Skip to main content
Category: Effects
Type: Nonconscious matching of another person’s postures, mannerisms, and facial expressions
Origin: Tanya L. Chartrand and John A. Bargh, Journal of Personality and Social Psychology 76(6), June 1999, pages 893–910. Both authors’ byline is New York University
Also known as: The paper’s own phrase for this matching. A deliberate sales script, a shared mood, and a crowd’s choice are separate pages
Quick Answer — The Chameleon Effect is nonconscious matching of another person’s postures, mannerisms, facial expressions, and other behavior, so that one’s own behavior changes to fit the person in the room. Chartrand and Bargh name it in a 1999 paper. In the first experiment, 35 students smiled 1.03 times a minute with a smiling partner and 0.36 times with a neutral partner. The foot-shaking test in that same experiment prints p = .06.

​
What is the Chameleon Effect?

The Chameleon Effect is nonconscious matching of another person’s postures, mannerisms, facial expressions, and other behavior, so that one’s own behavior changes to fit the current partner.
“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.”
That sentence opens the 1999 paper. The match is something the person did not plan and, in their debriefing, did not report. A sidewalk pair that falls into the same pace, without either person calling it, is the everyday picture. The counted version is a videotaped session with a stranger who was told which gesture to repeat. The authors suggest the mechanism is the perception-behavior link, and they cite Bargh, Chen, and Burrows, 1996, for the idea that merely seeing a behavior raises the chance of doing it. Priming Effect is a different count: two words on one screen, and a faster correct decision that both are words when the words are common associates. No walking time from 1996 is on this page. Matching a gesture is also distinct from Self-Fulfilling Prophecy, which is an expectation that helps produce the outcome it expected.

​
The Chameleon Effect in 3 Depths

  • Beginner: You fold your arms after the other person does, and you had not decided to. Write the gesture down before you call the meeting chemistry.
  • Practitioner: One line for the gesture, one line for whether the partner was doing it, and one line for whether anyone noticed. A smoothness rating is a third line, kept apart from the gesture count.
  • Advanced: Smiles per minute are 1.03 against 0.36, F(1, 34) = 20.31, p below .0001. The mannerism interaction is F(1, 34) = 9.36, p = .004. The foot-shaking test is F(1, 34) = 3.76, p = .06. Those three results stay inside one New York University sample.

​
Origin

The name is in the 1999 paper, not in an earlier textbook label this page has re-read. Tanya L. Chartrand and John A. Bargh, “The Chameleon Effect: The Perception-Behavior Link and Social Interaction,” Journal of Personality and Social Psychology, 1999, volume 76, number 6, pages 893–910. The copyright line reads 1999 by the American Psychological Association. The byline is New York University. The paper reports three experiments. The first asks whether students match a stranger’s gestures. The second asks whether being matched raises liking and smoothness ratings. The third asks which empathy score travels with the matching. The opening page also prints a literary picture of a man taking on the mannerisms of people around him without realizing it. That picture is not the definition and not a count. The definition is the sentence quoted above.

​
Key Points

Four readings keep a smile count, a foot test, a rating, and an empathy score from collapsing into one slogan.
1

The match was scored from videotape

Experiment 1 used two sessions and two partners. Each partner was a confederate, a person playing the role of another participant. In one session the confederate rubbed their face or shook their foot. In the other session the second confederate did the behavior the first had not done. One confederate smiled through the session. The other kept a neutral face. The order was counterbalanced. Students thought they were helping develop a projective measure made of photographs. The camera recorded them so later coders could score the gestures.
2

Smiling moved. Foot shaking printed p = .06

Thirty-nine introductory-psychology students at New York University took part for course credit. Three declined to release the videotape for coding. One, in debriefing, suspected the other person was a confederate and did not guess the hypothesis. The analyses use the remaining 35. Smiles per minute were 1.03 with the smiling confederate and 0.36 with the neutral confederate, F(1, 34) = 20.31, p below .0001. Face rubbing was higher beside the face-rubbing confederate, F(1, 34) = 5.71, p below .025. Foot shaking was higher beside the foot-shaking confederate, F(1, 34) = 3.76, p = .06. The interaction of confederate behavior and participant behavior is F(1, 34) = 9.36, p = .004. A later summary in the paper says significant mimicry was found for facial expressions and for two mannerisms. The foot-shaking line still prints p = .06.
3

Being copied raised two ratings

Experiment 2 used the same photograph task. The confederate either mirrored the participant’s posture and mannerisms or used neutral, nondescript mannerisms. The analyses use 72 people, 37 mirrored and 35 in the control. Liking, on a scale from 1 (extremely dislikable) to 9 (extremely likable), was 6.62 against 5.91, F(1, 70) = 5.55, p = .02. Smoothness, from 1 (extremely awkward) to 9 (extremely smooth), was 6.76 against 6.02, F(1, 70) = 4.08, p = .05. A joint test of the two ratings is F(2, 69) = 3.47, p = .04, and the paper says that effect was not moderated by which rating was used, interaction F below 1. The results paragraph refers to the confederate as her.
4

