On the Correspondence Between Physiological and Nonverbal Responses Nonverbal Behavior Accompanying

On the Correspondence Between Physiological and Nonverbal Responses Nonverbal Behavior Accompanying Running head: THE FACE AND THE VOICE OF THREAT On the Correspondence Between Physiological and Nonverbal Responses: Nonverbal Behavior Accompanying Challenge and Threat Max Weisbuch Tufts University Mark D. Seery University at Buffalo, The State University of New York Nalini Ambady Tufts University Jim Blascovich University of California, Santa Barbara In Press, Journal of Nonverbal Behavior Abstract Little is known regarding the covariance between physiological and nonverbal responses to “stressful” situations. We argue that physiological markers are especially likely to be accompanied by psychologically-meaningful nonverbal behavior. Within “stressful” motivated performance situations, complex patterns of cardiovascular (CV) reactivity mark challenge and threat motivational states relevant to confidence. We expected the CV challenge and threat markers to be distinguished on the basis of facial and vocal confidence. In a test of this hypothesis, participants’ cardiovascular (CV) responses were recorded during a videotaped social interaction. As expected, the CV challenge marker was associated with less vocal confidence and more facial confidence than the CV threat marker. These findings are related to the complexity of human responses to motivated performances. On the Correspondence Between Physiological and Nonverbal Responses: Nonverbal Behavior Accompanying Challenge and Threat Speculation about the relationship between nonverbal behavior and physiology has a long history. Charles Darwin (1872), William James (1890), and Walter Cannon (1915) agreed that responses to what we would today call “stressful situations” are characterized by specific patterns of physiologic and nonverbal activity. For example, Darwin writes, “…terror causes the body to tremble. The skin becomes pale, sweat breaks out, and the hair bristles…the heart beats quickly, wildly, and violently…” (p. 77). These early views suggest that physiological and nonverbal activity constitute an integrated response to environmental demands, implying covariance in physiological and nonverbal responses to stressful situations. Yet little evidence supports a reliable connection—there are few if any published investigations on the topic. One possible explanation is that no reliable relationship between physiological and nonverbal responses exists. For example, nonverbal expressions may serve a communicative function independent of physiological states (Fridlund, 1994). Alternatively, the difficulty of measuring complex physiological responses may account for the lack of physiological-nonverbal evidence. Simple physiological responses alone (e.g., heart rate) can reflect a variety of psychological processes (Cacioppo & Tassinary, 1990) and are thus unlikely to be reliably associated with psychologically meaningful nonverbal activity. We examined the extent to which a particular stressful situation elicited covariance between patterns of nonverbal and physiological activity. Specifically, we examined if and how the cardiovascular markers of challenge and threat were associated with nonverbal behavior. The Biopsychosocial Model of Challenge and Threat Context-general invariants and context-specific markers are physiological patterns that reliably and selectively index specific psychological processes in general or within a specified context, respectively, whereas correlates (e.g., heart rate) can be associated with a variety of psychological processes (Cacioppo & Tassinary, 1990). Accordingly, invariants and markers should be especially likely to covary with psychologically meaningful nonverbal responses. One set of empirically established markers are the cardiovascular patterns associated with the motivational states of challenge and threat (Blascovich & Tomaka, 1996; Blascovich, in press). The challenge and threat markers reveal their corresponding psychological constructs only in contexts that require instrumental cognitive responses to meet a self-relevant goal. Examples of such motivated performance situations include taking a test, giving a speech, and making a good impression. Hence, motivated performance situations include many of those situations that laypeople would consider “stressful.” Challenge and threat motivational states follow from evaluations of resources to cope with the demands of motivated performance situations. Threat occurs when the demands of the situation (e.g., required effort, danger, uncertainty) are evaluated as exceeding coping resources (e.g., skills, dispositions, external support). Challenge occurs when resource evaluations meet or exceed demand evaluations. Challenge and threat may thus roughly be understood as levels of context-specific “confidence.” The physiological markers of challenge and threat are based on Dienstbier’s (1989) model of physiological toughness. To mobilize energy for performance, the