Copyright 1999 by the American Psychological Association, Inc.

0022-3514/99/S3.00

Journal of Personality and Social Psychology

1999, Vol. 76, No. 6, 893-910

The Chameleon Effect: The Perception-Behavior Link

and Social Interaction
Tanya L. Chartrand and John A. Bargh
New York University
The chameleon effect refers to nonconscious mimicry of the postures, mannerisms, facial expressions,
and other behaviors of one's interaction partners, such that one's behavior passively and unintentionally
changes to match that of others in one's current social environment. The authors suggest that the
mechanism involved is Has. perception-behavior link, the recently documented finding (e.g., J. A. Bargh,
M. Chen, & L. Burrows, 1996) that the mere perception of another's behavior automatically increases the
likelihood of engaging in that behavior oneself. Experiment 1 showed that the motor behavior of
participants unintentionally matched that of strangers with whom they worked on a task. Experiment 2
had confederates mimic the posture and movements of participants and showed that mimicry facilitates
the smoothness of interactions and increases liking between interaction partners. Experiment 3 showed
that dispositionally empathic individuals exhibit the chameleon effect to a greater extent than do other
people.

We believe that the chameleon is a better one. In the motion

picture Zelig, Woody Allen plays a human chameleon who cannot
help but take on the behavior, personality, and values of whomever
he is with. Like a chameleon changing its color to match its current
surroundings, Zelig's behavior changes to match the norms and
values of the group with which he is currently involved. Although
Allen's film took this phenomenon to laughable extremes, it is
nevertheless a common experience to discover, after the fact, that
one has taken on the accent, speech patterns, and even behavioral
mannerisms of one's interaction partners. The naturalness and
nonconsciousness of this process was frequently commented on by
the author Georges Simenon, whose fictional Inspector Maigret
(the subject of the opening epigraph) routinely immersed himself
in the lives of murder victims as a favorite method for solving the
crimes.

He looked about his surroundings. They had become so familiar to

him that, without realizing it, he was beginning to take on some of the
mannerisms of the people who lived there.
Georges Simenon, Maigret and the Toy Village

As the saying goes, "Monkey see, monkey do." Primates, including humans, are quite good at imitation. Such imitation, in all
primates, has generally been considered to be an intentional,
goal-directed activityfor instance, mimicry helps one to learn
vicariously from the experience of conspecifics or to ingratiate
oneself to the other person (see Bandura, 1977; Galef, 1988;
Heyes, 1993; Piaget, 1946; Tomasello, Savage-Rumbaugh, &
Kruger, 1993). Recently, however, several studies have documented a passive, direct effect of social perception on social
behavior, an effect that is unintended and not in the service of any
discernible purpose (Bargh, Chen, & Burrows, 1996; Chen &
Bargh, 1997; Dijksterhuis, Spears, et al., 1998; Dijksterhuis & van
Knippenberg, 1998; Macrae et al., 1998; Mussweiler & Foerster,
1998). These findings suggest that imitation and mimicry effects in
humans might often be unintentional (Chen, Chartrand, Lee Chai,
& Bargh, 1998). As the popular meaning of the phrase "to ape" is
"to intentionally imitate," perhaps the monkey metaphor may not
be the most appropriate animal metaphor for the phenomenon.

Such a "chameleon effect" may manifest itself in different ways.

One may notice using the idiosyncratic verbal expressions or
speech inflections of a friend. Or one may notice crossing one's
arms while talking with someone else who has his or her arms
crossed. Common to all such cases is that one typically does not
notice doing these thingsif at alluntil after the fact.
Perceiving Is for Doing

The Perception-Behavior Link

This research was supported in part by National Science Foundation

Grants SBR-9409448 and SBR-9809000.
We thank Vinnie Chawla, Catherine Cordova, Elina Geskin, Nora
Guerra, Peter Karp, Marianna Moliver, and Christina Ungerer for serving
as experimenters and confederates and Annette Lee Chai, Ap Dijksterhuis,
Peter Gollwitzer, Katelyn McKenna, and Dan Wegner for helpful suggestions on an earlier version of this article.
Correspondence concerning this article should be addressed to Tanya L.
Chartrand, who is now at the Department of Psychology, Ohio State
University, 1885 Neil Avenue Mall, Columbus, Ohio 43210, or to John A.
Bargh, Department of Psychology, New York University, 6 Washington
Place, Seventh Floor, New York, New York 10003. Electronic mail may be
sent to [email protected] or to [email protected].

Throughout the history of psychology, many have argued that

the act of perceiving another person's behavior creates a tendency
to behave similarly oneself. To begin with, William James's principle of ideomotor action held that merely thinking about a behavior increases the tendency to engage in that behavior (James,
1890). This principle is in harmony with the proposed existence of
a perception-action link, if one assumes perceptual activity to be
one source of behavior-relevant ideation. Making just this assumption, Berkowitz (1984) invoked the principle of ideomotor action
in his revised theory of how violence portrayed in the mass media
increases the probability of aggression in the viewer. He argued
893

894

CHARTRAND AND BARGH

that activation spread automatically in memory from representations of the perceived violent acts to other aggressive ideas of the
viewer. This spreading activation to aggressive behavioral representations, he asserted, automatically led the viewer to behave in a
more aggressive manner.
Carver, Ganellen, Framing, and Chambers (1983) tested
Berkowitz's ideomotor account of modeling effects. They posited
that individuals use interpretive schemas for perceiving and interpreting behaviors and behavioral schemas for producing behaviors. Because these two schemas are assumed to have substantial
overlap in their semantic features, they should tend to become
active at the same times. Carver et al. predicted that perceiving a
hostile behavior in the environment would activate not only one's
hostile interpretive schema, but one's hostile behavioral schema as
well, so that the mere act of interpreting the behavior as hostile
would make the perceiver more likely to behave in a hostile
manner. Participants first were primed (or not) with hostile-related
stimuli and then, in an ostensibly unrelated study, were to give
shocks to another participant each time the latter made an error in
a learning task. Results supported the hypothesis: Relative to the
control group, participants who had been previously exposed to
hostility-related priming stimuli gave longer shocks to the
"learner."
Researchers in the area of language acquisition have also posited a "common-coding" principle to account for rapid language
acquisition in young children. In a seminal paper, Lashley (1951)
asserted that "the processes of language comprehension and language production have too much in common to depend on wholly
different mechanisms" (p. 120). Following Lashley, Prinz (1990)
hypothesized a common, or shared, representational system for
language comprehension and action codes. He further suggested
that the coding system for perceiving behaviors in others is the
same as for performing those behaviorsand if so, he argued, that
code cannot be used simultaneously in the service of perception
and of behavior.
In an experimental demonstration of Prinz's (1990) thesis,
Muesseler and Hommel (1997) instructed participants to reproduce
certain sequences of four left and right arrow key presses as
quickly as they could on each trial (the keys were labeled " < " and
" > " respectively; thus, on one trial the sequence might be "< <
> < " and on another trial "> < > <"). Participants practiced the
sequence until they were ready to perform it rapidly. As soon as
they made the first keypress of the sequence, however, the computer display briefly presented an additional left or right arrow key
that they had been instructed to append to the end of their practiced
sequence. The timing of this presentation was such that it occurred
precisely when the participant was pressing the second of the four
keys in the sequence. Which of the two keys ("<" or ">'') was to
be pressed at the end of the practiced sequence was manipulated to
be either the same or the opposite of the key actually being pressed
at that moment. As hypothesized, participants made more errors
(i.e., more often pressed the wrong extra key) if the presented
symbol corresponded to the one they were pressing at that moment
than when it was different. Apparently, the behavior of pressing
the right (or left) arrow key interfered with the ability to perceive
the right (or left) arrow key symbol, consistent with Prinz's position that the same representation is used for perceiving as for
behaving, and cannot be used for both at the same moment in time.

Priming of Social Behavior

The existence of an automatic, unintended, and passive effect of
perception on behavior has important ramifications for whether
social behavior can occur nonconsciously and without intention. If
the effect of perception on behavior is automatic, then direct
environmental causation of social behavior could be produced in a
two-step process. The first would involve automatic (i.e., not
effortful or consciously guided) perceptual categorization and interpretation of social behavior (environment to perception), with
this perceptual activation continuing on to activate corresponding
behavioral representations (perception to behavior). In this way,
the entire sequence from environment to behavior would occur
automatically, without conscious choice or guidance playing a role
(see Bargh & Chartrand, 1999).
Regarding the first stage of this hypothetical sequence, it is now
widely accepted that much of social perceptual activity is automated (i.e., immediate, efficient, and not consciously guided).
Many years of research have demonstrated the variety of ways in
which (a) behaviors are encoded spontaneously and without intention in terms of relevant trait concepts (e.g., Bargh & Thein, 1985;
Carlston & Skowronski, 1994; Uleman, Newman, & Moskowitz,
1996; Winter & Uleman, 1984), (b) contextual priming of trait
concepts changes the perceiver's interpretation of an identical
behavior through temporarily increasing their accessibility or
readiness to be used (see Bargh, 1989; Higgins, 1989, 1996; Wyer
& Srull, 1989, for reviews), and (c) stereotypes of social groups
become activated automatically upon the mere perception of the
distinguishing features of a group member (e.g., Bargh, 1994,
1999; Brewer, 1988; Devine, 1989).
ThusY" if the automatic activation of perceptual representations
continuously activates behavioral representations, the same priming manipulations that have been shown to influence social perception should also influence social behavior. In support of this
prediction, Bargh, Chen, et al. (1996) found that when stereotypes
or trait constructs were "primed," or nonconsciously activated in
the course of an unrelated task, the participant subsequently was
more likely to act in line with the content of the primed trait
construct or stereotype. In Experiment 1, in what was ostensibly a
language test, participants were exposed to words related to either
rudeness (e.g., "rude," "impolite," and "obnoxious"), politeness
(e.g., "respect," "considerate," and "polite"), or neither (in the
control condition); considerable previous work on impression formation using the same priming method (but with varying trait
content; e.g., Banaji, Hardin, & Rothman, 1993; Srull & Wyer,
1979, 1980) had shown it to activate the corresponding perceptual
trait constructs. On the basis of the hypothesized perceptionbehavior link, this activation was expected to continuously activate
the behavioral constructs of rudeness or politeness, increasing the
likelihood of such behavior.
After completing this priming task, participants encountered a
situation in which they could either behave in a rude fashion and
interrupt an ongoing conversation or behave in a polite fashion and
wait for the conversation to end on its ownwithout the participant's intervention, the conversation would continue on for 10
min. Results showed that significantly more participants in the
rude priming condition (67%) interrupted the conversation than
did those in the control condition (38%), whereas only 16% of
those primed with the polite condition interrupted itin other

895

CHAMELEON EFFECT

words, fully 84% of participants in the politeness priming condition waited the entire 10 min without interrupting.
Experiment 2 of Bargh, Chen, et al. (1996) extended these
findings to the case of stereotype (collections of group-related
traits, as opposed to single-trait concepts) activation. Participants were first primed either with words related to the stereotype of the elderly (e.g., "Florida," "sentimental," "wrinkle") or
with words unrelated to the stereotype. Importantly, none of the
primes was semantically related to slowness or weakness,
though these concepts are components of the stereotype. As
predicted, priming the stereotype caused participants to subsequently behave in line with the stereotype content; specifically,
they walked more slowly down the hallway after leaving the
experiment. Experiment 3 conceptually replicated this effect by
subliminally presenting faces of young male African Americans
to some participants, who then reacted to a provocation with
greater hostility (a component of the African American stereotype; see, e.g., Devine, 1989) than did control participants. This
latter effect was replicated and extended by Chen and Bargh
(1997).
Dijksterhuis and van Knippenberg (1998) have conceptually
replicated these findings by demonstrating that priming a stereotype or trait can affect subsequent performance on an intellectual
task. In several studies, these researchers primed participants with
a positive stereotype ("professor"), a negative stereotype ("soccer
hooligans"), a positive trait ("intelligent"), or a negative trait
("stupid"). Those participants primed with either the professor
stereotype or the "intelligent" trait showed enhanced performance
on a general knowledge scale (similar to Trivial Pursuit), whereas
those primed with the hooligan stereotype or the "stupid" trait
showed decreased performance.

