The authors express their appreciation to Harry Reis and three anonymous reviewers for their helpful comments on an earlier version of this article. They also gratefully acknowledge Michael Danko's assistance in collecting the data presented here.

Forty college students viewed videotaped excerpts of happiness/reassurance, anger/ threat, and fear/evasion expressive displays by President Reagan. Within each display condition one excerpt was presented in image-only and one in sound-plus-image format. Emotional reactions were assessed by facial electromyography (EMG) from the brow and cheek regions and by skin resistance and heart rate. Following each excerpt, subjects also reported verbally the intensity of eight emotions. Self-reported emotions were influenced strongly by both the expressive displays and prior attitude toward Reagan as well as by media condition. Facial EMG indicated smiling during happiness/reassurance displays and frowning during anger/threat and fear/evasion displays, especially during image-only presentations. Display effects were also found for skin resistance responses when the media conditions were combined and for heart rate changes in the sound-plus-image condition. In contrast to the self-report measures, expressive and autonomic differences did not reveal an interaction between prior attitude and display condition. These results indicate that expressive displays had a direct emotional impact on viewers and that prior attitudes influenced retrospective self-reports of emotion but did not affect autonomic or facial muscle responses during stimulus exposure.

Three studies are reported that examine the relationship between the nonverbal display of emotional affect and indices of the emotional state. Subjects were asked either to conceal or to exaggerate the facial display associated with the anticipation and reception of painful shocks that varied in intensity. Both self-reports of shock painfulness and skin conductance measures of emotional response showed significant changes paralleling the changes induced in expressive behavior; that is, the suppression of expressive responses decreased the magnitude of phasic skin conductance changes and subjective reports of painfulness as compared to the free expression or exaggeration of pain-related expressive response. The effects were obtained for shocks of varying intensities and for both male and female subjects. The findings support theories of emotion that assume that expressive responses serve a self-regulatory as well as a social-communicative function, and further suggest that the self-regulation is mediated neurally, rather than via a process of self-attribution. Finally, the results highlight the need for research on dissimulation in social interaction to consider the effects of acting upon the actor, as well as its effects upon the inferences of observers.

Two experiments explored the effects of observation by another on responses to painful stimuli. It was anticipated that the intensity of pain-related non-verbal expressivity decreases under observation, while indices of arousal (skin conductance and self-report) increase. In Experiment 1, subjects' expressive responses to shock were attenuated when subjects were observed as compared to when they were alone, but the anticipated augmentation of arousal did not occur. Rather, the attenuation of expressive behavior was accompanied by a general decrease in subjective and autonomic responses to the painful stimuli. A second experiment replicated the results of the first study and, in addition, found no evidence for a differential impact of sex of observer on the three measures of arousal. An interpretation is discussed for the effect of observation on expressive behavior and for the relationships observed among expressive, autonomic, and subjective indices of pain.

Twelve college-age males viewed a sequence of equally spaced and randomly ordered red and green lights, in which the red light signaled the advent of shock. Continuous skin-resistance measures were taken. Subjects' nonverbal responses to the red and green stimuli were video taped without their knowledge and were later viewed by themselves and five of the other subjects, individually, under conditions which required them to discriminate between shock and nonshock trials. The subjects' accuracy scores were above chance levels, and significant differences in error rates for stimulus persons but not for judges were found. Subjects who were proficient at the discrimination task were themselves poor stimuli for others and vice versa. The degree of physiological reactivity of the subjects was positively associated with the number of errors made to them as stimuli, but negatively related to their error scores as judges. A number of alternative explanations for this last result are discussed, and several suggestions for future research are made.It is now clearly established (e.g., Davitz, 1964b; Frijda, 1969) that information relevant to the affective state of one individual can be communicated to another via nonverbal behaviors. Facial expressions, for example, whether spontaneously emitted in the presence of emotionally loaded stimuli (Frijda, 1953; Munn, 1940) or posed in response to emotion-simulating instructions (Felcky, 1922), have been consistently found to accurately communicate feeling states, though the level of observed accuracy differs from study to study (Ekman, Friesen, & P^llsworth, f970). In the course of this research, individuals have been found to differ both in their ability to encode or display affective states, which can be recognized by others, and in their ability to decode or interpret the nonverbal displays of others. While these individual differences, particularly as they apply to the decoding process, have been examined in terms of the personality and cognitive factors associated with them (cf. Davitz, 1964a),