Perspective taking moved. Empathic concern did not

Experiment 3 used the Interpersonal Reactivity Index. The paper cites Davis, 1983, for that questionnaire and was not a re-read of Davis. A median split on perspective taking used a median of 19: 28 students above it and 22 below it. Across face rubbing and foot shaking, F(1, 48) = 3.85, p = .05. High scorers rubbed their face 1.30 times a minute and shook a foot 0.40 times. Low scorers were at 0.85 and 0.29. The type of gesture did not change that gap, p above .20. Empathic concern, median 21, with 28 above and 22 below, did not show a reliable difference. The abstract’s phrase “dispositionally empathic” is wider than this split.

​
Applications

Use the name after the gesture, the count, and the thing that was only a rating are written down. The figures belong to these laboratories.

Write the gesture before the chemistry

In a study group or a first meeting, note one repeated gesture and whether you had decided to copy it. None of the 35 students in Experiment 1 mentioned the confederate’s foot shaking or face rubbing. One mentioned hand motions. Two commented on a slouch. A feeling that the hour went well is the Experiment 2 rating, kept on its own line.

Keep a copied posture off a copied opinion

Bandwagon Effect is doing what a visible crowd is doing. Two strangers and one assigned gesture are not that crowd. In-Group Bias is a preference for one’s own group. The 1999 partners were strangers in a laboratory, and the students were not choosing a group.

Treat a planned mirror as a different act

The 1999 definition requires that the change be passive and unintended. A manager who is told to copy a client’s posture has left that definition. Experiment 2 still shows what happened when a confederate was instructed to mirror: liking 6.62 against 5.91 on one 9-point item. That gap is this questionnaire. It is a rating in this sample.

Leave the pen count in Nijmegen

Rick B. van Baaren, Rob W. Holland, Kerry Kawakami, and Ad van Knippenberg, Psychological Science 15(1), 2004, pages 71–74, asked a further question: after being mimicked, do people help. In their first study, 17 students at the University of Nijmegen took part. Students who had been mimicked picked up six dropped pens more often (100 percent) than students who had not (33 percent), chi-square (1, N = 17) = 8.24, p below .01. That pair of shares is this study. A service counter, a donation page, and a classroom are different rooms.

​
Case Study

The case is the first New York University experiment, with the rating experiment from the same paper kept beside it. Students in introductory psychology sat with a stranger and described photographs, as a cover story about building a projective measure. The stranger was a confederate. Across two sessions the confederates swapped jobs: face rubbing or foot shaking, and a smile or a neutral face. One printed order is 14 of the 35: the first confederate smiled and shook a foot, and the second kept a neutral face and rubbed their face. The paper does not print every other order cell on the lines read here. Coders later counted the student’s own smiles, face rubs, and foot shakes per minute. The analysis controlled for a baseline of each behavior. The paper’s discussion uses the abbreviation BL for that baseline. The length of the baseline is not on this page. The smile gap is the clear one: 1.03 against 0.36 times a minute. Face rubbing followed the face-rubbing confederate at p below .025. Foot shaking followed the foot-shaking confederate at p = .06. The interaction that crosses the two gestures is p = .004. Afterward, no student named the foot or the face rub. The match the coders scored was not a match the students reported. The second experiment turns the arrow around. A confederate either copied the student’s posture and mannerisms or stayed with neutral mannerisms, and no longer smiled, shook a foot, or rubbed a face as the assigned behavior. On the 9-point items, liking moved from 5.91 to 6.62 and smoothness from 6.02 to 6.76. The smoothness test prints p = .05. The paper also says four participants suspected that the other person was a confederate, and that none of them guessed the hypothesis. This page does not add those four to the 72, and does not subtract them. The 37 and the 35 are the analyzed counts the paper prints. A later laboratory asked whether help, rather than a rating, moved. At the University of Nijmegen, van Baaren and colleagues paid students 2 euros. In the first study the experimenter copied posture, arm position, and leg position for about six minutes while the student talked through ten advertisements, then dropped six pens. The 10-second window for picking them up belongs to that study: 100 percent against 33 percent, N = 17. Mood, on three 7-point scales with alpha = .89, did not show a mimicry effect, F(1, 15) = 0.58, marked n.s., and mood did not track how many pens were picked up, r = −.22, marked n.s. A second study, N = 42, had a new experimenter drop the pens: 84 percent against 48 percent, chi-square (1, N = 42) = 6.00, p below .02. A third study, N = 41 after three visiting students who were not native Dutch speakers were excluded, used four 50-cent coins and a locked box for CliniClowns, a charity that visits seriously ill children in hospital. Mimicked students donated more often, 76 percent against 43 percent, chi-square (1, N = 41) = 4.84, p below .03, and gave a higher mean amount, 0.79 euro against 0.38 euro, F(1, 37) = 4.26, p below .05. The same-experimenter means are 0.91 against 0.25 euro. The new-experimenter means are 0.72 against 0.56 euro. The interaction is F(1, 37) = 1.52, p = .23. The lesson is a videotaped match of smiles, a weaker foot test, a small shift on two rating scales, and a later helping count in a Dutch laboratory. The 1.03 smiles, the 6.62, and the 100 percent are three different measurements.

