r:

"

_b

Chapter Twenty ... Four

Social Processes
John T. Cacioppo Richard E. Petty

INTRODUCTION Different strategies can be employed to study social behavior. One, for instance, involves the use of gross observational procedures to catalog the vagaries of social behavior and to begin identifying the eliciting conditions. In a second, the emphasis is not on discovering yet more categories of social behavior, but rather on specifying the processes underlying the behaviors that gross observations have already catalogued. Choice of strategy will be dependent on the question that aparticular investigator seeks to answer. Some strategies (and tools) are better suited for cataloguing social behaviors, whereas others are better suited for studying the rudimentary processes underlying these behaviors. The application of psvchophvsiological recording techniques, for example, can be, and has often been, overly complicated and cumbersome when used as a means of discovering, describing, and cataloguing social behavior. In the present chapter, we are concerned with the task of delineating the foundations of social processes. Our premise is that social processes are composed of a finite number of distinct subprocesses, or "stages." A

stage, in turn, is viewed as consisting of a limited number of elements, or "constructs." Finally, each stage is viewed as possessing identifiable properties and propensities for combining to yield social behavior. We propose that psychophysiology can be useful in helping to identify these stages. The present approach to the study of social processes is not unlike the chemist's approach to the study of chemical processes. In this conception, processing stages roughly correspond to chemical elements; social processes correspond to chemical substances. A finite number of chemical elements combine in an orderly sequence to yield an almost infinite number of substances, and a finite number of elementary processing stages are assumed to combine in an orderly sequence to yield an almost infinite variety of social processes. Moreover, just as early investigations in chemistry sometimes revealed that isolated "elements" were in fact combinations of smaller elements, the isolation of certain fundamental processing stages does not preclude. but rather facilitates, the subsequent division of these components into yet more elementary stages. Finally, just as the total number of distinct chemical elements that exists

John T. Cacioppo. Department of Psychology, University of Iowa, Iowa City, Iowa. Richard E. Petty. Department of Psychology, University of Missouri-Columbia, Columbia, Missouri.

Social Processes

647
their properties and (3) explore the manner in which they combine. We begin by reviewing the historical roots of psychophysiological research on social processes.

is empirically defined, with the presumption being only that the total is finite, the total number of processing stages is defined empirically and is assumed to be finite. With few exceptions, the stages of social processes have been viewed as being conscious, though not necessarily reportable (Nisbett & Wilson, 1977; cf. lajonc, 1980b). These conceptions are due partially to the predominant paradigm in the social sciences. Generally, this research relies on people's self-reports to assess the efficacy of the experimental manipulations, the effects of these manipulations on verbal or overt behavioral responses, and the operation of some assumed intervening operation (Gerard. 1964). This methodology may provide details about those aspects of social processes that involve short-term memory (see Ericcson & Simon, 1980). but may be less sensitive to rudimentary elements and stages that are not conscious, only briefly conscious, or conscious but excised by subjects in experiments or interviews. In addition, inferences about the stages comprising social processes that are based solely on the study of subsequent verbal or behavioral responses may suffer, because these responses reflect upon an entire sequence of stages and lag differentially behind these constituent stages. Thus, although a number of informative functional relationships have been documented in the social sciences, the successful specification of the foundations of these relationships has been more elusive. This is in part because these individual elements and stages may have no overt behavioral component. and in part because the similar behavioral outcomes that emerge do not always derive from the same set and sequence of stages (e.g., Fazio, Zanna, & Cooper, 1977). Continuous recordings of physiological responses, though not without problems, can provide concurrent and independent data regarding social processes. Our aims in the present chapter are ( 1) to trace the history of psychophysiological applications to the study of social processes and behavior; (2) to outline two analytical frameworks for collecting and interpreting physiological data to study the rudiments of social processes; and (3) to illustrate these interpretive frameworks in a brief review of studies of social processes in which psychophysiological measures have been obtained. Most of the studies that we review are from social psychology or psychophysiology. but the analytical frameworks outlined and the principles described in this chapter are obviously applicable generally to the social sciences. Finally, it should be emphasized that the physiological data obtained in these studies are not viewed here as mere correlates of. or alternative expressions for, social processes. Rather, they are viewed as harboring information (limited only by the logic of the experimental design) with which to (1) isolate candidates for the fundamental stages and elements of social processes, (2) identify

