Abstract
Extraverts are better than introverts at building rapport, but it remains unknown what they do behaviorally to better connect with other individuals. We hypothesized that extraverts mimic more than introverts as a way to build rapport; however, we predicted that this social skillfulness of extraverts emerges only when they are motivated to affiliate. In Study 1, we found that extraversion predicted increased mimicry when an affiliation goal was present, but not when an affiliation goal was absent. In Study 2, we found that mimicry mediates the relationship between extraversion and rapport, but only when an affiliation goal is present. Our findings are the first to identify a behavior that extraverts engage in that helps them build rapport. Furthermore, our studies show that this skillfulness of extraverts emerges only when they are motivated to affiliate, providing evidence in favor of the reward-sensitivity-as-core model of extraversion over the sociability-as-core model of extraversion.
Social relationships are vital to human flourishing. Given their importance, it may not be surprising that extraverts tend to be happier than introverts 1 (e.g., Furnham & Cheng, 1999), a finding that holds cross-culturally—from Western societies that put a premium on extraverted traits to Eastern societies that strongly value introverted traits (Furnham & Cheng, 1999). Some researchers have suggested that extraverts are happier because they are more sociable (McCrae & Costa, 1987), are more socially skillful (as assessed by self-report measures; Argyle & Lu, 1990a), and contribute more to social interactions (Smillie, Wilt, Kabbani, Garratt, & Revelle, 2015). Indeed, research indicates that extraverts, compared with introverts, have higher-quality social interactions (Barrett & Pietromonaco, 1997; Barrick et al., 2012; Berry & Hansen, 2000) and are better at creating favorable impressions (Riggio & Friedman, 1986; Thorne, 1987) and establishing rapport (Barrick et al., 2012). However, a critical question remains: What behaviors do extraverts engage in that make them better at building rapport?
Extraverts have been found to talk faster and speak more energetically than introverts (Feldstein & Sloan, 1984; Scherer, 1978), but there is no evidence that this leads to their greater rapport with others. Some researchers have attempted to characterize a nonverbal profile of extraversion, but this work has yielded inconsistent results (Funder, 2001; La France, Heisel, & Beatty, 2004). Theoretically, extraverts, compared with introverts, should engage more frequently in nonverbal behaviors related to rapport, such as gazing, smiling, and nodding (Bernieri, Gillis, Grahe, & Davis, 1996), but research has not consistently found such a difference (La France et al., 2004). For example, in one study, even though extraversion was associated with higher-quality social interactions, it was unrelated to the nonverbal behaviors (i.e., visual attention, body openness, and gestures) that actually predicted the quality of social interaction (Berry & Hansen, 2000). Thus, no specific behavior, either verbal or nonverbal, has been shown to explain extraverts’ better ability to build rapport (Back, Schmukle, & Egloff, 2009; Wilt & Revelle, 2009).
Given that a behavioral profile of extraversion remains elusive, is it possible that extraverts engage in a subtler form of affiliation that generally goes undetected—a behavior that occurs automatically and nonconsciously when they are motivated to affiliate? Behavioral mimicry is such a behavior. Research indicates that it is a powerful affiliation tool, increasing liking and rapport between interactants (Chartrand & Bargh, 1999; Lakin & Chartrand, 2003). Furthermore, although mimicry occurs nonconsciously, people in fact mimic more when they are motivated to affiliate with others than when they are not so motivated (Lakin & Chartrand, 2003; Lakin, Chartrand, & Arkin, 2008). Perhaps behavioral mimicry is a nonverbal behavior that explains why extraverts are better than introverts at building affiliation (Barrick et al., 2012; Riggio & Friedman, 1986; Thorne, 1987).
Two competing models of extraversion make different predictions about the context in which the social skillfulness of extraverts should emerge. The sociability-as-core model of extraversion states that extraverts are happier than introverts because they are more sociable (Argyle & Lu, 1990b; McCrae & Costa, 1987), whereas the reward-sensitivity-as-core model of extraversion asserts that extraverts are more sociable than introverts because they are happier (Gray, 1973; Lucas, Diener, Grob, Suh, & Shao, 2000). From these models come different predictions regarding when extraverts should be more socially skillful than introverts.