​
Boundaries and Failure Modes

The pattern has a definition, a smile count, a foot test at p = .06, and a rating study. It does not have one number that travels. The authors’ discussion groups both mannerisms with the smiles as significant mimicry, after controlling for baseline. The foot-shaking simple effect still prints p = .06. A reader who needs both gestures to clear .05 does not have that pair. The face-rubbing effect does clear .025, and the crossed interaction is p = .004. Those are three statements. They stay in the sample of 35. The abstract says dispositionally empathic students show the effect more. Experiment 3 separates that claim. Perspective taking, the cognitive habit of imagining the other person’s point of view, shows F(1, 48) = 3.85, p = .05, with means 1.30 and 0.40 against 0.85 and 0.29. Empathic concern, the emotional subscale, does not show a reliable difference. People at the median are outside those 28-and-22 splits, because the paper classifies scores above the median and below it. This page does not invent a full head count by adding 28 and 22. The paper sets gender aside and does not discuss it further, so no gender result is printed here. Van Baaren’s pen and coin counts answer a different question: help after being copied. Experiment 1 of that paper is 17 students. The 100 percent is a share of one condition inside that 17, and the paper does not print the condition sizes as head counts. Mood did not explain the pens. Experiment 3’s 0.79 euro is a mean donation from the 2 euros just handed to the student, in a room with a locked charity box. The interaction with whether the requester was the same person is p = .23, so the page keeps the same-person gap, 0.91 against 0.25, beside the new-person gap, 0.72 against 0.56, and does not average them into one charity effect. A common misuse is to treat 1.03 smiles a minute, the liking gap from 5.91 to 6.62, or 100 percent of a 17-person study as the size of the Chameleon Effect in a shop, a clinic, a classroom, or a family dinner. Halo Effect is one known trait coloring the rest of a judgment. A 9-point liking item after a mirrored posture is not that coloring. The 1996 citation inside the 1999 mechanism paragraph was not re-read, and no second from it is on this page.

​
Common Misconceptions

Three retellings travel with the name. One widens an empathy score. One treats every gesture as equally clear. One turns a rating into a result outside the questionnaire.
The abstract says dispositionally empathic students show the Chameleon Effect more. The subscale test that moves is perspective taking, F(1, 48) = 3.85, p = .05: 1.30 face rubs and 0.40 foot shakes a minute against 0.85 and 0.29. Empathic concern, split at a median of 21, does not show a reliable difference. “Empathic” in the abstract and “perspective taking” in the table are not the same score.
Smiles per minute are 1.03 against 0.36, F(1, 34) = 20.31, p below .0001. Face rubbing is F(1, 34) = 5.71, p below .025. Foot shaking is F(1, 34) = 3.76, p = .06. The interaction of the two gestures is p = .004, and a later sentence in the paper calls both mannerisms significant. The foot line the paper prints is still p = .06. A summary that drops that p has left the table.
Experiment 2’s liking means are 6.62 and 5.91 on a 9-point item, F(1, 70) = 5.55, p = .02. Smoothness is 6.76 against 6.02, p = .05. Both are ratings by students who had just described photographs with a confederate. Van Baaren’s 100 percent is the share who picked up pens in one condition of a 17-student study, beside a 33 percent share in the other condition. A household, a sales floor, and a donation drive are different rooms.
The Chameleon Effect sits beside a word-pair speedup, an expectation that shapes an outcome, a trait that colors a judgment, a crowd’s visible choice, and a preference for one’s own group.

Priming Effect

A faster correct decision that two letter strings are words, when the words are common associates. The 1999 paper names the perception-behavior link as its suggested mechanism. The millisecond count lives on the priming page.

Self-Fulfilling Prophecy

An expectation that helps bring about the outcome it expected. A foot shake scored from videotape is a gesture count. It is an expectation only if someone wrote the expectation down.

Halo Effect

One known trait colors the rest of a judgment. The 6.62 is a liking rating after a confederate had mirrored posture and mannerisms. It is that item in that sample.

Bandwagon Effect

Doing what a visible crowd is already doing. The 1999 sessions are two people, and one of them was assigned a gesture. A crowd’s choice is a different design.

In-Group Bias

A preference for one’s own group over an outside group. The partners in Experiment 1 were strangers, and the task was a photograph description. Matching a stranger’s smile is not a vote for a group.

​
One-Line Takeaway

Before you call a meeting chemistry, write the gesture that was copied, whether anyone noticed, and whether the number you have is a smile count, a rating, or a helping share.