Mediational Evidence
The Bargh, Chen, et al. (1996) and Dijksterhuis and van Knippenberg (1998) studies showed that priming techniques produce
changes in behavior based on the hypothesis of an automatic
perception-behavior link. However, these studies (as well as that
of Carver et al., 1983) did not provide evidence that perceptual
activity mediated the effect of priming on behavior, because perception itself was never manipulated (or measured). It remains
possible that environmental events (which priming manipulations
simulate) directly activate perception and separately directly activate behavioral tendencies.
One way to show that passive perceptual activity automatically
causes behavior would be to show that manipulations known to
cause changes in perception and judgment produce corresponding
changes in behavior. Dijksterhuis and his colleagues (Dijksterhuis,
Aarts, Bargh, & van Knippenberg, 1998; Dijksterhuis, Spears, et
al., 1998) have conducted a series of such studies.
Assimilation and contrast effects in automatic behavior. Research in social perception has documented two main forms of
representation that moderate social judgments: trait categories
(e.g., honesty) and exemplars (representations of specific people
who exemplify the trait, such as Einstein for intelligence). In
general, the evidence shows that activated trait categories usually
produce assimilation effects; ambiguously relevant behavior is
assimilated into the category rather than contrasted against it. A

person whose trait category of honesty is in a heightened state of

accessibility or activation is more likely than the average person to
consider someone generous when he or she gives money to a
charity following his or her boss's request to do so (e.g., Higgins,
Rholes, & Jones, 1977; Srull & Wyer, 1979). But if a person is
thinking about exemplars of a given trait, such as Einstein for the
"intelligent" trait, then ambiguously relevant behaviors (getting a
B on a test) are seen as less, not more representative of that trait
(Herr, Sherman, & Fazio, 1984; Smith & Zarate, 1992; Stapel,
Koomen, & van der Pligt, 1997). The exemplar sets a high standard against which mundane trait-consistent behaviors pale in
comparison.
If perception mediates the ideomotor effects of the environment
on behavior, then one should find assimilation effects on behavior
with category priming and contrast effects on behavior with exemplar primingthe same effects one obtains on perceptual and
judgmental dependent measures. Confirming this prediction,
Dijksterhuis, Spears, et al. (1998) showed that priming (without
specific examples) the stereotype of professors versus that of
supermodels (the latter group being stereotypically viewed as
unintelligent by the participant population) produced assimilation
effects on behavior. Those participants primed with the professor
stereotype gave more correct answers on a subsequent knowledge
test than did those primed with the supermodel stereotype. But
when specific exemplars of the two categories served as the
priming stimuli (e.g., Albert Einstein and Claudia Schiffer), the
opposite pattern was obtained; that is, contrast effects on behavior
were observed.
Amount of experience mediates perception-behavior effects.
Another approach to gaining positive evidence of mediation by
perceptual activity is to assess individual differences regarding
how much contact the individual has had with that group. The
more contact, the stronger and more automatic the perceptual
representation, and thus the stronger and more likely the behavioral effect. Dijksterhuis, Aarts, et al. (1998) assessed how much
contact college-age experimental participants had per week with
the elderly. It was assumed that greater amounts of contact with the
elderly would correspond to stronger perceptual associations between being elderly and having relatively poor memory. In the
course of a lexical decision task, it was shown that the greater the
participant's amount of contact with the elderly, the stronger the
association between the concepts of the elderly and of forgetfulness. Moreover, a subsequent memory test for all of the target
stimuli in the lexical decision task showed that greater amounts of
contact with the elderly were related to poorer memory performance. Most importantly, however, the effect of contact on memory was entirely mediated by the strength of the perceptual association between the concepts elderly and forgetful. There was no
direct effect of amount of contact on behavior that was not mediated by the strength of the perceptual representation.

Research on Behavioral Coordination

Observations of and theories about nonconscious mimicry have
a long history (see Bandura, 1977; Bavelas, Black, Lemery, &
Mullett, 1987; Koffka, 1925; Piaget, 1946). Interestingly, most of
the early writers on the topic conceptualized mimicry in terms of
empathy. Adam Smith (1759/1966), for example, posited that

CHARTRAND AND BARGH

reflexive imitation occurs after one takes the perspective of the

other and realizes what he or she must feel, and Charles Darwin
(1872/1965) used the term sympathy to refer to imitation based on
reflex or habit. In fact, according to Gordon Allport (1968), the
original meaning of the term empathy was "objective motor mimicry"; it was only in the latter half of the 20th century that it came
to be used as a global term encompassing vicarious emotion, role
taking, and the ability to understand others.
Research on nonconscious mimicry began after a seminal
paper by Scheflen in 1964. He observed that postural configurations were a source of information about an ongoing social
interaction, as they communicated messages about liking and
understanding. Moreover, individuals were said to utilize this
postural information unconsciously to orient themselves within
a group. Three basic lines of research on behavioral coordination developed thereafter (see Bernieri & Rosenthal, 1991).
Research on rhythmic synchrony has included work on the
precise synchronization between the speech and body movements of the two interaction partners (Bernieri, 1988; Condon &
Ogston, 1966; Condon & Sander, 1974; Dittmann & Llewellyn,
1968, 1969; Kendon, 1970; cf. McDowall, 1978). Facial mimicry research has focused on neonates' mimicry of adult facial
expressions (Meltzoff & Moore, 1977, 1979, 1983; cf. Kaitz,
Meschulach-Sarfaty, Auerbach, & Eidelman, 1988). However,
no consensus developed from this research as to the mechanisms responsible for the effect (Anisfeld, 1979; Jacobson &
Kagan, 1979; Masters, 1979).
Facial mimicry has also been found in adults (Dimberg, 1982;
Vaughan & Lanzetta, 1980; Zajonc, Adelmann, Murphy, &
Niedenthal, 1987), although it is not clear from these studies
whether the observers actually experience the same emotions as
the other person or simply mimic his or her facial expressions.
The Zajonc et al. finding that couples grow to resemble each
other the longer they are together is especially intriguing given
the present hypothesis of a perception-behavior link, because
one reason for the increased resemblance could be the similar
facial lines left by many years of unconsciously mimicking the
perceived facial expressions of the partner.
The third type of behavioral coordination research, and the one
that most closely resembles the chameleon effect, is that on behavior matching, which occurs when people mimic behavior patterns by adopting similar postures or showing similar body configurations (La France, 1979, 1982; La France & Broadbent,
1976). The main focus of this research has been to link posture
similarity in naturalistic settings to rapport, which (though rarely
operationalized the same way twice) often includes measures of
involvedness, togetherness, being "in step," and compatibility (see
also Bavelas, Black, Chovil, Lemery, & Mullett, 1988; Bavelas,
Black, Lemery, & Mullett, 1986, 1987).
Despite the considerable amount of research on mimicry and
behavioral coordination, there has been relatively little attention
given to the mechanism responsible for it. The consensus position
appears to be that behavioral coordination is in some way related
to empathy, rapport, and liking, although some see mimicry as the
cause and others see it as the effect of empathic understanding.
That mimicry and behavioral coordination are said to serve the
adaptive function of facilitating social interaction and interpersonal bonding does not, however, answer the question of how
these effects are produced.

Four critical elements are missing from these observational

studies. First, although moderate posture sharing has been reported, there has been no baseline or control group with which to
compare the amount of mimicry observed; without this, one cannot
determine whether it occurs more often than chance would predict.
In fact, La France (1982) has stated that "posture mirroring is not
constant nor ubiquitous" (p. 290), and the results of one statistical
test of its existence suggested that it did not occur more often than
would be predicted by chance (Bernieri, 1988). Although there is
wide agreement that posture and body movement mimicking
do occur, it nonetheless remains an experimentally unproven
observation.
Second, there has been no test of the minimal conditions
under which behavior matching occurs. As noted before, research has shown that there is greater posture similarity when
the interactants like each other and feel more rapport (Charney,
1966; La France, 1979; La France & Broadbent, 1976; Scheflen, 1964). However, there has been no compelling test of
whether there is significant mimicry among unacquainted interaction partners. If the perception-behavior link is the mechanism underlying behavior matching, then it should occur even
among strangers. Furthermore, the chameleon effect is hypothesized to be an entirely passive and preconsciously automatic
process (i.e., it does not depend on the concurrent operation of
an intentional goal, such as ingratiation, during the interaction;
see Bargh, 1989). Thus, not only should it occur among strangers, but it should occur even without an active goal to get along
with and be liked by the interaction partner. To date, there have
been no tests of whether posture and behavior mimicry occur
under-such minimal conditions.
Third, the previous studies were correlational and did not
manipulate the postures and mannerisms of either interactant.'
This lack of experimental control over which mannerisms are
done and how long they are engaged in precludes one from
inferring causation. That is, one cannot conclude from these
studies that Person X was mimicking Person Y; rather, one can
only say that Persons X and Y were displaying the same
mannerisms or postures at a given time. For one thing, there
could be other, third factors that could spuriously lead to these
shared behaviors (e.g., a hot room causing all present to fan
their face). For a valid demonstration of the chameleon effect,
one would need to show that Person X first engages in a
particular behavior, and then Person Y mimics that behavior,
without intending or having any reason to do so.
Finally, just as chameleons change their coloring to blend in
with their current environment, an experimental demonstration of
a behavioral chameleon effect should incorporate, as a withinsubjects factor, variability in the behavior of interaction partners,
to show that the participant's behavior changes accordingly.
1

Although motor-mimicry researchers have manipulated confederates' behaviors, they were not interested in (and therefore did not
manipulate) general postures or behavioral mannerisms. Instead, the
experimenters created situations in which participants observed a confederate experiencing a specific event and emotion and then tested
whether participants reacted as if the experience were happening to
them (e.g., Bavelas et al., 1988; Bavelas, Black, Lemery, & Mullett,
1987).

CHAMELEON EFFECT

Again, to date, there has been no demonstration of such passive

behavior adaptations to multiple interaction partners.

897

Experiment 1: A Test of Unintentional Mimicry

Between Strangers
Method

The Chameleon Effect as Cause of Interpersonal Rapport

and Empathy
We propose that the chameleon effect is the mechanism behind
mimicry and behavioral coordination and thereby is the source of
the observed smoother social interaction and interpersonal bonding
produced by the (nonconscious) mimicry. In relating these formerly disparate areas of research, we hypothesize that the perception of another's behavior (be it facial expression, body posture,
mannerism, etc.) increases the tendency for the perceiver to behave
in a similar manner, and that this is an entirely passive and
nonconscious phenomenon. Thus, we argue that the perception of
another's behavior does not require or depend on the perceiver
having any interpersonal goal, such as ingratiation, toward the
person being perceived, nor does perception require the two interaction partners to have an already established relationship (i.e., a
preexisting state of rapport). Unlike the prior correlational accounts of mimicry and rapport, we posit a directional causal
sequence: Perception causes similar behavior, and the perception
of the similar behavior on the part of the other creates shared
feelings of empathy and rapport. In short, the widely documented
automatic link between perception and behavior exists, at least in
part, as a kind of natural "social glue" that produces empathic
understanding and even greater liking between people, without
their having to intend or try to have this happen.
As noted above, the studies that showed that the same priming
manipulations that influence social perception also influence social
behavior are suggestive, but not conclusive, evidence for automatic effects of perception on behavior. What is needed is a
demonstration, within a social interaction context, that the perceiver's behavior changes as a function of the behavior of the interaction partner, and that these changes occur without conscious
choice or guidance.
Thus, our first goal (Experiment 1) was to provide an experimental test of the existence of nonconscious mimicry of behavioral
mannerisms in a way that (a) determines whether it occurs at
greater-than-chance levels, (b) tests whether it occurs among
strangers when no affiliation goal is operating, (c) manipulates
mannerisms and behaviors of interaction partners (confederates of
the experimenter) to determine the direction of causality of the
effect, and (d) tests for a chameleon-type change in behavior as a
function of the behavior of the current interaction partner. Unlike
previous researchers, we did not observe individuals who were
already engaged in an interaction; rather, we created dyadic interactions between participants and confederates during which confederates varied their facial expressions and behavioral mannerisms. In Experiments 2 and 3, we sought to verify that these
automatic effects of social perception on social interaction produce
greater empathy and liking between the interaction partners; in
Experiment 3, we examined this issue by testing whether individual differences in empathy covary with individual differences in
the chameleon tendency.

Overview. Students participated in two consecutive dyadic sessions.