According to the sociability-as-core model (Argyle & Lu, 1990b; McCrae & Costa, 1987), extraverts should be more socially skillful across all social interactions. Given that mimicry leads to smoother social interactions (Chartrand & Bargh, 1999), one might therefore predict that extraverts should mimic more than introverts across the board. However, according to the reward-sensitivity-as-core model (Gray, 1973), the greater happiness of extraverts arises from their greater motivation to achieve goals (Smillie, Cooper, Wilt, & Revelle, 2012; Smillie, Geaney, Wilt, Cooper, & Revelle, 2013). Therefore, it follows that extraverts should be more socially skillful than introverts when they are motivated by an affiliation goal (Depue & Collins, 1999). On the basis of this model, one would predict that extraverts should engage in socially skillful behaviors such as mimicry more than introverts only when they have an affiliation goal. Thus, the sociability-as-core model predicts that extraverts will mimic more than introverts regardless of the presence or absence of an affiliation goal, whereas the reward-sensitivity-as-core model predicts that in the absence of an affiliation goal, extraverts and introverts will not mimic differently. In the research reported here, we tested these competing hypotheses by determining the contexts under which the social skillfulness of extraverts emerges.
The reward-sensitivity-as-core model has been gathering increasing empirical support; indeed, an abundance of research, including a large-scale cross-cultural study (Lucas et al., 2000), supports the idea that extraversion might reflect individual differences in the motivational system. Extraversion is associated with genetic variation in dopamine, a neurotransmitter that drives reward-seeking behavior, and this genetic variation predicts neural activation in a reward-seeking paradigm (Cohen, Young, Baek, Kessler, & Ranganath, 2005). In addition, extraversion is correlated with larger volume of brain structures involved in reward processing (DeYoung et al., 2010; Rauch et al., 2005), and functional MRI studies have shown that extraversion is linked to increases in activation of reward-related brain structures (e.g., Canli, Sivers, Whitfield, Gotlib, & Gabrieli, 2002; Cohen et al., 2005). Extraversion is also related to enhanced behavioral pursuit of reward (Cooper, Agocha, & Sheldon, 2000). Furthermore, behavioral studies have shown that extraverts experience increases in positive affect after exposure to situations and stimuli involving the pursuit of reward, but not after exposure to situations and stimuli that are merely pleasant (Smillie et al., 2012; Smillie et al., 2013).
Given this behavioral, genetic, and neural evidence in support of the reward-sensitivity-as-core model, we predicted that extraverts would mimic more than introverts only when they were motivated by an affiliation goal. Furthermore, we predicted that the mimicry that occurred in the presence of an affiliation goal would mediate the relationship between extraversion and rapport. These findings would provide evidence of a nonverbal behavior that helps extraverts build rapport, as well as of the context in which extraverts boost their rapport-building abilities.
To test these hypotheses, we designed two studies. Study 1 tested whether extraverts mimic more than introverts in the presence of an affiliation goal. Study 2 replicated and expanded on the design of Study 1 by assessing whether mimicry mediates the relationship between extraversion and rapport (as measured by an independent observer).
Method
Participants
For Study 1 (N = 95) and Study 2 (N = 102), we recruited participants from a university participant pool. All were female, 18 to 30 years old, and compensated $11 for participation in a 30-min study. All research activities were approved by Duke University’s institutional review board.
On the basis of sample sizes used in prior mimicry research (Cheng & Chartrand, 2003; Lakin & Chartrand, 2003; Lakin et al., 2008), we aimed to collect data for 20 participants per experimental condition. Initially, we had a 2 × 2 between-subjects design because we were testing an additional hypothesis, 2 so we aimed for a total of 80 participants. We anticipated that some participants might need to be excluded because of experimental errors, so we aimed beyond the target. After we excluded participants for whom mimicry data were missing, there were 84 participants remaining in Study 1 and 100 participants remaining in Study 2.
Procedure
Manipulation of affiliation goal
Participants were randomly assigned to an affiliation-goal condition (goal absent vs. present). They were told that they would be participating in a task with another participant (actually a confederate). In the affiliation-goal-present condition, before bringing in the confederate, the experimenter suggested to the participant that the next task had the best results when both people got along well. In the affiliation-goal-absent condition, participants were not told this.
Measure of behavioral mimicry
The standard paradigm for measuring behavioral mimicry involves measuring how much the participant mimics the confederate while they engage in a diversion task together. We used different diversion tasks in Studies 1 and 2.
