From body motion to cheers: Speakers' body movements as predictors of applause.

Appearance cues and brief displays of behavior are related to people's personality, to their performance at work and to the outcomes of elections. Thus, people present themselves to others on different communication channels, while their interaction partners form first impressions on the basis of the displayed cues. In the current study we examined whether people are able to read information from politicians' body motion. For a rating experiment we translated short video clips of politicians giving a speech into animated stick-figures and had these animations rated on trustworthiness, dominance, competence and the Big Five personality dimensions. Afterwards we correlated the ratings with the applause and the hecklings that the speakers received throughout their entire speech. This revealed that speakers whose body movements were perceived as high on dominance, as high on extraversion and as low on agreeableness received more applause. Although the results obtained need support from additional studies they indicate that body motion is an informative cue in real life settings.

Introduction

Appearance cues and brief displays of behavior (so called “thin slices”) are a sufficient source of information for forming quite accurate impressions of other people. To a certain degree, measures of such first impressions predict job performances, financial performances of companies, leadership effectiveness and a stranger’s personality (Ambady et al., 2000, Borkenau et al., 2004, Harms et al., 2012, Hecht and LaFrance, 1995, Kenny et al., 1992, Olivola et al., 2014, Rule and Ambady, 2008, Wong et al., 2011). Consequently, people seem to verbally and nonverbally communicate their abilities and personality to their social environment while their social environment, in turn, uses this information to create an impression (Ambady et al., 2000).

Given such evidence it is not surprising that appearance and other nonverbal cues also play a role in the domain of politics. For instance, politicians or leaders that show facial micro-expressions of facial affect or a heightened overall nonverbal expressiveness influence the emotional state of their audience as well as the impressions this audience forms of their leaders (Cherulnik et al., 2001, Stewart et al., 2009). Moreover, people readily attribute trustworthiness, competence, dominance, and other personality traits to facial photographs of political candidates and some of these ratings are reliable predictors of actual and hypothetical voting decisions (Little et al., 2012, Olivola and Todorov, 2010, Oosterhof and Todorov, 2008, Poutvaara et al., 2009).

In the current study we extended the research on first impressions of politicians. We explored whether people’s ratings of socially relevant traits can be predictors of the behavioral responses a politician might receive from the plenary in the parliament. Our focus was on dynamic cues such as gestures and body motion because people appear to be able to read affective states from motion or to attribute different personalities to different motion cues (Clarke et al., 2005, Hugill et al., 2011, Pollick et al., 2001, Thoresen et al., 2012). For this reason we translated short video clips of politicians into stick figure animations in order to create abstract representations of the speakers’ body movements that diminish the influence of confounding variables such as appearance cues and the speakers’ gender (see also Koppensteiner & Grammer, 2011). These animations were then rated on dominance, competence, trustworthiness and the Big Five personality dimensions.

Previous studies have already found that people ascribe personality traits to the body movements of speakers (Koppensteiner, 2013, Koppensteiner and Grammer, 2010). The current study investigated whether trait ratings of the speakers’ body movements are coupled to the amount of applause or hecklings the speakers received throughout their entire speech. We thus intended to demonstrate that people make sense of parsimonious nonverbal cues and that judgments based on such cues can serve as predictors of behavioral outcomes in a real life setting of high ecological validity. Other “thin slices” studies have already linked job performances or election results to certain behaviors or the appearance of a person. Such variables, however, provide no insight into the direct impact of nonverbal cues on human communication. In contrast to that our research not only focused on body motion but also examined its relationship to behavioral responses that occur in a direct interaction between an audience and a speaker. We provide evidence that motion cues, indeed, reflect socially relevant information that affects behavioral responses arising in interpersonal communication processes.

To sum up, by using trait ratings as predictors of real life outcomes (i.e., audience reactions) we show that people not only read meaning into body motion but also infer relevant social information from it.

Method

Stimuli

We randomly selected 60 speeches (30 male and 30 female) from three parliamentary sessions of the German parliament. From these speeches, we extracted brief, randomly chosen video segments with an average length of 15 s.

To create stick-figure movies of the speakers’ performances, we used the computer program SpeechAnalyzer that enabled us to run through a movie frame by frame and to position landmarks on the speakers’ major joints and their heads (Koppensteiner, 2013, Koppensteiner and Grammer, 2010). To capture body movements these landmarks were repositioned according to the position shifts of a speaker’s body. Thus, landmark positions were translated into time series of two dimensional coordinates on which basis we created stick figure movies we used for our rating experiments.

