of persons in various situations were
shown along with brief descriptive sentences (the items).
For each item, participants indicated whether in their
opinion this sentence applied to the picture by pressing
the respective key.
The items are phrased such that they tap either the
power, achievement, or affiliation motive. For instance,
an item indicating hope of power is “Here one wants to
have influence” and an item indicating hope of achieve
ment is “Here one has trust in one’s success.” Further
more, to make the test shorter, each picture measures
only the motives it is sensitive to: only one, two, or all
three. Because one motive is always assessed with two
items for a picture, there can be either two, four, or six
items for one picture. For each motive, 6 pictures ap
plied and therefore 12 items had to be answered in total.
Because the test simply counts yes answers (applies to the
picture), the resulting score for each motive can range
from 0 to 12.
Mood and arousal were measured in the same way as
in Study 1. In addition, participants were asked for their
current emotional state by asking them whether they
were proud, satisfied, relieved, annoyed, angry, frustrated,or
anxious. All items had to be answered on scales ranging
from 1 (not at all) to 5 (fully).
Recruiting and gesture manipulation followed Study
1, with a slightly different cover story: Participants were
told that we were interested in how distracting they
would find a manual task in addition to their main task.
Instructions and pictures were presented by a computer.
After the hand position manipulation, pictures and
items were shown. Each picture first appeared for 7 s
without an item on the screen, with the instruction to
take the perspective of one person in the picture. Then,
the items were shown one by one underneath the picture
and participants decided whether “this thought applied
to this situation” by pressing the yes or the no key. Pictures
and items were presented in the fixed order used in the
original MMG. After finishing the MMG, participants
rated their mood, arousal, and emotions and indicated
age, gender, and ideas about the purpose of the study. At
the end, participants received a full debriefing, a report
of their average scores, and their candy.
Achievement and power affirmations were summed
to a total hope for control score and divided by two to
762 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
make them comparable to the number of affirmed hope
for affiliation items. Table 2 displays the means. Hope for
control and hope for affiliation scores were submitted to
two 2 (hand position) × 2 (gender) ANOVAs. For hope
for control, the expected interaction was significant, F(1,
70) = 4.88, p = .030, η
= .07. Figure 2 shows that the pat
tern was as predicted. Simple effects analyses confirmed
that women expressed less hope for control when they
made a fist (M = 6.67, SD = 1.63) compared to the neutral
gesture, (M = 7.63, SD = 1.64), F(1, 70) = 3.21, p = .039
= .04. Men, however, expressed slightly
more hope for control when they made a fist (M = 7.44,
SD = 1.62) compared to the neutral gesture (M = 6.61, SD
= 2.06), but this difference was only marginal, F(1, 70) =
1.88, p = .088 (one-tailed), ηp
= .03. There were no main
effects of hand position or gender on hope for control
and no effects at all on hope for affiliation, all Fs < 1.
Because the scores combined very different pictures,
additional analyses explored which pictures contributed
most to the effect on hope for control. For these analy
ses, affirmations for control motives were averaged per
picture. Two out of the eight power-relevant pictures
contributed particularly to the interaction, η
= .05 and
.09. The first showed dancing couples in a bar and the
second showed a beach scene, with several persons sun
bathing and one person playing ball by himself. Al
though for the first picture a connection to control and
power is plausible, such an association is not obvious for
the second picture, for which nevertheless a significant
crossover interaction was found, F(1, 70) = 6.54, p = .013;
men expressed more and women less hope for control
when they made a fist.
MOOD, AROUSAL, AND EMOTIONS
For both tense and energetic arousal, the respective
3-item scales were internally consistent, α = .73 and α =
.89, respectively. To determine whether mood, arousal,
tense or energetic arousal, or any of the specific emo
tions mediated the obtained interaction effect, Gender ×
Hand Position ANOVAs were conducted on these vari
ables. For none of them, a significant interaction effect
emerged, all Fs < 1. The only significant main effect of
hand position emerging from these analyses indicated
that tense arousal was lower in the fist condition (M =
1.69, SD = .84) than in the neutral gesture condition (M =
2.27, SD = 1.13), F(1, 70) = 6.08, p = .016, η
Bodily feedback from making a gesture of bodily
force results in activation of the concept of power but the
direction of these effects differs between men and
women, as the significant interaction of the manipula
tion and gender in Study 2 showed: Women who made a
fist perceived less hope for control in the pictures than
women who did not make a fist. In contrast, men who
made a fist saw marginally more hope for control than
those men who made a neutral gesture. A more detailed
exploratory analysis suggests that strong effects were not
only due to clearly power-related pictures but also to
rather ambiguous and not clearly power-related pic
tures. A detailed view on the underlying process is post
poned to the General Discussion.
