Abstract thinking increases one’s sense of power

Abstract

One’s subjective sense of power often has greater influence on behavior than the amount of power one actually possesses. We propose that this sense of power may be determined in part by one’s style of information processing. As abstract thought is less constraining than concrete thought, and having power leads to more abstract thought [Smith, P. K., & Trope, Y. (2006). You focus on the forest when you’re in charge of the trees: Power priming and abstract information processing. Journal of Personality and Social Psychology, 90, 578–596.], we predicted that thinking more abstractly would make one feel more powerful. Indeed, in four experiments, abstract thought led to a greater sense of power, greater preference for high-power roles, and more feelings of control over the environment, relative to both a concrete-thought and a control condition. This bidirectional relationship between power and abstract thinking suggests one way in which power hierarchies may be unintentionally perpetuated.

Full text

Abstract thinking increases one’s sense of power
Pamela K. Smith
a,*
, Danie
¨l H.J. Wigboldus
b
, Ap Dijksterhuis
c
a
Social and Organizational Psychology, Leiden University, P.O. Box 9555, 2300 RB Leiden, The Netherlands
b
Department of Social Psychology, Behavioural Science Institute, Radboud University Nijmegen, P.O. Box 9104, 6500 HE Nijmegen, The Netherlands
c
Social Psychology Program, University of Amsterdam, Roetersstraat 15, 1018 WB Amsterdam, The Netherlands
Received 20 June 2006; revised 6 December 2006
Available online 30 December 2006
Communicated by John Skowronski
Abstract
One’s subjective sense of power often has greater influence on behavior than the amount of power one actually possesses. We propose
that this sense of power may be determined in part by one’s style of information processing. As abstract thought is less constraining than
concrete thought, and having power leads to more abstract thought [Smith, P. K., & Trope, Y. (2006). You focus on the forest when
you’re in charge of the trees: Power priming and abstract information processing. Journal of Personality and Social Psychology, 90,
578–596.], we predicted that thinking more abstractly would make one feel more powerful. Indeed, in four experiments, abstract thought
led to a greater sense of power, greater preference for high-power roles, and more feelings of control over the environment, relative to
both a concrete-thought and a control condition. This bidirectional relationship between power and abstract thinking suggests one way
in which power hierarchies may be unintentionally perpetuated.
2007 Elsevier Inc. All rights reserved.
Keywords: Subjective power; Sense of power; Concrete vs. abstract thought; Bidirectional relationship; Unintentional maintenance of hierarchies
How do we know our place in society, whether we are a
top dog or the low man on the totem pole? In daily life,
people’s sense of power is more of a determinant of their
behavior than their actual power (e.g., Haidt & Rodin,
1999). Extending recent work on power and information
processing, the present research explores how thinking
styles might affect one’s subjective sense of power.
Power is a primary dimension of relationships and
broader societal dynamics (Fiske, 1992; Mazur, 1985).
Those who have power have more access to resources
and control how these are distributed to those without
power (e.g., De
´pret & Fiske, 1993; Keltner, Gruenfeld, &
Anderson, 2003; Thibaut & Kelley, 1959). That is, the pow-
erful control the powerless. The amount of power individ-
uals possess determines what behavior is acceptable,
whether they may ‘‘be themselves’’ or must follow social
norms (Keltner et al., 2003). Thus, it is critical for individ-
uals to know how much power they have.
Faced with this dilemma, one might seek objective infor-
mation about one’s level of power. Though such informa-
tion may illuminate the official power structure, it may
not accurately predict how people think and behave.
Instead, individuals’ subjective sense of power generally
drives the psychological effects of actual power (e.g.,
Anderson & Berdahl, 2002; Haidt & Rodin, 1999). When
objective and subjective appraisals of one’s power conflict,
subjective appraisals dominate and guide behavior (Bugen-
tal, Lyon, Krantz, & Cortez, 1997).
But what determines this sense of power? We propose
that it may be determined in part by one’s style of informa-
tion processing, by whether one thinks abstractly vs. con-
cretely. Consider the nature of abstract vs. concrete
thinking. Because abstract thinking moves beyond the
details of a stimulus (e.g., Levy, Freitas, & Salovey, 2002;
Liberman, Sagristano, & Trope, 2002), it is less constrain-
0022-1031/$ - see front matter 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jesp.2006.12.005
*
Corresponding author. Present address: Department of Social
Psychology, Behavioural Science Institute, Radboud University Nijmegen,
P.O. Box 9104, 6500 HE, Nijmegen, The Netherlands.
E-mail address: p.smith@psych.ru.nl (P.K. Smith).
www.elsevier.com/locate/jesp
Available online at www.sciencedirect.com
Journal of Experimental Social Psychology 44 (2008) 378–385

ing than concrete thinking. Abstract thinking involves gen-
eralization, which allows for more freedom and flexibility.