body activates the sympathetic-adrenal-medullary (SAM) and hypothalamic-pituitary-adrenocortical (HPA or PAC) axes. Whereas SAM activation can provide an efficient spike of energy mobilization, HPA activation results in a prolonged response. Exhibiting a fast onset and offset of SAM activation coupled with little HPA activation in response to stressors is associated with favorable outcomes, including better task performance and lower anxiety. The biopsychosocial model describes cardiovascular patterns sensitive to relative activation of SAM and HPA axes. A constellation of cardiovascular measures differentiates challenge from threat: heart rate (HR); cardiac output (CO), the amount of blood in liters pumped by the heart per minute; and total peripheral resistance (TPR), an index of vascular constriction/dilation. During challenge, heightened SAM activation leads to increased HR from a resting baseline, dilation of arteries (lower TPR), and greater blood flow (higher CO). During threat, SAM activation increases HR, but HPA activation, inhibits the CO increase and TPR decrease that would otherwise occur. Over 30 studies establish these cardiovascular indexes as markers of challenge/threat motivational states. Initially, studies showed that within (and only within) motivated performance situations, self-reports of challenge and threat (a) were correlated with the cardiovascular patterns, (b) were not caused by the cardiovascular patterns, and (c) increased in parallel with the cardiovascular patterns following challenging versus threatening instructional sets (Tomaka, Blascovich, Kelsey, & Leitten, 1993; Tomaka, Blascovich, Kibler, & Ernst, 1997; for a review, see Blascovich & Tomaka, 1996). Providing convergent evidence, these cardiovascular markers respond in a manner that is consistent with predictions from theories of social comparison, social facilitation, intergroup interaction, affective priming, emotional disclosure, social stigma, self-esteem, and more (Blascovich et al, 1999; Blascovich, Mendes, Hunter, Lickel, & Kowai-Bell, 2001; Blascovich, Mendes, & Seery, 2002; Mendes, Blascovich, Lickel, & Hunter, 2002; Mendes, Blascovich, Major, & Seery, 2001; Mendes, Reis, Seery, & Blascovich, 2003; Seery, Blascovich, Weisbuch, & Vick, 2004; Weisbuch-Remington, Mendes, Seery, & Blascovich, 2005). Finally, the challenge/threat markers have predictive validity. The challenge motivational state includes more confidence in task coping and greater energy mobilization than threat, factors which should enhance performance. Indeed, challenge predicts better performance than threat, and this predictive power is greater than with any single cardiovascular measure (Blascovich, Seery, Mugridge, Norris, & Weisbuch, 2004; Seery, Weisbuch, Hetenyi, & Blascovich, 2008). In summary, the biopsychosocial model of challenge and threat describes a theoretically-based, empirically-validated, and psychologically-relevant cardiovascular pattern. Importantly, challenge and threat are indexed as responses to a specific type of “stressful situation”—a motivated performance. The Current Research Cardiovascular patterns of challenge and threat may correspond to equally complex patterns of nonverbal behavior. In comparison to early theories of nonverbal behavior (e.g., Darwin 1872), recent theories (Ekman & Friesen, 1969) emphasize the role of higher-order cognitive processes in satisfying social constraints on nonverbal behavior. Facial behavior appears to be especially amenable to conscious control (Ekman & Friesen); consequently, facial responses may reflect social desirability more than subjective or physiological experience. Indeed, positive facial expressions can sometimes be more likely among people experiencing negative affect than among people experiencing neutral affect (Cole 1986; Ansfield, 2007). In contrast, the voice appears to be substantially less amenable to conscious control and may be a more veridical indicator of psychological experience (Ekman & Friesen). Thus, facial and vocal channels may differ with regard to their relationship to physiological processes. The face and the voice were examined here with regard to their relationship to the confidence. In the context of a socially-situated task (i.e., colloquial equivalent of challenge— interacting with and trying to make a good impression on a stranger), the face should portray socially-desirable confidence and perhaps especially among individuals experiencing threat. Those experiencing threat may put particular effort into creating a confident facial expression (cf. Ansfield, 2007; Ekman & Friesen, 1969) and may therefore appear more confident than challenged individuals. In contrast, the voice should portray confidence more veridically, such that individuals experiencing challenge sound more confident than those experiencing threat. This conflicting pattern should yield a complex picture in which cardiovascular reactivity consistent with threat is associated with more confidence in the face than in the voice, as compared to individuals experiencing