These effects are not attributable to simple in-
fluences of mood: No interaction was found for mood,
arousal, or any of the self-reported emotions. Because
the cover story obscured the fact that the gesture was re
lated to bodily force and because the word fist was never
used, it seems safe to conclude that the obtained effects
are due to bodily feedback and not to a process in which
participants observed and categorized themselves as
Schubert / BODILY FEEDBACK OF POWER 763
Hope for Control
Figure 2 Mean number of expressed hope for control (+ SE)asa
function of participants’ hand position and gender, Study 2.
NOTE: Scores can range from 0 to 12.
Perceived Applicability of Hope for Control and Hope
for Affiliation Statements to Pictures, Depending on
Participants’ Hand Position and Gender, Study 2
Motive Hand Position MSDM SD
Hope for control Neutral 7.63 1.64 6.61 2.06
Fist 6.67 1.63 7.44 1.62
Hope for affiliation Neutral 5.38 1.58 5.36 1.82
Fist 5.11 1.88 5.61 1.54
NOTE: Values reflect summed affirmations of respective items and can
range from 0 to 12.
making a fist, as a self-perception account would suggest
(Laird, 1974). In other words, effects were not driven by
priming the word fist through instruction or through
STUDY 3: WHEN YOUR BODY TELLS YOU
WHETHER DONALD IS HOSTILE OR KIND
Experiences of power are central to how we relate to
other persons, and therefore, changing power apprais
als through bodily feedback also should affect judg
ments of other persons. Study 3 investigates the effects of
bodily force on judgments that are influenced by power
relations, namely, judgments of hostility of others. How
are appraisals of powerfulness (in men) and powerless
ness (in women) going to influence perceptions of
Whether we judge the actions of another person as ag
gressive depends on a complex attribution process that
hinges on assumptions about the other’s intentions
(Crick & Dodge, 1996; Löschper, Mummendey,
Linneweber, & Bornewasser, 1984). Power may influ
ence attribution because it focuses attention on possible
environmental threats or chances (Keltner et al., 2003;
Schwartz, Dodge, & Coie, 1993). Powerless individuals
are more likely to think about hostility and aggressive-
ness of others and perceive others in this light. Powerful
individuals are more likely to think about chances and
their own assertiveness. Support for this hypothesis
comes from a study by Dodge and Coie (1987). They in-
vestigated boys’ attributions of hostility to ambiguous
acts and distinguished between boys who were either ha-
bitually proactively aggressive or reactively aggressive.
Proactive and reactive aggression roughly correspond to
the concepts of instrumental and expressive aggression
described earlier. Dodge and Coie found that boys only
high in reactive aggression, but not boys only high in
proactive aggression, showed a bias toward interpreting
a provocateur’s intentions as hostile.
To investigate the idea that bodily feedback influ
ences attributions of hostility, Study 3 used the classic
Donald paradigm introduced by Higgins, Rholes, and
Jones (1977) and Srull and Wyer (1979). Again, making
a fist served as a manipulation of bodily force. Partici
pants had to judge whether an ambiguously acting target
was hostile. As explained above, bodily force has more
expressive functions in women, whereas it is instrumen
tal for men. This means that women tend to activate
these behavior programs in situations where they see
others as behaving aggressively. Thus, when activating
these motor programs, they should tend to see another
person as behaving more aggressively. Men, on the other
hand, should activate these motor programs in situa
tions where they are dominating. They should thus see
another person as behaving less aggressively.
While participants either made a fist or a neutral ges
ture, they read the description of an ambiguously acting
male person called Frank, judged Frank on positive and
negative adjectives that were either related or unrelated
to hostility, and rated their mood and arousal. Thus, the
full design was a 2 (hand position: fist vs. neutral) × 2
(gender) × 2 (trait valence: positive vs. negative) × 2
(trait hostility: hostility related vs. not hostility) factorial
design with the first two factors varied between and the
last two factors varied within subjects. We expected ef
fects on traits related to hostility but not on traits
unrelated to hostility (Srull & Wyer, 1979).
Nineteen women and 18 men took part in the study,
mean age was 21.3 (SD = 2.2, ranging from 19 to 27).
None of them suspected an effect of the hand position
on their judgments.