HISTORICAL BACKGROUND
Traditionally, psychophysiological investigations of social processes were defined by the use of a social factor as an independent variable and of a polygraph to obtain at least one of the dependent measures. This most often meant recording galvanic skin responses (GSRs) during what in all other respects would be regarded as a traditional social-psychological study (Schwartz & Shapiro, 1973). Research of this kind has gone by various names over the years, including "interpersonal physiology" (Dilvlascio, Boyd, & Greenblatt, 1957), "sociophvsiologv" (Boyd & DtMascio, 1954; Waid, 1984), "the psychobiology of social psychology" (Leiderman & Shapiro, 1964), and "social psychophysiology" (Cacioppo, 1982; Cacioppo & Petty, 1983b; Schwartz & Shapiro, 1973). The application of psychophysiology to the study of social processes, however, can no longer be characterized by any single measure or measurement instrument. The relationships between social processes and bodily states have now been advanced using misattribution procedures (e.g., lanna & Cooper, 1976; Zillmann, 1978), cardiac pacemakers (Cacioppo, 1979), drugs (e.g., J. Cooper, Zanna, & Taves, 1978; Schachter & Singer, 1962), hypnosis (Maslach, 1979), expressive and postural manipulations (e.g., Laird, 1974; Petty, Wells, Heesacker, Brock, & Cacioppo, 1983; Vaughan & Lanzetta, 1981), operant training procedures for shaping physiological activity (Cacioppo, Sandman, & Walker, 1978), and exercise (e.g., Pennebaker & Lightner, 1980; Zillmann, 1983). Moreover, a full array of psychophysiological measures, ranging from heart rate (HR) to electrocortical activity. has been employed (see Cacioppo & Petty, 1983a; Shapiro & Crider, 1969). The utility of adopting a psychophysiological perspective as well as a developed methodology to delineate social processes is in part responsible for this broader conception. For example, although most social psychologists have focused on the role of situational factors in shaping social behavior, it is now evident that human association can markedly influence the organismic environment (Christie & Mellett, 1981; Kaplan, 1979) and that physiological factors are important determinants of people's exchanges and interactions with one another (Cacioppo & Petty, 1983b; Waid, 1984). We use the term "social psychophysiology" to refer to this area of research in the present chapter, because it best reflects the origins of much of the early research; it is the term most com-

,. '-

648
monlv used in the contemporary literature; and, most importantly, it is defined more generally than its alternatives to include any use of noninvasive procedures to study the relationships between actual and perceived physiological events (I.e., signs, symptoms, undetected physiological responses) and social cognition and behavior (Cacioppo & Petty, 1983a).

PROCESSES

Early Observations The perspective on human behavior epitomized by social psychophysiology is quite old, dating back at least to the third century B.C., when a young man named Antiochus had the fortune of being the son of one of Alexander the Great's leading generals and the misfortune of falling in love with his father's young and beautiful bride (Mesulam & Perry, 1972). The son knew that his love for his stepmother should never be realized. He tried to control his manner when in her presence, but apparently wished privately to lapse. The young man soon began manifesting what appeared to be a mysterious illness. He would on occasion suffer from debilitating attacks, characterized by changes including increased heart beat variability, stammering speech, profuse perspiring, weakening muscles, and pallor. Several local physicians had unsuccessfully attended to the young man when the well-known Greek physician Erasistratos arrived. To isolate what particular stimulus triggered the attacks, Erasistratos observed the peripheral physiological responses of the young man during the course of the day. Based upon the distinctive syndrome of physiological responses that was evoked by the stepmother's visits to the young man, Erasistratos deduced that the young man suffered from lovesickness. The father gave up his new bride, and the son recovered. This case is, to our knowledge, the first recorded "socialpsychophysiological" inquiry. Articles bearing this perspective began appearing in the psychological literature in the 1920s. For example, Riddle (1925) studied the changes in people's breathing when they were bluffing during a game of poker. For the next 35 years, most of the empirical investigations were conducted by psychiatrists intere~ted in obtaining "objective" measures of therapeutic effectiveness (Shapiro & Crider, 1969). One study, however, conducted by C. E. Smith in 1936 and reported in the Journal of Abnormal and Social Psychology, is quite contemporary in its attempt to focus on the foundations of social processes using Psyc?~physiology (d. Crider, 1983). Smith began by admInistering a questionnaire in a class of college students The q uestlonnalre was desi '. · esigne d to assess each student' .d 20 d'ff s attltu es and strength of conviction on (e I ,,~~nt and, at that time, controversial issues .g., f rvorce should be granted only in the case of grave 0 fense ""M d . , 0 emtsnc art should be enthusias-