Viewing a chair as a piece of furniture leaves open more
possibilities for interpretation and action (e.g., Let’s use
it to reach that burned-out light bulb) than viewing that
same chair as a La-Z-Boy recliner. In this way, thinking
abstractly allows a person to take more control of the envi-
ronment. Indeed, Vallacher and Wegner (1989) found that
describing actions in more abstract terms is related to a
more internal locus of control. Concrete thinking, in con-
trast, narrows one’s focus and ties one to the particular
details in the environment.
Concrete thought is also associated with thinking more
about feasibility than abstract thought (e.g., Liberman &
Trope, 1998). Individuals who think more concretely are
more concerned with the difficulty of a course of action.
They are less likely to choose to do something that is hard,
even if it would result in a desired outcome, than individu-
als who think more abstractly. It is not surprising, then,
that being in an eager promotion state (i.e., focusing on
hopes and aspirations) is correlated with using more
abstract language than being in a vigilant prevention state
(i.e., focusing on duties and responsibilities; Semin, Hig-
gins, de Montes, Estourget, & Valencia, 2005).
Thus, thinking abstractly should lead individuals to feel
more powerful than thinking concretely. This prediction
also follows from recent work on the relationship between
power and abstract thought. Smith and Trope (2006) dem-
onstrated that the concept of having power is inherently
linked to abstract thinking, and the concept of lacking
power to concrete thinking. Priming people with having
power caused them to think more abstractly than priming
them with lacking power, even though these individuals
were unaware that the concept of power was activated.
Given that power and abstract thinking are so intimately
linked, the converse may also be true: abstract thinking
may lead individuals to feel more powerful.
In short, we propose that the link between power and
abstract thinking is bidirectional. Just as activating the
concept of power activates the associated representation
of abstract thinking (Smith & Trope, 2006), inducing peo-
ple to think more abstractly should activate a representa-
tion of abstract thinking, in turn activating the associated
construct of power. Our theorizing is similar to that of
Mussweiler (2006) in the domain of the perception-behav-
ior link. He demonstrated the bidirectionality of the rela-
tionship between stereotype activation and stereotype-
consistent behavior. Not only does activating a stereotype
lead to stereotype-consistent behavior (Dijksterhuis &
Bargh, 2001), but inducing individuals to behave in a ste-
reotype-consistent way also activates the associated stereo-
type. Such reversals are found in a variety of domains. For
example, the relationship between expectancies and lan-
guage abstraction is bidirectional. When a target person‘s
behavior is consistent with people’s expectancies, they
describe the behavior in more abstract terms (e.g., Maass,
Salvi, Arcuri, & Semin, 1989), a phenomenon known as
the linguistic expectancy bias (LEB). Wigboldus, Semin,
and Spears (2000) demonstrated that the reverse also
occurs: When information about a target is described in
abstract terms, people make more dispositional inferences,
thus reinforcing their expectancies.
In addition to contributing to our limited understanding
of the basis of people’s sense of power, such a bidirectional
link has important implications regarding the stability of
hierarchies. Though shifts and upheavals do occur, power
hierarchies are often stable (Sidanius & Pratto, 1999). The-
ories regarding this stability traditionally rely on deliberate,
intentional explanations. For example, system justification
theory posits that people are motivated to view existing
social arrangements as legitimate, even when this justifies
their own disadvantaged positions (Jost & Hunyady,
2002; Pratto, Sidanius, Stallworth, & Malle, 1994). Such
theories place a certain degree of blame or responsibility
on the shoulders of the powerful, the powerless, or both.
In contrast, the bidirectional relationship between power
and abstract thinking suggests that such hierarchies may
also be unintentionally maintained. Those with power
should automatically engage in more abstract thought than
those without power. If our hypothesis proves correct,
those with power should automatically feel more powerful
than those without power due to these differences in
thought, thus reinforcing already existing power differenc-
es. That is, hierarchies may be more tenacious than previ-
ously thought.
The following four experiments explore this link
between abstract thinking and power. Concrete or abstract
thinking was always primed in a purportedly unrelated task
before participants’ feelings of power were assessed. We
used three different measures: self-reported standing on
power-relevant traits, preference for higher- vs. lower-
power roles, and sense of control over the environment.
These measures allow us to explore the effects of different
styles of thinking on both general preferences and concrete
experiences.
Experiment 1
Actions can be construed at varying levels of abstrac-
tion. Thinking about how to perform an action is more
concrete, whereas thinking about why one would perform
an action is more abstract (Trope & Liberman, 2003; Vall-
acher & Wegner, 1987). To prime concrete vs. abstract
thought in this experiment, participants wrote repeatedly
about either how to pursue a given goal or why one would
pursue that same goal (Freitas, Gollwitzer, & Trope, 2004).
Then they rated themselves on a series of traits, including
traits that measured their sense of power.