The Donald paragraph from Srull and Wyer (1979)
was translated into German and Donald was called Frank.
The text described a day in Frank’s life and mentioned
several behaviors that could be interpreted as either ag-
gressive or assertive, for instance, that he refuses to pay
his rent until his apartment is renovated, that he first
buys a tool and then tries to give it back immediately, and
that he refuses to donate blood. As the dependent mea-
sure, participants rated Frank on 12 adjectives, using
scales from 0 (not at all)to10(extremely so). There were
three negative and three positive adjectives related to
hostility, measuring Frank’s hostility (negative, hostility-
related) and friendliness (positive, hostility-related), re
spectively. Hostility adjectives were unpleasant, hostile,
and unfriendly; friendliness adjectives were kind, friendly,
and considerate. In addition, three negative and three
positive adjectives unrelated to hostility were used: bor
ing, narrow-minded, and vain and intelligent, reliable, and
Adjectives were randomized and presented
in fixed order. Mood and general arousal were assessed
as in previous studies.
Recruiting, procedure, and cover story followed
Study 2. While holding the nondominant hand in the
assigned position, participants read the story about
Frank with the instruction to form an impression of him,
rated him on the 12 adjectives, and rated their own
mood and arousal thereafter. After giving information
on gender and age and what they thought the actual pur
pose of the study was, they were debriefed and given
764 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
JUDGMENTS OF FRANK
The scales for hostility (α = .69), friendliness (α = .74),
and unrelated positivity (α = .60) displayed satisfactory
internal consistencies, but the unrelated negativity scale
was not internally consistent (α = .34). Apparently, the
German translation of narrow-minded was rather associ-
ated with hostility, which was not intended. However, be-
cause this works rather against the current hypothesis
and because exploratory analyses revealed no changes in
the results when the item was removed, the scale was
used as originally planned. For each category, ratings
were averaged to one score.
We predicted that hand position would affect ratings
of Frank on hostility-related, but not on nonhostility-
related adjectives, and that when making a fist, men
would rate Frank as less hostile and more friendly,
whereas women would rate Frank as more hostile and
less friendly. In a 2 (hand position) × 2 (gender) × 2 (trait
valence) × 2 (trait hostility) mixed-model ANOVA, this
prediction translates into a four-way interaction, which
was significant, F(1, 33) = 7.97, p = .008, η
are displayed in Table 3.
To explore this interaction, two separate 2 (hand posi
tion) × 2 (gender) × 2 (trait valence) mixed-model
ANOVAs with repeated measures on the last factor were
computed for hostility-related and nonhostility-related
adjectives. For hostility-related adjectives, only a main
effect of trait valence, F(1, 33) = 73.03, p < .001, and the
expected three-way interaction, F(1, 33) = 6.58, p = .015,
= .17, emerged, all other effects Fs < 1.9, ps > .10, η
0.53. Figure 3 shows that the pattern was as predicted.
Simple effects analyses showed that men judged Frank as
more friendly when they made a fist (M = 3.17, SD = .95)
than when they made a neutral gesture (M = 1.52, SD =
1.31), F(1, 33) = 6.62, p = .008 (one-tailed), η
= .17. For
women, this effect was reversed—they judged Frank as
less friendly when they made a fist (M = 2.08, SD = 1.31)
than when they made a neutral gesture (M = 3.37, p =
2.08), F(1, 33) = 3.77, p = .030 (one-tailed), η
= .10. The
equivalent (reversed) pattern was apparent for hostility,
but here the simple effects did not reach significance:
Figure 3 shows that women judged Donald as slightly
more hostile when they made a fist (M = 7.17, SD = 2.19)
compared to a neutral gesture (M = 5.93, SD = 1.39), F(1,
33) = 2.15, p = .076 (one-tailed), η
= .06, although there
was no difference for men on this dimension, F < 1.
The second 2 × 2 × 2 ANOVA was conducted on non-
hostility-related adjectives. Besides a nonsignificant
main effect of trait valence, F(1, 33) = 2.76, p = .106, a sig
nificant gender effect was found, F(1, 33) = 5.01, p = .032,
which was qualified by an unexpected marginal Hand
Position × Gender interaction, F(1, 33) = 3.80, p = .060,
= .10; all other Fs < 1 (see Table 3). This interaction in
dicated that irrespective of trait valence, men gave lower
ratings when they made a fist, whereas women gave
higher ratings when they made a fist. Because this inter-
action was unexpected, we explored it and the involved
simple main effects further, although the three-way in-
teraction was not significant (F < 1). It seemed that the
interaction was mainly caused by an unexpected simple
main effect: Women judged Frank somewhat lower on
nonhostile negative adjectives when they made a fist
(M =5.33, SD = 1.32) than when they did not (M = 6.60,
SD = 1.97), F(1, 33) = 2.46, p = .126, η
= .07. This un-
expected pattern will be explained in the Discussion.