tically encouraged"). Four weeks later, subjects were asked to participate in an experiment. Subjects (all males) were tested individually, and GSR recordings were obtained. For half of the issues, a subject was led to believe that the majority of his peers agreed with him, whereas he was led to believe that the majority disagreed with him on the other half. Smith found that (1) disagreement with an opinion statement was accompanied by greater GSRs than was agreement; (2) the responses of subjects who knew they were in the minority rather than the majority were accompanied by greater GSRs; (3) the magnitude of the GSRs varied with the degree of conviction reported, excluding indifference and absolute conviction; and (4) the GSRs accompanying absolute conviction tended to be smaller than the response accompanying the degree of conviction immediately preceding the absolute. Thus, Smith compared the GSRs obtained across clearly constructed experimental conditions to generate hypotheses regarding the processes underlying conformity, and in so doing, developed the notion that unless individuals have a very strong conviction in an attitude, they undergo conflict when they realize that they hold a deviant (i.e., nonnormative) position on an issue (see also Murray, 1938). The first summary of empirical research in social psychophysiology was published by Kaplan and Bloom in 1960. This review focused on the psychophysiological concomitants of social status, social sanction, definition of the situation, and empathy. Kaplan and Bloom expressed an optimism that this approach had come of age: "In recent years sociological and social psychological concepts have been applied in physiological studies at an ever increasing rate. The acceptance and utilization of such concepts have been said to form the basis for a relatively new field of inquiry" (Kaplan & Bloom, 1960, p. 133). Their review appeared in the Journal of Nervous and Mental Diseases rather than a social science journal, perhaps in part because most of the research reviewed by Kaplan and Bloom was conducted to assess thera- peutic effectiveness. For example, neither the study by Riddle ( 1925) nor any of C. E. Smith's studies (C. E. Smith, 1936; C. E. Smith & Diven, cited in Murray, 1938) was discussed. At about the same point in time, John Lacey's seminal critique of psychophysiological measures of therapeutic effectiveness appeared in a volume on research in psychotherapy. Lacey (1959) argued that there was little consistency in the literature upon which to build bridges between psychophysiological data and therapeutic outcomes. He went on to argue that the phenomena of individual and stimulus-response stereotypies were largely responsible for the apparent inconsistencies in the psychophysiological data, and that these stereotypies were so pervasive that they, rather than complex therapeutic factors, should be the focus of research in the coming years (cf. Coles, Jennings, & Stern, 1984). Not surpris-

Social Processes

649
either, as a chapter entitled "Social Psychophysiology" (Schwartz & Shapiro, 1973) appeared early in this decade in an edited book on electrodermal activity (EDA) that was compiled largely for a psychophysiological audience (Prokasy & Raskin, 1973). Also. Kaplan (1972, 1979) reviewed early research in the area for the field of medical sociology.

inglv, the search for psychophysiological indices of psychotherapeutic effectiveness slowed following Lacey's review. Investigations of the reciprocal influence between social and physiological systems, however, began to broaden in scope and increase in number. In 1962, Schachter and Singer published their influential paper. "Cognitive, Social, and Physiological Determinants of Emotional State." The thrust of their twofactor theory of emotions was that the sensations derived from a large and unexpected increase in diffuse physiological arousal could be experienced as widely different emotions, depending upon the circumstances covarying with these sensations.' Two years later, Leiderman and Shapiro ( 1964) published a small edited book, Psychobiological Approaches to Social Behavior, which generally depicted a different vein of research: Evidence was presented for the dramatic impact that social factors such as conformity pressures and group interactions can have on physiological responding. The book was interesting and informative, but has been overlooked by most social psychologists. Moreover, psychophysiologists who were aware of this research apparently concluded that social factors were potentially powerful contaminants in psychophysiological studies (d. Schwartz & Shapiro, 1973; Shapiro & Crider, 1969). Hence, much of the research in psychophysiology during the 1960s was conducted with care in eliminating potentially complicating social factors from empirical studies and, consequently, from psychophysiological theorizing. By.the next decade, psychophysiological recording had become relatively standardized. The discipline of psychophysiology had its own textbook (Sternbach, 1966), handbook (Greenfield & Sternbach, 1972), journal, and society. The second edition of The Handbook of Social Psychology was published at about this point in time. containing an interesting chapter by Shapiro and Crider (1969) on "Psychophysiological Approaches to Social Psychology"; a year later, in the Annual Review of Psychology, a chapter entitled "Psychophysiological Contributions to Social Psychology" (Shapiro & Schwartz, 1970) appeared. This coverage of social-psychophysiological research was not directed solely to a social-psychological audience,

Early Obstacles A major emphasis in these reviews was on applying psychophysiological procedures to validate social-psychological constructs and measures. Their utility in this endeavor proved considerable:
At the very simplest methodological level. the attraction of social psychologists to physiological techniques is not hard to understand. The techniques provide nonverbal, objective [and] relatively bias-free indices of human reaction that have some of the same appeal as gestural, postural, and other indicators of overt response. (Shapiro & Schwartz, 1970, pp. 88-89)

1. According to Schachter (1964). "the present formulation ... maintains simply that cognitive and situational factors determine the labels applied to any of a variety of states of phvstological arousal" (p. 71). The physiological features of these arousal states were never explicitly specified. but investigators came to focus 01\ the following prerequisite reportable features of this physiol~cal state: (1) the perception of a state of diffuse physiological arousal. (2) experienced immediately as being neither pleasant nor unpleasant. and (3) emerging in the absence of an immediate and plausible cause. Recent reviews of the present status of this formulation can be found in Cotton ( 1981). Manstead and Wagner (1981). Cacioppo and Petty (in press). and Zillmann (1978).