challenge. In summary, we expected covariance in cardiovascular and nonverbal responses to a motivated performance task. This covariance should only be observable with complex cardiovascular patterns that index psychological processes. Method Participants Ninety female students from an introductory psychology course at the University of California at Santa Barbara received partial course credit for participation. Of these 90 participants, 30 were selected for nonverbal analysis on the basis of their cardiovascular patterns (see “selection of target participants”). Although social interaction is engaging for both males and females, it is most likely to be considered a motivated performance for females as the latter are often more sensitive interpersonal feedback (e.g., Williams & Sommer, 1997). Measures Physiological measures. Physiological measurement devices were used to noninvasively record cardiac and hemodynamic signals allowing the recording or calculation of HR, CO, and arterial resistance (total peripheral resistance; TPR). A Minnesota Impedance Cardiograph (Model 304B), a Cortronics continuously inflated blood pressure monitor (Model 7000) and a Coulbourn EKG amplifier/coupler (Model S75-11) provided physiological signals. Coulborn amplifiers (Model S79-02) conditioned the impedance signals. The Cortronics blood pressure monitor provided continuous blood pressure readings. Cardiac performance was monitored by electrocardiographic (EKG) and impedance cardiographic (ZKG) recordings. A Standard Lead II configuration (left leg, right arm, and right leg ground) or the ZKG provided input to the EKG amplifier. The ZKG uses a tetrapolar aluminum/mylar tape electrode system. An interactive software program was used to record and subsequently score the data. Nonverbal measures. Participant behavior was recorded with a microphone and a video camera hidden behind tinted glass (directly facing the participant). From these recordings, two sets of video clips were created. In the first set, silent video clips were cropped at the participant’s chin in order to isolate nonverbal behavior in the face (facial expression clips). In the second set, video was removed, as were high-frequency sounds (digitally), so that prosody, but not individual words, could be perceived (cf., Rogers, Scherer, & Rosenthal, 1971). This set of “content-filtered” clips was created in order to isolate paraverbal behavior. Twenty-four undergraduate students were randomly assigned to judge either video clips or content-filtered clips for the self-confidence and dominance (on 0-5 scales) exhibited by each target participant. Technical problems were encountered by the first six judges of content-filtered clips; the data from these judges were eliminated and the problems were resolved for the final six judges. Hence, twelve judges rated the video clips and six judges rated the content-filtered clips. Inter-rater reliability was satisfactory for all ratings: facial self-confidence (α = .81), vocal self-confidence (α = .59), facial dominance (α = .74), and vocal dominance (α = .56). Consequently, ratings were averaged across judges; resulting facial self-confidence and facial dominance scores (r = .52, p = .004) were averaged to form a facial confidence index, whereas vocal self-confidence and vocal dominance scores (r = .89, p < .001) were averaged to form a vocal confidence index. Procedure Participants completed the experiment individually. Upon arrival at the laboratory, an experimenter greeted the participant and escorted her to a preparation room. Sensors necessary for physiological recording were applied and the participant was brought into a control room and seated upright in a comfortable upholstered chair. A 5-minute rest period began when the experimenter left the room, during which baseline levels of physiological responses were assessed. After the rest period, the participant was informed that a member of the research team would now enter the chamber to engage the participant in a “getting to know you” exercise. A female experimenter unfamiliar to the participant entered the recording room, introduced herself, and sat down to start a conversation. The conversation lasted three minutes, during which physiological measurements were recorded. The interviewer asked the participant a pre-determined set of questions about herself. Questions included, “Tell me about your hometown,” “How do you like college so far?” “What do you like to do in your spare time?” “What are your plans after college?” and “How do you think people that know you well would describe you?” This format encouraged the participant to actively engage in conversation. Upon completion, the sensors were removed and the participant was debriefed. Selection of Target Participants Mean HR, CO, and TPR values were calculated for each minute within the rest and task periods. Cardiovascular reactivity values were calculated for each measure by taking the value from the last minute of the baseline period and subtracting it from the value obtained for the social