For Study 1, we used a photo description task (Chartrand & Bargh, 1999; Cheng & Chartrand, 2003; Lakin & Chartrand, 2003; Lakin et al., 2008). The participant and the confederate took turns describing five photographs each for approximately 1 min per photograph. The participant and confederate sat in chairs facing one another at a 90° angle. While they described the photographs, the confederate engaged in easily mimicked behaviors, continuously touching her hair and face and moving her foot.
For Study 2, we used a word-listing task adapted from Rachl, Leander, and Van Yperen (2015). The participant and confederate took turns coming up with words within a given category. For each category, they went back and forth listing words for 30 s until a timer beeped. They then switched to the next category until they had covered all 12 categories (total of 6 min). The setup again involved the participant and confederate sitting in chairs facing one another at a 90° angle, but unlike in Study 1, they were sitting at a table. While they were listing words, the confederate continuously touched her hair and face.
During these tasks, the participants were filmed with a hidden camera, and the videos were coded for how much time the participants spent touching their hair and face (Studies 1 and 2) and moving their foot (Study 1). One observer (blind to the participants’ conditions) coded all videos, and a second observer (also blind to the participants’ conditions) coded 15% (Study 1) or 20% (Study 2) of the videos so that we could assess coder reliability (Chartrand & Bargh, 1999; Lakin & Chartrand, 2003). For videos coded by both observers, we assessed the interrater reliability using a two-way mixed, consistency, average-measures intraclass correlation (ICC; McGraw & Wong, 1996). The resulting correlation was excellent for both Study 1, ICC = .98, and Study 2, ICC = .99 (Cicchetti, 1994). Then, we calculated separate mimicry rates from the amount of time each participant (a) touched her hair and face and (b) moved her foot during the social interaction (seconds of mimicry/minute of interaction).
Measure of extraversion
At the end of the study, participants completed a Qualtrics survey assessing trait extraversion using Goldberg’s Mini-Markers, an eight-item extraversion scale (Saucier, 1994). Participants rated themselves on each adjective (bold, energetic, extraverted, talkative, quiet, shy, bashful, and withdrawn) using a 7-point Likert scale. After four items (quiet, shy, bashful, and withdrawn) were reverse-scored, scores on all items were averaged to obtain a composite measure of extraversion.
Measure of rapport
In Study 2, we assessed rapport within the social interactions. One observer (blind to the hypotheses, design of the study, and participants’ conditions) coded all the videos for the rapport between the participant and confederate, and a second observer coded 20% of the videos. For videos coded by both observers, we assessed the interrater reliability using a two-way mixed, consistency, average-measures intraclass correlation (McGraw & Wong, 1996). The resulting correlation between coders was excellent, ICC = .91 (Cicchetti, 1994).
Results 3
Given that extraversion was measured after our manipulations and we were hoping to capture trait rather than state extraversion, we wanted to ensure that our manipulations did not affect extraversion scores. We therefore ran an analysis of variance with affiliation-goal condition (goal absent vs. present) as the independent variable and extraversion score as the dependent variable. There was no statistically significant difference in extraversion scores between groups in either Study 1, F(1, 82) = 0.04, p = .83, or Study 2, F(1, 98) = 2.30, p = .13. Note that although there was a marginal group difference in Study 2, the trend was in the unexpected direction. That is, if our manipulation had affected extraversion scores, they should have been higher in the affiliation-goal-present than in the affiliation-goal-absent condition. However, we actually obtained the opposite pattern (goal absent: M = 4.86; goal present: M = 4.49). Thus, we conclude that our measure of chronic trait extraversion was not affected by our manipulation.
We did not find significant effects for mimicry as measured by foot moving, which is consistent with recent mimicry research (Duffy, Stanton, Chartrand, & Harris, 2015; Liu, Gleibs, Chartrand, & Shih, 2015; Rachl & Leander, 2015). Thus, all the analyses reported in this article focused on the rate of hair and face touching. We performed a square-root transformation on hair- and face-touching mimicry in order to correct the normality of the distribution.