Procedure

At locations throughout the University of Vienna we recruited 60 persons (33 females and 27 males; age M = 22.5 years, SD = 3.7) for the stick figure rating experiment. Participants performed the rating task on their own using a computer-controlled interface. Stimuli were presented on the left-hand side of the user interface; rating scales with the items dominant, trustworthy, and competent and items from a German version of a brief questionnaire measuring the Big Five personality domains (i.e., Ten-Item Personality Inventory, TIPI) were presented on the right hand side (Gosling et al., 2003, Muck et al., 2007). The scales were divided into 200 subunits with 0 indicating strongly disagree and 200 strongly agree. Each participant rated a subset of 20 randomly selected stick figure animations. Participants received financial compensation of €5.

Analyses

The German parliament provides transcripts of the parliamentary sessions. These transcripts contain the original wording of given speeches and how often speakers received applause or were heckled. For statistical analysis applause per speech length (in seconds) and heckling per speech length were correlated with stick figure ratings.

Results and discussion

The number of trait ratings for the stick figure clips ranged from 18 to 22. Each personality dimension of the Big Five questionnaire (i.e., TIPI) consisted of two items. For this reason we used simple bivariate correlations to measure the reliability of the scales (Table 1). Analyses revealed high reliabilities for extraversion and agreeableness, a moderate reliability for conscientiousness and a relatively low one for openness. Reliability for emotional stability was unacceptably low. For this reason we did separate analyses for both items of emotional stability.

Table 1

Bivariate correlations between corresponding items of the Big Five.

	Extraversion	Conscientiousness	Openness	Agreeableness	Emotional stability
Reversed scored items	Reserved, quiet	Disorganized, careless	Conventional, uncreative	Critical, quarrelsome	Anxious, easily upset
Extraverted, enthusiastic	.89⁎⁎⁎
	[.83, .93]
Dependable, self-disciplined		.70⁎⁎⁎
		[.54, .81]
Open to new experiences, complex			.53⁎⁎⁎
			[.32, .70]
Sympathetic, warm				.85⁎⁎⁎
				[.76, .90]
Calm, emotionally stable					.03
					[−.22, .28]

Notes: Algebraic signs of the reversed scored items’ scores were inverted before correlation analysis. Numbers are Pearson correlation coefficients, numbers in brackets are 95% confidence intervals; N = 60.

p ⩽ .001.

Trait ratings were averaged for each speaker. Correlations between ratings revealed a wide range of interdependencies (Table 2). The prominent intercorrelations between dominance, agreeableness, and extraversion were of special importance, because ratings in these categories were noteworthy predictors of the applause the speakers received throughout their speeches (Table 3). More precisely, speakers whose stick-figures were perceived as being high on dominance and high on extraversion but low on agreeableness received more applause from their colleagues in the plenum.

Table 2

Bivariate correlations between traits measured by the questionnaire.

	Competence	Dominance	Trustworthiness	Extraversion	Agreeableness	Calm, emotionally stable	Anxious, easily upset	Conscientiousness
Competence	1
Dominance	−.11	1
	[−.35, .15]
Trustworthiness	.76⁎⁎⁎	−.57⁎⁎⁎	1
	[.63, .85]	[−.72, −.37]
Extraversion	−.09	.93⁎⁎⁎	−.54⁎⁎⁎	1
	[−.33, .17]	[.89, .96]	[−.70, −.33]
Agreeableness	.42⁎⁎⁎	−.88⁎⁎⁎	.82⁎⁎⁎	−.86⁎⁎⁎	1
	[.19, .61]	[−.93, −.80]	[.71, .89]	[−.92, −.78]
Calm, emotionally stable	.71⁎⁎⁎	−.53⁎⁎⁎	.82⁎⁎⁎	−.55⁎⁎⁎	.77⁎⁎⁎	1
	[.56, .82]	[−.69, −.32]	[.72, .89]	[−.71, −.35]	[.65, .86]
Anxious, easily upset	.31⁎	.67⁎⁎⁎	−.03	.68⁎⁎⁎	−.40⁎⁎	0.03	1
	[.06, .52]	[.50, .79]	[−.28, .23]	[.52, .80]	[−.59, −.16]	[−.22, .28]
Conscientiousness	.72⁎⁎⁎	−.43⁎⁎⁎	.81⁎⁎⁎	−.41⁎⁎	.67⁎⁎⁎	.81⁎⁎⁎	.14	1
	[.57, .82]	[−.61, −.19]	[.70, .88]	[−.60, −.18]	[.50, .79]	[.70, .88]	[−.12, .38]
Openness	.64⁎⁎⁎	−.14	.60⁎⁎⁎	−.02	.41⁎⁎	.55⁎⁎⁎	.31⁎	.54⁎⁎⁎
	[.46, .77]	[−.38, .12]	[.41, .74]	[−.28, .23]	[.18, .60]	[.35, .71]	[.06, .52]	[.34, .70]