MOOD AND AROUSAL
Mood and arousal scores were submitted to 2 (hand
position) × 2 (gender) ANOVAs. None of the effects
reached significance, all Fs < 1.10, ps > .310.
In Study 3, bodily feedback from making a gesture re
lated to bodily force influenced impressions of a target
whose actions could be interpreted as either hostile or
assertive. Of importance, as in the previous study, gender
moderated the bodily feedback effect. Women who
made a fist judged the target as less kind than women
who made a neutral gesture. Men who made a fist judged
Frank as kinder than men who made a neutral gesture.
Thus, the bodily feedback of power had effects on social
information processing. Making a gesture related to
bodily force elicited in women a hostile attribution bias
(Crick & Dodge, 1996), presumably because the be
havior facilitated cognitions on threats and dangers
(Keltner et al., 2003). This result is compatible with our
basic hypothesis that performing acts of bodily force im
plies for women a loss of control over the environment.
Schubert / BODILY FEEDBACK OF POWER 765
TABLE 3: Judgments of Ambiguously Acting Male Person,
Depending on Participants’ Hand Position and Gender,
and Trait Valence and Trait Hostility, Study 3
Trait Hostility Valence Position MSDMSD
Hostility- Positive Neutral 3.37 1.84 1.52 1.31
related (friendliness) Fist 2.08 1.31 3.17 0.95
Negative Neutral 5.93 1.39 6.41 2.30
(hostility) Fist 7.17 2.19 5.97 1.07
Nonhostility- Positive Neutral 5.47 1.91 3.82 2.37
related Fist 5.13 1.78 4.83 1.73
Negative Neutral 6.60 1.97 5.00 1.94
Fist 5.33 1.32 5.40 1.43
NOTE: Scores range from 0 to 10.
For men, however, making a gesture related to bodily
force actually increased ratings of kindness of the ambig-
uously acting target. This fits the contention that for
men, bodily force facilitates thoughts about being asser-
tive and in power.
The effects were stronger for friendliness than for
hostility. In this regard, it is interesting to note that rat-
ings on hostility were in total rather high, and much
higher than ratings on kindness. Apparently, Frank’s
actions were seen as rather hostile. This could be the
reason that the effects turned out to be stronger on the
positive traits: Increased thinking about (positive) con
trol in men might not have allowed denial of Frank’s
meanness but caused the addition of positive traits. For
women, there was in fact a marginal increase in hostility
ratings for Frank when they made a fist, but in addition,
women who made a fist judged him as clearly less
friendly than those who made a neutral gesture.
An unexpected difference emerged on the negative
traits that were unrelated to hostility. Women showed
slightly, although not significantly, lower ratings on these
traits, which resulted in a marginal interaction. At first
glance, this seems to indicate that they rated Frank as less
negative when they made a fist, which would contradict
the above findings. However, this effect seems to hinge
on the precise negative, nonhostile traits used. On two
out of the three traits, women showed decreasing ratings
when they made a fist: vain and boring. For boring, this
simple effect was even significant, F(1, 31) = 7.24, p =
.011. But it seems clear that this drop does not indicate a
more positive attitude toward Frank; in contrast, it prob
ably indicates stress and annoyance. (Or is a mean dog
barking at you boring?)
We asked whether and how the performance of a
gesture related to bodily force influences power-related
cognitions. Study 1 established that bodily feedback
from making a gesture of bodily force indeed activated
the concept of power, as evidenced by the fact that it took
both men and women longer to name the color of words
related to power when they made a fist. Furthermore, it
was hypothesized that given the gender differences in
use and meaning of bodily force, bodily feedback effects
of such a gesture would be moderated by gender; this
was found in Studies 2 and 3. When women made a fist
under the disguise of a cover story that obscured the rela
tion to bodily force, they perceived less hope for control
in hypothetical situations and judged an ambiguously
acting male target as less kind and more hostile. In con
trast, when men made a fist, they perceived marginally
more hope for control in hypothetical situations and
they judged an ambiguously acting male target as kinder.