Session 1 consisted of a 10-min interaction with 1 other "participant"
(Confederate 1; Cl), during which they took turns describing various
photographs. Participants then repeated this photograph description task in
Session 2 with a 2nd "participant" (Confederate 2; C2).
Confederates varied their mannerisms throughout the interactions. During Session 1, Cl either rubbed his or her face or shook his or her foot.
During Session 2, C2 did whichever behavior Cl did not do. Facial
expressions varied as well; Cl either smiled or had a neutral expression
(i.e., did not smile) throughout Session 1. During Session 2, C2 smiled if
Cl had not smiled, and did not smile if Cl had smiled. The order of
mannerisms and facial expressions was counterbalanced, and C2 always
did the mannerism and facial expression that Cl did not do. A video
camera recorded participants during both sessions so that coders could later
judge the extent to which participants mimicked the mannerisms and facial
expressions of the 2 confederates.
Participants. Thirty-nine male and female students enrolled in an
introductory psychology course at New York University participated in the
experiment in partial fulfillment of a course requirement. Data from 4 of
these participants were excluded from subsequent analyses for the following reasons: (a) 3 participants chose to not sign the consent form giving us
permission to code and analyze their videotape, and (b) during debriefing, 1 participant expressed suspicion that the other participant was in fact
a confederate. However, neither she nor any of the other participants
accurately guessed our hypothesis.
Thus,-we computed all analyses on responses from a final sample of 35
participants. For 14 participants, Cl smiled and shook his or her foot and
C2 did not smile and rubbed his or her face. Because the possibility existed
that encountering the smiling confederate first would affect participants'
interactions with the nonsmiling C2, it was important to counterbalance the
order of facial expressions by having the nonsmiling confederate interact
first with some of the participants. Thus, Cl did not smile with 21
participants (of these, Cl rubbed his or her face with 8 and shook his or her
foot with 13).
Apparatus and materials. Two male and two female assistants served
as experimenters and confederates, rotating in the roles of experimenter or
confederate. The experiment room had one chair for the experimenter at the
front of the room, behind a desk in which the materials and stimuli for the
experiment were kept. The room also contained two chairs for the participant and confederate that were placed approximately 1.2 m apart. These
two chairs were half-facing each other and half-facing the experimenter's
desk. With this arrangement, the participants could see the confederates'
mannerisms during the interaction but could not see the experimenter's,
whose body was effectively hidden by the desk.
Participants were videotaped throughout both sessions by means of a
video camera on a shelf in the corner of the room. The camera was
focused on the participant's chair, resulting in a clear view of the
participant's entire seated body. To ensure that coders of the videos
were blind to condition, we did not videotape the confederates. Thus,
when judging a particular participant's responses, the raters did not
know the corresponding mannerisms or facial expressions of the 2
confederates.
Color photographs for the experiment were chosen from magazines such
as Newsweek, Time, and Life. The photos were cut out of the magazines and
mounted on heavy black cardboard. Twelve photos were chosen that
ranged somewhat in emotional content, amount of action involved, and

898

CHARTRAND AND BARGH

ambiguity of what was being portrayed in the photo.2 These variables were
not manipulated systematically, but the photographs were rotated so that
participants did not always describe the same type of photo when with the
smiling or nonsmiling confederate (e.g., only describing somewhat "happy" photos when with the smiling confederate). Thus, although 6 of the 12
photos were reserved for the confederates (so they could memorize a
prepared script for each) and the other 6 were reserved for the participant,
the order of the photos within each set varied.
Procedure. Each participant completed the experiment individually.
Prior to each session, the experimenter turned on the video camera that
would record the participant throughout the session. The experimenter then
brought the participant into the laboratory room and seated him or her in
the participant's chair. The experimenter then left the participant alone in
the room for 1 min (ostensibly to retrieve copies of a needed form from
another room), during which time the participant was videotaped to obtain
a baseline measure. This baseline period was later coded to determine the
extent to which the participant was already rubbing his or her face, shaking
his or her foot, or smiling before interacting with any confederate.
The experimenter reentered the room and delivered the cover story. It
was explained that the purpose of the study was to test a new projective
measure being created by some psychologists in the department. (The
assumptions underlying the use of projective measures were briefly explained to those participants unfamiliar with them.) The participant was
informed that some researchers were trying to develop a revised version of
one of the more common measures (the Thematic Apperception Test) that
(a) could be administered to more than one person at a time and (b) would
use photographs instead of picture drawings.
The participant was told that the researchers were in the initial stage of
creating working sets of photographs to serve as the stimuli for the
projective test. Toward this end, they were first testing various sets of
photos on a "normal" (i.e., nonpatient) population. Specifically, college
students were being recruited to describe what they saw in the various
photographs. Participants could discuss the visual aspects of the photo, or
free associate and say whatever came to mind (including what the people
in the photos were thinking or feeling), or both. Importantly, the experimenter emphasized to the participant that responses would not be analyzed
by any of the psychologists (or anyone else), so there was no need to be
concerned about the content of his or her responses. Instead, the ease with
which the students described and generated responses to the photos would
ostensibly be taken as the indicator of the usefulness of those particular
photos. Accordingly, the participant was told that at the conclusion of the
experiment, he or she would be asked about the experience of describing
the photographs (e.g., how easy it was to generate responses for them).
The participant was further informed that several sets of photographs
had already been gathered and tested on students 1 at a time in individual
sessions. The photos were now being tested in group settings, beginning
with groups of 2 students at a time. The participant was then given a
consent form to sign and told that he or she would be involved in two
separate group sessions, each with 1 other participant. It was explained that
another session was being conducted concurrently in another room, and
that 1 of the participants from a previous session there would be the 1st
partner. The experimenter then brought in the 1st other participant (Cl) and
seated him or her in the confederate's chair. The participant and Cl were
each given a set of three photos facing down. The experimenter explained
that the two sets of photos were different and reminded them that their task
was to take turns describing what they saw in each photograph. They were
told to describe each photo in any way they wished for approximately 1
min.
The experimenter suggested that Cl turn over the first photo and begin.
Cl described the photograph, following a memorized script to ensure that
responses were standardized across different confederates and different
experimental sessions. It should be noted that the confederates were trained
to deliver the responses with natural hesitation, including pauses, umms,

and hmms. One example of a scripted response refers to a photo of a man

holding in his arms a small dog with a leg cast:
This is a picture of a man holding a small dogmaybe a chihuahua
but I'm not sure. The dog's leg is in a cast, so I guess it's broken. I
don't know how dogs' legs get broken, but maybe it got stuck
somewhere, like in those gutters outside or something. So then it was
probably crying or making a lot of noise and this man heard it. The
man looks like a pretty nice guy, so he probably felt sorry for the dog
and wanted to help it. This picture looks like it's taken at a vet's
office, so the man probably brought the dog to the vet and then they
put the cast on the leg. And then this picture was taken right after that.
The man didn't know who the dog belonged to, so he's having people
take pictures of the dog so that the owner can come pick him up.
The experimenter then asked the participant to turn over his or her first
photo and begin describing it. After the participant finished, Cl and the
participant continued alternating turns until both completed their sets of
three photographs. During the interaction, Cl made minimal eye contact
with the participant to minimize the possibility that any personal relationship between the two would be established. Cl was either smiling or not
smiling and either rubbing his or her face or shaking his or her foot.
Behaviors were always performed throughout the interactions.
After all photographs had been described, the experimenter told participants that they would now be switching partners. One of the participants
would be brought to the other laboratory room to join another participant,
and the other would stay in the current room to meet a new partner. The
experimenter escorted Cl out of the laboratory room and approximately 1
min later brought in C2 to join the participant. The experimenter gave the
participant and C2 a different set of three photographs each, and once again
they alternated talcing turns describing them. This time, however, C2 was
displaying the mannerism (rubbing his or her face or shaking his or her
foot) and facial expression (smiling or not smiling) that Cl had not.
Following the session with C2, the experimenter said that the debriefing
would take place individually, and that C2 would be taken to the other
laboratory room where he or she would be debriefed by the other experimenter. The experimenter escorted C2 out of the room and returned alone
approximately 30 s later. The experimenter then queried the participant in
a "funneled" question sequence (i.e., from general to increasingly specific
questions about awareness of hypotheses; see Bargh & Chartrand, in press)
to determine if he or she (a) was suspicious that the other participants were
in fact confederates, (b) noticed that the confederates each displayed
certain mannerisms throughout the session, or (c) thought that the purpose
of the experiment was anything other than what the cover story indicated.
Finally, the hypotheses and purpose of the study were explained to the
participant. The participant was asked to sign a video release form allowing
the researchers to examine the data and was thanked for his or her
participation in the study.

Results
Interjudge reliability. Videotapes were coded by two independent judges blind to the condition of participants. Three time
periods were coded for each participant: 1 min of baseline before
interacting with confederates (BL), the time spent with Cl (Tl),
and the time spent with C2 (T2). The coding procedure yielded the
following dependent variables: (a) the number of times the partic2

Although an effort was made to avoid photographs with strong emotional content, at the same time we needed to choose photos that would (a)
be convincing as stimuli for a projective measure and (b) be able to
stimulate 1-2 min of description as well as conjecture as to what was being
thought or felt by the people in the photographs.

899

CHAMELEON EFFECT

ipant smiled, (b) the number of times the participant rubbed his or
her face, and (c) the number of times the participant shook his or
her foot.3
The following are the interjudge reliabilities: For the number of
times smiling, the reliability for the three ratings (BL, TI, and T2)
ranged from r = .79 to 1.00, with mean r = .89. For number of
times participants shook their foot, the three ratings ranged from
r = .53 to .79, with mean r = .68. For number of times participants
rubbed their face, the interjudge reliabilities ranged from r = .33
to .60, mean r = .50.4 All reliabilities were significant atp < .001.
The mean of the two judges' ratings was taken to form a single
rating for each behavior. Ratings for Tl and T2 were then divided
by the number of minutes (to the nearest second) that the interaction lasted to arrive at a rate per minute. (This method had the
further advantages of equating Tl and T2 with BL so that the
numbers would all be in the same metric and ensuring that any
differences would not be artifactually due to somewhat longer or
shorter interactions in Tl vs. T2.)
For both the smiling and behavioral measures, a repeated measures analysis of variance (ANOVA) was conducted on the number of times each action occurred per minute. For each analysis, we
included the baseline rating as a covariate to adjust for individual
differences in performing the key behaviors in the absence of
another person. Neither the participant's gender nor the order in
which the confederates enacted the various behaviors affected the
results, so neither of these variables is discussed further.
Facial expression. As predicted, there was a significant effect
of confederate expression, F(l, 34) = 20.31, p < .0001. Participants smiled more times per minute when with the smiling confederate (M = 1.03) than with the neutral confederate (M = 0.36).
This result suggests that participants did indeed mimic the facial
expression of the confederates.
Behavioral measures. We next conducted a repeated measures
ANOVA on the number of times participants engaged in the
mannerisms per minute. Confederate behavior (foot shaking vs.
face rubbing) and participant behavior (foot shaking vs. face
rubbing) were the two within-subject variables. Whereas there
were no main effects for confederate behavior (F < 1) or participant behavior (p > .25), the predicted interaction between the
two was, in fact, reliable, F(l, 34) = 9.36, p = .004 (see Figure
1). Our hypothesized chameleon effect specifically predicts that
participants should engage in face rubbing (or foot shaking) more
in the presence of the confederate engaging in that behavior than
in the presence of the confederate not engaging in that behavior.
Consistent with this prediction are our findings that participants
rubbed their face more times in the presence of the face-rubbing
confederate than when with the foot-shaking confederate, F( 1, 34)
= 5.71, p < .025, and shook their foot more times when with the
foot-shaking confederate than with the face-rubbing confederate,
F ( l , 34) = 3.76, p = .06. These results, in conjunction with the
facial expression findings, support our hypothesis that individuals
passively take on the mannerisms and facial expressions of those
around them without the intention or reason to do so.
Liking as potential mediator. If the perception-behavior link
is, as we argue, a completely nonconscious, non-goal-dependent
mechanism that produces the chameleon effect, mimicry of others
should occur even in the absence of a reason to do so, such as
pursuing an affiliation goal. In the present study, with one smiling
and one nonsmiling confederate, it is reasonable to suppose that

0.8

Confederate Rubs Face

Confederate Shakes Foot
0.73

0.7

0.6

CD

0.5

.a

0.57

0.4
0.3

Participant Rubs Face

Participant Shakes Foot

Figure 1. Number of times participants rubbed their face and shook their
foot per minute when with a confederate who was rubbing his or her face
and a confederate who was shaking his or her foot.

participants would be more likely to have an affiliation goalif

they had one at all, which the design of the experiment attempted
to minimizewith the smiling than with the nonsmiling confederate. Thus, one could conceptualize the smiling and nonsmiling
confederates as a likability manipulation. The question becomes,
did the participants mimic the foot-shaking and face-rubbing behaviors of the nonsmiling confederate or only those of the smiling
confederate?
In the following analysis, whichever of the two behaviors the
nonsmiling confederate performed was the key behavior for a
given participant. We compared how much participants engaged in
the key behavior with the nonsmiling confederate with how much
they engaged in that same behavior with the smiling confederate
(who was doing the other behavior). A repeated measures ANOVA
was conducted on the number of times the action occurred per
minute. A significant effect of confederate behavior was obtained,
F ( l , 34) = 4.16, p = .05. Participants performed the key action
more times with the nonsmiling confederate doing that key behavior (M = .56) than with the smiling, likable confederate doing the
other behavior (M = .40). It was not the case that participants only
mimicked the behavior of the smiling, apparently friendly confederate and not the nonsmiling, apparently less friendly confederate.
Were the mimicry effects greater in the presence of the smiling
confederate? We next compared the extent to which participants
mimicked the behavior (either foot shaking or face rubbing) of the
3
We also coded number of seconds participants spent smiling, rubbing
their face, and shaking their foot. The correlations between these seconds
measures and the number of times measures was high (for smiling, r = .92;
for face rubbing, r = .88; for foot shaking, r = .94). Because of this
redundancy, we report only the number of times analyses in the text.
However, we computed all analyses on the number of seconds as well, and
the results showed the identical pattern and significance level as the
number of times analyses.
4
Reliability between judges was higher for the foot-shaking than for the
face-rubbing measures. Because there are many physical gestures that can
be made in the facial area (e.g., scratching an itch, playing with an earring,
fixing hair), a detailed coding key was created and used by both raters.
However, judgment calls had to be made, which reduced reliability. It
should be noted that in Experiment 2 the reliability for this measure was
substantially higher.