interaction. Although change scores (of which reactivity is one example) are sometimes discouraged on psychometric grounds, their use is preferable in psychophysiological research (Llabre, Spitzer, Saab, Ironson, & Schneiderman, 1991). As in previous research (Blascovich, Seery, Mugridge, Norris, & Weisbuch, 2004), TPR reactivity scores were standardized and subtracted from standardized CO reactivity scores to form a challenge/threat (CT) index, in which higher scores indicate greater challenge and lower scores indicate greater threat. Because both challenge and threat patterns require increases from baseline in HR, only participants whose HR increased from baseline were considered. Of the 90 initial participants, 88 exhibited increased HR and overall HR reactivity was greater than 0, t(89) = 16.14, p < .001, r = .86. Given HR increases, the 15 lowest and 15 highest CT scores were selected as prototypical of threat and challenge, respectively (see Table 1 for descriptive statistics). This strategy made it possible to examine the clearest instances of challenge versus threat without incorporating participants who exhibited an intermediate state. Results and Discussion We predicted that participants who exhibited threat—consistent with experiencing lower confidence—should sound less confident but look more confident than challenged participants, a pattern that should result in an interaction. A two-way ANOVA with repeated measures on nonverbal channel revealed only a significant interaction, such that vocal confidence was lower and facial confidence was greater among threatened (vs. challenged) participants, F(1, 28) = 5.25, p = .03, r = .91. Figure 1 depicts the marginal means responsible for the interaction (Rosenthal & Rosnow, 1985). No effects for HR reactivity approached significance (ps > .7). For participants who exhibited threat, these results are consistent with attempting to mask underlying lack of confidence—indicated by their vocal responses—with a relatively controllable facial nonverbal display. In contrast, the nonverbal pattern associated with challenge is consistent with actually experiencing higher confidence, coupled with a lack of concern for appearing confident. More generally, the discrepancy between vocal and facial confidence that was perceptually available to observers indicated a cardiovascular pattern normally observable only through sophisticated measurement devices. It is noteworthy that these observable nonverbal patterns occurred within a situation likely to elicit considerable impression management motives. In general, the relationship between psychological processes and nonverbal communication may be moderated by impression management motives in a channel-specific manner. Cole’s (1986) study provides indirect evidence for this claim, in that children’s facial responses to a disappointing toy changed markedly on the departure of the experimenter. Examining the conditions most likely to give rise to vocal-facial discrepancies may be a fruitful topic for future research. More broadly, these results demonstrate that there is covariance in physiological and nonverbal responses but that this covariance may be best observed with a meaningful pattern of physiological activity. Conversely, facial expressions alone may often be misleading, especially as compared to vocal expressions (cf. Ansfield, 2007; Cole, 1986; Ekman & Friesen, 1969) and may be particularly subject to motivational concerns, such as impression management. These considerations highlight the importance of utilizing complex, empirically-validated patterns of physiological reactivity in examining the relationship between physiological and nonverbal responses. References Ansfield, M. E. (2007). Smiling when distressed: When a smile is a frown turned upside down. Personality and Social Psychology Bulletin, 33, 763-775. Blascovich, J. (in press). Challenge and threat. In A. J. Elliot (Ed.), Handbook of approach and avoidance motivation. Mahwah, NJ: Lawrence Erlbaum Associates. Blascovich, J., & Tomaka, (1996). The biopsychosocial model of arousal regulation. Advances in Experimental Social Psychology, 28, 1-51. Blascovich, J., Mendes, W. B., Hunter, S. B., Lickel, B., & Kowai-Bell, N. (2001). Perceiver threat in social interactions with stigmatized others. Journal of Personality and Social Psychology, 80, 253-267. Blascovich, J., Mendes, W. B., & Seery, M. D. (2002). Intergroup encounters and threat: A multi-method approach. In D. M. Mackie & E. R. Smith (Eds.), From prejudice to intergroup emotions: Differentiated reactions to social groups (pp. 89-109). New York: Psychology Press. Blascovich, J., Seery, M. D., Mugridge, C. A., Norris, R. K., & Weisbuch, M. D. (2004). Predicting athletic performance from cardiovascular indexes of challenge and threat. Journal of Experimental Social Psychology, 683-688. Cacioppo, J. T., & Tassinary, L. G. (1990). Inferring psychological significance from physiological signals. American Psychologist, 45, 16-28. Cannon, W. B. (1915). Bodily changes in pain, hunger, fear, and rage. New