This evidence demonstrates that the different meanings
men and women attribute to performing acts of bodily
force implicate differential effects of performing these
behaviors on social information processing. Merely mak
ing the gesture, without noticing its relation to bodily
force at all, is sufficient to change perceptions of situa
tions and others in line with the conceptions one’s gen
der has about physical aggression.
766 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
Figure 3 Mean (+ SE) perceived kindness (left) and hostility (right) of an ambiguously acting male target as a function of participants’ hand
position and gender, Study 3.
NOTE: Scores can range from 1 to 10.
The conclusion that bodily force facilitates the under
standing of the world for men in terms of having power
and for women in terms of powerlessness is disturbing
when its consequences in interactions between men and
women are concerned. Men’s potential to associate
bodily force with having power may lead them to sponta
neously initiate such behavior, sustain its performance
once it is initiated, or actually use bodily force in argu
ments, resulting in violence. When for women their own
use of bodily force facilitates thoughts about powerless
ness, this then creates a compatibility that strengthens
the dominance of male actors and weakens opposition
However, it needs to be added that although we found
general differences between the genders, it is very likely
that there is a large within-gender variance, that is, for
some women, bodily force might activate having power
instead of powerlessness, whereas for some men, bodily
force might activate powerlessness instead of having
power. Of interest, such differences within the genders
also have been found for instrumental and expressive
views of aggression. Campbell and Muncer (1994) found
that both male and female nurses had a more expressive
view of their own aggression than both male and female
soldiers, although the total gender difference prevailed.
The same differences may be observable for bodily feed-
back from behaviors related to bodily force; this hypoth-
esis awaits empirical testing.
Processes Mediating Effects of Bodily Feedback
Concerning the precise mediator of these effects, at
least two mechanisms discussed in the literature may
apply. First, it may be that bodily force is part of the men
tal representation of powerful or powerless situations
and thereby facilitates the construction of such a repre
sentation. In terms of Barsalou et al. (in press), forceful
behavior activates a situated conceptualization and facil
itates a conceptualization in terms of powerfulness or
powerlessness. As cited above, evidence for effects of
bodily feedback on mental capacity supports this ac
count (Förster & Stepper, 2000; Förster & Strack, 1996).
It could be that in the present studies, bodily force facili
tated (a) processing of those elements of the stimuli
(pictures and story) that indicated compatible concep
tualizations of one’s power and (b) completion of the
stimuli with compatible ideation of one’s power.
A second possible mechanism is that the bodily feed
back created feelings of power or powerlessness, simi
lar to affective feeling of amusement created by a smile
in conjunction with comics (Strack et al., 1988), or the
nonaffective feeling of mental effort created by furrow
ing the brow in conjunction with memory recall (Strack
& Neumann, 2000). When the participants then judged
whether a thought would apply to a given picture in
Study 2, or whether Frank was acting hostile or kind in
Study 3, this feeling would then have been used as a cue
in a heuristic (Strack & Neumann, 2000).
The two accounts differ in that feelings of powerful
ness or powerlessness can be an (additional) outcome in
the first account, whereas they are expected to mediate
the effects in the second account. Two arguments speak
for the first account. It is more parsimonious because it
also can explain the development of feelings, whereas
the second account cannot easily explain effects on men
tal capacity. Second, and more important, the second ac
count is unlikely because it assumes that either the
bodily feedback can directly instigate feelings, for which
there is not much evidence in the literature (Adelmann
& Zajonc, 1989), or that the conditions of Studies 2 and 3
were such that feelings of power or powerlessness would
be provoked and then increased by the bodily feedback
(Stepper & Strack, 1993). However, this was not the case.
The studies did not create situations in which the partici
pants’ power would have mattered, there was no possi
ble influence on others, and there was no influence by
others. In sum, it seems more likely that the effects of
performing a forceful gesture were not mediated by a
feeling of power but that they facilitated situated concep-
tualizations of power or powerlessness.
The facial feedback literature offers another media-
tor of bodily feedback effects, namely, a direct neurologi-
cal or visceral feedback that is not mediated by cognitive
processes. Laird (1974) used the analogy of depth per-
ception that receives input from the angular relation of
the eyes. Similarly, one might speculate whether such a
feedback is also provided by bodily behavior. However,
this account seems unlikely for the present effects be
cause (a) it is not clear why there should be a gender dif
ference in such a biological variable and (b) the evolu
tion of such mechanisms is only likely for such frequently
used and basic bodily feedback as eye parallax, but not
Power, Pride, and Anger
All three studies also investigated mood and arousal
as possible mediators of the present effects, and in none
of the studies was there evidence for such processes.
