- Email: [email protected]
Moving speeches: Dominance, trustworthiness and competence in body motion
Personality and Individual Differences 94 (2016) 101–106
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Moving speeches: Dominance, trustworthiness and competence in body motion Markus Koppensteiner ⁎, Pia Stephan, Johannes Paul Michael Jäschke Department of Anthropology, University of Vienna, Althanstraße 14, 1090, Vienna, Austria
a r t i c l e
i n f o
Article history: Received 11 August 2015 Received in revised form 6 January 2016 Accepted 8 January 2016 Available online 18 January 2016 Keywords: Nonverbal communication Motion cues Social cognition Politics Ethology Impression formation
a b s t r a c t People read dominance, trustworthiness and competence into the faces of politicians but do they also perceive such social qualities in other nonverbal cues? We transferred the body movements of politicians giving a speech onto animated stick-figures and presented these stimuli to participants in a rating-experiment. Analyses revealed single body postures of maximal expansiveness as strong predictors of perceived dominance. Also, stick-figures producing expansive movements as well as a great number of movements throughout the encoded sequences were judged high on dominance and low on trustworthiness. In a second step we divided our sample into speakers from the opposition parties and speakers that were part of the government as well as into male and female speakers. Male speakers from the opposition were rated higher on dominance but lower on trustworthiness than speakers from all other groups. In conclusion, people use simple cues to make equally simple social categorizations. Moreover, the party status of male politicians seems to become visible in their body motion. © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
1. Introduction Nonverbal cues affect impression formation (Ambady, Bernieri, & Richeson, 2000; Borkenau, Mauer, Riemann, Spinath, & Angleitner, 2004) and decision making in the public arena (Rule & Ambady, 2011; Zebrowitz & Montepare, 2008). For instance, attributions of dominance, trustworthiness, competence and other personality traits to specific facial features of political candidates can be reliable predictors of hypothetical and actual election outcomes (Antonakis & Dalgas, 2009; Chen, Jing, & Lee, 2014; Little, Roberts, Jones, & DeBruine, 2012; Olivola & Todorov, 2010; Oosterhof & Todorov, 2008; Poutvaara, Jordahl, & Berggren, 2009). Apart from facial and other nonverbal cues people also read socially relevant information from and into body motion. They recognize emotions in arm movements (Pollick, Paterson, Bruderlin, & Sanford, 2001), in whole body gestures (Atkinson, Tunstall, & Dittrich, 2007), and in movements displayed during interpersonal dialog (Clarke, Bradshaw, Field, Hampson, & Rose, 2005). Moreover, they perceive personality traits in dance movements (Hugill, Fink, Neave, Besson, & Bunse, 2011) and in walking behaviors (Thoresen, Vuong, & Atkinson, 2012), and health related cues and personality in the body movements of politicians giving a speech (Koppensteiner, 2013; Koppensteiner & Grammer, 2010; Kramer, Arend, & Ward, 2010). Humans and animals appear to use expansive body postures, expressive body movements, and broad gestures to display power and
⁎ Corresponding author. E-mail address: [email protected] (M. Koppensteiner).
dominance (Carney, Hall, & LeBeau, 2005; De Waal, 2007; Eisenberg & Reichline, 1939; Mehrabian, 1972; Tiedens & Fragale, 2003). Also, people adopting open and expansive postures (i.e., power posing) have an enhanced sense of power (Carney, Cuddy, & Yap, 2010; Huang, Galinsky, Gruenfeld, & Guillory, 2011; Park, Streamer, Huang, & Galinsky, 2013). All this implies that there is a link between dominance and expansiveness in body postures and body movements. Communicating dominance as well as building connections to followers are vital abilities for leaders and politicians. Skilled selfpresenters communicate their dominance by reassuring followers in noncompetitive contexts while threatening those who would jeopardize group stability (Stewart, Salter, & Mehu, 2009; Stewart, Waller, & Schubert, 2009). Consequently, speakers may not only present themselves differently according to their personality and their rhetorical skills but also according to situational factors such as their role in parliament. In the present study we selected brief video clips of politicians and mapped the body movements of the speakers onto animated stick-figures to control for appearance features. Then we asked people to judge these stimuli on dominance as well as on two additional basic social categories, namely trustworthiness and competence (Fiske, Cuddy, & Glick, 2007; Oosterhof & Todorov, 2008). In line with previous research we examined to what degree expansiveness in body postures and in body motion is related to judgments of dominance. To track down the relative contributions of body motion and expansiveness, we also investigated the influence of the quantity of motion. Analyses of the relationship of perceived trustworthiness and perceived competence with nonverbal cues were exploratory, because there is no research upon which to derive clear hypotheses. Our measures were
http://dx.doi.org/10.1016/j.paid.2016.01.013 0191-8869/© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
102
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
simple because first impressions of speakers' body movements seem to be guided by simple and salient cues (Koppensteiner, 2013). This study pursued several aims. In accordance with previous research we intended to show that people's impressions of a speaker's dominance are guided by expansive body postures and expansive body movements. In addition, we investigated whether the quantity of body motion also influences ratings of dominance. We also explored whether ratings of trustworthiness and competence, which have been shown to be important qualities in judgments of faces, are related to our “nonverbal measures”. Males tend to challenge someone else's status by dominance contests (Mazur & Booth, 1998). Moreover, male speakers appear to display motion behaviors that lead to higher ratings of extraversion than the motion behaviors displayed by female speakers (Koppensteiner & Grammer, 2011) and perceived extraversion of speakers' motion behaviors is positively related to perceived dominance (Koppensteiner, Stephan, & Jäschke, 2015). Thus, we expected male speakers to show more dominance displays than female speakers. Finally, we examined whether speakers (i.e., their stick-figure animations) from the opposition and speakers from the government are judged differently on dominance, trustworthiness and competence and whether such differences show an interaction with gender. 2. Method 2.1. Participants At locations throughout the University we recruited 60 participants (33 females and 27 males; age M = 24.2 years, SD = 4.1) to take part in our rating experiment (see also Koppensteiner et al., 2015). Participants received a financial compensation of €5. 2.2. Stimulus preparation Using a random number generator we selected 60 speeches from parliamentary sessions (29.11.2012, 30.11.2012, 14.12.2012) of the German Parliament. Deviations from random selection were necessary to reach equal numbers of male and female speakers (i.e., 30 different males and females) and nearly equally sized groups representing the parties. From each of the selected speeches we extracted brief, randomly chosen video segments with an average length of 15 s. Thirty-two speakers belonged to opposition parties (i.e., SPD, Bündnis 90/Die Grünen, Die Linke) and 28 speakers belonged to the government (i.e., CDU, FDP). To encode behavior we used the program SpeechAnalyzer, which runs through a movie frame by frame. In the first frame of each video clip, so called landmarks were positioned on the speaker's forehead, the hollow of the throat between the collar bones, ears, shoulders, elbows, hands, a spot in the middle of the body near the navel, and at the corners of the lectern (see also Koppensteiner, 2013; Koppensteiner & Grammer, 2010). Shifts in the positions of these body regions were automatically tracked by software routines based on optical flow (e.g., landmark of left shoulder in frame one was moved to position of left shoulder in frame two). As these software routines are prone to error, landmark positions often had to be corrected doing drag and drop operations with the computer mouse. This procedure of behavior encoding on the basis of landmark shifts yielded a time series of two dimensional marker positions that were used to create stick-figure animations (Fig. 1) representing the speakers' body movements. We only used every third frame in the encoding process; linear interpolation was used to fill in missing frames. 2.3. Procedure Participants were brought to our laboratory and asked to rate stickfigure animations of speakers. They received instructions on how to use the rating program and performed the rating tasks on their own (i.e., no
Fig. 1. Landmarks that have been used to extract horizontal and vertical movements of different body parts. Variation in distances between landmarks yields time series of amplitudes that were used to create measures of expansiveness.