Effect of extraversion and presence of an affiliation goal on mimicry
We predicted that extraverts would mimic more than introverts, but that this effect would be moderated by the presence of an affiliation goal. To test our predictions, we ran a moderated multiple regression analysis on square-root-transformed mimicry. We set extraversion score, affiliation-goal condition (goal absent = 0, goal present = 1), and their interaction as the independent variables. The resulting model was significant for both Study 1, adjusted R2 = .08, F(3, 80) = 3.4, p = .02, and Study 2, adjusted R2 = .08, F(3, 96) = 3.9, p = .01. The interaction between extraversion and affiliation-goal condition was significant in Study 1, b = 0.54, t(80) = 2.1, p = .04, and in Study 2, b = 0.65, t(96) = 2.1, p = .04. The simple slope of extraversion in the affiliation-goal-present condition was significant for both Study 1, b = 0.43, t(80) = 2.55, p = .01, and Study 2, b = 0.54, t(96) = 2.39, p = .02. In contrast, the simple slope of extraversion in the affiliation-goal-absent condition was not significant in either Study 1, b = −0.11, t(80) = −0.59, p = .56, or Study 2, b = −0.11, t(96) = −0.55, p = .59 (see Fig. 1). Thus, higher extraversion predicted greater mimicry, but only in the presence of an affiliation goal.

Best-fitting regression lines showing the effect of extraversion on mimicry in the affiliation-goal-present and affiliation-goal-absent conditions of Study 1 and Study 2.
In order to substantiate the claim that the null hypothesis is more likely than the alternative hypothesis for the simple slope of extraversion in the affiliation-goal-absent condition, we calculated the Bayes factor for each study. The Bayes factor indicates the probability of the data under the null hypothesis relative to the probability of the data under the alternative hypothesis, indicating the likelihood of one model relative to the other (Rouder, Speckman, Sun, Morey, & Iverson, 2009). For Study 1, the Bayes factor was .0005, which means that the null hypothesis was 2,000 times more likely than the alternative hypothesis. For Study 2, the Bayes factor was .0003, which means that the null hypothesis was 3,333 times more likely than the alternative hypothesis. For both studies, the magnitude of the effect was large enough to be considered decisive evidence in favor of the null hypothesis (Jeffreys, 1961).
Effect of extraversion on rapport
Previous research has demonstrated that extraverts are better than introverts at building rapport within social interactions (Barrick et al., 2012). Because we ultimately wanted to test whether mimicry helps extraverts build rapport in social interactions, we first needed to demonstrate that the extraverts in Study 2 indeed built more rapport within their social interactions than the introverts did. To test this, we ran a standard regression analysis. We set extraversion score as the independent variable and rapport (as rated by an independent observer) as the dependent variable. The resulting model was significant, adjusted R2 = .04, F(1, 98) = 4.7, p = .03, and indicated that extraversion significantly predicted rapport, 4 b = 0.18, t(98) = 2.2, p = .03 (see Fig. 2). Replicating previous research, we found that extraversion leads to greater rapport with other people. However, because the focus of this study was to understand how and when extraverts build rapport, we next ran a moderated mediation model.

Best-fitting regression line showing the effect of extraversion on rapport in Study 2.
Is the effect of extraversion on rapport mediated by mimicry when an affiliation goal is present?
In Study 2, we tested the hypothesis that when extraverts (but not introverts) are given an affiliation goal, they engage in more behavioral mimicry, which, in turn, mediates the relationship between extraversion and rapport. To test this hypothesis, we constructed a moderated mediation model using the PROCESS macro (Hayes, 2013). This procedure generated a 95% confidence interval (CI) for the indirect effect using 50,000 bootstrap resamples. Confidence intervals that do not contain a zero indicate a significant effect. Using Model 7, we set extraversion score as the independent variable, square-root-transformed mimicry as the mediator, affiliation-goal condition (goal present vs. goal absent) as the moderator of the path between extraversion and mimicry, and rapport as the outcome variable.