Method
Participants
One hundred sixteen undergraduate students from the
University of Amsterdam participated in the experiment
P.K. Smith et al. / Journal of Experimental Social Psychology 44 (2008) 378–385 379

as part of a course requirement or for €7. Four participants
were dropped from the analyses for not following direc-
tions. Thus, 112 participants (37 males, 75 females)
1
were
included in the final analyses. Average age was 20.81 years
(SD = 3.14).
Procedure and materials
Participants began with the how/why task (Freitas et al.,
2004). Concrete-thought participants were told this was a
thought exercise in which people think about how their
ultimate life goals can be expressed through specific
actions. An initial example walked participants, step-by-
step, through the process of how one might find happiness
in life, ending with the very concrete step of ‘‘participating
in a psychology experiment.’’ Participants were then asked
to think themselves, step-by-step, about how they might
improve and maintain their health. First they were asked,
‘‘How do you improve and maintain good physical
health?’’ After answering, they were asked how they would
do this. For example, a participant who responded, ‘‘Exer-
cise regularly,’’ to the first question was then asked, ‘‘How
do you exercise regularly?’’ Their response to this second
question was then used for another ‘‘how’’ question. In this
way, concrete-thought participants provided four ‘‘how’’
responses.
Abstract-thought participants were told this was a
thought exercise in which people think about how their
actions relate to their ultimate life goals. An initial example
walked participants, step-by-step, through the reasons why
one might participate in a psychology experiment, ending
with the very abstract reason of ‘‘to find happiness in life.’’
Participants were then asked to think themselves, step-by-
step, about why they might improve and maintain their
health. First they were asked, ‘‘Why do you improve and
maintain good physical health?’’ After answering, they
were asked why they would do this. For example, a partic-
ipant who responded, ‘‘To lose weight,’’ to the first ques-
tion was then asked, ‘‘Why do you want to lose weight?’’
Their response to this second question was then used for
another ‘‘why’’ question. In this way, abstract-thought par-
ticipants provided four ‘‘why’’ responses.
Participants next rated themselves on 25 items. Each
item consisted of a 9-point scale, anchored on each end
by a trait. For example, ‘‘boring’’ and ‘‘fun’’ anchored
one item’s scale. Above the scale was the stem ‘‘To what
extent would you say you are:’’ Scattered throughout these
items were 7 trait pairs related to power: submissive–dom-
inant, passive–active, unassertive–assertive, timid–firm,
uncertain–certain, insecure–confident, and dependent–in-
dependent. These traits have been used in previous research
to measure individuals’ sense of power or dominance (e.g.,
Stapel & Van der Zee, 2006; Tiedens & Jimenez, 2003; Wig-
gins, Trapnell, & Phillips, 1988).
Next participants indicated on 9-point scales (0 = not at
all,8=very much) how difficult, interesting and enjoyable
the how/why task was. Finally, they were probed for suspi-
cion and debriefed.
Results
Self ratings on power-relevant traits
Responses to the 7 trait pairs related to power were
averaged together (a= .83). Abstract-thought participants
(M= 6.08, SD = 1.03) rated themselves higher on these
power-relevant traits than concrete-thought participants
(M= 5.61, SD = 1.14), F(1, 110) = 5.27, p= .02, g2
p
¼:05.
Perhaps our thought manipulation simply led to a
response bias so that participants who thought abstractly
rated themselves higher on all traits. Most of the remaining
trait pairs (e.g., young–old, short–tall, and ugly–beautiful)
were fillers unrelated to each other. However, 5 of these
remaining trait pairs—unpleasant–pleasant, unlikeable–
likeable, unfriendly–friendly, cold–warm, and nice–mean
(reverse-coded)—tapped into the general construct of
sociability (a= .82). If our effects are specific to power,
thought condition should not have affected participants’
self ratings on these sociability items. Indeed, thought con-
dition had no effect on the average of these items, F<1.
Additional measures
Thought condition had no effects on the additional mea-
sures, ps > .11.
Experiment 2
Individuals who first thought abstractly expressed an
elevated sense of power. One possible consequence is that
these individuals may also be more interested in higher-
power roles. Such a pattern occurs with personality domi-
nance: Individuals high in dominance are much more likely
to want to become a leader than are individuals low in
dominance (Fleischer & Chertkoff, 1986). Given that
abstract thought makes an individual feel more powerful
or dominant, that person might then feel more comfortable
taking on a high-power job. To test this idea, in Experi-
ment 2 participants completed the how/why task from
Experiment 1. Then they read a series of three scenarios,
each describing two available roles within a business. One
role was relatively low in power, the other relatively high
in power. Participants rated which of the two roles they
preferred to have. We predicted that participants who first
thought abstractly would show greater relative preference
for the high-power role than participants who first thought
concretely.
Method
Participants
Ninety-six undergraduate students from the University
of Amsterdam took part in the experiment as part of a
1
In all experiments, participant gender did not significantly moderate
thought effects.