Study 2 also assessed whether participants felt several
emotions, among them pride and anger, and found no
effects either. This failure of the procedure to instigate
emotions may be attributable to the same causes that
were just discussed for feelings of power. Bodily feedback
probably only in rare cases causes feelings or emotions
by itself, although it is a powerful modulator of feelings
and emotions that are caused by external stimuli
(Adelmann & Zajonc, 1989; Neumann & Strack, 2000c).
The present studies did not provide such external stim
uli, and therefore, emotions did not arise.
Schubert / BODILY FEEDBACK OF POWER 767
Nevertheless, it is interesting to note that making a fist
is used not only as a gesture accompanying thoughts and
claims of power but is also a common expression of both
pride and anger (Darwin, 1872; Eibl-Eibesfeldt, 1989),
which is observable in many sporting events. Pride and
anger are associated with the need for power: fulfillment
of this need instigates pride, whereas blocking of a power
goal instigates anger (Zurbriggen & Sturman, 2002). Al
though I know of no systematic investigation, it seems
that men express pride in sportive achievements more
often by showing a fist than women (this difference, how
ever, seems attenuated in some sports, e.g., tennis).
Thus, it may be the case that the different meanings of
bodily force also stretch to which expressions are used
when the associated motive is frustrated or fulfilled. To
display the achievement of power (i.e., pride), bodily
forceful behavior may only be used by those who see it as
a valid instrument to reach it.
The Embodiment of Power
On a more general level, the current research adds a
new perspective to the research on embodiment of
power. This long tradition of research has mainly inves-
tigated how dominance or submission is expressed and
communicated to others and how those others react.
Dimensions used for inferring power are, among others,
elevation and erectness of posture, facial expressions
and eye contact, relaxation, and how much space a per-
son occupies (Argyle, 1988; Ellyson & Dovidio, 1985).
Tiedens and Fragale (2003) demonstrated the effects of
spatial behavior. In a dyad interaction, the more space a
confederate occupied (sitting in an expansive manner
instead of sitting in a constricted manner), the more
dominant the confederate was judged. Furthermore,
interaction partners reacted to these power signals by
taking complementary positions without being aware of
doing so. But although the reactions to the display of
power and dominance have always been understood as
at least partly automatic and unconscious, it has to my
knowledge not been suggested that such displays of
power also feed back to the actor. It may be that power
moves have as much influence on those who perform
them as on observers, whether they create feelings of
power or influence conceptualization of situations as in
the present research.
1. In German, beherrschen, gewinnen, erreichen, Einfluss, mächtig,
Autorität, kraftvoll, stark, einflussreich; Angriff, Haß, Gewalt, Mord, brutal,
schlagen, aggressiv, Streit, Kampf.
2. In German, aktiv, energetisch, and träge for energetic and nervös,
unruhig, and ruhig for tense arousal.
denotes partial Eta squared computed by SPSS 11.0.
4. In German, unangenehm, feindselig, unfreundlich; liebenswürdig,
freundlich, fürsorglich; langweilig, engstirnig, eitel; intelligent, verlässlich,
5. Less relevant to the central hypothesis was a significant trait
valence effect, F(1, 33) = 32.85, p < .001, a significant trait hostility
effect, F(1, 33) = 7.72, p = .009, and an interaction of these two factors,
F(1, 33) = 48.54, p < .001.
Adelmann, P. K., & Zajonc, R. B. (1989). Facial efference and the
experience of emotion. Annual Review of Psychology, 40, 249-280.
Anderson, C., & Berdahl, J. L. (2002). The experience of power:
Examining the effects of power on approach and inhibition ten
dencies. Journal of Personality and Social Psychology, 83, 1362-1377.
Argyle, M. (1988). Bodily communication (2nd ed.). London: Methuen.
Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral and
Brain Sciences, 22, 577-609.
Barsalou, L. W., Niedenthal, P. M., Barbey, A., & Ruppert, J. (in press).
Social embodiment. In B. Ross (Ed.), The psychology of learning and
Bettencourt, B. A., & Miller, N. (1996). Gender differences in aggres
sion as a function of provocation: A meta-analysis. Psychological
Bulletin, 119, 422-447.
Bugental, D. B., & Lewis, J. C. (1999). The paradoxical misuse of
power by those who see themselves as powerless: How does it hap
pen? Journal of Social Issues, 55, 51-64.