experimenter present) using a computer-controlled interface. Stickfigure video clips were presented on the left-hand side of the user interface; rating scales that were named dominant, trustworthy, and competent were displayed on the right hand side. Participants completed their ratings by dragging a track bar control to the right pole (i.e. named strongly disagree) or the left pole (i.e., named strongly agree) of the rating scales using a computer mouse. The scales were divided into 200 subunits, with − 100 being the minimum value and + 100 being the maximum value (i.e., similar to a visual analog scale). Time to complete the ratings was unrestricted. Each participant rated a subset of 20 randomly selected stick-figure animations (i.e., each participant rated her/his own set of stimuli, which were presented in randomized order). All video clips were presented without sound. 2.4. Analysis Coordinate data obtained during the behavior encoding was used for analyses of body motion. Previous studies have shown that the horizontal and vertical components of body motion affect impression formation differently (Koppensteiner, 2013; Koppensteiner & Grammer, 2010). For this reason we decomposed the speakers' body movements into horizontal (x) and vertical (y) components by calculating distances between the coordinates of different landmarks. We determined the magnitude of vertical hand movements: the sum of lectern(y) – right hand(y) and lectern(y) – left hand(y), and vertical body movements: throat(y) – lectern(y). In a second step we determined horizontal hand movements: throat(x) – right hand(x) and left hand(x) – throat(x) and horizontal body movements: throat(x) – origin(x). This gave four time series of changing landmark distances from which we extracted the amplitudes between successive local maxima and local minima. The sum of all vertical amplitudes served as an estimate of the overall vertical distances a moving body produced (i.e., overall vertical expansiveness in motion) while the sum of all horizontal amplitudes served as an estimate of the overall horizontal distances a body produced (i.e., overall horizontal expansiveness in motion). The sum of the number of local minima and maxima, on the other hand, served as an estimate of the vertical quantity of motion and as an estimate of
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
103
Table 2 Bivariate correlations between trait ratings and nonverbal measures. Trait ratings Measures
Dominance
Trustworthiness
Competence
Vertical expansiveness
Vertical amplitudes
.48⁎⁎⁎ [.26, .64] .29⁎ [.03, .52] .68⁎⁎⁎
−.25⁎ [−.47, .02] −.17 [−.42, .07] −.34⁎⁎
Horizontal amplitudes
[.53, .77] .52⁎⁎⁎
[−.52, −.12] −.33⁎⁎
[.32, .69] .54⁎⁎⁎ [.22, .72] .55⁎⁎⁎ [.36, .68] .70⁎⁎⁎
[−.55, −.03] −.33⁎⁎ [−.55, −.03] −.45⁎⁎⁎ [−.61, −.25] −.38⁎⁎
[.55, .80]
[−.57, −.13]
.03 [−.21, .29] −.02 [−.27, .22] −.06 [−.28, .20] −.12 [−.39, .16] −.15 [−.39, .07] −.19 [−.39, .03] −.09 [−.34, .17]
Horizontal expansiveness
Vertical quantity of motion Fig. 2. Position shifts of landmarks changes distances between landmarks (see Fig. 1). This gives curves with local minima and local maxima (e.g., hand going up and down). Distances between successive local maxima (Max) and local minima (Min) yields amplitudes. Summing the amplitudes of such a curve (or sequence of behaviors) provides a measure of a body movements' expansiveness. Dashed lines give three examples of amplitudes between successive local extremes.
the horizontal quantity of motion without including the influence of motion amplitudes (see Figs. 1 and 2). In addition to this we also used single positions of the speakers' bodies in our analyses. The maximal horizontal distance between the speakers' left and right hand served as a measure of maximal horizontal expansiveness. Similarly, the sum of the maximal vertical distances between the speakers' hands and the lectern and between the speakers' throats and the lectern served as measures of vertical expansiveness. In summary, we created three types of measures: (1) a measure that captures both expansiveness and the quantity of motion, (2) a measure that only captures the quantity of motion, and (3) a measure that captures expansiveness without including the quantity of motion. Amplitudes and distance measures were corrected for variation in body height by dividing them by the maximum vertical distance between the lectern and the speakers' foreheads (i.e., stick-figure height). Each stimulus was rated by a range from 18 to 22 participants (M = 20). For the statistical analyses we averaged these ratings. Motion variables tend to be highly interdependent, which affects the interpretation of regression coefficients of multiple regressions. For this reason we used simple bivariate correlations (with bootstrapped confidence intervals) to examine the relationships between the variables. We also divided the sample of stimuli into four categories: male and female speakers belonging to the opposition and male and female speakers belonging to the government. The differences between these groups in trait ratings and motion measures were assessed by calculating means with bootstrapped confidence intervals. As we expected interactions between speaker gender and role in Parliament (opposition or government), we also calculated a series of two-factorial Type III ANOVAs.
Horizontal quantity of motion Overall motion
Note. Values in brackets are 95% confidence intervals based on bootstrap samples with 9999 replicates. Expansiveness measures are maximal distances between hands (horizontal) or maximal distances between hands plus throat and lectern (vertical). Amplitudes are sum of distances produced by a moving body. Quantity of motion is the sum of the minima and the maxima a moving body produces (i.e., measure is independent of amplitude). Overall motion is a composite measure of the sum of vertical and horizontal amplitudes. N = 60. ⁎ p b .05. ⁎⁎ p ≤ .01. ⁎⁎⁎ p ≤ .001.
All statistical analyses were carried out in the program R (R Core Team, 2013).
3. Results Correlations between the indices of body motion and expansiveness (Table 1) revealed a wide range of interdependencies. For instance, horizontal and vertical measures of motion were strongly linked. Specifically, speakers that showed a great a deal of expansive movements along the vertical axis (up and down movements) also showed expansive movements (e.g., hands moving from the left to the right) along the horizontal axis. A similar relationship was found between horizontal and vertical quantity of motion. Measures of static expansiveness (i.e., maximal distance between hands and hands held up high) were strongly correlated to measures of expansiveness of motion (i.e., amplitudes). Consequently, speakers who produced expansive body postures also tended to produce expansive movements and speakers who produced expansive movements also showed a high quantity of motion. Because our nonverbal measures showed strong interdependencies (Table 1) we also created a
Table 1 Bivariate correlations between measures of motion and expansiveness.* Measures
Vert. expan.
Vertical expansiveness Horizontal expansiveness
– .21 [−.02, .42] .71⁎⁎⁎ [.54, .83] .37⁎⁎ [.16, .57] .24 [−.02, .46] .36⁎⁎
Vertical amplitudes Horizontal amplitudes Vertical quantity of motion Horizontal quantity of motion
[.03, .61]
Hori. expan.
Vert. ampl.
Hori. ampl.
Vert. quant.
– .13 [−.09, .35] .42⁎⁎⁎ [.23, .61]
– .55⁎⁎⁎ [.38, .70]
–
−.11 [−.32, .13] .00 [−.23, .22]
.61⁎⁎⁎ [.45, .71]
.49⁎⁎⁎ [.28, .64]
–
.66⁎⁎⁎ [.48, .79]
0.46⁎⁎⁎ [.19, .66]
.65⁎⁎⁎ [.50, .75]
Note. Values in brackets are 95% confidence intervals based on bootstrap samples with 9999 replicates. Expansiveness is a measure based on a single posture comprising the maximal distances between hands (horizontal) or maximal distances between body plus hands and lectern (vertical). Amplitudes are sum of distances covered by a moving body. Quantities of motion are sum of minima and maxima a moving body produces (i.e., measure is independent of amplitude). N = 60. ⁎⁎ p ≤ .01. ⁎⁎⁎ p ≤ .001.