York: D. Appleton and Co. Cole, P. M. (1986). Children’s spontaneous control of facial expression. Child Development, 57, 1309-1321. Darwin, C. (1872). The expression of the emotions in man and animals. London: J Murray. Dienstbier, R. A. (1989). Arousal and physiological toughness: Implications for mental and physical health. Psychological Review, 96, 84-100. Ekman, P., & Friesen, W. V. (1969). Nonverbal leakage and clues to deception. Psychiatry, 32, 88-105. Fridlund, A. J. (1994). Human facial expression: An evolutionary view. San Diego, CA: Academic Press. James, W. (1890/1908). The principles of psychology, Volume 2. New York: Henry Holt and Co. Llabre, M. M., Spitzer, S. B., Saab, P. G., Ironson, G. H., & Schneiderman, N. (1991). The reliability and specificity of delta versus residualized change as measures of cardiovascular reactivity to behavioral challenges. Psychophysiology, 28, 701-711. Mendes, W. B., Blascovich, J., Lickel, B., & Hunter, S. (2002). Challenge and threat during social interactions with white and black men. Personality and Social Psychology Bulletin, 28, 939-952. Mendes, W. B., Blascovich, J., Major, B., & Seery, M. D. (2001). Challenge and threat responses during downward and upward social comparisons. European Journal of Social Psychology, 31, 477-497. Mendes, W. B., Reis, H. T., Seery, M. D., & Blascovich, J. (2003). Cardiovascular correlates of emotional expression and suppression: Do content and gender context matter? Journal of Personality and Social Psychology, 84, 771-792. Rogers, P. L., Scherer, K. R., & Rosenthal, R. (1971). Content filtering human speech: A simple electronic system. Behavioral Research Methods and Instrumentation, 3, 16-18. Rosenthal, R., & Rosnow, R. L. (1985). Contrast analysis: Focused comparisons in the analysis of variance. New York: Cambridge University Press. Seery, M. D., Blascovich, J., Weisbuch, M., & Vick, S. B. (2004). The relationship between self-esteem level, self-esteem stability, and cardiovascular reactions to performance feedback. Journal of Personality and Social Psychology, 87, 133-145. Seery, M. D., Weisbuch, M., Hetenyi, M., & Blascovich, J. (2008). Cardiovascular challenge and threat markers predict academic test performance. Unpublished manuscript. rdTabachnick, B. G. & Fidell, L. S. (1996). Using multivariate statistics. (3 ed.). New York: HarperCollins. Tomaka, J., Blascovich, J., Kelsey, R. M., & Leitten, C. L. (1993). Subjective, physiological, and behavioral effects of threat and challenge appraisal. Journal of Personality and Social Psychology, 65, 248-260. Tomaka, J., Blascovich, J., Kibler, J., & Ernst, J. M. (1997). Cognitive and physiological antecedents of threat and challenge appraisal. Journal of Personality and Social Psychology, 73, 63-72. Weisbuch-Remington, M., Mendes, W. B., Seery, M. D., & Blascovich, J. (2005). The nonconscious influence of religious symbols in motivated performance situations. Personality and Social Psychology Bulletin, 31, 1203-1216. Williams, K. D., & Sommer, K. L. (1997). Social ostracism by coworkers: Does ostracism lead to loafing or compensation? Personality and Social Psychology Bulletin, 23, 693-706. Author Note Max Weisbuch, Department of Psychology, Tufts University; Mark D. Seery, University at Buffalo, The State University of New York; Nalini Ambady, Department of Psychology, Tufts University; Jim Blascovich, University of California, Santa Barbara. This research was supported in part by a National Institute of Health National Research Service Award (5F32MH078350-02) to the first author and National Institute of Health Grant MH070833-01A1 to the third author. Correspondence concerning this article should be addressed to Max Weisbuch, Department of Psychology, Tufts University, Medford, Massachusetts, 02155. E-mail: max.weisbuch@tufts.edu Table 1 Cardiovascular Baseline and Reactivity Data for the “Challenge” and “Threat” Groups Variable “Challenge” group “Threat” group t (28) Reactivity Cardiac Output .76 (.50) -1.62 (.47) -13.43* Total Peripheral Resistance -64.33 (41.6) 174.92 (93.6) 9.05* Heart Rate 13.35 (8.09) 13.41 (6.64) .02 Baseline Cardiac Output 8.93 (3.04) 9.26 (2.34) .33 Total Peripheral Resistance 821.72 (378.92) 780.99 (210.97) -.36 Heart Rate 68.21 (13.93) 87.30 (11.13) 4.14* Note. Means are reported with standard deviation in parentheses. Univariate outliers (those exceeding 3.3 standard deviations from the grand mean) were transformed by assigning the deviant raw score to a value one unit larger or smaller than the next most extreme score (Tabachnick & Fidell, 1996). To generate t statistics, mean values from the challenge and threat groups were tested against each other. When CT index scores were adjusted for baseline values, one participant in the challenge group no longer exhibited a strong challenge response; dropping this participant from analyses had no substantive effect on results. *p < .05. Figure Caption Figure 1. Judged confidence as a function of cardiovascular pattern and nonverbal channel. Bars represent marginal means. FIGURE 1