900

CHARTRAND AND BARGH

smiling confederate more than the behavior of the nonsmiling

confederate. Behavior mimicked (face rubbing or foot shaking)
when the participant was with the smiling confederate was the
between-subjects variable, and confederate expression (smiling
versus nonsmiling) was the within-subjects variable. There was no
significant main effect for confederate expression across the two
behaviors being mimicked, nor was the interaction significant
(Fs < 1). Thus, there was no evidence in our study that the
obtained effects were goal dependent.
Participants' awareness of having engaged in behavioral mimicry. For the chameleon effect to be considered passive and
automatic, it must be demonstrated that participants were not
aware of having mimicked the confederates. Although intuitively
it seems unlikely that participants would want to purposefully
mimic the confederates' mannerisms, it is possible that participants
believed that shaking their foot or rubbing their face simultaneously with the confederate would be beneficial for the interaction in some way, and they mimicked for these conscious, motivated reasons. However, we have evidence that this was not the
case. Participants were asked during the funneled debriefing at the
conclusion of the experiment whether anything about either of the
confederates stood out to them. Participants were then asked
whether either of the confederates had any particular mannerisms
or ways of speaking that they noticed or that seemed distinctive.
One participant (out of 35) mentioned that 1 of the confederates
made hand motions while speaking, and 2 others commented on
the slouching posture of 1 confederate. However, none of the
participants mentioned noticing that the confederates were shaking
their foot or rubbing their face. (When asked, most reported that
they "hadn't noticed" the mannerisms of the confederate.) Thus, it
seems that not only were participants not consciously trying to
imitate the mannerisms of the confederates, but they did not even
pay attention to these mannerisms in the first place.

mannerisms between participants and confederates could not have

been due to any third factor.
Moreover, because participants interacted with 2 different partners, each of whom engaged in different facial expressions
and behavioral mannerisms, the results demonstrated the true
chameleon-like nature of the perception-behavior effect, as the
participants' behavior changed as a function of the behavior of
their current interaction partner. Thus, they rubbed their face when
interacting with the face-rubbing confederate but then reduced face
rubbing and increased foot shaking during their interaction with
the foot-shaking confederate. No previous study had demonstrated
how an individual's behavior naturally adapts to changes in social environmental settings by blending in to each of them
successively.
Because the perception-behavior link is preconscious and not
goal dependent, for it to be the cause of the chameleon effect, the
effect should occur among strangers when no affiliation goal is
present. This was found to be the case; participants mimicked the
behavior of strangers, even nonsmiling ones who never made eye
contact with them. In designing the experiment, we sought to
minimize the possibility that participants would choose to pursue
an affiliation or other social goal toward the confederates that
would cause them, in pursuit of that goal, to engage in behavioral
mimicry (even at a nonconscious, goal-dependent, automatic level). Thus, confederates were instructed to not make eye contact
with the participants, and when serving as the smiling confederate,
to never smile at the participants. Finally, that the obtained behavioral mimicry occurred just as much in the presence of the nonsmiling as the smiling confederate is further evidence against the
goal-dependent alternative account.

Discussion

What is the adaptive function served by the chameleon effect,

the nonconscious tendency to behave with others as those others
are behaving? As reviewed above, there is consensus among
researchers that behavior matching is related to greater liking and
rapport between the interactants. Our second goal for the present
research was to test whether behavior matching does in fact
increase liking and create a sense of smoother interactions. Our
hypothesis that automatic effects of perception on behavior serve
adaptive functions is part of a larger research effort that traces the
"downstream" consequences of a variety of immediate, preconscious reactions to the social environment. For example, recent
research on the downstream effect of the tendency to automatically
evaluate perceived stimuli as either good or bad (e.g., Bargh,
Chaiken, Govender, & Pratto, 1992; Bargh, Chaiken, Raymond, &
Hymes, 1996; Fazio, Sanbonmatsu, Powell, & Kardes, 1986) has
demonstrated direct effects on behavioral predispositions toward
those stimuli (Chen & Bargh, 1999), as well as mood effects that
reflect the average valence of automatic evaluations made over
time in a given environment (Chartrand & Bargh, 1999). The
chameleon effect, as another variety of a preconscious automatic
process, also likely exists for a useful, adaptive reason.

Researchers have long been interested in nonconscious mimicry,

yet there has been little attention given to identifying the mechanism underlying the phenomenon. We have argued that the
perception-behavior link can provide such a mechanism. The
perception-behavior link posits the existence of a natural and
nonconscious connection between the act of perceiving and the act
of behaving, such that perceiving an action being done by another
makes one more likely to engage in that same behavior. This
mechanism can account for the chameleon effect, the tendency of
people to take on the postures and mannerisms of those around
them.
In Experiment 1, we sought to provide an experimental test of
the chameleon effect in which the mannerisms and facial expressions of interaction partners were manipulated and standardized
across participants. Analyses revealed that behavioral mimicry did
in fact occur at significantly greater than chance levels. Significant
mimicry was found for facial expressions and for two different
behavioral mannerisms, after controlling for BL measures of each
behavior. Furthermore, the design of the experiment, in which the
confederates' behavior was predetermined and standardized and so
it was clear who was mimicking whom, enables conclusions to be
drawn for the first time about the causal direction of the effect.
Unlike previous studies, this one showed that the similarity in

Experiment 2: The Adaptive Function of the

Chameleon Effect

It is plausible that the chameleon effect serves the basic human

need to belong. In a recent review, Baumeister and Leary (1995)
argued that according to the existing evidence, the human need to
belong is a powerful, fundamental, and extremely pervasive mo-

CHAMELEON EFFECT

tivation.5 We desire frequent, nonaversive interactions with others

and want to form and maintain strong, stable interpersonal relationships. We try to orient toward fellow human beings in a way
that is relatively free from conflict and negative affect. To the
extent that two interactants are similar to each other and have
things in common (even at the level of behavioral mannerisms),
such a smooth, conflict-free interaction will be more likely to
occur. Moreover, automatically behaving in a manner similar to
other group membersincluding having similar facial reactions to
eventshelps prevent an individual member from standing out as
different, and so it would help to prevent ostracism and social
distance from other group members (see Brewer, 1991).
Researchers of elementary motor mimicry have posited a very
specific function served by motor mimicry that is consistent with
this analysis. Recall that motor mimicry is a subset of behavior
matching that refers to an individual reacting to another person
going through a specific, emotion-laden incident (e.g., wincing at
the other's pain). The individual reacts as if he or she were
experiencing and feeling the same thing as the other person.
Bavelas and her colleagues (Bavelas et al., 1988; Bavelas, Black,
Lemery, & Mullett, 1986, 1987) take a strong stand that motor
mimicry is not an overt manifestation of an intrapersonal process,
such as vicarious emotion or cognitive role taking, but rather is an
important communication tool that relays the message "I am like
you" or "I feel as you do" to the other person.
Over 20 years earlier, Scheflen (1964) similarly suggested that
mimicry might serve a communicative function without a person's
awareness or intent: "Human behavior can be communicative
whether or not it is intended to communicate . . . . The intent of an
interactant and the function that a behavior actually has in a
group process must be conceptually distinguished (italics in original; p. 318). We suggest that behavior matching serves this same
function, and individuals use behavior mimicry as a communication tool on a completely nonconscious level.
Although behavior-matching researchers have not discussed its
use as a communication tool per se, the notion is consistent with
the proposed link between behavior matching and rapport. Scheflen (1964) originally posited that people in a group often mirror
one another's posture and that this reflects a shared viewpoint.
Bernieri and Rosenthal (1991) pointed out that people seem to get
along better when their behaviors are well coordinated: "Interpersonal coordination and synchrony may eventually explain how it is
that we 'hit it off immediately with some people and never 'get it
together' with others" (p. 429). Tickle-Degnen and Rosenthal
(1987) also reviewed the evidence for a link between interpersonal
coordination and rapport and suggested that it is quite strong.
Empirical evidence supporting the link between social rapport
and interpersonal coordination comes primarily from the work on
posture mirroring. In a typical study, La France (1982) found that
students frequently displayed the same postural configuration as
that of the teacher, and the extent of posture similarity was positively correlated with the students' ratings of rapport, involvement,
and togetherness. Interestingly, La France has discovered that
posture mirroring (e.g., one person lifting his or her right arm and
another person lifting his or her left arm in a "mirror image") is
related to rapport, although posture mimicking (e.g., both individuals lifting their right arm) is not (La France & Broadbent, 1976).
Additional studies have found a relationship between behavior
matching and self-reported rapport and involvement (Charney,

901

1966; La France, 1979; Trout & Rosenfeld, 1980). Hatfield, Cacioppo, and Rapson (1994) also argued that behavioral mimicry
leads to emotional convergence between interaction partners.
Thus, there is consensus among researchers that behavior
matching is related to greater liking and rapport. However, there
has been disagreement over the causal direction. Some researchers
have conceptualized various types of behavioral coordination as
by-products or outgrowths of preexisting emotional rapport or
liking (Levenson & Ruef, 1997; Scheflen, 1964). However, others
have argued for the reverse causal direction. La France (1982), for
instance, suggested that posture mirroring may not only reflect
shared viewpoints and harmony but may actually be instrumental
to achieving them.
Evidence for the mimicry-to-rapport causal direction has been
mixed. In a correlational study, La France (1979) used a cross-lag
technique to assess causality and found that posture similarity
seems to lead to rapport slightly more than vice versa, although
there was some evidence that the effect was bidirectional. In a
study of the impact of gesture similarity on persuasion and interpersonal influence, Dabbs (1969) manipulated movement similarity by having a confederate "interviewee" mimic the gestures and
mannerisms of 1 of 2 participant "interviewers" in the room.
Results were equivocal; whereas the participant who was mimicked did not report liking the confederate more than did the
participant who was not mimicked, mimicry did cause the confederate to be evaluated more favorably on other dimensions (e.g., he
was considered to be well informed and to have sound ideas). In a
second experiment, some participants were trained to be confederates 10 min before the start of the experiment and were told to
either mimic a 2nd participant or to "antimimic" him (i.e., do the
opposite of what he did). Results were again unclear as to the
effect of mimicry, but they did suggest that antimimicry could
have a negative effect in certain circumstances. Finally, Maurer
and Tindall (1983) focused on whether perceptions of a counselor's empathy partially depend on nonverbal cues such as having
similar behavioral mannerisms. They found that when counselors
mimicked the body positions of their clients, the clients perceived
a greater level of expressed empathy on the part of the counselor.
In Experiment 2, we sought to test whether manipulated variations in posture similarity produce variations in liking between
interaction partners. We especially wanted to test the extent to
which posture similarity affects liking when there is no overarching interpersonal goal held by the interactants toward each other.
In both the Dabbs (1969, Experiment 1) and Maurer and Tindall
(1983) experiments, there was a role-power differential between
the confederate and participant, and so interpersonal goals (e.g.,
ingratiation) may have affected their results. Our hypothesis, however, is that the chameleon effect operates in a passive, non-goaldependent manner to create greater liking and ease of interaction.
Hence, mimicry of one interaction partner by the other should
cause the former to like the partner more and to experience greater

5
Brewer's (1991) model of optimal distinctiveness is consistent with
this argument and puts it in a larger framework by bringing in a second,
opposing need. In this model, social identity is viewed as a reconciliation
between the two needs: On the one hand, we have a need for validation,
similarity to others, and a sense of belonging, and on the other, we have a
need for uniqueness, individuation, and a sense of distinctiveness.

902

CHARTRAND AND BARGH

ease of interacting, even when the two are strangers or new

acquaintances who are not seeking to establish a relationship.