Notes: Numbers are Pearson correlation coefficients, numbers in brackets are 95% confidence intervals. Items measuring emotional stability were depicted separately because of low internal consistency (see Table 1); N = 60.

p ⩽ .05.

p ⩽ .01.

p ⩽ .001.

Table 3

Bivariate correlations of stick figure ratings of dominance, trustworthiness, competence and the Big Five personality dimensions with applause and hecklings per second.

	Hecklings/s		Applause/s
	r	95% CI	r	95% CI
Competence	−.01	[−.27, .24]	−.13	[−.37, .13]
Dominance	.13	[−.13, .37]	.39⁎⁎	[.15, .59]
Trustworthiness	−.09	[−.33, .17]	−.25	[−.47, .01]
Extraversion	.16	[−.10, .39]	.32⁎⁎	[.07, .53]
Agreeableness	−.11	[−.35, .15]	−.39⁎⁎	[−.59, −.15]
Emotional stability
Calm, emotionally stable	−.19	[−.42, .07]	−.27⁎	[−.49, −.02]
Anxious, easily upset	.04	[−.22, .29]	.26⁎	[.01, .48]
Conscientiousness	−.14	[−.38, .12]	−.11	[−.35, .15]
Openness	−.19	[−.42, .07]	−.27⁎	[−.49, −.02]

p ⩽ .05.

p ⩽ .01.

Less pronounced but still non-negligible relationships were found between both items of emotional stability (i.e., calm, emotionally stable and anxious, easily upset) and applause and between trustworthiness and applause. Thus, to a certain degree speakers who received more applause were perceived as less calm and emotionally stable, as more anxious and easily upset, and as less trustworthy. No effects of importance were found between trait ratings and hecklings.

Our findings indicate that some of the trait ratings we collected are more than mere attributions. They have ecological validity because they in part reflected how the audience in the plenary reacted to the speakers. In other words, abstract displays of a speaker’s body movements can be a sufficient source of information to make predictions about real life outcomes. This underlines that people are sensitive to motion cues and are able to use them for quick judgments in social encounters.

Dominance is frequently associated with acts or displays of forcefulness and assertiveness (Buss & Craik, 1980) and appears to express itself in behaviors, which are clearly visible and affect the social environment. A similar reasoning applies to extraversion. It is also a personality trait that is clearly visible in nonverbal behaviors (e.g., Kenny et al., 1992). Hence, it was plausible to expect that dominance and extraversion have an impact on audience reactions. Moreover, we were able to demonstrate that ratings of agreeableness are negatively related to the amount of applause a speaker received. To conclude, findings suggest that body movements perceived as dominant were also perceived as extraverted and as unfriendly or aggressive (i.e., low agreeableness).

We were not able to determine whether applause triggers “certain displays” or “certain displays” trigger applause. Future research, therefore, should analyze whether certain behaviors occur more often after people have applauded. This could clarify the causal direction of the relationship between nonverbal displays and applause. In addition, with the presented experimental set-up we were unable to reveal how verbal content and information from other communication channels are related to body motion. It is very plausible that “aggressive” body movements are coupled to an “aggressive” language that is aimed at political opponents. This also needs to be investigated in future studies.

Conclusions

As already demonstrated in previous work, body motion appears to be an important nonverbal communication channel that conveys affective and social information. In the current study we found that people’s attributions of dominance, extraversion, and agreeableness to speakers’ body movements can provide sufficient information to predict the amount of applause the speakers received throughout their entire speech. Nonverbal displays expressing qualities such as dominance might be important for those who strive for leadership positions while potential followers might benefit from easily recognizing who has the ability to be a leader. Consequently, such information of social relevance might be legible from different nonverbal and verbal communication channels including body motion.