104
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
Table 3 Interactions between the speakers' role in parliament (opposition or government) and the speakers' gender on the basis of Type III ANOVA's. Gender Ratings or motion measures Dominance Trustworth. Competence Overall motion
F(1,56) .02 1.26 .09 .57
Opposition or government η2p .00 .02 .00 .01
F(1,56) .18 .16 .70 .73
Interaction
η2p
F(1,56)
η2p
.00 .00 .01 .00
5.13⁎ 4.00⁎
.08 .07 .01 .01
.30 .38
Note.; Trustworth. = trustworthiness; Horz. = horizontal; Vert. = vertical; η2p = partial eta-squared. Overall motion is a composite of horizontal and vertical amplitudes. ⁎ p b .05.
with the ratings of males and females belonging to the government, this result was impressive (confidence intervals show no overlap). Findings for trustworthiness were similar to findings of dominance. Again, the ratings of males belonging to the opposition stood out (Table 2). They received the lowest ratings on trustworthiness of all groups. The composite measure of overall motion provided no significant interaction between gender and party position (Table 3). Means and their confidence intervals showed that overall motion was highest for males from the opposition (Table 4). This was in line with this group's ratings on dominance. However, the differences in trait ratings were not convincingly reflected in our motion measure. Therefore, people's ratings must be guided by additional nonverbal cues. 4. Discussion
composite measure of body motion by adding up vertical and horizontal amplitudes. Stick-figures rated high on dominance tended to be rated low on trustworthiness (r(58) = −.57), those rated high on trustworthiness were also rated high on competence (r(58) = .76), and no relationship (r(58) = −.11) was found between dominance and competence (see also Koppensteiner et al., 2015). Thus, in part dominance and trustworthiness were mutually exclusive categories. Indices of body motion, expansiveness and our composite measure of motion showed profound relationships with ratings of dominance and trustworthiness. Perceived competence did not yield a meaningful relationship with any of our measures (Table 2). Stick-figures adopting an expansive body posture, displaying expansiveness in overall motion behavior and a great deal of body movements were rated high on dominance. These findings indicate that people not only ascribe dominance to static cues of expansiveness but also to motion cues. Low to moderate interrelations between expansiveness of body postures and the quantity of motion – a motion measure independent of amplitude – further supported such an interpretation (see Table 1). The negative relationship between trait ratings of dominance and trustworthiness was also reflected in correlations between motion measures and trustworthiness. There were negative relationships throughout (Table 2). Overall, this can be condensed into a simple formula, namely that high ratings of trustworthiness were linked to low activity in body motion. In the second step we analyzed whether the stick-figures were judged differently on dominance, trustworthiness, and competence depending on the speakers' gender and their party status (i.e., opposition or government). ANOVAs yielded significant interactions between gender and the politicians' role in parliament (i.e., opposition or government) for perceived dominance and perceived trustworthiness (Table 3). Inspection of the mean ratings for dominance provided detailed insights into the differences between the four groups we investigated (Table 4 and Fig. 3). Perceived dominance was highest for males belonging to the opposition parties. In particular, when compared
To display dominance humans and animals make broad gestures or adopt expansive body postures (e.g., De Waal, 2007; Mehrabian, 1972). We investigated whether people perceive such dominance cues in the body motion of politicians giving speeches. For this purpose we translated short excerpts from different speeches into stick-figure animations and extracted the “maximum expansiveness” (i.e., expansiveness of a single posture), the overall “expansiveness of motion” and “the quantity of motion”. All three measures were strongly interrelated and good predictors of perceived stick-figure dominance. This indicates that speakers who adopt expansive postures also produce such expansive postures frequently. Moreover, expansive postures lead to impressions of dominance and perceiving such expansive postures frequently may even reinforce such impressions. In contrast to dominance, trustworthiness was associated with low expansiveness in motion and a low number of movements. Such a negative relationship between dominance and trustworthiness was also reflected in the trait ratings. Therefore, displays of high dominance have a negative impact on impressions of trustworthiness. Taking a closer look at the results revealed that our “nonverbal measures” were not as strongly related to trustworthiness than to dominance. In other words, people perceived dominance more easily than trustworthiness in the nonverbal cues we extracted. This is not surprising because dominance displays are not subtle in general and intended to attract attention. Stick-figures perceived as trustworthy were also perceived as competent, yet we found no relationship between our “nonverbal measures” and competence. Despite being one of the basic social categories in judgments of politicians (Fiske et al., 2007) competence was not conveyed by the cues we extracted. It is conceivable that competence is associated with motion patterns our simple measures failed to capture. However, it is also possible that people, although able to perceive competence in facial cues, are unable to perceive competence in body motion. First impressions may represent a cognitive adaptation that helps to decide whether to approach or to avoid someone (Oosterhof &
Table 4 Mean trait ratings and means of motion measures grouped by the speaker's role in parliament (government or opposition) and the speakers' gender. Government
Opposition
Ratings or motion M.
Female
Male
Female
Male
Dominance
−3.46 (26.66) [−18.03, 10.66] −.56 (14.60) [−8.29, 7.56] 9.6 (12.47) [2.56, 15.98] 13.35 (5.61) [11.18, 17.87]
−4.75 (21.31) [−14.45, 5.74] 5.90 (13.73) [−1.04, 11.99] 11.13 (13.63) [4.94, 17.78] 14.56 (5.52) [12.05, 17.23]
.43 (29.19) [−11.48, 14.85] −2.79 (15.85) [−10.07, 3.98] 5.39 (13.74) [−.50, 11.90] 14.06 (6.60) [11.50, 17.56]
28.39 (19.82) [13.48, 35.61] −12.04 (15.71) [−19.31, −3.81] 3.08 (13.65) [−3.64, 9.83] 17.83 (3.66) [15.92, 19.62]
Trustworth. Competence Overall motion
Note. Values in parenthesis are standard deviations. Values in brackets are bootstrapped 95% confidence intervals on the basis of 9999 replicates. Trustworth. = trustworthiness; Overall motion is a composite of horizontal and vertical amplitudes. Rating scales ranged from −100 to +100. Government/male n = 16; Government/female n = 12; Opposition/male n = 14; Opposition/female n = 18.
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
105
Fig. 3. Interaction plots showing mean ratings of dominance and trustworthiness grouped by party status and gender (rating scales ranged from −100 to +100).
Todorov, 2008). The results obtained may be interpreted in a similar way and the ratings may be classified as “positive versus negative”. However, previous studies using stick-figure stimuli show that the body movements of speakers allow more than such simple categorizations (e.g., Koppensteiner et al., 2015; Koppensteiner, 2013). In the present study we only extracted conspicuous dominance cues, which may have a predominant influence on first impressions, but more in depth social evaluations may be based on more complex motion cues. Also, it is not clear whether social categories such as those applied here or the Big Five personality dimensions sufficiently capture the information people perceive in motion cues (see also Thoresen et al., 2012). To clarify this follow-up studies are needed. Politicians change their rhetoric depending on the situation or their role in Parliament (Pancer, Hunsberger, Pratt, Boisvert, & Roth, 1992; Tetlock, 1981). This study is a first step toward investigating whether such flexibility in self-presentation also becomes apparent in body motion. Although the results obtained need to be backed by additional studies, they support such an assumption regarding male speakers in the opposition. Stick-figures representing male politicians in the opposition were perceived as more dominant but less trustworthy than stick-figures representing politicians of the government. However, to provide more clarity future work needs to elaborate on this and test whether former opposition members change their nonverbal performance when they become members of the government. Although a previous study using stick-figure stimuli revealed differences between male and female speakers for perceived extraversion and emotional stability (Koppensteiner & Grammer, 2011), in the present study we only found noteworthy gender differences when including party status. The results obtained suggest that dominance displays, that negatively affect perceptions of trustworthiness, are predominantly shown by male speakers expected to control and criticize the work of those who are in power. Male and female speakers of the government,
on the other hand, may intend to communicate integrity and on the level of body motion this might look similar for both genders. 4.1. Conclusions Snap judgments are not only guided by the outward appearance and facial expressions but also by salient and simple motion cues. We show that speakers displaying expansive movements and a great deal of body activity are rated high on dominance and low on trustworthiness. In addition, females and male members of the ruling parties appear to communicate their views in a different way than males from the opposition and such differences appear to be discernable in body motion. This hints that politicians not only express their positions with words but also with their bodies. Acknowledgments This research was funded by the Austrian Science Fund (FWF): P 25262-G16. We thank the reviewers for their constructive feedback. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.paid.2016.01.013. References Ambady, N., Bernieri, F. J., & Richeson, J. A. (2000). Toward a histology of social behavior: Judgmental accuracy from thin slices of the behavioral stream. Advances in Experimental Social Psychology, 32, 201–271. http://dx.doi.org/10.1016/S00652601(00)80006-4. Antonakis, J., & Dalgas, O. (2009). Predicting elections: Child's play! Science, 323, 11831183. http://dx.doi.org/10.1126/science.1167748.