380 P.K. Smith et al. / Journal of Experimental Social Psychology 44 (2008) 378–385

course requirement or for €7. Two participants were
dropped from analyses for not following directions. Thus,
94 participants (27 males, 67 females) were included in
the final analyses. Average age was 21.93 years
(SD = 2.93).
Procedure and materials
First participants completed the how/why task as in
Experiment 1. Immediately afterwards, they reported how
they felt on an 11-point scale (5=very bad,+5=very
good). Then they read and responded to a series of three
scenarios involving a construction company, a toy compa-
ny, and an art gallery. Each scenario described two roles
within that group (e.g., supervisor and employee for the
construction company). One was a high-power role that
involved supervising and evaluating others, assigning tasks
to others, and making final decisions. The other was a low-
power role. The tasks for the low-power role varied by sce-
nario (e.g., construction company employees performed
‘‘both construction/architecture tasks and interior/exterior
design tasks, depending on their personal skills and inter-
ests’’), but this role always involved being supervised and
evaluated by the person in the high-power role. The scenar-
ios were carefully designed so that the low- and high-power
roles were similar in desirability. After reading each scenar-
io, participants rated which role they would prefer to have
on a single 5-point scale anchored by the two roles
(1 = definitely Role A,2=maybe Role A,3=don’t know,
4=maybe Role B,5=definitely Role B). The order of
the anchors (low-power role first vs. high-power role first)
and the order of the scenarios were counterbalanced.
Finally, participants answered several questions on
9-point scales (0 = not at all,8=very much) to rule out
alternative explanations, such as differences in mood and
motivation between thought conditions. They rated how
difficult, interesting and enjoyable the how/why and sce-
nario tasks were. They also indicated how much effort they
put into the scenario task, as well as how they felt
(5=very bad,+5=very good). Finally, they were probed
for suspicion and debriefed.
Results
Role preference
Responses were recoded so that higher numbers indicated
greater preference for the high-power role, and then aver-
aged together. Concrete-thought participants (M= 2.79,
SD = 0.86) showed more preference for the low-power role
(relative to the high-power role) than abstract-thought par-
ticipants (M= 3.21, SD = 0.89), F(1, 92) = 5.22, p= .02,
g2
p
¼:05.
Additional measures
Abstract-thought participants (M= 3.60, SD = 1.95)
found the how/why task more difficult than concrete-
thought participants (M= 2.70, SD = 1.72), F(1, 92) =
5.55, p= .02, g2
p
¼:06. However, difficulty was not related
to role preferences, r(92) = .13, p> .19, and including dif-
ficulty as a covariate somewhat strengthened the effect of
thought condition on role preferences, F(1, 91) = 7.40,
p= .008, g2
p
¼:08. Thought condition did not affect the
other measures, Fs<1.
Experiment 3
Experiments 1 and 2 demonstrated that having partici-
pants think in terms of ‘‘why’’ made them feel more pow-
erful and made them more interested in a high-power
role than having them think in terms of ‘‘how.’’ In both
experiments, abstract-thought priming involved thinking
about high-level goals and values, whereas concrete-
thought priming involved thinking about details of specific
procedures. However, the high-power roles in Experiment
2 also involved monitoring and pursuing the larger goals
of the group (e.g., supervising and evaluating others),
and the low-power roles involved working on specific tasks
(e.g., selecting furniture). After completing the thought
task, participants’ ability to think about either high-level
goals (abstract thought) or details (concrete thought) might
have been salient, and participants then picked their roles
accordingly. Such a direct mapping of procedures from
one task to another is not the same as our proposed general
effect of concrete or abstract thought on experienced
power.
We addressed this alternative explanation in Experiment
3 by using a perceptual manipulation of concrete/abstract
thought involving hierarchical figures, larger figures made
up of an arrangement of smaller figures, such as an Omade
up of Ts(Navon, 1977). Abstract or holistic thinking may
be primed by having participants focus on the overall shape
of these figures (the O), and concrete thinking by having
them focus on the smaller components (the Ts; Macrae &
Lewis, 2002). This procedure does not directly map onto
the duties described for the low- and high-power roles.
Again we predicted that abstract-thought participants
would show a greater relative preference for high-power
roles than concrete-thought participants. Additionally, we
added a control condition to explore the direction of the
effects.
Method
Participants
One hundred forty-eight undergraduate students from
the University of Amsterdam took part in the experiment
as part of a course requirement or for €7. Four participants
were dropped from the analyses for not following direc-
tions. Thus, 144 participants (52 males, 92 females) were
included in the final analyses. Average age was 21.04 years
(SD = 2.32).