Method
Overview. Participants had one 15-min session with another "participant" (a confederate). During this session, the participant and confederate
took turns describing what they saw in various photographs. Confederates
either mirrored the behavioral mannerisms of the participant throughout the
interaction (the experimental condition) or engaged in neutral, nondescript
mannerisms (the control condition). When the interaction was over, participants completed a questionnaire on which they were asked to report (a)
how much they liked the confederate and (b) how smoothly the interaction
had gone.
Participants. Seventy-eight male and female students enrolled in an
introductory psychology course participated in the experiment in partial
fulfillment of a course requirement. Data from 6 of these participants were
excluded from analyses for the following reasons: 2 participants in the
control condition sat in the same neutral position as the confederates,
making it equivalent to the experimental condition in which body language
and mannerisms are in synchrony. Four participants suspected that the
other participant was in fact a confederate. It should be noted, however,
that none of these participants were able to guess our hypothesis. Thus, we
computed all analyses on responses from a final sample of 72 participants,
with 37 in the mimicking (mirroring) condition and 35 in the control
condition.
Apparatus and materials. The experiment room was the same as used
in Experiment 1. The same color photographs from Experiment 1 were also
used for Experiment 2. There were 4 female assistants who served as
confederate and experimenter, and they alternated roles. All assistants were
trained to mirror the body language and mannerisms of the participants.
Although the confederates were kept blind to the specific hypothesis of
the experiment, they were necessarily aware of the manipulation involved
and of the participant's assigned experimental condition. It is therefore
possible that they could have, intentionally or unintentionally, behaved
differently toward the participants who were in the experimental condition
(e.g., acted more friendly or likable toward them). To address this possibility, 22 of the sessions (11 of the control condition and 11 of the
experimental condition) were videotaped in their entirety to later assess,
through the ratings of outside judges, whether the confederates were
behaving differently (other than in the mimicry itself) toward participants
in the mimicry versus no-mimicry conditions. Both the participant and
confederate were visible through the lens of the camera so that judges
would be able to see and code the confederate's behavior toward the
participant.
The dependent measures were ratings from participants on liking for the
confederate and smoothness of the interaction. The key items read, "How
likable was the other participant?" and "How smoothly would you say your
interaction went with the other participant?" To help camouflage the
hypothesis of the study, we embedded these two items among eight other
questions that asked about the task itself and the group format (e.g., how
easy or difficult it was for them to generate responses to the photos, and
whether they thought the various photographs went well together as a
single "set"). All items were rated on 9-point scales (for the smoothness
item, 1 = extremely awkward, 9 = extremely smooth; for the likability
item, 1 = extremely dislikable, 9 = extremely likable).
Procedure. The procedure was the same as for Experiment 1, with
participants working with confederates to ostensibly help develop the
projective measure involving sets of photographs, except that the confederates no longer smiled (or not), shook their foot, or rubbed their face.
Instead, during the interaction, the confederate avoided eye contact with
the participant and maintained a neutral facial expression. Furthermore, in
the mimicry condition, the confederate mirrored the posture, movements,
and mannerisms displayed by the participant. In the control condition, the

confederate sat in a neutral relaxed position, with both feet on the floor and
both hands holding the photos (or resting in the lap).6
When the participant and confederate had completed the photograph
descriptions, the experimenter explained that they would next complete the
questionnaire about the task. Because it was necessary to complete it
independently and privately, they would be separated and seated in different rooms. The experimenter asked the confederate to complete the survey
in an adjoining room and escorted her there. Then, the experimenter
returned to the laboratory room, gave the participant the questionnaire to
complete, and told him or her to come to the hallway outside when
finished. At this point, the experimenter queried the participant to determine whether he or she was suspicious that (a) the other participant was in
fact a confederate, (b) the confederate was mirroring his or her own
behaviors, or (c) the purpose of the experiment was anything other than
what the cover story indicated. Finally, the purpose and hypotheses of the
study were explained to the participant. (Those who were videotaped were
asked to sign a video consent form.) The participant was thanked for his or
her participation.

Results
Liking and smoothness as a function of being mimicked. We
predicted that relative to those in the control condition, participants
in the experimental condition would report (a) finding the confederate more likable and (b) having smoother interactions with her.
To test these hypotheses, a multivariate analysis of variance
(MANOVA) was conducted on the liking and smoothness variables, with mimicking of participants by confederates (yes vs. no)
as the between-subjects variable. Gender was also included as a
between-subjects variable in this and all subsequent analyses, but
no reliable main effect for gender or interaction between gender
and mimicking emerged, and so the gender variable is not discussed further. In addition, we initially included as an additional
between-subjects variable in the MANOVA whether the experimental session had been videotaped, but this variable also did not
interact with any of the effects, Fs < 1. Therefore, the sessions that
were videotaped were representative of the larger sample; the
liking and smoothness ratings of the participants in these sessions
did not differ from the ratings of the participants who were not
videotaped.
As predicted, there was an overall effect of mimicking across
the two dependent measures, F(2, 69) = 3.47, p = .04. This effect
was not moderated by type of dependent measure, interaction
F < 1. We also conducted separate univariate tests on the liking
and smoothness ratings. Participants in the experimental condition
reported liking the confederate more (M = 6.62) than did those in
the control condition (M = 5.91), F ( l , 70) = 5.55, p = .02.
Furthermore, they reported that the interaction went more

6
It was important that the confederates in the nonmimicking condition
not come across as stiff and awkward, while the mimicking confederates
came across as relaxed, mobile, and animated. This potential confounded
difference in behavior might provide an alternative explanation for our
results: The participants liked the confederate in the mimicking condition
more not because they were being mimicked by her, but rather because she
seemed more relaxed, at ease, animated, and interesting than the confederate in the neutral condition. Consequently, we instructed all confederates
to sit in a relaxed (i.e., not stiff and upright) position in both the mimicking
and nonmimicking conditions; the only difference was that in the mimicking condition the relaxed position happened to mirror the participant's
position, whereas in the nonmimicking condition it did not.

903

CHAMELEON EFFECT

smoothly {M = 6.76) than did those in the control condition

(M = 6.02), F(l, 70) = 4.08, p = .05. Thus, the results support
the hypothesis that mimicry increases liking and fosters smooth,
harmonious interactions. Although previous, correlational research
showed liking and rapport to be related to posture similarity, this
is the first demonstration that mimicry causes greater liking and
smoother interactions.
Confederates' behavior toward participants. It is important to
consider an alternative explanation for these findings. Although we
believe that mimicry by the confederate produced the greater
liking and smoothness ratings by participants in that condition,
relative to the no-mimicry condition, it is possible that some
associated difference in the behavior of the confederates in the two
conditions produced the effects. For obvious reasons, it was not
possible to keep the confederates blind to the participant's assigned condition (mimicry vs. no-mimicry). Although confederates were kept blind to the specific hypothesis in the study, it
remains possible that they unwittingly behaved differently toward
the participants in the mimicry versus no-mimicry conditions; for
example, they may have behaved in a more friendly manner
toward those they mimicked or, more subtly, engaged in greater
smiling or made more eye contact with them. If so, this would
provide an alternative reason for the participants liking the confederates more in this conditionone having nothing to do with
mimicry. Hence, we sought to determine whether there were any
such differences in confederate behavior in the two conditions.
As described in the Method section, we videotaped a sample
(n = 22) of the experimental sessions for precisely this reasonto
collect evidence germane to this alternative explanation. These
videotapes were then independently coded by two judges blind to
the experimental hypothesis. For each interaction, the following
behaviors were coded: (a) how much eye contact the confederate
made with the participant, (b) how much the confederate smiled at
the participant, (c) how friendly the confederate acted toward the
participant, and (d) how much the confederate appeared to like the
participant. All items were rated on a 6-point scale (1 = low, 6 =
high). The reliability between the two judges for the four items
combined was quite high, r = .96. (The interjudge correlations for
each of the individual scale items ranged from r .72 to r = .91.)
Accordingly, ratings from the two coders were averaged to form a
single index for each measure.
The means of each of the four ratings are presented in
Table 1. No significant differences in eye contact, smiling,
friendliness, or liking were observed between the experimental
and control conditions (all ps > .20). In fact, three of the four
measures were actually (but nonsignificantly) lower in the
mimicking than the no-mimicking condition. It does not appear
that confederates behaved differently toward the participants in
the mimicry versus no-mimicry conditions, other than in the
mimicry manipulation itself, and so we can more confidently
attribute the observed differences in liking for the confederate
and for the rated level of smoothness of the interaction to the
effects of mimicry.7
Participants' awareness of having been mimicked. Participants were asked during the funneled debriefing whether they
noticed anything in particular about the confederate's behavior or
mannerisms and whether anything about the confederate's behavior made them feel awkward or uncomfortable. One participant
reported that the confederate kept her head down and did not make

Table 1
Outside Judges' Ratings (1 = Low, 6 = High) of Confederate's
Openness and Friendliness to Participant as a Function of
Experimental Condition (Experiment 2)
No mimicking

Mimicking

Measure

SD

SD

Eye contact
Smiling
Friendliness
Liking participant

1.63
1.75
2.94
3.25

0.52
0.53
0.18
0.46

1.41
1.45
3.00
3.23

0.49
0.52
0.00
0.41

eye contact with her. A 2nd participant reported that the confederate was crossing her legs (as was the participant), but she
remarked that it "seemed normal and did not make me feel uncomfortable." Thus, only 1 out of 37 participants in the mimicking
condition noticed that the confederate had a similar mannerism,
but it was not interpreted by that participant as mimicry.

Discussion
After it was demonstrated in Experiment 1 that the perceptionbehavior link produced chameleon-like passive behavioral mimicry of interaction partners, we sought in Experiment 2 to assess
the possible adaptive value of this effect. On the basis of past
research linking mimicry to rapport, we hypothesized that the
chameleon effect serves the adaptive function of fostering liking
between people and creating smooth, harmonious interactions. It
follows that if an individual's movements and postures are purposefully mirrored by an interaction partner, that individual should
report that the interaction went more smoothly and that the partner
was more likable compared with individuals whose movements
were not mirrored. The results of Experiment 2 confirmed that,
compared with control condition participants, those participants
whose movements were mirrored by the confederate both experienced the interaction as having gone more smoothly and liked the
confederate significantly more.
It should be noted that this link between mimicking and liking
contradicts some previous findings. For instance, La France found
that posture similarity and rapport were positively correlated when
the interactants were acquainted with each other and involved in an
ongoing interaction (La France, 1979, 1982; La France & Broadbent, 1976) but negatively correlated when the interactants were
unacquainted (La France & Ickes, 1981; see Bernieri, 1988, for a
similar finding). This latter finding implies that the relation between mimicry and rapport should hold only for people involved in

Ideally, one would obtain the judges' blind ratings of the likeability of
the confederates per sethat is, how likeable a person who is not being
mimicked considers the confederate to be. Such a rating would correspond
more directly to the liking ratings made by the participants. However, the
same confederate interacted with many different participantssometimes
mimicking them and sometimes not. Thus, an overall likeability rating for
a given confederate would necessarily include both mimicking and nonmimicking sessions. To avoid this problem and to obtain separate ratings
for the mimicking versus nonmimicking conditions, we opted instead to
have judges rate the confederates' likeableness toward each individual
participant.

904

CHARTRAND AND BARGH

an ongoing interaction. The most crucial difference between the La

France and Ickes study and ours is that in the former, participants
were not interacting at all; rather, they were simply sitting in the
same waiting room at the same time. Thus, the positive effects of
chameleon-like mimicry for ease of interaction and liking may
only accrue within the context of a social interaction and not
between strangers who do not interact at all. To us, however, this
is an inconsequential constraint that would not diminish the adaptive value of the chameleon effect as a kind of social glue that
helps to bind interaction partners together.

Experiment 3: Individual Differences in

Nonconscious Mimicry
Although we believe nonconscious behavior mimicry to be a
pervasive and ubiquitous phenomenon, we also expect there to be
individual differences in the extent to which an individual engages
in such behavioral and posture mimicry. Certainly not everyone
engages in the chameleon effect to the same degree as did Woody
Allen's Zelig. What might determine whether one is more or less
likely to nonconsciously mimic others? Surprisingly, no one has
thus far posited any personality or individual difference variables
as moderators of the chameleon effect, to our knowledge. In
Experiment 3, we focused on one such potential moderator.
On the basis of the relation established in Experiment 2 between
behavior mimicry on the one hand and liking and interaction
smoothness on the other, one individual difference likely to be
related to the chameleon effect is empathy. Theoretical distinctions
have been made between various components of the empathic
response, but research has distilled two major forms. The first is
based on cognitive, intellectual reactions, such as the ability to take
and understand the other person's perspective. The second is based
on visceral, emotional reactions to the others' situation (see Davis,
1983).
We suggest that the cognitive facet of empathy (i.e., perspective
taking) is more relevant to the chameleon effect, because, as was
demonstrated in Experiment 1, the mechanism that produces the
effect is the perception-behavior link. The cause of the chameleon
effect is therefore a purely passive, cognitive mechanism that is not
associated with or dependent on any particular affective or emotional state. Thus, the most likely candidate for an individualdifference moderator of the chameleon effect would be one concerned with differences in how much attention and thought are
paid to one's interaction partners. In other words, a person will be
more susceptible to the effects of perception on behavior if he or
she engages in greater perceptual activity directed at the other
person. Taking the perspective of others is a perceptual, cognitive
process that is likely to lead to greater perception of an interaction
partner, which in turn leads to more mimicking. Moreover, if it is,
as we argue, the passive perception-behavior link that produces
the chameleon effect and its consequent benefits for social interaction, then individual differences in the emotional or affectivebased form of empathy should not be related to differences in the
chameleon effect.
In harmony with this argument is Davis's (1983) finding that
perspective taking but not empathic concern (the affective component of empathy) was consistently related to various measures of
interpersonal functioning: "Perspective-taking ability should allow
an individual to anticipate the behavior and reactions of others,

therefore facilitating smoother [italics added] and more rewarding

interpersonal relationships" (p. 115). Given that (a) we believe
social functioning to be one of the adaptive consequences of the
chameleon effect, and (b) in Experiment 2 it was demonstrated that
mimicry led to smoother interactions, individual differences in
perspective taking should be related to individual differences in the
extent of the chameleon effect.
We note that many theorists, Mead (1934) and Piaget (1932)
among them, have argued that possessing and using an ability to
take another's perspective is responsible for much of human social
capacity. Among other benefits, well-developed perspectivetaking abilities help an individual gain more satisfying interpersonal relations. In a similar vein, Cialdini, Brown, Lewis, Luce,
and Neuberg (1997) asserted that the merging of self with other is
influenced by perspective taking. It is likely no coincidence that
these consequences of frequent perspective taking parallel the
consequences of behavioral mimicry we found in Experiment 2.
That is, both behavioral mirroring and perspective taking lead to
smoother interactions and greater liking. Perhaps, then, one of the
reasons why those with a greater tendency to take the perspective
of others have greater social functioning and compassion for others
is because they engage in more behavioral mimicry; that was our
prediction in Experiment 3.