As we reported earlier, the total effect of extraversion on rapport was significant, b = 0.18, t(98) = 2.2, p = .03. However, we wanted to test whether mimicry mediated this effect, specifically in the presence of an affiliation goal. We found that the effect of extraversion on mimicry as moderated by affiliation-goal condition was significant, b = 0.65, t(96) = 2.1, p = .04. The effect of mimicry on rapport (controlling for extraversion) was also significant, b = 0.12, t(97) = 2.3, p = .02. After we controlled for mimicry, there remained a significant direct effect of extraversion on rapport, b = 0.17, t(97) = 2.0, p = .05. Furthermore, our results for the indirect effect of the highest-order interaction indicated that the effect of extraversion on rapport was significantly mediated through mimicry, b = 0.08, 95% bias-corrected CI = [0.01, 0.20] (see Fig. 3). In addition, we assessed the conditional indirect effect at different levels of the moderator. In the presence of an affiliation goal, the effect of extraversion on rapport was significantly mediated through mimicry, b = 0.06, 95% bias-corrected CI = [0.008, 0.17]. In the absence of an affiliation goal, as expected, there was no significant conditional indirect effect of extraversion on rapport through mimicry, b = −0.01, 95% bias-corrected CI = [−0.07, 0.02].

Moderated mediation model of the effect of extraversion on rapport in Study 2 as mediated by mimicry and moderated by affiliation-goal condition (goal present vs. goal absent). Unstandardized regression coefficients are shown; the value in parentheses is the coefficient for the path from extraversion to rapport when the model controlled for mimicry (i.e., the direct effect). Asterisks indicate significance of the coefficients (*p < .05).
Discussion
The results of these studies suggest that extraverts mimic other people most when it is highly adaptive to do so; they rely on mimicry as a way to build rapport particularly when they have a goal to affiliate. Previous research indicated that extraverts are better than introverts at building affiliation in social interactions (Thorne, 1987), but the behaviors driving this difference remained unidentified. The current findings suggest one behavior that extraverts engage in more effectively—mimicry. Thus, we have identified a context-dependent behavioral phenotype of extraverts that can help explain why and how they are so socially adept. Our studies address a long-standing gap in the literature and further substantiate the claim that traits such as extraversion can predict actual (i.e., not self-reported) behaviors, despite a lack of such research in the literature 5 (Back et al., 2009; Funder, 2001; Wilt & Revelle, 2009).
In addition to exploring how extraverts build rapport, we examined when they build rapport. In so doing, we provided a critical test of two competing models of extraversion, the sociability-as-core and the reward-sensitivity-as-core models. The sociability-as-core model suggests that extraverts should mimic more than introverts across all situations. However, our results do not corroborate this position. We found that when there was no affiliation goal present, introverts and extraverts did not differ in their mimicking of a confederate, providing evidence against the widely held notion that extraverts are more generally socially skillful (Argyle & Lu, 1990a; McCrae & Costa, 1987). Instead, we found that extraverts used mimicry to connect with other people only when they were motivated by an affiliation goal. Our findings are the first evidence that extraverts’ ability to connect with other people is motivationally dependent, and they support the reward-sensitivity-as-core model of extraversion (Depue & Collins, 1999; Gray, 1973). Although we did successfully replicate our finding that extraverts mimic more than introverts when an affiliation goal is present, we should note that our sample sizes were rather small, and our studies were likely somewhat underpowered. Future research should test our hypotheses using larger sample sizes.
Overall, our results demonstrate a behavioral mechanism by which extraverts boost rapport. Specifically, extraverts mimic more when they want to get along with another person, and this mimicry mediates the relationship between extraversion and rapport. The processes underlying mimicry occur automatically and nonconsciously, which reinforces the idea that mimicry may represent an evolutionarily hardwired system of affiliating that is dialed up and down by motivational goals (Lakin, Jefferis, Cheng, & Chartrand, 2003). Thus, extraverts are not always more socially skillful than introverts—they are more skillful only when they are motivated to be.
Footnotes
Acknowledgements
We would like to thank Brock Knapp, Kaitlyn Batt, Courtney Jones, Bree Gray-Jordan, Devon Gagliardi, Lauren Jamiolkowski, Olivia Chen, and Colin Wierengo for coding behavioral videos and helping to run the study. We extend special thanks to Gavan Fitzsimons, Vivian Qin, Steven Shepherd, and Peggy Liu for helping with statistics.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.
Open Practices
All data have been made publicly available via Open Science Framework and can be accessed at https://osf.io/wc3sj/. The complete Open Practices Disclosure for this article can be found at http://pss.sagepub.com/content/by/supplemental-data. This article has received the badge for Open Data. More information about the Open Practices badges can be found at https://osf.io/tvyxz/wiki/1.%20View%20the%20Badges/ and
.
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