106
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
Atkinson, A. P., Tunstall, M. L., & Dittrich, W. H. (2007). Evidence for distinct contributions of form and motion information to the recognition of emotions from body gestures. Cognition, 104, 59–72. http://dx.doi.org/10.1016/j.cognition.2006.05.005. Borkenau, P., Mauer, N., Riemann, R., Spinath, F. M., & Angleitner, A. (2004). Thin slices of behavior as cues of personality and intelligence. Journal of Personality and Social Psychology, 86, 599–614. http://dx.doi.org/10.1037/0022-3514.86.4.599. Carney, D. R., Cuddy, A. J. C., & Yap, A. J. (2010). Power posing: Brief nonverbal displays affect neuroendocrine levels and risk tolerance. Psychological Science, 21, 1363–1368. http://dx.doi.org/10.1177/0956797610383437. Carney, D., Hall, J., & LeBeau, L. (2005). Beliefs about the nonverbal expression of social power. Journal of Nonverbal Behavior, 29, 105–123. http://dx.doi.org/10.1007/ s10919-005-2743-z. Chen, F. F., Jing, Y., & Lee, J. M. (2014). The looks of a leader: Competent and trustworthy, but not dominant. Journal of Experimental Social Psychology, 51, 27–33. http://dx.doi. org/10.1016/j.jesp.2013.10.008. Clarke, T. J., Bradshaw, M. F., Field, D. T., Hampson, S. E., & Rose, D. (2005). The perception of emotion from body movement in point-light displays of interpersonal dialogue. Perception, 34, 1171–1180. http://dx.doi.org/10.1068/p5203. De Waal, F. (2007). Chimpanzee politics: Power and sex among apes. Baltimore: JHU Press. Eisenberg, P., & Reichline, P. B. (1939). Judging expressive movement: II. Judgments of dominance-feeling from motion pictures of gait. The Journal of Social Psychology, 10, 345–357. http://dx.doi.org/10.1080/00224545.1939.9713373. Fiske, S. T., Cuddy, A. J., & Glick, P. (2007). Universal dimensions of social cognition: Warmth and competence. Trends in Cognitive Sciences, 11, 77–83. http://dx.doi.org/ 10.1016/j.tics.2006.11.005. Huang, L., Galinsky, A. D., Gruenfeld, D. H., & Guillory, L. E. (2011). Powerful postures versus powerful roles. Which is the proximate correlate of thought and behavior? Psychological Science, 22, 95–102. http://dx.doi.org/10.1177/0956797610391912. Hugill, N., Fink, B., Neave, N., Besson, A., & Bunse, L. (2011). Women's perception of men's sensation seeking propensity from their dance movements. Personality and Individual Differences, 51, 483–487. http://dx.doi.org/10.1016/j.paid.2011.05.002. Koppensteiner, M. (2013). Motion cues that make an impression: Predicting perceived personality by minimal motion information. Journal of Experimental Social Psychology, 49, 1137–1143. http://dx.doi.org/10.1016/j.jesp.2013.08.002. Koppensteiner, M., & Grammer, K. (2010). Motion patterns in political speech and their influence on personality ratings. Journal of Research in Personality, 44, 374–379. http://dx.doi.org/10.1016/j.jrp.2010.04.002. Koppensteiner, M., & Grammer, K. (2011). Body movements of male and female speakers and their influence on perceptions of personality. Personality and Individual Differences, 51, 743–747. http://dx.doi.org/10.1016/j.paid.2011.06.014. Koppensteiner, M., Stephan, P., & Jäschke, J. P. M. (2015). From body motion to cheers: Speakers' body movements as predictors of applause. Personality and Individual Differences, 74, 182–185. http://dx.doi.org/10.1016/j.paid.2014.10.019. Kramer, R. S., Arend, I., & Ward, R. (2010). Perceived health from biological motion predicts voting behaviour. The Quarterly Journal of Experimental Psychology, 63, 625–632. http://dx.doi.org/10.1080/17470210903490977. Little, A. C., Roberts, S. C., Jones, B. C., & DeBruine, L. M. (2012). The perception of attractiveness and trustworthiness in male faces affects hypothetical voting decisions
differently in wartime and peacetime scenarios. The Quarterly Journal of Experimental Psychology, 65, 2018–2032. http://dx.doi.org/10.1080/17470218.2012. 677048. Mazur, A., & Booth, A. (1998). Testosterone and dominance in men. Behavioral and Brain Sciences, 21, 353–363. Mehrabian, A. (1972). Nonverbal communication. Chicago: Aldine. Olivola, C. Y., & Todorov, A. (2010). Elected in 100 milliseconds: Appearance-based trait inferences and voting. Journal of Nonverbal Behavior, 34, 83–110. http://dx.doi.org/ 10.1007/s10919-009-0082-1. Oosterhof, N. N., & Todorov, A. (2008). The functional basis of face evaluation. Proceedings of the National Academy of Sciences, 105, 11087–11092. http://dx.doi.org/10.1073/ pnas.0805664105. Pancer, S. M., Hunsberger, B., Pratt, M. W., Boisvert, S., & Roth, D. (1992). Political roles and the complexity of political rhetoric. Political Psychology, 13, 31–43. Park, L. E., Streamer, L., Huang, L., & Galinsky, A. D. (2013). Stand tall, but don't put your feet up: Universal and culturally-specific effects of expansive postures on power. Journal of Experimental Social Psychology, 49, 965–971. http://dx.doi.org/10.1016/j. jesp.2013.06.001. Pollick, F. E., Paterson, H. M., Bruderlin, A., & Sanford, A. J. (2001). Perceiving affect from arm movement. Cognition, 82, B51–B61. http://dx.doi.org/10.1016/S0010-0277(01)00147-0. Poutvaara, P., Jordahl, H., & Berggren, N. (2009). Faces of politicians: Babyfacedness predicts inferred competence but not electoral success. Journal of Experimental Social Psychology, 45, 1132–1135. http://dx.doi.org/10.1016/j.jesp.2009.06.007. R Core Team (2013). R: A language and environment for statistical computing. Rule, N. O., & Ambady, N. (2011). Face and fortune: Inferences of personality from managing partners' faces predict their law firms' financial success. The Leadership Quarterly, 22, 690–696. http://dx.doi.org/10.1016/j.leaqua.2011.05.009. Stewart, P. A., Salter, F. K., & Mehu, M. (2009a). Taking leaders at face value: Ethology and the analysis of televised leader displays. Politics and the Life Sciences, 28, 48–74. http:// dx.doi.org/10.2990/28_1_48. Stewart, P. A., Waller, B. M., & Schubert, J. N. (2009b). Presidential speechmaking style: Emotional response to micro-expressions of facial affect. Motivation and Emotion, 33, 125–135. http://dx.doi.org/10.1007/s11031-009-9129-1. Tetlock, P. E. (1981). Pre-to postelection shifts in presidential rhetoric: Impression management or cognitive adjustment. Journal of Personality and Social Psychology, 41, 207–212. http://dx.doi.org/10.1037/0022-3514.41.2.207. Thoresen, J. C., Vuong, Q. C., & Atkinson, A. P. (2012). First impressions: Gait cues drive reliable trait judgements. Cognition, 124, 261–271. http://dx.doi.org/10.1016/j. cognition.2012.05.018. Tiedens, L. Z., & Fragale, A. R. (2003). Power moves: Complementarity in dominant and submissive nonverbal behavior. Journal of Personality and Social Psychology, 84, 558–568. http://dx.doi.org/10.1037/0022-3514.84.3.558. Zebrowitz, L. A., & Montepare, J. M. (2008). Social psychological face perception: Why appearance matters. Social and Personality Psychology Compass, 2, 1497–1517. http://dx. doi.org/10.1111/j.1751-9004.2008.00109.x.