Cacioppo, J. T., Priester, J. R., & Berntson, G. G. (1993). Rudimentary
determinants of attitudes II: Arm flexion and extension have dif
ferent effects on attitudes. Journal of Personality and Social Psychol-
ogy, 65, 5-17.
Campbell, A., & Muncer, S. (1994). Sex differences in aggression:
Social representation and social roles. British Journal of Social Psy-
chology, 33, 233-240.
Campbell, A., Muncer, S., & Coyle, E. (1992). Social representation of
aggression as an explanation of gender differences: A preliminary
study. Aggressive Behavior, 18, 95-108.
Chen, M., & Bargh, J. A. (1999). Consequences of automatic evalua-
tion: Immediate behavioral predispositions to approach or avoid
the stimulus. Personality and Social Psychology Bulletin, 25, 215-224.
Crick, N. R., & Dodge, K. A. (1996). Social information-processing
mechanisms in reactive and proactive aggression. Child Develop
ment, 67, 993-1002.
Crick, N. R., & Grotepeter, J. K. (1995). Relational aggression, gen
der, and social-psychological adjustment. Child Development, 66,
Darwin, C. (1872). The expressions of emotions in man and animals.New
York: Oxford University Press.
Dodge, K. A., & Coie, J. D. (1987). Social-information-processing fac
tors in reactive and proactive aggression in children’s peer
groups. Journal of Personality and Social Psychology, 53, 1146-1158.
Eibl-Eibesfeldt, I. (1989). Human ethology. New York: Aldine de Gruyer.
Ellyson, S. L., & Dovidio, J. F. (1985). Power, dominance, and nonver
bal behavior: Basic concepts and issues. In S. L. Ellyson & J. F.
Dovidio (Eds.), Power, dominance, and nonverbal behavior (pp. 1-27).
New York: Springer-Verlag.
Feingold, A. (1994). Gender differences in personality: A meta-analysis.
Psychological Bulletin, 116, 429-456.
Felson, R. B. (2002). Violence & gender reexamined. Washington, DC:
American Psychological Association.
Förster, J., & Stepper, S. (2000). Compatibility between approach/
avoidance stimulation and valenced information determines
residual attention during the process of encoding. Journal of Per
sonality and Social Psychology, 30, 853-871.
Förster, J., & Strack, F. (1996). Influence of overt head movements on
memory for valenced words: A case of conceptual-motor compati
bility. Journal of Personality and Social Psychology, 71, 421-430.
Forster, K. I., & Forster, J. C. (2003). DMDX [Computer software].
Developed at Monash University and the University of Arizona;
retrieved from http://www.u.arizona.edu/~kforster/dmdx/
768 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
Friedman, R. S., & Förster, J. (2000). The effects of approach and
avoidance motor actions on the elements of creative insight. Jour
nal of Personality and Social Psychology, 79, 477-492.
Geen, R. G. (1998). Aggression and antisocial behavior. In D. T.
Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psy
chology (4th ed., pp. 317-356). New York: McGraw-Hill.
Gitin, S. R. (1970). A dimensional analysis of manual expression. Jour
nal of Personality and Social Psychology, 15, 271-277.
Glenberg, A. M. (1997). What memory is for. Behavioral and Brain Sci
ences, 20, 1-55.
Glenberg, A. M., & Kaschak, M. P. (2003). Grounding language in
action. Psychonomic Bulletin & Review, 9, 558-565.
Haidt, J., & Rodin, J. (1999). Control and efficacy as interdisciplinary
bridges. Review of General Psychology, 3, 317-337.
Higgins, E. T., Rholes, W. S., & Jones, C. R. (1977). Category accessi
bility and impression formation. Journal of Experimental Social Psy
chology, 13, 141-154.
Keltner, D., Gruenfeld, D., & Anderson, C. P. (2003). Power,
approach, and inhibition. Psychological Review, 110, 265-284.
Klatzky, R. L., Pellegrino, J. W., McCloskey, B. P., & Doherty, S. (1998).
Can you squeeze a tomato? The role of motor representations in
semantic sensibility judgments. Journal of Memory and Language,
Laird, J. D. (1974). Self-attribution of emotion: The effects of expres
sive behavior on the quality of emotional experience. Journal of
Personality and Social Psychology, 29, 475-486.
Löschper, G., Mummendey, A., Linneweber, V., & Bornewasser, M.
(1984). The judgement of behaviour as aggressive and sanction
able. European Journal of Social Psychology, 14, 391-404.