Method
Participants. Fifty-five students in an introductory psychology course
participated in this study in partial fulfillment of a course requirement.
Three of these participants suspected that the confederate was part of the
experimental setup, 1 had general suspicions regarding the study, and 1
was not videotaped because of equipment malfunction (again, none of the
participants accurately guessed our hypothesis). Data from these 5 participants were excluded from further analysis. Thus, data from 50 participants
remained in final analyses.
Apparatus and materials. The experiment room was the same as that
used in Experiments 1 and 2. Four female assistants alternated serving as
confederate and experimenter. Assistants were trained to continually shake
their foot and rub their face throughout each interaction as the confederate.
The same color photographs from Experiments 1 and 2 were used. The
same video camera setup was used as in Experiment 1, such that only the
participants (and not the confederates) were visible through the camera
lens.
To measure perspective taking, we used the perspective-taking subscale
of Davis's (1980) Interpersonal Reactivity Index (IRI). The IRI also
conveniently includes a subscale for empathic concern, which represents
the emotional concern-for-others facet of empathy. Thus, administering the
IRI allowed us to test our hypothesis that it is the cognitive, perspectivetaking component of empathy and not the emotional, empathic-concern
facet that moderates the perception-behavior link.
The perspective-taking subscale assesses the tendency to spontaneously
adopt the psychological point of view of others. Sample items include
"When I'm upset at someone, I usually try to 'put myself in his/her shoes'
for a while," "I believe that there are two sides to every question and try
to look at them both," and "I sometimes try to understand my friends better
by imagining how things look from their perspective." The empathic
concern subscale assesses "other-oriented" feelings of sympathy and concern for unfortunate others, and sample items include "I often have tender,
concerned feeling for people less fortunate than me"; "I am often quite
touched by things that I see happen"; and "Other people's misfortunes do
not usually disturb me a great deal." All items are rated on a 5-point scale
(A = does not describe me well; E = describes me very well). There are
seven items on each subscale, some of which are reverse-coded. The alpha

905

CHAMELEON EFFECT
coefficient for perspective taking is .71 for men and .75 for women; for
empathic concern, the alpha is .68 for men and .73 for women.
Procedure. Each participant completed the experiment individually.
The confederate was always sitting in the waiting area before the participant arrived. The experimenter brought them both into the laboratory room,
seating them in the two chairs reserved for them.
The procedure was essentially the same photograph-description task
used in Experiments 1 and 2. The major change was that the confederate
engaged in two different mannerisms throughout the interaction: rubbing
her face and shaking her foot. As in Experiment 2, the confederate avoided
eye contact with the participant whenever possible and maintained a
neutral facial expression.
As soon as the participant and confederate completed the photograph
descriptions, the experimenter asked if they would mind completing a
questionnaire that another psychologist in the department was planning to
use in a future experiment. All participants agreed to fill out the questionnaire (the IRI). The experimenter explained that because it was necessary
to complete the scale independently, they would be separated from each
other at this time and seated in separate rooms. The experimenter chose the
confederate to complete the survey in an adjoining room and escorted her
there. Then the experimenter returned to the laboratory room, gave the
participant the IRI scale, and told him or her to come to the hallway outside
when the questionnaire was completed. At that point, the experimenter
queried the participant as to any suspicions that (a) the other participant
was in fact a confederate or (b) the purpose of the experiment was anything
other than what the cover story indicated. Next, the purpose and hypotheses
of the study were divulged to the participant. The participant was asked to
sign a video consent form. Finally, the participant was thanked for his or
her participation.

Results
Interjudge reliability. Videotapes were coded by a judge for
the number of times participants rubbed their face and shook their
foot. Approximately half (23) of the videotapes were then coded
by a second judge. Reliability between the two judges was very
high: for the number of times participants rubbed their face, r
.97, and for the number of times they shook their foot, r = .82,
both significant atp < .001. Ratings between the two judges were
averaged to form a single rating for face rubbing and foot shaking.
Ratings were then divided by the time duration of the interaction
(to the nearest second) to arrive at a rate for behavior per minute.
Perspective taking. To test the hypothesis that individuals who
are high perspective-takers nonconsciously mimic others to a
greater extent, we categorized participants into those who scored
high and those who scored low on perspective taking. We computed the median on the perspective-taking subscale of the IRI
(median = 19) and classified those participants above the median
(n = 28) into the high-perspective-taking category and those
below it ( = 22) into the low-perspective-taking category.
A repeated measures ANOVA was performed with number of
times participants rubbed their face and number of times participants shook their foot as a within-subject variable and perspective
taking (high vs. low) as a between-subjects variable. Gender was
included as an additional between-subjects variable in this and all
subsequent analyses, but no significant main effect for gender or
interaction between gender and perspective taking was revealed.
Thus, the gender variable is not discussed further. As predicted,
however, there was a significant main effect of perspective taking
across the two types of mimicking, F(l, 48) = 3.85, p = .05. This
main effect was not moderated by an interaction with type of
behavior (face rubbing vs. foot shaking), p > .20. Specifically,

high-perspective takers rubbed their face (M = 1.30) and shook

their foot (M = 0.40) more times per minute than did lowperspective takers (Ms = 0.85 and 0.29, respectively). These
results support our hypothesis that those individuals who have a
greater tendency to take the perspective of others also are more
likely to engage in behavioral mimicry.
Empathic concern. To test our hypothesis that the emotional
facet of empathy would not moderate the chameleon effect, we
also performed a median split on the Empathic Concern subscale
scores (median = 21). Participants with scores above the median
(n = 28) were classified into the high-empathic-concern category,
and those with scores below it (n = 22) were placed in the
low-empathic-concern category.
A repeated measures ANOVA was performed with number of
times participants rubbed their face and number of times participants shook their foot as a within-subject variable and empathic
concern (high vs. low) as a between-subjects variable. As predicted, there was no main effect of empathic concern across the
two types of mimicking, F < 1, nor was there an interaction
between empathic concern and type of behavior (face rubbing vs.
foot shaking), F < 1. In fact, the means showed a slight trend for
there to be more mimicry of foot shaking and face rubbing among
those low in empathic concern than those high on this subscale.

Discussion
Our third goal in this research was to test a personality variable
that may moderate the extent to which one engages in behavior
mimicry. Because of the link among perspective taking and social
skills, empathy with others, and compassion for others, individuals
who often take the perspective of others are more likely to have
positive, smooth interactions. High-perspective takers may be the
ones who are better at nonconsciously guiding social interactions
and automatically doing the things that ensure smooth and easy
interactions. Part of this may entail mimicking the behavioral
mannerisms of interaction partners.
In Experiment 3, we tested whether those who take the perspective of others have more strongly developed this covert mechanism
for attaining smooth, positive interactions. Specifically, we predicted that high-perspective takers would be more likely to mimic
the mannerisms of another person. We also predicted that the
emotional facet of empathy (operationalized as the Empathic Concern subscale of the IRI) would not similarly moderate the cognitive perception-behavior link. As predicted, high-perspective takers mimicked the mannerisms of a confederate more so than did
low-perspective takers, and, also as predicted, participants who
scored low and participants who scored high on empathic concern
did not significantly differ in the extent to which they mimicked
the confederate. This supports our prediction that chronic differences in perspective taking would be related to chronic differences
in nonconscious mimicking tendencies.
General Discussion
We have argued that the perception-behavior link, through
which merely perceiving an action performed by another can lead
one to perform that action, is the mechanism behind the often
observed behavior mimicry and consequent empathic understanding within social interactions. In Experiment 1, we tested the

906

CHARTRAND AND BARGH

existence of the chameleon effect in an experimental demonstration that supported the perception-behavior link as its proximal
cause: Changes in a confederate's behavior caused changes in the
participant's behavior, in the absence of the participant's awareness of this influence. Experiment 2 provided an explicit test of the
commonly held belief that nonconscious mimicry serves the adaptive function of facilitating smooth interactions and fostering liking. In line with this prediction is the finding that individuals
whose postures and movements were mirrored by a confederate
liked that partner more and thought the interaction went more
smoothly compared with those whose behaviors were not mirrored, again without being aware of the true source of this increased empathic understanding and liking. Finally, in Experiment 3, we tested perspective taking as a individual difference that
moderates the extent to which one engages in behavior mimicry.
As hypothesized, those who frequently take the perspective of
interaction partners mimicked the mannerisms of a confederate to
a greater extent than did those who less often take the perspective
of others, as would be expected if social-perceptual activity mediated the effect.
The present experiments go beyond other recent perceptionbehavior studies in showing, for the first time, automatic behavioral effects mediated by actual, in-person perception of the partner's behavior (as opposed to priming manipulations that could
have influenced behavior directly). They also represent an advance
over existing mimicry-empathy research by providing an experimental instead of a correlational demonstration of the effect, by
ruling out the need for a purposive interaction goal in order for the
effect to occur, and by providing for the first time a mechanism for
the effect (viz., the perception-behavior link). Finally, the present
investigation shows that two formerly separate effects, previously
studied in isolation from one another, are actually outcomes of the
same underlying process.
Our conclusion that the effect of perception on behavior is an
automatic process that does not depend on conscious choice is
consistent with recent neuropsychological findings as well. One
telling piece of evidence is the fact that the frequency of direct
effects of perception on action is increased in pathological states in
which strategic conscious control over behavior is impaired or
nonexistent (Prinz, 1990, p. 176). Such "echo-reactions" as the
unintentional repetition of the words used by another (echolalia) or
unintentional imitation of another's actions (echopraxia) are commonly observed in patients with aphasia, apraxia, mental retardation, and brain damage whose ability to consciously and intentionally self-regulate is severely impaired. Thus, in the absence of
intentional forms of action control, the perception-behavior link
remains intact, arguing against the role of conscious choice as a
mediator.
Our conclusion is also in harmony with Hilgard's (1965) account of hypnotic suggestion. According to Hilgard, the directives
given by the hypnotist are first perceived by the person being
hypnotized, and then, because of the suspension of the will that is
characteristic of the hypnotic state, passive effects of perception on
action are left free to operate. In other words, the suggestions made
by the hypnotist have a direct automatic effect on behavior because
of the abdication of conscious control by the hypnotized person; in
other words, it is an instance of James's (1890) ideomotor action
effect in which the ideation is externally induced by the hypnotist
(see also Wegner & Wheatley, 1999).