Contents lists available at ScienceDirect
Personality and Individual Differences journal homepage: www.elsevier.com/locate/paid
Moving speeches: Dominance, trustworthiness and competence in body motion Markus Koppensteiner ⁎, Pia Stephan, Johannes Paul Michael Jäschke Department of Anthropology, University of Vienna, Althanstraße 14, 1090, Vienna, Austria
a r t i c l e
i n f o
Article history: Received 11 August 2015 Received in revised form 6 January 2016 Accepted 8 January 2016 Available online 18 January 2016 Keywords: Nonverbal communication Motion cues Social cognition Politics Ethology Impression formation
a b s t r a c t People read dominance, trustworthiness and competence into the faces of politicians but do they also perceive such social qualities in other nonverbal cues? We transferred the body movements of politicians giving a speech onto animated stick-figures and presented these stimuli to participants in a rating-experiment. Analyses revealed single body postures of maximal expansiveness as strong predictors of perceived dominance. Also, stick-figures producing expansive movements as well as a great number of movements throughout the encoded sequences were judged high on dominance and low on trustworthiness. In a second step we divided our sample into speakers from the opposition parties and speakers that were part of the government as well as into male and female speakers. Male speakers from the opposition were rated higher on dominance but lower on trustworthiness than speakers from all other groups. In conclusion, people use simple cues to make equally simple social categorizations. Moreover, the party status of male politicians seems to become visible in their body motion. © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
1. Introduction Nonverbal cues affect impression formation (Ambady, Bernieri, & Richeson, 2000; Borkenau, Mauer, Riemann, Spinath, & Angleitner, 2004) and decision making in the public arena (Rule & Ambady, 2011; Zebrowitz & Montepare, 2008). For instance, attributions of dominance, trustworthiness, competence and other personality traits to specific facial features of political candidates can be reliable predictors of hypothetical and actual election outcomes (Antonakis & Dalgas, 2009; Chen, Jing, & Lee, 2014; Little, Roberts, Jones, & DeBruine, 2012; Olivola & Todorov, 2010; Oosterhof & Todorov, 2008; Poutvaara, Jordahl, & Berggren, 2009). Apart from facial and other nonverbal cues people also read socially relevant information from and into body motion. They recognize emotions in arm movements (Pollick, Paterson, Bruderlin, & Sanford, 2001), in whole body gestures (Atkinson, Tunstall, & Dittrich, 2007), and in movements displayed during interpersonal dialog (Clarke, Bradshaw, Field, Hampson, & Rose, 2005). Moreover, they perceive personality traits in dance movements (Hugill, Fink, Neave, Besson, & Bunse, 2011) and in walking behaviors (Thoresen, Vuong, & Atkinson, 2012), and health related cues and personality in the body movements of politicians giving a speech (Koppensteiner, 2013; Koppensteiner & Grammer, 2010; Kramer, Arend, & Ward, 2010). Humans and animals appear to use expansive body postures, expressive body movements, and broad gestures to display power and
⁎ Corresponding author. E-mail address: [email protected] (M. Koppensteiner).
dominance (Carney, Hall, & LeBeau, 2005; De Waal, 2007; Eisenberg & Reichline, 1939; Mehrabian, 1972; Tiedens & Fragale, 2003). Also, people adopting open and expansive postures (i.e., power posing) have an enhanced sense of power (Carney, Cuddy, & Yap, 2010; Huang, Galinsky, Gruenfeld, & Guillory, 2011; Park, Streamer, Huang, & Galinsky, 2013). All this implies that there is a link between dominance and expansiveness in body postures and body movements. Communicating dominance as well as building connections to followers are vital abilities for leaders and politicians. Skilled selfpresenters communicate their dominance by reassuring followers in noncompetitive contexts while threatening those who would jeopardize group stability (Stewart, Salter, & Mehu, 2009; Stewart, Waller, & Schubert, 2009). Consequently, speakers may not only present themselves differently according to their personality and their rhetorical skills but also according to situational factors such as their role in parliament. In the present study we selected brief video clips of politicians and mapped the body movements of the speakers onto animated stick-figures to control for appearance features. Then we asked people to judge these stimuli on dominance as well as on two additional basic social categories, namely trustworthiness and competence (Fiske, Cuddy, & Glick, 2007; Oosterhof & Todorov, 2008). In line with previous research we examined to what degree expansiveness in body postures and in body motion is related to judgments of dominance. To track down the relative contributions of body motion and expansiveness, we also investigated the influence of the quantity of motion. Analyses of the relationship of perceived trustworthiness and perceived competence with nonverbal cues were exploratory, because there is no research upon which to derive clear hypotheses. Our measures were
http://dx.doi.org/10.1016/j.paid.2016.01.013 0191-8869/© 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
102
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
simple because first impressions of speakers' body movements seem to be guided by simple and salient cues (Koppensteiner, 2013). This study pursued several aims. In accordance with previous research we intended to show that people's impressions of a speaker's dominance are guided by expansive body postures and expansive body movements. In addition, we investigated whether the quantity of body motion also influences ratings of dominance. We also explored whether ratings of trustworthiness and competence, which have been shown to be important qualities in judgments of faces, are related to our “nonverbal measures”. Males tend to challenge someone else's status by dominance contests (Mazur & Booth, 1998). Moreover, male speakers appear to display motion behaviors that lead to higher ratings of extraversion than the motion behaviors displayed by female speakers (Koppensteiner & Grammer, 2011) and perceived extraversion of speakers' motion behaviors is positively related to perceived dominance (Koppensteiner, Stephan, & Jäschke, 2015). Thus, we expected male speakers to show more dominance displays than female speakers. Finally, we examined whether speakers (i.e., their stick-figure animations) from the opposition and speakers from the government are judged differently on dominance, trustworthiness and competence and whether such differences show an interaction with gender. 2. Method 2.1. Participants At locations throughout the University we recruited 60 participants (33 females and 27 males; age M = 24.2 years, SD = 4.1) to take part in our rating experiment (see also Koppensteiner et al., 2015). Participants received a financial compensation of €5. 2.2. Stimulus preparation Using a random number generator we selected 60 speeches from parliamentary sessions (29.11.2012, 30.11.2012, 14.12.2012) of the German Parliament. Deviations from random selection were necessary to reach equal numbers of male and female speakers (i.e., 30 different males and females) and nearly equally sized groups representing the parties. From each of the selected speeches we extracted brief, randomly chosen video segments with an average length of 15 s. Thirty-two speakers belonged to opposition parties (i.e., SPD, Bündnis 90/Die Grünen, Die Linke) and 28 speakers belonged to the government (i.e., CDU, FDP). To encode behavior we used the program SpeechAnalyzer, which runs through a movie frame by frame. In the first frame of each video clip, so called landmarks were positioned on the speaker's forehead, the hollow of the throat between the collar bones, ears, shoulders, elbows, hands, a spot in the middle of the body near the navel, and at the corners of the lectern (see also Koppensteiner, 2013; Koppensteiner & Grammer, 2010). Shifts in the positions of these body regions were automatically tracked by software routines based on optical flow (e.g., landmark of left shoulder in frame one was moved to position of left shoulder in frame two). As these software routines are prone to error, landmark positions often had to be corrected doing drag and drop operations with the computer mouse. This procedure of behavior encoding on the basis of landmark shifts yielded a time series of two dimensional marker positions that were used to create stick-figure animations (Fig. 1) representing the speakers' body movements. We only used every third frame in the encoding process; linear interpolation was used to fill in missing frames. 2.3. Procedure Participants were brought to our laboratory and asked to rate stickfigure animations of speakers. They received instructions on how to use the rating program and performed the rating tasks on their own (i.e., no
Fig. 1. Landmarks that have been used to extract horizontal and vertical movements of different body parts. Variation in distances between landmarks yields time series of amplitudes that were used to create measures of expansiveness.