However, the attractiveness of applying psychophysiological procedures to help isolate and identify the elementary constructs and stages of social processes was tempered, it would appear, by three barriers: (1) the physiological background, technical sophistication, and elaborate instrumentation that are necessary for collecting, reducing, and analyzing interpretable psychophysiological data in already complex social-psychological paradigms; (2) the absence of analytical frameworks within which to create treatment comparisons and interpret physiological data bearing upon social processes, resulting in a paucity of conceptuallinks with clear ranges of validity between the physiological data and social-psychological constructs; and (3) the inevitable pitting of social-psychological and psychophysiological procedures against each other in studies of construct validation (Cacioppo, 1982; Shapiro & Crider, 1969). Three distinct responses to these barriers can be found in the literature: (1) the simple dismissal of physiological factors as being irrelevant, at least at present, to the study of social processes, and the dismissal of the study of social factors as too molar to contribute to an understanding of psvchophvslological relationships; (2) adoption of the view that diffuse and perceptible changes in physiological arousal are the major physiological processes relevant to social cognition and behavior; and (3) the narrowing of the breadth of the social issue under investigation while increasing the depth (levels) of the analysis to entertain highly specific, potentially reciprocal, and biologically aptive (i.e., adaptive and exaptive; cf. Gould & Vrba, 1982) influences between social and physiological systems.

-;~
650
Dismissal of Social-Psychophysiological Influences
When physiological data have been used successfully to validate a social-psychological construct, it has occasionally been reasoned that there is no further utility in using psychophysiological procedures to investigate the phenomena involving the construct, since the simpler, less expensive, and less time-consuming verbal or behavioral measures traditionally used by social psychologists have been documented to be generally just as informative. On the other hand, when application of a psychophysiological perspective has failed to confirm the validity of a verbal or behavioral measure of a social construct, the theoretical inclination of the investigator has influenced the decision as to whether the theoretical construct, the verbal or behavioral measure, or the psychophysiological validation procedure is suspect. Moreover, in pointing to powerful situational determinants and consequences of social behavior, investigators have occasionally deemphasized the role of organismic factors. Valins (1966), for example, proposed a modification of Schachter and Singer's ( 1962) theory of emotions when he argued that the actual sensation of a change in physiological arousal is unimportant; rather, the simple belief that a change has occurred is both necessary and sufficient to initiate an emotional reaction by evoking the search for an emotional label. Several years later, Bern (1972) argued more generally that people are typically insensitive to internal sources of information and reasoned that people infer their feelings toward issues and events by observing their overt behavior just as an observer would do. These theoretical stances, of course, imply that there is no need to try to surmount the obstacles to social-psychophysiological research.
PROCESSES

Focus on

Arousal

The second and perhaps predominant response, however, has been to view the physiological factor important in the study of social processes as consisting of diffuse, perceptible changes in physiological arousal. In its simplest form, this view has provided the justification for research being conducted with little or no psychophysiological recording equipment and expertise, since it is reasoned that any single physiological response (e.g., see Krugman, 1981), or even simple self-reports of bodily sensations, can index a person's physiological arousal at that moment in time (e.g., Thayer, 1970). Hence, psychophysiological procedures as simple as occasionally palpating the radial artery in a subject's wrist to measure HR or monitoring anyone of a number of electrodermal responses (EDRs) have been used in various studies as if they were equivalent measures of physiological functionmg. Interestingly, this viewpoint has led to the development or refinement of assessment procedures (e.g.,