Preconscious Automatic Processes as Adaptive

and Beneficial
The perception-behavior link is one of several routes through
which the environment can influence behavior without one's
awareness, intent, or control. With this particular route, perceptual
activity nonconsciously spreads to behavioral representations, increasing the likelihood of behaving similarly to others in the
current environment. There has also been research on automatic
routes from environment to behavior via the nonconscious activation of motivations and goals (see Bargh, 1990, 1997; Bargh &
Gollwitzer, 1994; Chartrand & Bargh, 1996) and via nonconscious
evaluation of environmental stimuli (see Bargh, Chaiken, et al.,
1996; Chen & Bargh, 1999). Uncovering the adaptive purpose of
the perception-behavior link is in harmony with a recent trend in
social automaticity research of identifying the adaptive purposes of
these various preconscious determinants of behavior; at the same
time, it stands in contrast with those theorists who hold that such
nonconscious effects are uniformly negative and maladaptive (e.g.,
Bandura, 1986; Langer, 1989, 1997; Locke & Latham, 1990;
Mischel, Cantor, & Feldman, 1996).
For instance, a person's chronic goals within a situation become
linked in memory to the representation of that situation, such that
entering that environment automatically causes that goal to become active and to operate without the individual's awareness of
its activation or guiding role in subsequent behavior (Bargh, 1990;
Chartrand & Bargh, 1996). This nonconscious reaction has been
conceptualized as an adaptive mechanism because it reflects the
individual's history of goal choice within the situation and increases the probability of goal pursuit. It also eliminates the need
to consciously choose the goal on each occasion, when attention
and thought may be on other matters at the time. Positive, selfactualizing goals, such as achievement, and positive aspects of
motivational states, such as persistence and overcoming obstacles
to attain the desired goal, have all been shown to occur with
nonconscious goal activation and pursuit just as they do with
conscious goal pursuit (Bargh & Chartrand, 1999; Bargh, Gollwitzer, Lee Chai, & Barndollar, 1998).
Automatic evaluation research has documented the pervasive
tendency for people to classify all environmental stimuli as either
positive or negative. This process, too, has been shown to produce
adaptive consequences. For one thing, it alerts us to what is
beneficial and helpful and what is dangerous in our environment
when conscious attention and thought are elsewhere, and it signals
the valence of the current environment by automatically affecting
the individual's mood (Chartrand & Bargh, 1999). Moreover,
approach and avoidance behavioral tendencies are put into motion
immediately by positive instead of negative automatic evaluations
(Chen & Bargh, 1999), readying the individual to react in an
appropriate manner, yet through an entirely nonconscious
mechanism.
In the present research we have continued this trend by focusing
on the adaptive function of the chameleon effect. Nonconscious
behavior mimicry was found to increase liking for the partner and
the reported smoothness of the interactions, and individuals who
often take the perspective of others engage in it more than do other
people.
It should be acknowledged that previous demonstrations of the
perception-behavior link did not produce such positive social

907

CHAMELEON EFFECT

effects. For instance, in the original Bargh, Chen, et al. (1996)

studies, individuals engaged in stereotype-consistent behavior
(e.g., hostility) following automatic activation of that stereotype
(e.g., for African Americans). Subsequent research has found that
such nonconsciously produced stereotype-consistent behavior can
produce a self-fulfilling prophecy (Chen & Bargh, 1997), in that
one's interaction partner reacts to one's behavior in kind, yet one
is not aware of the effect of one's own behavior in causing that
stereotype-confirming behavioral response.
That the stereotype version of the perception-behavior effect
can produce negative outcomes should come as no surprise, because stereotype effects on perception and judgment are also
largely negative. But stereotypes are categories gone awrythey
take the perceiver beyond the information actually present in the
other person's behavior. This does not mean that categories per se
are maladaptive or problematic; to the contrary, they are absolutely
essential for normal, moment-to-moment functioning, to simplify
the world, to give it meaning, and to furnish anticipations about
what is likely to happen next (e.g., Barsalou, 1992; Smith &
Medin, 1981). It follows that the typical form of the chameleon
effectbehavior tendencies generated nonconsciously from the
perceived behavior of one's interaction partneris, unlike the
stereotype version, largely adaptive and of high social utility. This
is what we sought to demonstrate in Experiments 2 and 3. The
usual form of the chameleon effect, we assert, is to enhance the
positivity of social interactions.

may be that newly formed groups would benefit from nonconscious mimicry and imitation as well as would established groups.
We suspect that the chameleon effect contributes to effective
behavior coordination among members of a group. The synchrony
and immediacy of such behavior coordination in moving schools
of fish or flocks of birds, for example, are the result of an
automatic, direct effect of perception on behavior (Breder, 1976;
Pitcher, 1979; Reynolds, 1987, 1993)one that clearly does not
require conscious choice or reflection to operate. Moreover, the
positive effects of empathy, liking, and bonding that occur automatically because of the chameleon effect would likely benefit
most newly formed groups in which relationships among the
members do not yet exist or are fragileit would also tend to
shape initial feelings among group members in a positive direction.
Such speculations aside, the chameleon effect is clearly a basic
and important social psychological phenomenon, one to which all
can relate on a personal level. It is our hope that research will
continue to elucidate the conditions under which the effect is
augmented or diminished. Extending the paradigms into more
complex and dynamic group settings seems to us to be an important next step to this end. It seems unlikely to us that such
pervasive, nonconscious effects on social behavior as the chameleon effect arose by accident, and such effects are more likely than
not to have positive, desirable effects for the individual and for the
groups to which he or she belongs.

References

Individual Differences in Nonconscious Mimicry

In Experiment 3, we focused on one personality variable that
moderated the chameleon effect; however, we do not mean to
suggest that no other moderators exist. Further research may well
uncover additional individual difference variables that can increase
or decrease the extent to which individuals nonconsciously mimic
those around them. One such possibility is the communal/exchange
orientation dimension proposed by Clark and colleagues (Clark &
Mills, 1979; Clark, Mills, & Powell, 1986). Individuals with a
communal orientation towards others might exhibit more nonconscious mimicry than those with an exchange orientation, because
communally oriented people are, by definition, more perceptually
attuned to the needs of others. Moreover, by the same logic, there
may be greater frequency of chameleon-like social behavior in
collectivistic versus individualistic societies, because the former
more than the latter are characterized by interdependence (e.g.,
Markus & Kitayama, 1991). Therefore, collectivistic cultures are
likely to be characterized by a relatively intensified attentional and
perceptual focus by individuals on the behavior of others, and in
light of the present experimental findings, this increased perception of others' behavior would be expected to produce greater
rapport and smoother social interactions.

Implications for Group Processes

At the level of the social group, then, to the extent that members
are mimicking each others' facial expressions, postures, mannerisms, and other behaviors, there is likely to be greater cohesion and
liking within the group. In Experiment 1, we found that mimicry
occurred even in the most minimal circumstances in which the
interactants were unacquainted and had no goal to affiliate; thus, it

Allport, G. W. (1968). The historical background of modern social psychology. In G. Lindzey & E. Aronson (Eds.), Handbook of social
psychology (2nd ed., Vol. 1, pp. 1-80). Reading, MA: Addison-Wesley.
Anisfeld, M. (1979, July 13). Response to Meltzoff and Moore (1977).
Science, 205, 214.
Banaji, M. R., Hardin, C , & Rothman, A. J. (1993). Implicit stereotyping
in person judgment. Journal of Personality and Social Psychology, 65,
272-281.
Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice
Hall.
Bandura, A. (1986). Social foundations of thought and action: A social
cognitive theory. Englewood Cliffs, NJ: Prentice Hall.
Bargh, J. A. (1989). Conditional automaticity: Varieties of automatic
influence in social perception and cognition. In J. S. Uleman & J. A.
Bargh (Eds.), Unintended thought (pp. 3-51). New York: Guilford Press.
Bargh, J. A. (1990). Auto-motives: Preconscious determinants of thought
and behavior. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of
motivation and cognition (Vol. 2, pp. 93-130). New York: Guilford
Press.
Bargh, J. A. (1994). The four horsemen of automaticity: Awareness,
efficiency, intention, and control in social cognition. In R. S. Wyer, Jr.,
& T. K. Srull (Eds.), Handbook of social cognition (2nd ed., pp. 1-40).
Hillsdale, NJ: Erlbaum.
Bargh, J. A. (1997). The automaticity of everyday life. In R. S. Wyer, Jr.
(Ed.), Advances in social cognition (Vol. 10, pp. 1-61). Mahwah, NJ:
Erlbaum.
Bargh, J. A. (1999). The cognitive monster. In S. Chaiken & Y. Trope
(Eds.), Dual process theories in social psychology (pp. 361-382). New
York: Guilford Press.
Bargh, J. A., Chaiken, S., Govender, R., & Pratto, F. (1992). The generality
of the automatic attitude activation effect. Journal of Personality and
Social Psychology, 62, 893-912.
Bargh, J. A., Chaiken, S., Raymond, P., & Hymes, C. (1996). The automatic evaluation effect: Unconditionally automatic attitude activation

908

CHARTRAND AND BARGH

with a pronunciation task. Journal of Experimental Social Psychology, 32, 104-128.

Bargh, J. A., & Chartrand, T. L. (1999). The unbearable automaticity of
being. American Psychologist, 54, 462-479.
Bargh, J. A., & Chartrand, T. L. (in press). Studying the mind in the
middle: A practical guide to priming and automaticity research. In H.
Reis & C. Judd (Eds.), Research methods for the social sciences. New
York: Cambridge University Press.
Bargh, J. A., Chen, M., & Burrows, L. (1996). Automaticity of social
behavior: Direct effects of trait construct and stereotype activation on
action. Journal of Personality and Social Psychology, 71, 230244.
Bargh, J. A., & Gollwitzer, P. (1994). Environmental control of goaldirected action: Automatic and strategic contingencies between situations and behavior. In W. D. Spaulding (Ed.), Nebraska Symposium on
Motivation (Vol. 41, pp. 71-124). Lincoln: University of Nebraska
Press.
Bargh, J. A., Gollwitzer, P., Lee Chai, A., & Barndollar, K. (1998).
Bypassing the will: Nonconscious self-regulation through automatic
goal pursuit. Manuscript submitted for publication, New York University.
Bargh, J. A., & Thein, R. D. (1985). Individual construct accessibility,
person memory, and the recall-judgment link: The case of information
overload. Journal of Personality and Social Psychology, 46, 1129-1146.
Barsalou, L. W. (1992). Cognitive psychology: An overview for cognitive
scientists. Hillsdale, NJ: Erlbaum.
Baumeister, R. R, & Leary, M. R. (1995). The need to belong: Desire for
interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497-529.
Bavelas, J. B., Black, A., Chovil, N., Lemery, C. R., & Mullett, J. (1988).
Form and function in motor mimicry: Topographic evidence that the
primary function is communication. Human Communication Research, 14, 275-299.
Bavelas, J. B., Black, A., Lemery, C. R., & Mullett, J. (1986). "I show how
you feel": Motor mimicry as a communicative act. Journal of Personality and Social Psychology, 50, 322-329.
Bavelas, J. B., Black, A., Lemery, C. R., & Mullett, J. (1987). Motor
mimicry as primitive empathy. In N. Eisenberg & J. Strayer (Eds.),
Empathy and its development (pp. 317-338). Cambridge, England: Cambridge University Press.
Berkowitz, L. (1984). Some effects of thoughts on anti- and prosocial
influences of media events: A cognitive-neoassociation analysis. Psychological Bulletin, 95, 410-427.
Bernieri, F. J. (1988). Coordinated movement and rapport in teacherstudent interactions. Journal of Nonverbal Behavior, 12, 120-138.
Bernieri, F. J., & Rosenthal, R. (1991). Interpersonal coordination: Behavior matching and interactional synchrony. In R. S. Feldman & B. Rime
(Eds.), Fundamentals of nonverbal behavior (pp. 401432). Cambridge,
England: Cambridge University Press.
Breder, C. M. (1976). Fish schools as operational structures. Fishery
Bulletin, 74, 471-502.
Brewer, M. B. (1988). A dual process model of impression formation. In
R. S. Wyer, Jr., and T. K. Srull (Eds.), Advances in social cognition
(Vol. 1, pp. 1-36). Hillsdale, NJ: Erlbaum.
Brewer, M. B. (1991). The social self: On being the same and different at
the same time. Personality and Social Psychology Bulletin, 17, 475482.
Carlston, D. E., & Skowronski, J. J. (1994). Savings in the relearning of
trait information as evidence for spontaneous inference generation. Journal of Personality and Social Psychology, 66, 840-856.
Carver, C. S., Ganellen, R. J., Framing, W. J., & Chambers, W. (1983).
Modeling: An analysis in terms of category accessibility. Journal of
Experimental Social Psychology, 19, 403421.
Charney, E. J. (1966). Psychosomatic manifestations of rapport in psychotherapy. Psychosomatic Medicine, 28, 305-315.