experimenter present) using a computer-controlled interface. Stickfigure video clips were presented on the left-hand side of the user interface; rating scales that were named dominant, trustworthy, and competent were displayed on the right hand side. Participants completed their ratings by dragging a track bar control to the right pole (i.e. named strongly disagree) or the left pole (i.e., named strongly agree) of the rating scales using a computer mouse. The scales were divided into 200 subunits, with − 100 being the minimum value and + 100 being the maximum value (i.e., similar to a visual analog scale). Time to complete the ratings was unrestricted. Each participant rated a subset of 20 randomly selected stick-figure animations (i.e., each participant rated her/his own set of stimuli, which were presented in randomized order). All video clips were presented without sound. 2.4. Analysis Coordinate data obtained during the behavior encoding was used for analyses of body motion. Previous studies have shown that the horizontal and vertical components of body motion affect impression formation differently (Koppensteiner, 2013; Koppensteiner & Grammer, 2010). For this reason we decomposed the speakers' body movements into horizontal (x) and vertical (y) components by calculating distances between the coordinates of different landmarks. We determined the magnitude of vertical hand movements: the sum of lectern(y) – right hand(y) and lectern(y) – left hand(y), and vertical body movements: throat(y) – lectern(y). In a second step we determined horizontal hand movements: throat(x) – right hand(x) and left hand(x) – throat(x) and horizontal body movements: throat(x) – origin(x). This gave four time series of changing landmark distances from which we extracted the amplitudes between successive local maxima and local minima. The sum of all vertical amplitudes served as an estimate of the overall vertical distances a moving body produced (i.e., overall vertical expansiveness in motion) while the sum of all horizontal amplitudes served as an estimate of the overall horizontal distances a body produced (i.e., overall horizontal expansiveness in motion). The sum of the number of local minima and maxima, on the other hand, served as an estimate of the vertical quantity of motion and as an estimate of
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
103
Table 2 Bivariate correlations between trait ratings and nonverbal measures. Trait ratings Measures
Dominance
Trustworthiness
Competence
Vertical expansiveness
Vertical amplitudes
.48⁎⁎⁎ [.26, .64] .29⁎ [.03, .52] .68⁎⁎⁎
−.25⁎ [−.47, .02] −.17 [−.42, .07] −.34⁎⁎
Horizontal amplitudes
[.53, .77] .52⁎⁎⁎
[−.52, −.12] −.33⁎⁎
[.32, .69] .54⁎⁎⁎ [.22, .72] .55⁎⁎⁎ [.36, .68] .70⁎⁎⁎
[−.55, −.03] −.33⁎⁎ [−.55, −.03] −.45⁎⁎⁎ [−.61, −.25] −.38⁎⁎
[.55, .80]
[−.57, −.13]
.03 [−.21, .29] −.02 [−.27, .22] −.06 [−.28, .20] −.12 [−.39, .16] −.15 [−.39, .07] −.19 [−.39, .03] −.09 [−.34, .17]
Horizontal expansiveness
Vertical quantity of motion Fig. 2. Position shifts of landmarks changes distances between landmarks (see Fig. 1). This gives curves with local minima and local maxima (e.g., hand going up and down). Distances between successive local maxima (Max) and local minima (Min) yields amplitudes. Summing the amplitudes of such a curve (or sequence of behaviors) provides a measure of a body movements' expansiveness. Dashed lines give three examples of amplitudes between successive local extremes.
the horizontal quantity of motion without including the influence of motion amplitudes (see Figs. 1 and 2). In addition to this we also used single positions of the speakers' bodies in our analyses. The maximal horizontal distance between the speakers' left and right hand served as a measure of maximal horizontal expansiveness. Similarly, the sum of the maximal vertical distances between the speakers' hands and the lectern and between the speakers' throats and the lectern served as measures of vertical expansiveness. In summary, we created three types of measures: (1) a measure that captures both expansiveness and the quantity of motion, (2) a measure that only captures the quantity of motion, and (3) a measure that captures expansiveness without including the quantity of motion. Amplitudes and distance measures were corrected for variation in body height by dividing them by the maximum vertical distance between the lectern and the speakers' foreheads (i.e., stick-figure height). Each stimulus was rated by a range from 18 to 22 participants (M = 20). For the statistical analyses we averaged these ratings. Motion variables tend to be highly interdependent, which affects the interpretation of regression coefficients of multiple regressions. For this reason we used simple bivariate correlations (with bootstrapped confidence intervals) to examine the relationships between the variables. We also divided the sample of stimuli into four categories: male and female speakers belonging to the opposition and male and female speakers belonging to the government. The differences between these groups in trait ratings and motion measures were assessed by calculating means with bootstrapped confidence intervals. As we expected interactions between speaker gender and role in Parliament (opposition or government), we also calculated a series of two-factorial Type III ANOVAs.
Horizontal quantity of motion Overall motion
Note. Values in brackets are 95% confidence intervals based on bootstrap samples with 9999 replicates. Expansiveness measures are maximal distances between hands (horizontal) or maximal distances between hands plus throat and lectern (vertical). Amplitudes are sum of distances produced by a moving body. Quantity of motion is the sum of the minima and the maxima a moving body produces (i.e., measure is independent of amplitude). Overall motion is a composite measure of the sum of vertical and horizontal amplitudes. N = 60. ⁎ p b .05. ⁎⁎ p ≤ .01. ⁎⁎⁎ p ≤ .001.
All statistical analyses were carried out in the program R (R Core Team, 2013).
3. Results Correlations between the indices of body motion and expansiveness (Table 1) revealed a wide range of interdependencies. For instance, horizontal and vertical measures of motion were strongly linked. Specifically, speakers that showed a great a deal of expansive movements along the vertical axis (up and down movements) also showed expansive movements (e.g., hands moving from the left to the right) along the horizontal axis. A similar relationship was found between horizontal and vertical quantity of motion. Measures of static expansiveness (i.e., maximal distance between hands and hands held up high) were strongly correlated to measures of expansiveness of motion (i.e., amplitudes). Consequently, speakers who produced expansive body postures also tended to produce expansive movements and speakers who produced expansive movements also showed a high quantity of motion. Because our nonverbal measures showed strong interdependencies (Table 1) we also created a
Table 1 Bivariate correlations between measures of motion and expansiveness.* Measures
Vert. expan.
Vertical expansiveness Horizontal expansiveness
– .21 [−.02, .42] .71⁎⁎⁎ [.54, .83] .37⁎⁎ [.16, .57] .24 [−.02, .46] .36⁎⁎
Vertical amplitudes Horizontal amplitudes Vertical quantity of motion Horizontal quantity of motion
[.03, .61]
Hori. expan.
Vert. ampl.
Hori. ampl.
Vert. quant.
– .13 [−.09, .35] .42⁎⁎⁎ [.23, .61]
– .55⁎⁎⁎ [.38, .70]
–
−.11 [−.32, .13] .00 [−.23, .22]
.61⁎⁎⁎ [.45, .71]
.49⁎⁎⁎ [.28, .64]
–
.66⁎⁎⁎ [.48, .79]
0.46⁎⁎⁎ [.19, .66]
.65⁎⁎⁎ [.50, .75]
Note. Values in brackets are 95% confidence intervals based on bootstrap samples with 9999 replicates. Expansiveness is a measure based on a single posture comprising the maximal distances between hands (horizontal) or maximal distances between body plus hands and lectern (vertical). Amplitudes are sum of distances covered by a moving body. Quantities of motion are sum of minima and maxima a moving body produces (i.e., measure is independent of amplitude). N = 60. ⁎⁎ p ≤ .01. ⁎⁎⁎ p ≤ .001.