misattribution paradigms) that allow researchers to consider the influence of perceived physiological events (e.g., "felt arousal") on social processes without the use of psychophysiological recording equipment. For instance, in the misattribution paradigm, subjects ingest a placebo that they believe has specific experimenter-specified side effects. This assessment procedure is based upon the notion that people search for the cause of an unexplained change in their bodily sensations, such as a sudden feeling of tension or arousal (cf Zanna & Cooper, 1976). In their search, people consider external as well as internal causes and sometimes misartribute the true cause of a felt bodily reaction (e.g., increased sympathetic activity due to a transgression they have committed) to something that seems to be a reasonable cause at the time (e.g., a pill they have recently ingested). Of course, only when a stimulus is believed to have the same or nearly the same effects on individuals as they are experiencing at the moment will misattribution to the stimulus be likely (cf. Zillmann, 1983). Thus, in the misattribution paradigm, subjects are either exposed or not exposed to social conditions that are believed to affect physiological responding, and are either exposed or not exposed to a possible external cause for experiencing specific bodily sensations. The nature of the internal state that people are experiencing in the various social conditions is then inferred indirectly by determining what type of "side effect" elicits misattribution. In an illustrative experiment by Higgins, Rhodewalt, and lanna (1979), subjects were given a drug and were told that the experiment concerned the effects of the drug on learning. Subjects were instructed that the pill might cause them to feel "unpleasantly sedated" (unpleasant-unaroused), "relaxed" (pleasant-unaroused), "tense" (unpleasant-aroused), "pleasantly excited" (pleasant-aroused), or nothing in particular (no-side-effects control group). Subjects were also told that any side effects would occur quickly (e.g., 5 min), and that the effects of the drug on learning would not occur fully for some time (e.g., 30 min). Subjects were asked to assist the experimenter in another "unrelated" study to allow time for the drug's possible effect on learning to take full effect. Subjects in this study had, in fact, been given a placebo, a pill made of milk powder, which had absolutely no perceptible physiological effects. While subjects were "waiting" for the pill to reach its full effect, they were induced to write a counterattitudinal message under conditions designed to arouse cognitive dissonance (see pp. 670-673). Higgins er al. (1979) replicated previous research findings that a pill described as having the side effect of making the person feel "tense," in contrast to a pill with no side effects, attenuated attitude change. This suggests that subjects misartnbuted the sensations aroused by cognitive dissonance to the pill, rather than resolving the dissonance by changing their attitudes (cf, lanna &

SociaL Processes

651
sations or states may best be viewed as svrnptomatology data whose relationship to physiological reactions is open to empirical study. In a review of relevant literature, Blascovich and Katkin (1983) indicated that verbal reports of specific bodily states (e.g., resting HR) and reactions (e.g., increased HR) are generally inaccurate, and Pennebaker (1982) concluded that the causes of bodily reactions are typically reported incorrectly by people. Thus, verbal reports of specific physiological activity cannot be assumed to be redundant with actual physiological measures. Although the apparent dissociation between verbal and psychophysiological measures of specific bodily reactions quelled research in this area for decades, the recent spate of investigations on the operant conditioning of visceral activity has led to a renewed interest in the relationship between proprioception or interoception and physiological reactions (Brener, 1977). From the point of view of operant conditioning theorists, an individual's detection of a specific physiological change need not be accurate, as long as a perceptible event is correlated with the physiological change of interest to the investigator. For instance, individuals may not be able to discriminate between large changes in HR during mild exertion, but nevertheless may reliably detect increases in HR from momentarily correlated and perceptible changes in cardiac stroke volume. More recently, Mackay ( 1980) has contended that adjective checklists (e.g., the stress-arousal checklist) provide an integrative measure of physiological activity that is more representative of general physiological arousal than any single psychophysiological measure. In a review of verbal measures of physiological activity, Mackay (1980) cites two lines of evidence for this proposition. The first derives from research showing that verbal reports of arousal (e.g., as assessed using adjective checklists) correlate more highly with phvsiological composites (e.g., HR and skin conductance) than do pairs of physiological variables. The second line of evidence comes from a deductive rather than an inductive research strategy: Subjects are exposed to discernible external stimuli (e.g., noise) that are hvpothesized to alter physiological arousal, and verbal measures of physiological arousal are administered. Generally, self-reports of arousal have covaned with the expected effects of the independent variables on physiological activation. The validity and sensitivity of verbal measures of general physiological arousal are still-questionable for two reasons, however. First, self-report measures of arousal may be more highly correlated with the average of several physiological measures of arousal than with any single physiological measure of arousal, or than with any two physiological measures of arousal, because of reduction in measurement error that accompanies the use of multiple measures to gauge a single construct. Second, Zillmann and his colleagues

Cooper, 1974). Importantly, Higgins et al. (1979) found that the pill described as possibly causing a feeling of unpleasant sedation blocked attitude change, too. Overall, the feature of felt unpleasantness was significantly more effective in eliciting the rnisattribution of cognitive dissonance to the pill than was the feature of felt physiological arousal. These data suggest that the social conditions postulated to instigate cognitive dissonance create a perceptible organismic reaction that is felt to be unpleasant. Perhaps the oldest and most straightforward procedure, however, is to ask subjects to describe their bodily sensations, and this procedure has also been employed in research on social processes. For exampie, Shaffer (1975) was interested in assessing the symptomatological effects characterizing cognitive dissonance; however, rather than employing a rnisattribution paradigm, he simply asked subjects to report how much "frustration/mental discomfort" they felt. Some subjects had written an essay consonant with their initial attitude (low-dissonance group), and others had written an essay dissonant with their initial attitude (high-dissonance group). Shaffer found that subjects who had written the counterattitudinal essay reported feeling more frustration and mental discomfort than subjects who had written the proattitudinal essa y. Moreover, the within-cell correlation between reported frustration/mental discomfort and attitude evidenced a strong association (r =; .57); this is consistent with the notion that cognitive dissonance is subjectively perceived as an unpleasant state that is followed by the modification of attitudes in an attempt to eliminate the conditions (e.g., attitude-behavior discrepancy) that have caused the dissonance. Verbal reports have been promoted as measures of specific physiological changes as well. The JamesLange theory of emotions, for instance, assumed that people can at a minimum distinguish the idiosyncratic bodily reactions that accompany divergent emotional reactions. However, studies of the relationship between the physiological and verbal covariants of emotion revealed large discrepancies, and the accuracy and validity of verbal reports were called into question (e.g., Svz, 1926). Nisbett and Wilson (1977) have argued more recently that people cannot accurately report the causes (antecedents) of their emotions or behaviors. Their critique of verbal reports as data about psychological processes applies to stimuli arising from within an individual's body as well as to those emanating from the external environment. Although the experiments and interpretations offered by Nisbett and Wilson ( 1977) have been criticized on various grounds (see Ericcson & Simon, 1980; Sabini & Silver, 1981; E. R. Smith & Miller, 1978; White, 1980), their caveat regarding the interpretation of verbal reports nevertheless has important implications. Specifically, verbal reports concerning an individual's bodily sen-