Procedure and materials
Participants in the concrete- and abstract-thought con-
ditions first completed a letter-identification task (Macrae
P.K. Smith et al. / Journal of Experimental Social Psychology 44 (2008) 378–385 381

& Lewis, 2002; Navon, 1977). They were presented with a
series of 112 composite letters, large letters composed of
smaller letters. The small letters always differed from the
larger overall letter (e.g., an Scomposed of Es). For each
figure, concrete-thought participants reported ‘‘the small
letter that the figure is made up of,’’ whereas abstract-
thought participants reported ‘‘the large letter formed by
the overall shape of the figure.’’ Control participants com-
pleted unrelated filler tasks for an equivalent amount of
time (2–3 min). All participants then reported how they felt
(5=very bad,+5=very good). Next they read the sce-
narios from Experiment 2. After each scenario, partici-
pants rated separately how interested they were in each
of the two roles (0 = not at all,8=very much). Question
order and scenario order were counterbalanced.
Finally, participants answered several questions on 9-
point scales (0 = not at all,8=very much). Concrete- and
abstract-thought participants rated how difficult, interest-
ing and enjoyable the letter-identification task was. All par-
ticipants also indicated how difficult, interesting and
enjoyable the scenario task was and how much effort they
put into it. Finally, participants were probed for suspicion
and debriefed.
Results
Role preferences
A 3 (Thought condition: concrete vs. control vs.
abstract) ·2 (Role: low-power vs. high-power) mixed-
model ANOVA was run on ratings of interest in the roles,
with the last factor within participants. Only the two-way
interaction was significant, F(2, 141) = 3.27, p= .04,
g2
p
¼:04. The means are listed in Table 1. Concrete-thought
and control participants expressed equal interest in having
the low-power and high-power roles, Fs < 1.5. Abstract-
thought participants, however, were more interested in
the high-power role than the low-power role,
F(1, 46) = 4.73, p= .03, g2
p
¼:09. In short, abstract-
thought participants showed greater relative preference
for the high-power role than both control, p= .05, and
concrete-thought participants, p= .02, who did not differ,
p= .68.
Looking at the two job types separately, abstract-
thought participants were more interested in the high-
power job than both control, p= .003, and concrete-
thought participants, p= .02, who did not differ, p= .49.
The three conditions did not differ in their interest in a
low-power job, F<1.
Additional measures
Control participants (M= 2.75, SD = 1.33) felt better
than concrete- (M= 1.40, SD = 2.38) and abstract-thought
participants (M= 1.83, SD = 1.88), F(2, 141) = 6.57,
p= .002, and abstract-thought participants (M= 1.53,
SD = 1.65) found the letter task more difficult than con-
crete-thought participants (M= 0.93, SD = 1.19),
F(1,90) = 3.93, p= .05. However, these variables were
uncorrelated with role preferences, except for a marginal
relationship between mood and interest in having a low-
power role, r(142) = .15, p= .07. Including either variable
as a covariate in the above mixed-model ANOVA did
not reduce the effect of thought condition on job preferenc-
es. Thought condition did not affect the other measures,
ps > .22.
Experiment 4
Experiments 2 and 3 demonstrated that thinking more
abstractly made participants express more interest in exer-
cising power over others. Such effects reflect social power,
which involves relationships between people or groups.
However, power also involves intrapersonal control or per-
sonal power (e.g., reflexive control: Thibaut & Kelley,
1959), one’s ability to control one’s own outcomes and
environment (see Overbeck & Park, 2001 for more on the
distinction between social and personal power). This final
experiment explored whether priming concrete vs. abstract
thought can influence personal power.
Participants completed the how/why task from Experi-
ments 1 and 2. They then did a lexical decision task where
they tried to respond to a letter string before the computer
erased this string from the screen (Dijksterhuis, Preston,
Wegner, & Aarts, in press). We predicted that participants
who first thought abstractly would be more likely to think
they had removed the word themselves, and thus had more
control over the situation, than participants who first
thought concretely.
Method
Participants
Sixty-five native Dutch speakers from the University of
Amsterdam took part in the experiment as part of a course
requirement or for €7. Seven participants were dropped
from the analyses: five for not following directions and
two due to computer crashes. Thus, 58 participants (12
males, 46 females) were included in the final analyses.
Average age was 20.44 years (SD = 1.69).
Procedure and materials
First participants completed the how/why task. The lex-
ical decision task followed (Dijksterhuis et al., in press).
Participants were told they would classify letter strings as
Table 1
Interest in holding low- and high-power jobs by primed mindset,
Experiment 3
Job Control Concrete Abstract
M SD M SD M SD
Low power 4.99
a
1.36 5.01
a
1.49 4.82
a
1.57
High power 4.64
a
1.30 4.84
a
1.56 5.49
b
1.28
Note. Means in the same row or column that do not share subscripts differ
at p< .05.
382 P.K. Smith et al. / Journal of Experimental Social Psychology 44 (2008) 378–385

words or nonwords by pressing the Dor Kkey as quickly
as possible. Pressing a key removed a letter string from the
screen. However, the instructions also explained that the
computer could remove the letter string before they
responded. Participants were told to try to respond quickly
enough to beat the computer. After each trial they would
rate who removed the string: themselves or the computer.