104
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
Table 3 Interactions between the speakers' role in parliament (opposition or government) and the speakers' gender on the basis of Type III ANOVA's. Gender Ratings or motion measures Dominance Trustworth. Competence Overall motion
F(1,56) .02 1.26 .09 .57
Opposition or government η2p .00 .02 .00 .01
F(1,56) .18 .16 .70 .73
Interaction
η2p
F(1,56)
η2p
.00 .00 .01 .00
5.13⁎ 4.00⁎
.08 .07 .01 .01
.30 .38
Note.; Trustworth. = trustworthiness; Horz. = horizontal; Vert. = vertical; η2p = partial eta-squared. Overall motion is a composite of horizontal and vertical amplitudes. ⁎ p b .05.
with the ratings of males and females belonging to the government, this result was impressive (confidence intervals show no overlap). Findings for trustworthiness were similar to findings of dominance. Again, the ratings of males belonging to the opposition stood out (Table 2). They received the lowest ratings on trustworthiness of all groups. The composite measure of overall motion provided no significant interaction between gender and party position (Table 3). Means and their confidence intervals showed that overall motion was highest for males from the opposition (Table 4). This was in line with this group's ratings on dominance. However, the differences in trait ratings were not convincingly reflected in our motion measure. Therefore, people's ratings must be guided by additional nonverbal cues. 4. Discussion
composite measure of body motion by adding up vertical and horizontal amplitudes. Stick-figures rated high on dominance tended to be rated low on trustworthiness (r(58) = −.57), those rated high on trustworthiness were also rated high on competence (r(58) = .76), and no relationship (r(58) = −.11) was found between dominance and competence (see also Koppensteiner et al., 2015). Thus, in part dominance and trustworthiness were mutually exclusive categories. Indices of body motion, expansiveness and our composite measure of motion showed profound relationships with ratings of dominance and trustworthiness. Perceived competence did not yield a meaningful relationship with any of our measures (Table 2). Stick-figures adopting an expansive body posture, displaying expansiveness in overall motion behavior and a great deal of body movements were rated high on dominance. These findings indicate that people not only ascribe dominance to static cues of expansiveness but also to motion cues. Low to moderate interrelations between expansiveness of body postures and the quantity of motion – a motion measure independent of amplitude – further supported such an interpretation (see Table 1). The negative relationship between trait ratings of dominance and trustworthiness was also reflected in correlations between motion measures and trustworthiness. There were negative relationships throughout (Table 2). Overall, this can be condensed into a simple formula, namely that high ratings of trustworthiness were linked to low activity in body motion. In the second step we analyzed whether the stick-figures were judged differently on dominance, trustworthiness, and competence depending on the speakers' gender and their party status (i.e., opposition or government). ANOVAs yielded significant interactions between gender and the politicians' role in parliament (i.e., opposition or government) for perceived dominance and perceived trustworthiness (Table 3). Inspection of the mean ratings for dominance provided detailed insights into the differences between the four groups we investigated (Table 4 and Fig. 3). Perceived dominance was highest for males belonging to the opposition parties. In particular, when compared
To display dominance humans and animals make broad gestures or adopt expansive body postures (e.g., De Waal, 2007; Mehrabian, 1972). We investigated whether people perceive such dominance cues in the body motion of politicians giving speeches. For this purpose we translated short excerpts from different speeches into stick-figure animations and extracted the “maximum expansiveness” (i.e., expansiveness of a single posture), the overall “expansiveness of motion” and “the quantity of motion”. All three measures were strongly interrelated and good predictors of perceived stick-figure dominance. This indicates that speakers who adopt expansive postures also produce such expansive postures frequently. Moreover, expansive postures lead to impressions of dominance and perceiving such expansive postures frequently may even reinforce such impressions. In contrast to dominance, trustworthiness was associated with low expansiveness in motion and a low number of movements. Such a negative relationship between dominance and trustworthiness was also reflected in the trait ratings. Therefore, displays of high dominance have a negative impact on impressions of trustworthiness. Taking a closer look at the results revealed that our “nonverbal measures” were not as strongly related to trustworthiness than to dominance. In other words, people perceived dominance more easily than trustworthiness in the nonverbal cues we extracted. This is not surprising because dominance displays are not subtle in general and intended to attract attention. Stick-figures perceived as trustworthy were also perceived as competent, yet we found no relationship between our “nonverbal measures” and competence. Despite being one of the basic social categories in judgments of politicians (Fiske et al., 2007) competence was not conveyed by the cues we extracted. It is conceivable that competence is associated with motion patterns our simple measures failed to capture. However, it is also possible that people, although able to perceive competence in facial cues, are unable to perceive competence in body motion. First impressions may represent a cognitive adaptation that helps to decide whether to approach or to avoid someone (Oosterhof &
Table 4 Mean trait ratings and means of motion measures grouped by the speaker's role in parliament (government or opposition) and the speakers' gender. Government
Opposition
Ratings or motion M.
Female
Male
Female
Male
Dominance
−3.46 (26.66) [−18.03, 10.66] −.56 (14.60) [−8.29, 7.56] 9.6 (12.47) [2.56, 15.98] 13.35 (5.61) [11.18, 17.87]
−4.75 (21.31) [−14.45, 5.74] 5.90 (13.73) [−1.04, 11.99] 11.13 (13.63) [4.94, 17.78] 14.56 (5.52) [12.05, 17.23]
.43 (29.19) [−11.48, 14.85] −2.79 (15.85) [−10.07, 3.98] 5.39 (13.74) [−.50, 11.90] 14.06 (6.60) [11.50, 17.56]
28.39 (19.82) [13.48, 35.61] −12.04 (15.71) [−19.31, −3.81] 3.08 (13.65) [−3.64, 9.83] 17.83 (3.66) [15.92, 19.62]
Trustworth. Competence Overall motion
Note. Values in parenthesis are standard deviations. Values in brackets are bootstrapped 95% confidence intervals on the basis of 9999 replicates. Trustworth. = trustworthiness; Overall motion is a composite of horizontal and vertical amplitudes. Rating scales ranged from −100 to +100. Government/male n = 16; Government/female n = 12; Opposition/male n = 14; Opposition/female n = 18.
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
105
Fig. 3. Interaction plots showing mean ratings of dominance and trustworthiness grouped by party status and gender (rating scales ranged from −100 to +100).
Todorov, 2008). The results obtained may be interpreted in a similar way and the ratings may be classified as “positive versus negative”. However, previous studies using stick-figure stimuli show that the body movements of speakers allow more than such simple categorizations (e.g., Koppensteiner et al., 2015; Koppensteiner, 2013). In the present study we only extracted conspicuous dominance cues, which may have a predominant influence on first impressions, but more in depth social evaluations may be based on more complex motion cues. Also, it is not clear whether social categories such as those applied here or the Big Five personality dimensions sufficiently capture the information people perceive in motion cues (see also Thoresen et al., 2012). To clarify this follow-up studies are needed. Politicians change their rhetoric depending on the situation or their role in Parliament (Pancer, Hunsberger, Pratt, Boisvert, & Roth, 1992; Tetlock, 1981). This study is a first step toward investigating whether such flexibility in self-presentation also becomes apparent in body motion. Although the results obtained need to be backed by additional studies, they support such an assumption regarding male speakers in the opposition. Stick-figures representing male politicians in the opposition were perceived as more dominant but less trustworthy than stick-figures representing politicians of the government. However, to provide more clarity future work needs to elaborate on this and test whether former opposition members change their nonverbal performance when they become members of the government. Although a previous study using stick-figure stimuli revealed differences between male and female speakers for perceived extraversion and emotional stability (Koppensteiner & Grammer, 2011), in the present study we only found noteworthy gender differences when including party status. The results obtained suggest that dominance displays, that negatively affect perceptions of trustworthiness, are predominantly shown by male speakers expected to control and criticize the work of those who are in power. Male and female speakers of the government,
on the other hand, may intend to communicate integrity and on the level of body motion this might look similar for both genders. 4.1. Conclusions Snap judgments are not only guided by the outward appearance and facial expressions but also by salient and simple motion cues. We show that speakers displaying expansive movements and a great deal of body activity are rated high on dominance and low on trustworthiness. In addition, females and male members of the ruling parties appear to communicate their views in a different way than males from the opposition and such differences appear to be discernable in body motion. This hints that politicians not only express their positions with words but also with their bodies. Acknowledgments This research was funded by the Austrian Science Fund (FWF): P 25262-G16. We thank the reviewers for their constructive feedback. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.paid.2016.01.013. References Ambady, N., Bernieri, F. J., & Richeson, J. A. (2000). Toward a histology of social behavior: Judgmental accuracy from thin slices of the behavioral stream. Advances in Experimental Social Psychology, 32, 201–271. http://dx.doi.org/10.1016/S00652601(00)80006-4. Antonakis, J., & Dalgas, O. (2009). Predicting elections: Child's play! Science, 323, 11831183. http://dx.doi.org/10.1126/science.1167748.