~-

..--------------~
652
(Zillmann, 1979) have dearly demonstrated that reported arousal subsides more quickly following exercise than do changes in blood pressure and HR. Hence, independent variables that excite specific effector systems may also alter verbal reports of arousal, but this convergence is apparently lost soon afterward in most instances (see also Cacioppo, MarshallGoodell, & Gormezano, 1983; Lang, 1971; Mewborn & Rogers, 1979). Finally, even if the change in reportable states is assumed to be based in a physiological change. it would be an error to equate the two. since numerous _ physiological changes go unnoticed (e.g., Cacioppo & Petty, 1982a). Furthermore. situational and social factors can alter the bodily sensations that are heeded and encoded (Scheier. Carver. & Matthews, 1983). and the same reportable state may be accompanied by a variety of physiological changes (e.g., Schwartz. 1982). Hence, the validity of self-report and misattribution measures of actual physiological activity in studies of social processes cannot be accepted without strong reservations at present. In sum, the view that arousal is the sine qua non of physiological influences on social processes has stimulated interesting research on nonelectrophysiological measures of bodily processes. The accumulated research suggests that verbal measures of physiological activity (or activation), as well as assessments derived from the use of the rnisattribution paradigm, cannot be taken as indices of physiological activity per se; rather, they stand as indices of a reportable state that mayor may not be highly related to the targeted physiological process. Thus, the low intercorrelations among physiological, reportable, and behavioral responses to experimental treatments suggest that each of these types of responses harbors distinctive information that must be tied empirically and theoretically to social processes.
PROCESSES

'_

,
down the distinction between the arousal aspect of "drives" or "motives" and that of "emotion," and to suggest instead a continuum in the degree of activation of the individual. (p. 273) Malmo (1958) made a similar proposal, and both he and Duffy viewed the relationship between drivel arousal and performance as being in the form of an inverted U. Lindsley (1951) noted the electroencephalogram (EEG) activity obtained during sleeping and waking states roughly mirrors the behavioral activity of the organism. He suggested that the general arousal system of the brain may serve as the basis for behavioral arousal. The early research on the reticular activating system (RAS) indicated that this diffuse subcortical tract was intimately related to electrocortical activity; it was considered to be the instrument of the brain's arousal system (Moruzzi & Magoun, 1949; but see Vanderwolf & Robinson, 1981). Hebb (1955) related the RAS to generalized drive states, arguing that the former represented an anatomical mechanism for energizing behavior in a nonspecific fashion, whereas the latter was a hypothetical construct representing the same nonspecific energization of behavior. Hence, Hebb (1955) suggested that the neural basis of generalized drive was the RAS. Physiological responses (e.g., the EEG) were believed to reflect the level of activity in the RAS, and therefore an individual's level of drive at any given moment in time. Research spearheaded by the Laceys (cf. Lacey, 1967), however, indicated that the extent of covariation is low between indices of autonomic and behavioral activity, and between measures of autonomic and electrocortical excitation. Specifically, electrocortical, autonomic, and behavioral measures tend not to act in a unitary fashion except across vast differences in activity levels (e.g., deep sleep to ecstatic states), following highly unusual and salient external events (e.g., following an unexpected gunshot), or in metabolically demanding situations (e.g., during strenuous exercise). Even physiological measures obtained during physically non demanding tasks from effectors within a single system (e.g., the autonomic or somatic nervous system) do not covary highly (Cacioppo & Petty, 1981a; Fridlund, Cottam, & Fowler, 1982; Lacey, Kagan, Lacey, & Moss, 1963). Finally, as noted above, there can be large discrepancies between the facts of and the feelings arising from physiological changes. The same physiological reaction in two people can lead to distinctive reports (symptoms), because the individuals differ in their sensitivity and attention to detectable physiological reactions (L. C. Miller, Murphy, & Buss, 1981), the aspects of the detectable physiological responses they heed (Pennebaker & Skelton, 1981), their allocation of attention to detectable physiological reactions and external demands when performing a task (Scheier et al., 1983),