Chartrand, T. L., & Bargh, J. A. (1996). Automatic activation of impression formation and memorization goals: Nonconscious goal priming
reproduces effects of explicit task instructions. Journal of Personality
and Social Psychology, 71, 464-478.
Chartrand, T. L., & Bargh, J. A. (1999). Consequences of automatic
evaluation for mood. Manuscript in preparation, New York University.
Chen, M., & Bargh, J. A. (1997). Nonconscious behavioral confirmation
processes: The self-fulfilling consequences of automatic stereotype activation. Journal of Experimental Social Psychology, 33, 541-560.
Chen, M., & Bargh, J. A. (1999). Consequences of automatic evaluation:
Immediate behavioral predispositions to approach or avoid the stimulus.
Personality and Social Psychology Bulletin, 25, 215-224.
Chen, M., Chartrand, T. L., Lee Chai, A., & Bargh, J. A. (1998). Priming
primates: Human and otherwise. Behavioral and Brain Sciences, 21,
685-686.
Cialdini, R. B., Brown, S. L., Lewis, B. P., Luce, C , & Neuberg, S. L.
(1997). Reinterpreting the empathy-altruism relationship: When one
into one equals oneness. Journal of Personality and Social Psychology, 73, 481-494.
Clark, M. S., & Mills, J. (1979). Interpersonal attraction in exchange and
communal relationships. Journal of Personality and Social Psychology, 37, 12-24.
Clark, M. S., Mills, J., & Powell, M. C. (1986). Keeping track of needs in
communal and exchange relationships. Journal of Personality and Social Psychology, 51, 333-338.
Condon, W. S., & Ogston, W. D. (1966). Sound film analysis of normal
and pathological behavior patterns. Journal of Nervous and Mental
Disease, 143, 338-347.
Condon, W. S., & Sander, L. W. (1974). Synchrony demonstrated between
movements of the neonate and adult speech. Child Development, 45,
456-462.
Dabbs, J..-M. (1969). Similarity of gestures and interpersonal influence.
Proceedings of the 77th Annual Convention of the American Psychological Association, 4, 337-339.
Darwin, C. (1965). The expression of the emotions in man and animals.
Chicago: University of Chicago Press. (Original work published 1872)
Davis, M. H. (1980). A multidimensional approach to individual differences in empathy. Catalog of Selected Documents in Psychology, 10, 85.
Davis, M. H. (1983). Measuring individual differences in empathy: Evidence for a multidimensional approach. Journal of Personality and
Social Psychology, 44, 113-126.
Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and
controlled components. Journal of Personality and Social Psychology, 56, 5-18.
Dijksterhuis, A., Aarts, H., Bargh, J. A., & van Knippenberg, A. (1998).
Intergroup contact and automatic behavior. Manuscript submitted for
publication, University of Nijmegen, the Netherlands.
Dijksterhuis, A., Spears, R., Postmes, T., Stapel, D. A., Koomen, W., van
Knippenberg, A., & Scheepers, D. (1998). Seeing one thing and doing
another: Contrast effects in automatic behavior. Journal of Personality
and Social Psychology, 75, 862-871.
Dijksterhuis, A., & van Knippenberg, A. (1998). The relation between
perception and behavior, or how to win a game of Trivial Pursuit.
Journal of Personality and Social Psychology, 74, 865-877.
Dimberg, U. (1982). Facial reactions to facial expressions. Psychophysiology, 19, 643-647.
Dittmann, A. T., & Llewellyn, L. G. (1968). Relationship between vocalizations and head nods as listener responses. Journal of Personality and
Social Psychology, 9, 79-84.
Dittmann, A. T., & Llewellyn, L. G. (1969). Body movement and speech
rhythm in social conversation. Journal of Personality and Social Psychology, 11, 98-106.
Fazio, R. H., Sanbonmatsu, D. M., Powell, M. C , & Kardes, F. R. (1986).

CHAMELEON EFFECT
On the automatic activation of attitudes. Journal of Personality and
Social Psychology, 50, 229-238.
Galef, B. G. (1988). Imitation in animals: History, definition and interpretation of data from the psychological laboratory. In T. Zentall & B. G.
Galef (Eds.), Comparative social learning (pp. 3-28). Hillsdale, NJ:
Erlbaum.
Hatfield, E., Cacioppo, J. T., & Rapson, R. L. (1994). Emotional contagion. Cambridge, England: Cambridge University Press.
Herr, P. M., Sherman, S. J., & Fazio, R. H. (1984). On the consequences
of priming: Assimilation and contrast effects. Journal of Experimental
Social Psychology, 19, 323-340.
Heyes, C. M. (1993). Imitation, culture and cognition. Animal Behaviour, 46, 999-1010.
Higgins, E. T. (1989). Knowledge accessibility and activation: Subjectivity
and suffering from unconscious sources. In J. S. Uleman & J. A. Bargh
(Eds.), Unintended thought (pp. 75-123). New York: Guilford Press.
Higgins, E. T. (1996). Knowledge activation: Accessibility, applicability,
and salience. In E. T. Higgins & A. W. Kruglanski (Eds.), Social
psychology: Handbook of basic principles (pp. 133-168). New York:
Guilford Press.
Higgins, E. T., Rholes, W. S., & Jones, C. R. (1977). Category accessibility
and impression formation. Journal of Experimental Social Psychology, 13, 141-154.
Hilgard, E. R. (1965). Hypnotic susceptibility. New York: Harcourt, Brace
& World.
Jacobson, S. W., & Kagan, J. (1979, July 13). Response to Meltzoff and
Moore (1977). Science, 205, 215.
James, W. (1890). Principles of psychology. New York: Holt.
Kaitz, M., Meschulach-Sarfaty, O., Auerbach, J., & Eidelman, A. (1988).
A reexamination of newborns' ability to imitate facial expressions.
Developmental Psychology, 24, 3-7.
Kendon, A. (1970). Movement coordination in social interaction: Some
examples described: Ada Psychologica, 32, 1-25.
Koffka, K. (1925). Die grundlagen der psychischen entwicklung [The
foundations of psychological development], Osterwieck, Germany:
Zickfeldt.
La France, M. (1979). Nonverbal synchrony and rapport: Analysis by the
cross-lag panel technique. Social Psychology Quarterly, 42, 66-70.
La France, M. (1982). Posture mirroring and rapport. In M. Davis (Ed.),
Interaction rhythms: Periodicity in communicative behavior (pp. 279298). New York: Human Sciences Press.
La France, M., & Broadbent, M. (1976). Group rapport: Posture sharing as
a nonverbal indicator. Group and Organization Studies, 1, 328-333.
La France, M., & Ickes, W. (1981). Posture mirroring and interactional
involvement: Sex and sex typing effects. Journal of Nonverbal Behavior, 5, 139-154.
Langer, E. (1989). Mindfulness. Reading, MA: Addison-Wesley.
Langer, E. (1997). The power of mindful learning. Reading, MA: AddisonWesley.
Lashley, K. S. (1951). The problem of serial order in behavior. In L. A.
Jeffress (Ed.), Cerebral mechanisms in behavior: The Hixon Symposium
(pp. 112-136). New York: Wiley & Sons.
Levenson, R. W., & Ruef, A. M. (1997). Physiological aspects of emotional knowledge and rapport. In W. Ickes (Ed.), Empathic accuracy (pp.
44-73). New York: Guilford Press.
Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task
performance. Englewood Cliffs, NJ: Prentice Hall.
Macrae, C. N., Bodenhausen, G. V., Milne, A. B., Castelli, L., Schloerscheidt, A. M., & Greco, S. (1998). On activating exemplars. Journal of
Experimental Social Psychology, 34, 330354.
Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications
for cognition, emotion, and motivation. Psychological Review, 98, 224253.

909

Masters, J. C. (1979, July 13). Response to Meltzoff and Moore (1977).

Science, 205, 215.
Maurer, R. E., & Tindall, J. H. (1983). Effect of postural congruence on
client's perception of counselor empathy. Journal of Counseling Psy-

chology, 30, 158-163.

McDowall, J. J. (1978). Interactional synchrony: A reappraisal. Journal of
Personality and Social Psychology, 36, 963-975.
Mead, G. H. (1934). Mind, self, and society. Chicago: University of
Chicago Press.
Meltzoff, A. N., & Moore, M. K. (1977, October 7). Imitation of facial and
manual gestures by human neonates. Science, 198, 75-78.
Meltzoff, A. N., & Moore, M. K. (1979, July 13). Note responding to
Anisfeld, Masters, and Jacobson and Kagan's comments on Meltzoff
and Moore (1977). Science, 205, 217-219.
Meltzoff, A. N., & Moore, M. K. (1983). Newborn infants imitate adult
facial gestures. Child Development, 54, 702709.
Mischel, W., Cantor, N., & Feldman, S. (1996). Principles of self-regulation: The nature of willpower and self-control. In E. T. Higgins & A. W.
Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp.
329-360). New York: Guilford Press.
Muesseler, J., & Hommel, B. (1997). Blindness to response-compatible
stimuli. Journal of Experimental Psychology: Human Perception and
Performance, 23, 861-872.
Mussweiler, T., & Foerster, J. (1998). The sex-^aggression link: A
perception-behavior dissociation. Manuscript submitted for publication, Universitat Wurzburg.
Piaget, J. (1932). The moral judgment of the child. London: Kegan, Paul,
Trench, Trubner.
Piaget, J. (1946). La formation du symbole chez Venfant [Symbol formation in the child]. Paris: Delachaux & Niestte.
Pitcher, T. J. (1979). Sensory information and the organization of behavior
in a shoaling cyprinid fish. Animal Behavior, 27, 126-149.
Prinz, W.'(1990). A common coding approach to perception and action. In
O. Neumann & W. Prinz (Eds.), Relationships between perception and
action (pp. 167-201). Berlin: Springer-Verlag.
Reynolds, C. W. (1987). Flocks, herds, and schools: A distributed behavioral model. Computer Graphics, 21, 25-34.
Reynolds, C. W. (1993). An evolved, vision-based behavioral model of
coordinated group motion. In J.-A. Meyer, H. L. Roitblat, & S. W.
Wilson (Eds.), From animals to animals 2 (pp. 384-392). Cambridge,
MA: MIT Press.
Scheflen, A. E. (1964). The significance of posture in communication
systems. Psychiatry, 27, 316-331.
Smith, A. (1966). The theory of moral sentiments. New York: Augustus M.
Kelley. (Original work published 1759)
Smith, E. R., & Medin, D. L. (1981). Categories and concepts. Cambridge,
MA: Harvard University Press.
Smith, E. R., & Zarate, M. A. (1992). Exemplar-based model of social
judgment. Psychological Review, 99, 3-21.
Srull, T. K., & Wyer, R. S., Jr. (1979). The role of category accessibility
in the interpretation of information about persons: Some determinants
and implications. Journal of Personality and Social Psychology, 37,
1660-1672.
Srull, T. K., & Wyer, R. S., Jr. (1980). Category accessibility and social
perception: Some implications for the study of person memory and
interpersonal judgment. Journal of Personality and Social Psychology, 38, 841-856.
Stapel, D. A., Koomen, W., & van der Pligt, J. (1997). Categories of
category accessibility: The impact of trait versus exemplar priming on
person judgments. Journal of Personality and Social Psychology, 74,
878-893.
Tickle-Degnen, L., & Rosenthal, R. (1987). Group rapport and nonverbal
behavior. Review of Personality and Social Psychology, 9, 113-136.
Tomasello, M., Savage-Rumbaugh, E. S., & Kruger, A. C. (1993). Imita-

910

CHARTRAND AND BARGH

tive learning of actions on objects by children, chimpanzees, and enculturated chimpanzees. Child Development. 64, 1688-1705.
Trout, D.. & Rosenfeld, H. M. (1980). The effect of postural lean and body
congruence on the judgment of psychotherapeutic rapport. Journal of
Nonverbal Communication. 4. 176-190.
Uleman. J. S., Newman. L. S., & Moskowitz, G. B. (1996). People as
flexible interpreters: Evidence and issues from spontaneous trait inference. In M. P. Zanna (Ed.), Advances in experimental social psychology
(Vol. 28, pp. 211-279). New York: Academic Press.
Vaughan, K. B., & Lanzetta, J. T. (1980). Vicarious instigation and
conditioning of facial expressive and autonomic responses to a model's
expressive display of pain. Journal of Personality and Social Psychology, 38, 909-923.
Wegner, D. M., & Wheatley, T. P. (1999). Apparent mental causation:

Sources of the experience of will. American Psychologist, 54, 4 8 0 492.

Winter, L., & Uleman, J. S. (1984). When are social judgments made?
Evidence for the spontaneousness of trait inferences. Journal of Personality and Social Psychology, 47, 237-252.
Wyer. R. S., Jr., & Srull, T. K. (1989). Memory and cognition in its social
context. Hillsdale, NJ: Erlbaum.
Zajonc, R. B., Adelmann, K. A., Murphy, S. T., & Niedenthal, P. M.
(1987). Convergence in the physical appearance of spouses. Motivation
and Emotion, 11, 335-346.

Received April 30, 1998

Revision received December 17, 1998
Accepted December 17, 1998

Low Publication Prices for APA Members and Affiliates

Keeping you up-to-date. All APA Fellows, Members, Associates, and Student Affiliates
receiveas part of their annual duessubscriptions to the American Psychologist and
APA Monitor. High School Teacher and International Affiliates receive subscriptions to
the APA Monitor, and they may subscribe to the American Psychologist at a significantly
reduced rate. In addition, all Members and Student Affiliates are eligible for savings of up
to 60% (plus a journal credit) on all other APA journals, as well as significant discounts on
subscriptions from cooperating societies and publishers (e.g., the American Association for
Counseling and Development, Academic Press, and Human Sciences Press).
Essential resources. APA members and affiliates receive special rates for purchases of
APA books, including the Publication Manual of the American Psychological Association,
and on dozens of new topical books each year.
Other benefits of membership. Membership in APA also provides eligibility for
competitive insurance plans, continuing education programs, reduced APA convention fees,
and specialty divisions.
More information. Write to American Psychological Association, Membership Services,
750 First Street, NE, Washington, DC 20002-4242.