MacLeod, C. (1991). Half a century of research on the Stroop effect:
An integrative review. Psychological Bulletin, 109, 163-203.
McClelland, D. C. (1975). Power: The inner experience.NewYork:John
Muncer, S. J., & Campbell, A. (2000). Comments on “Sex differences
in beliefs about aggression: Opponent’s sex and the form of
aggression” by J. Archer and A. Haigh. British Journal of Social Psy-
chology, 39, 309-311.
Mussweiler, T., & Förster, J. (2000). The sex → aggression link: A
perception-behavior dissociation. Journal of Personality and Social
Psychology, 79, 507-520.
Neumann, R., Förster, J., & Strack, F. (2003). Motor compatibility:
The bidirectional link between behavior and evaluation. In
J. Musch & K. C. Klauer (Eds.), The psychology of evaluation (pp. 371-
391). Mahwah, NJ: Lawrence Erlbaum.
Neumann, R., & Strack, F. (2000a). Approach and avoidance: The
influence of proprioceptive and extereoceptive cues on encoding
of affective information. Journal of Personality and Social Psychology,
Neumann, R., & Strack, F. (2000b). Experiential and non-experiential
routes of motor influences on affect and evaluation. In H. Bless &
J. P. Forgas (Eds.), The message within: The role of subjective experience
in social cognition and behavior (pp. 52-68). Philadelphia: Psychol
Neumann, R., & Strack, F. (2000c). Mood contagion: The automatic
transfer of mood between persons. Journal of Personality and Social
Psychology, 79, 211-223.
Schultheiss, O. C., & Brunstein, J. C. (1999). Goal imagery: Bridging
the gap between implicit motives and explicit goals. Journal of Per
sonality, 67, 1-38.
Schwartz, D., Dodge, K. A., & Coie, J. D. (1993). The emergence of
chronic peer victimization in boys’ play groups. Child Development,
Skinner, E. A. (1996). A guide to constructs of control. Journal of Per
sonality and Social Psychology, 71, 549-570.
Sokolowski, K., Schmalt, H. D., Langens, T. A., & Puca, R. M. (2000).
Assessing achievement, affiliation, and power motives all at once:
TheMulti-MotiveGrid(MMG).Journal of Personality Assessment, 74,
Srull, T. K., & Wyer, R. S. (1979). The role of category accessibility in
the interpretation of information about persons: Some determi
nants and implications. Journal of Personality and Social Psychology,
Stepper, S., & Strack, F. (1993). Proprioceptive determinants of emo
tional and nonemotional feelings. Journal of Personality and Social
Psychology, 64, 211-220.
Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitat
ing conditions of the human smile: A nonobtrusive test of the
facial feedback hypothesis. Journal of Personality and Social Psychol
ogy, 54, 768-777.
Strack, F., & Neumann, R. (2000). Furrowing the brow may under
mine perceived fame: The role of facial feedback in judgments of
celebrity. Personality and Social Psychology Bulletin, 26, 762-768.
Tedeschi, J. T., & Felson, R. B. (1994). Violence, aggression, and coercive
actions. Washington, DC: American Psychological Association.
Tiedens, L. Z., & Fragale, A. R. (2003). Power moves:
Complementarity in dominant and submissive nonverbal behav-
ior. Journal of Personality and Social Psychology, 84, 558-568.
White, J. W. (2001). Gendered aggression across the lifespan. In
J. Worell (Ed.), Encyclopedia of gender.SanDiego,CA:Academic
Woike, B. (1995). Most memorable experiences: Evidence for a link
between implicit and explicit motives and social cognitive pro-
cesses in everyday life. Journal of Personality and Social Psychology, 68,
Woike, B., Lavezzary, E., & Barsky, J. (2001). The influence of implicit
motives on memory processes. Journal of Personality and Social Psy
chology, 81, 935-945.
Zajonc, R. B., & Markus, H. (1984). Affect and cognition: The hard
interface. In C. Izard, J. Kagan, & R. B. Zajonc (Eds.), Emotion, cog
nition, and behavior (pp. 73-102). Cambridge, UK: Cambridge Uni
Zurbriggen, E. L., & Sturman, T. S. (2002). Linking motives and emo
tions: A test of McClelland’s hypotheses. Personality and Social Psy
chology Bulletin, 28, 521-535.
Received August 9, 2003
Revision accepted November 3, 2003
Schubert / BODILY FEEDBACK OF POWER 769