Participants first completed 12 practice trials, then 72
experimental trials. Each trial began with a 300 ms fixation
(XXX), followed by a letter string. In half the trials, the
string was a random letter string. In the remaining half,
it was a 4–7 letter, medium to high frequency Dutch word
(e.g., BERG [mountain]). The string was automatically
removed either after the participant had responded or after
a maximum word time, whichever came first. The maxi-
mum time was 450, 500, 550, 600, 650, or 700 ms. Each
maximum time was used on 2 practice trials and 12 exper-
imental trials, counterbalanced between words and non-
words. To ensure participants actually saw each string,
participants had to respond to the string by pressing a
key even if the computer had removed it first.
After each response participants were asked, ‘‘Was it
you or was it the computer that removed the letter string?’’
Responses were on a 6-point scale (1 = I’m sure it was me,
2=I think it was me,3=If I would have to guess I’d say it
was me,4=If I would have to guess I’d say it was the com-
puter,5=I think it was the computer,6=I’m sure it was
the computer).
After the lexical decision task, participants also reported
what percentage of time (0–100%) they thought they
removed the letter strings themselves. Several additional
questions were asked on 9-point scales (0 = not at all,
8=very much). Participants rated how difficult, interesting,
and enjoyable the how/why and lexical decision tasks were.
They answered additional questions about the lexical deci-
sion task: how well they thought they did, how much effort
they put into it, and how important it was for them to do
well and beat the computer. Participants also reported how
they felt (5=very bad,+5=very good). Finally, they
were probed for suspicion and debriefed.
Results and discussion
Performance on lexical decision task
Concrete- and abstract-thought participants did not dif-
fer in their percentage of correct responses or their average
response time, ps > .15. Abstract-thought participants
(M= 98.0%, SD = 3.4) tended to remove the letter strings
from the screen themselves a greater percentage of the time
than did concrete-thought participants (M= 96.5%,
SD = 2.3), F(1, 56) = 3.71, p= .06, g2
p
¼:06.
Sense of control
Ratings of who removed the string were averaged across
the 72 experimental trials. Lower numbers indicate a great-
er sense that the participant controlled the removal of the
word. Abstract-thought participants (M=2.29, SD =0.83)
were more certain that the letter string had been removed
by themselves, than concrete-thought participants (M=
2.80, SD = 0.98), F(1, 56) = 4.39, p= .04, g2
p
¼:07. This
was not moderated by whether the letter string was a word
or nonword, F<1.
Of course, since abstract-thought participants indeed
tended to remove the strings themselves more often, this
heightened sense of control may have merely reflected reali-
ty. However, when only the trials in which participants
removed the string themselves were examined,
2
abstract-
thought participants (M= 2.23, SD = 0.85) still rated them-
selves as more definitely in control than concrete-thought
participants (M= 2.72, SD = 1.00), F(1,56) = 4.02,
p< .05, g2
p
¼:07. Furthermore, actual performance did not
mediate the relationship between thought and rated sense
of control according to a Sobel test (Baron & Kenny,
1986), z=1.17, p= .24.
At the end of the experiment, abstract-thought partici-
pants (M= 66.59, SD = 19.56) also said they removed a
higher percentage of the words themselves than concrete-
thought participants (M= 53.74, SD = 24.86), F(1, 56) =
4.68, p= .03, g2
p
¼:08. Actual performance did not mediate
this effect according to a Sobel test, z= 0.49, p= .62.
Additional measures
Thought condition had no effects on the additional mea-
sures, ps > .09.
General discussion
Across two concrete/abstract thought manipulations
and three measures of perceived power, priming partici-
pants with abstract thought made them feel more powerful
than priming them with concrete thought, or not priming
them at all. These results cannot be explained by changes
in mood or motivation, or by a simple mapping of the pro-
cedures in the thought task onto the procedures of the
power measures. Instead, the less constraining nature of
abstract thought in itself increased participants’ sense of
power. These experiments extend the finding of Smith
and Trope (2006) that merely priming people with having
power made them think more abstractly than priming them
with lacking power. The causality also appears to work in
the opposite direction.
This research sheds new light on what determines peo-
ple’s sense of power. We know that feeling like a top dog
or bottom beagle can be based on information in the envi-
ronment that clearly relates to power, such as level of
2
Nineteen participants removed the word themselves in all trials. If
trials where the computer removed the word are analyzed for the other 39
participants, concrete- and abstract-thought participants do not differ in
their ratings of control, F< 1. This result suggests that the feelings of
control engendered by the thought conditions may be bounded by reality.
However, since the computer removed the word on an average of 2 trials
per participant, and even the slowest participants were beaten by the
computer on only 8 trials, it may also be an artifact of the limited data
available.