Focus

on

Response Syndromes and Systems

The third major response to the barriers to psychophysiological investigations of social processes emerged from research attacking the presumed pervasiveness and influence of general and diffuse physiological arousal. One of the major conceptions of physiological arousal derived from Cannon's early work on the flight-or-fight response. Cannon observed that organisms exhibit generalized discharges in the sympathetic nervous system in response to threatening stimuli. In part from Cannon's (1927) observations, Duffy (e.g., 1951, 1972) developed notions about ~tressors, autonomic arousal, and (adaptive) behavioral activation. According to Duffy (1957), Measurement of the physiological indicants of arousal affords. when other factors are constant. a direct measure of the "motivating" or "emotional" value of the Sltuatlon of the indiVidual. The concept serves to break

Social Processes

653
continued use of psychological explanatory constructs both superfluous and misleading" (p. 434). Although we do not anticipate the rejection of psychological constructs in favor of biological constructs to explain social processes, at least in the foreseeable future. there has been a clear decline in the dismissal of physiological factors as irrelevant. Moreover, the nonelectrophysiological procedures developed by investigators to study the effects of diffuse and perceptible physiological arousal on social processes have raised interesting questions regarding the actual physiological basis for the obtained data and observed individual differences. In turn, such questions have stimulated research, for example, on the symptoms and sensations people associate with various patterns of somatic and autonomic responses (Blascovich & Katkin, 1983; Leventhal & Mosbach, 1983; Pennebaker, 1982; Scheier er aL., 1983). Finally, it should be noted that psvchophvsiologists are taking an increasing interest in social factors not because they necessarily want to identify rudimentary social processes, but rather because of the dramatic effects social factors can exert on an organism's reactions to nonsocial stimuli (e.g., Gantt, Newton, Royer, & Stephens, 1966; Lynch & McCarthy, 1969; d. Pribram, 1962). This interest has grown for two reasons. First, since social factors may alter the psychophysiological reactions to nonsocial stimuli, comprehensive psychophysiological models must deal with these factors (Cacioppo & Petty, 1983c). Second, consideration of social factors has contributed to the construction of more sensitive, artifact-free psychophysiological experimentation. Gale and his colleagues, for instance, have consistently emphasized the importance of creating a comfortable psychosocial environment for psychophysiological experimentation-one that is free of demand characteristics, experimenter effects, and evaluation apprehension (Gale, 1973; Gale & Baker, 1981; Gale & Smith, 1980). Since this research represents more the application of social psychology to psychophysiological experimentation than the application of psychophysiology to the study of social processes, it is not reviewed here. Interested readers may wish to consult discussions of this work in Christie and Todd ( 1975) and Gale and Baker (1981).

and the hypotheses they carry about what internal sensations represent physiological changes such as "arousal" and "hypertension" (Leventhal & Nerenz, 1981 ). Psychophysiologists and social psychologists alike have wrestled with integrating these viewpoints: The effect of attempting to assimilate all of these traditions to a single arousal theory was to create a model in which the reticular activating system was assumed to serve as a generalized arousal mechanism which responded to sensory input of all kinds, energized behavior. and produced both EEG and sympathetic nervous system activation. . . . As is well-known, this model failed the empirical test rather badly. (Fowles. 1980, p. 88) The reactions of many frustrated investigators are represented in Kiesler and Pallak's (1976) comments: We use the terms motivation (drive) and arousal loosely and interchangeably .... We recognize the continuing controversy regarding these concepts in the literature .... Our simplistic use of the terms does not imply a theoretical stance on our part, but rather reflects the state of the art in social psychology. (p. 101S) Multicomponent arousal models emerged to account for these nuances (e.g .• Fowles. 1980, 1982; Routtenberg, 1968), and the conceptualizations of independent dimensions of physiological activation accent the dubious nature of using single physiological indices in social-psychological studies to assess "physiological arousal.' , Therefore, despite the past importance of the nonspecific, unidimensional concept of arousal in psychophysiological studies of social processes, there is a decreasing tendency for investigators to assume that general physiological arousal. specific physiological measures, reportable levels of felt arousal, and overt behavioral intensity necessarily or typically covary. Instead, these constructs are increasingly being treated as elements of perhaps independent stages whose interrelationships, and whose positions in social processes, are subject to empirical and theoretical investigation. Studies of the incipient and transient patterning of facial muscles during social interaction (e.g., see reviews by Fridlund & Izard, 1983; Schwartz, 1975) and excitation transfer (e.g., see Zillmann, 1979, 1984) are illustrative. Convergence of Approaches There is already evidence of a convergence among the three research strategies outlined above. Schachter (1978), for example, has recently explained that the essential point of his early position on emotion and arousal was that "most complex behavior cannot yet be understood in purely biological terms" (p. 433); however, he has come to believe "that there are areas about which we know sufficient biology to render the