106
M. Koppensteiner et al. / Personality and Individual Differences 94 (2016) 101–106
Atkinson, A. P., Tunstall, M. L., & Dittrich, W. H. (2007). Evidence for distinct contributions of form and motion information to the recognition of emotions from body gestures. Cognition, 104, 59–72. http://dx.doi.org/10.1016/j.cognition.2006.05.005. Borkenau, P., Mauer, N., Riemann, R., Spinath, F. M., & Angleitner, A. (2004). Thin slices of behavior as cues of personality and intelligence. Journal of Personality and Social Psychology, 86, 599–614. http://dx.doi.org/10.1037/0022-3514.86.4.599. Carney, D. R., Cuddy, A. J. C., & Yap, A. J. (2010). Power posing: Brief nonverbal displays affect neuroendocrine levels and risk tolerance. Psychological Science, 21, 1363–1368. http://dx.doi.org/10.1177/0956797610383437. Carney, D., Hall, J., & LeBeau, L. (2005). Beliefs about the nonverbal expression of social power. Journal of Nonverbal Behavior, 29, 105–123. http://dx.doi.org/10.1007/ s10919-005-2743-z. Chen, F. F., Jing, Y., & Lee, J. M. (2014). The looks of a leader: Competent and trustworthy, but not dominant. Journal of Experimental Social Psychology, 51, 27–33. http://dx.doi. org/10.1016/j.jesp.2013.10.008. Clarke, T. J., Bradshaw, M. F., Field, D. T., Hampson, S. E., & Rose, D. (2005). The perception of emotion from body movement in point-light displays of interpersonal dialogue. Perception, 34, 1171–1180. http://dx.doi.org/10.1068/p5203. De Waal, F. (2007). Chimpanzee politics: Power and sex among apes. Baltimore: JHU Press. Eisenberg, P., & Reichline, P. B. (1939). Judging expressive movement: II. Judgments of dominance-feeling from motion pictures of gait. The Journal of Social Psychology, 10, 345–357. http://dx.doi.org/10.1080/00224545.1939.9713373. Fiske, S. T., Cuddy, A. J., & Glick, P. (2007). Universal dimensions of social cognition: Warmth and competence. Trends in Cognitive Sciences, 11, 77–83. http://dx.doi.org/ 10.1016/j.tics.2006.11.005. Huang, L., Galinsky, A. D., Gruenfeld, D. H., & Guillory, L. E. (2011). Powerful postures versus powerful roles. Which is the proximate correlate of thought and behavior? Psychological Science, 22, 95–102. http://dx.doi.org/10.1177/0956797610391912. Hugill, N., Fink, B., Neave, N., Besson, A., & Bunse, L. (2011). Women's perception of men's sensation seeking propensity from their dance movements. Personality and Individual Differences, 51, 483–487. http://dx.doi.org/10.1016/j.paid.2011.05.002. Koppensteiner, M. (2013). Motion cues that make an impression: Predicting perceived personality by minimal motion information. Journal of Experimental Social Psychology, 49, 1137–1143. http://dx.doi.org/10.1016/j.jesp.2013.08.002. Koppensteiner, M., & Grammer, K. (2010). Motion patterns in political speech and their influence on personality ratings. Journal of Research in Personality, 44, 374–379. http://dx.doi.org/10.1016/j.jrp.2010.04.002. Koppensteiner, M., & Grammer, K. (2011). Body movements of male and female speakers and their influence on perceptions of personality. Personality and Individual Differences, 51, 743–747. http://dx.doi.org/10.1016/j.paid.2011.06.014. Koppensteiner, M., Stephan, P., & Jäschke, J. P. M. (2015). From body motion to cheers: Speakers' body movements as predictors of applause. Personality and Individual Differences, 74, 182–185. http://dx.doi.org/10.1016/j.paid.2014.10.019. Kramer, R. S., Arend, I., & Ward, R. (2010). Perceived health from biological motion predicts voting behaviour. The Quarterly Journal of Experimental Psychology, 63, 625–632. http://dx.doi.org/10.1080/17470210903490977. Little, A. C., Roberts, S. C., Jones, B. C., & DeBruine, L. M. (2012). The perception of attractiveness and trustworthiness in male faces affects hypothetical voting decisions
differently in wartime and peacetime scenarios. The Quarterly Journal of Experimental Psychology, 65, 2018–2032. http://dx.doi.org/10.1080/17470218.2012. 677048. Mazur, A., & Booth, A. (1998). Testosterone and dominance in men. Behavioral and Brain Sciences, 21, 353–363. Mehrabian, A. (1972). Nonverbal communication. Chicago: Aldine. Olivola, C. Y., & Todorov, A. (2010). Elected in 100 milliseconds: Appearance-based trait inferences and voting. Journal of Nonverbal Behavior, 34, 83–110. http://dx.doi.org/ 10.1007/s10919-009-0082-1. Oosterhof, N. N., & Todorov, A. (2008). The functional basis of face evaluation. Proceedings of the National Academy of Sciences, 105, 11087–11092. http://dx.doi.org/10.1073/ pnas.0805664105. Pancer, S. M., Hunsberger, B., Pratt, M. W., Boisvert, S., & Roth, D. (1992). Political roles and the complexity of political rhetoric. Political Psychology, 13, 31–43. Park, L. E., Streamer, L., Huang, L., & Galinsky, A. D. (2013). Stand tall, but don't put your feet up: Universal and culturally-specific effects of expansive postures on power. Journal of Experimental Social Psychology, 49, 965–971. http://dx.doi.org/10.1016/j. jesp.2013.06.001. Pollick, F. E., Paterson, H. M., Bruderlin, A., & Sanford, A. J. (2001). Perceiving affect from arm movement. Cognition, 82, B51–B61. http://dx.doi.org/10.1016/S0010-0277(01)00147-0. Poutvaara, P., Jordahl, H., & Berggren, N. (2009). Faces of politicians: Babyfacedness predicts inferred competence but not electoral success. Journal of Experimental Social Psychology, 45, 1132–1135. http://dx.doi.org/10.1016/j.jesp.2009.06.007. R Core Team (2013). R: A language and environment for statistical computing. Rule, N. O., & Ambady, N. (2011). Face and fortune: Inferences of personality from managing partners' faces predict their law firms' financial success. The Leadership Quarterly, 22, 690–696. http://dx.doi.org/10.1016/j.leaqua.2011.05.009. Stewart, P. A., Salter, F. K., & Mehu, M. (2009a). Taking leaders at face value: Ethology and the analysis of televised leader displays. Politics and the Life Sciences, 28, 48–74. http:// dx.doi.org/10.2990/28_1_48. Stewart, P. A., Waller, B. M., & Schubert, J. N. (2009b). Presidential speechmaking style: Emotional response to micro-expressions of facial affect. Motivation and Emotion, 33, 125–135. http://dx.doi.org/10.1007/s11031-009-9129-1. Tetlock, P. E. (1981). Pre-to postelection shifts in presidential rhetoric: Impression management or cognitive adjustment. Journal of Personality and Social Psychology, 41, 207–212. http://dx.doi.org/10.1037/0022-3514.41.2.207. Thoresen, J. C., Vuong, Q. C., & Atkinson, A. P. (2012). First impressions: Gait cues drive reliable trait judgements. Cognition, 124, 261–271. http://dx.doi.org/10.1016/j. cognition.2012.05.018. Tiedens, L. Z., & Fragale, A. R. (2003). Power moves: Complementarity in dominant and submissive nonverbal behavior. Journal of Personality and Social Psychology, 84, 558–568. http://dx.doi.org/10.1037/0022-3514.84.3.558. Zebrowitz, L. A., & Montepare, J. M. (2008). Social psychological face perception: Why appearance matters. Social and Personality Psychology Compass, 2, 1497–1517. http://dx. doi.org/10.1111/j.1751-9004.2008.00109.x.