P.K. Smith et al. / Journal of Experimental Social Psychology 44 (2008) 378–385 383

dependence (Bacharach & Lawler, 1976; Hegtvedt, 1988)
and others’ deference behaviors (e.g., Ellyson & Dovidio,
1985). This previous research tended to imply that a per-
son’s sense of power is overtly calculated or derived based
on outside input. The present studies represent the first
attempt to examine how one’s own thought processes can
unintentionally influence one’s sense of power. Based on
their responses to our funnel debriefings, our participants
saw our manipulations of concrete/abstract thought as
completely unrelated to our dependent measures. Thus,
these studies demonstrate an implicit basis for a sense of
power, thereby tying into a broader array of work on sub-
tle signs and signals of power. Power cues do not have to be
as obvious as a corner office or a king’s crown. Even some-
thing as simple as vertical position is related to perceptions
of power (Schubert, 2005): Placing stimuli higher on a com-
puter screen gains them more respect.
This research also adds to our burgeoning knowledge of
how power hierarchies are perpetuated. The bidirectional
relationship between power and abstract thinking suggests
that hierarchies may be unintentionally maintained. When
Joan is promoted into a higher power position, she will
start thinking more abstractly due to increased power
(Smith & Trope, 2006). Based on the present research, this
increase in abstract thought will also make her feel more
powerful, thus leading again to abstract thought, and so
on and so forth.
Our description of Joan’s situation raises the additional,
untested question of how one person’s abstract thinking
affects other people’s perceptions of that person’s power.
When Joan’s coworkers and subordinates perceive her
using more abstract language and generally taking a ‘‘big
picture’’ view, will this also make them view her as more
powerful? If so, even if their perception of Joan has only
changed implicitly, they should be more likely to respond
to her in a more subordinate manner (Tiedens & Fragale,
2003), thus further solidifying her place in the hierarchy.
This research also suggests one potential way to subvert
existing hierarchies. Martorana, Galinsky, and Rao (2005)
propose that one critical component necessary for subordi-
nates to fight the system is a sense of power. If people in
low-power positions nonetheless feel powerful, they are
more likely to attempt to change the system. But overt
manipulations of a sense of power are less likely to work
when one clearly has little power. Subtle manipulations,
such as simply taking a more abstract perspective, may
be the first step that helps the powerless challenge the
powerful.
References
Anderson, C., & Berdahl, J. L. (2002). The experience of power:
examining the effects of power on approach and inhibition tendencies.
Journal of Personality and Social Psychology, 83, 1362–1377.
Bacharach, S. B., & Lawler, E. J. (1976). Perception of power. Social
Forces, 55, 123–134.
Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable
distinction in social psychological research: Conceptual, strategic, and
statistical considerations. Journal of Personality and Social Psychology,
51, 1173–1182.
Bugental, D. B., Lyon, J. E., Krantz, J., & Cortez, V. (1997). Who’s the boss?
Differential accessibility of dominance ideation in parent–child rela-
tionships. Journal of Personality and Social Psychology, 72, 1297–1309.
De
´pret, E., & Fiske, S. T. (1993). Social cognition and power: some
cognitive consequences of social structure as a source of control
deprivation. In G. Weary, F. Gleicher, & K. L. Marsh (Eds.), Control
motivation and social cognition (pp. 176–202). New York: Springer.
Dijksterhuis, A., & Bargh, J. A. (2001). The perception-behavior
expressway: automatic effects of social perception on social behavior.
In M. P. Zanna (Ed.). Advances in experimental social psychology (Vol.
33, pp. 1–40). San Diego, CA: Academic Press.
Dijksterhuis, A., Preston, J., Wegner, D.M., & Aarts, H. Effects of
subliminal priming of self and God on self-attribution of authorship
for events. Journal of Experimental Social Psychology, 44, 2–9.
Ellyson, S. L., & Dovidio, J. F. (1985). Power, dominance, and nonverbal
behavior. New York: Springer.
Fiske, A. P. (1992). The four elementary forms of sociality: framework
for a unified theory of social relations. Psychological Review, 99,
689–723.
Fleischer, R. A., & Chertkoff, J. M. (1986). Effects of dominance and sex
on leader selection in dyadic work groups. Journal of Personality and
Social Psychology, 50, 94–99.
Freitas, A. L., Gollwitzer, P., & Trope, Y. (2004). The influence of
abstract and concrete mindsets on anticipating and guiding others’
self-regulatory efforts. Journal of Experimental Social Psychology, 40,
739–752.
Haidt, J., & Rodin, J. (1999). Control and efficacy as interdisciplinary
bridges. Review of General Psychology, 3, 317–337.
Hegtvedt, K. A. (1988). Social determinants of perception: power, equity,
and status effects in an exchange situation. Social Psychology Quar-
terly, 51, 141–153.
Jost, J. T., & Hunyady, O. (2002). The psychology of system justification
and the palliative function of ideology. European Review of Social
Psychology, 13, 111–153.
Keltner, D., Gruenfeld, D. H., & Anderson, C. (2003). Power, approach,
and inhibition. Psychological Review, 110, 265–284.
Levy, S. R., Freitas, A. L., & Salovey, P. (2002). Construing action
abstractly and blurring social distinctions: Implications for perceiving
homogeneity among, but also empathizing with and helping, others.
Journal of Personality and Social Psychology, 83, 1224–1238.
Liberman, N., Sagristano, M. D., & Trope, Y. (2002). The effect of
temporal distance on level of mental construal. Journal of Experimen-
tal Social Psychology, 38, 523–534.
Liberman, N., & Trope, Y. (1998). The role of feasibility and desirability
considerations in near and distant future decisions: A test of temporal
construal theory. Journal of Personality and Social Psychology, 75,
5–18.
Maass, A., Salvi, D., Arcuri, L., & Semin, G. R. (1989). Language use in
intergroup contexts: The linguistic intergroup bias. Journal of Person-
ality and Social Psychology, 57, 981–993.
Macrae, C. N., & Lewis, H. L. (2002). Do I know you?: Processing
orientation and face recognition. Psychological Science, 13, 194–196.
Martorana, P. V., Galinsky, A. D., & Rao, H. (2005). From system
justification to system condemnation: Antecedents of attempts to
change power hierarchies. In E. A. Mannix, M. A. Neale, & M.
Thomas-Hunt (Eds.). Research on managing in teams and groups (Vol.
7, pp. 285–315). Greenwich, CT: Elsevier Science.
Mazur, A. (1985). A biosocial model of status in face-to-face primate
groups. Social Forces, 64, 377–402.
Mussweiler, T. (2006). Doing is for thinking! Stereotype activation by
stereotypic movements. Psychological Science, 17, 17–21.
Navon, D. (1977). Forest before trees: The precedence of global features in
visual perception. Cognitive Psychology, 9, 353–383.
Overbeck, J. R., & Park, B. (2001). When power does not corrupt:
Superior individuation processes among powerful perceivers. Journal
of Personality and Social Psychology, 81, 549–565.
384 P.K. Smith et al. / Journal of Experimental Social Psychology 44 (2008) 378–385

Pratto, F., Sidanius, J., Stallworth, L. M., & Malle, B. F. (1994). Social
dominance orientation: A personality variable predicting social and
political attitudes. Journal of Personality and Social Psychology, 67,
741–763.
Schubert, T. W. (2005). Your highness: Vertical positions as perceptual
symbols of power. Journal of Personality and Social Psychology, 89,
1–21.
Semin, G. R., Higgins, T., de Montes, L. G., Estourget, Y., & Valencia, J.
F. (2005). Linguistic signatures of regulatory focus: How abstraction
fits promotion more than prevention. Journal of Personality and Social
Psychology, 89, 36–45.
Sidanius, J., & Pratto, F. (1999). Social dominance: An intergroup theory of
social hierarchy and oppression. Cambridge, UK: Cambridge Univer-
sity Press.
Smith, P. K., & Trope, Y. (2006). You focus on the forest when you’re
in charge of the trees: Power priming and abstract information
processing. Journal of Personality and Social Psychology, 90,
578–596.
Stapel, D. A., & Van der Zee, K. I. (2006). The self salience model of
other-to-self effects: Integrating principles of self-enhancement, com-
plementarity, and imitation. Journal of Personality and Social Psy-
chology, 90, 258–271.
Thibaut, J. W., & Kelley, H. H. (1959). Power and dependence. In The
social psychology of groups (pp. 100–125). New York: Wiley & Sons.
Tiedens, L. Z., & Fragale, A. R. (2003). Power moves: Complementarity
in dominant and submissive nonverbal behavior. Journal of Personality
and Social Psychology, 84, 558–568.
Tiedens, L. Z., & Jimenez, M. C. (2003). Assimilation for affiliation and
contrast for control: Complementary self-construals. Journal of
Personality and Social Psychology, 85, 1049–1061.
Trope, Y., & Liberman, N. (2003). Temporal construal. Psychological
Review, 110, 403–421.
Vallacher, R. R., & Wegner, D. M. (1987). What do people think they’re
doing? Action identification and human behavior. Psychological
Review, 94, 3–15.
Vallacher, R. R., & Wegner, D. M. (1989). Levels of personal agency:
Individual variation in action identification. Journal of Personality and
Social Psychology, 57, 660–671.
Wigboldus, D. H. J., Semin, G. R., & Spears, R. (2000). How do we
communicate stereotypes? Linguistic bases and inferential consequenc-
es. Journal of Personality and Social Psychology, 78, 5–18.
Wiggins, J. S., Trapnell, P., & Phillips, N. (1988). Psychometric and
geometric characteristics of the Revised Interpersonal Adjective Scales
(IAS-R). Multivariate Behavioral Research, 23, 517–530.
P.K. Smith et al. / Journal of Experimental Social Psychology 44 (2008) 378–385 385