ANALYTIC FRAMEWORKS
Although a social-psychophysiological perspective is not a privileged pathway to understanding social or physiological processes, contemporary research has highlighted several advantages that the applications of psychophysiological procedures can yield when used in clearly conceived experimental studies of social processes. First, social-psychophysiological research

654
has helped to advance our understanding of the determinants of people's physiological responses and the operation of physiological mechanisms by expanding the set of independent and dependent variables in psychophysiology to include powerful social factors (e.g., Cacioppo & Petty, 1983a; Shapiro & Crider, 1969). Second, this viewpoint can lead to the discovery and ultimately to the explanation of instances of complex human behavior that are shaped by a combination of social, dispositional, and physiological factors (e.g., Leventhal, 1980). Third psychophysiological procedures can provide means for assessing the construct validity of theoretical concepts in the social sciences (e.g., Cook & Selltiz, 1964; Tursky & [amner, 1983). Fourth, a general social-psychophysiological perspective can lead to discoveries in applied areas-for example, behavioral medicine-as the regulation (or deregulation) of the human organism is viewed within a broader social context (Henry & Stephens, 1977; Schwartz, 1983; Spitzer & Rodin, 1983). Finally, this orientation can lead to a greater specification and refinement of existing theories of social processes and to the development of new theories as the elementary stages of social processes are isolated and identified. As we have noted, it is this last focus with which we are primarily concerned in this chapter.

PROCESSES

In Search of Elementary Stages of Social Processes Psychophysiology as Description: The Molar Method
In one of the earliest empirical studies of social interaction, Boyd and DiMascio (1954) evaluated a psychotherapeutic interview of a single patient (diagnosed "schizophrenic, paranoid type") using the Bales Interaction Scoring System and measures of HR, GSR, and facial temperature. Boyd and DiMascio reported that positive and negative verbal exchanges increased while neutral exchanges decreased as the interview proceeded. They also noted that skin resistance dropped; HR increased, then slowed; and facial temperature rose, then dropped during the interview. The investigators suggested, At the beginning of the interview sympathetic tension was elevated immediately and was accompanied by profuse "neutral" interaction: as the interview proceeded the patient expressed more feelings (cf less neurral) and these were accompanied by reduced productivity and sympathetic relaxation. (Boyd & DiMascio, 1954, p.211) However, these physiological changes were not clearly related to changes in the behavioral interaction as the interview progressed, and no comparison conditions ~ere included to determine what aspects of the social lnteraction caused particular physiological and behavioral responses, what aspects were reliable, or what

aspects were unique to persons with thought disorders. The significance of these data for understanding the rudimentary processes operating in psychotherapeutic interviews (or social exchanges generally) is meager, although these data do furnish a description of the psychophysiology of a highly limited and possibly idiosyncratic event. How, then, are the constituent stages of various social processes to be analyzed? Note, first, that the term "stage" is used here in an operational rather than a functional sense (cf. A. F. Sanders, 1980). Although any complex social process is assumed to consist of more than one sequentially ordered processing stage, it is always possible that a stage is, in fact, comprised of a set of yet more elementary operations (e.g., unidentified substages), which mayor may not be ordered sequentially. Hence, a stage can be viewed as an empirically identified link in a chain of intervening components of a social process. Not surprisingly, the identification of the fundamental stages of various social processes through the application of psychophysiological concepts and techniques depends upon the reigning experimental logic. For example, randomization, replication, and multiple operationalization are as fundamental to social-psychophysiological investigations as they are to social and behavioral experimentation (cf, Campbell & Stanley, 1963). Investigators may wish to interpret an observed physiological response as "a nonreactive index," "the physiological basis," or "an integral component" of a social process. However, because social constructs, stages, and processes are abstractions based upon observable (e.g., physiological, verbal, behavioral) events, it is generally necessary to utilize an experimental design that includes several conditions, none of which uniquely identifies the effect of the stage under study, but which taken together define such an effect (cf. Pachella, 1974). In this section, two general analytic frameworks are outlined for designing and interpreting psychophysiological investigations of the stages of social processes. With regard to the design of the experiments, the first framework, based upon the subtractive method, begins with a theoretical sequence of stages that is thought to influence some social behavior. The second framework, based upon the additive-factors method, makes no assumptions about the nature of the constituent stages of a social process. It is possible, however, that one might design an experiment using one framework, but interpret the obtained data using another. For example, even though the logic of the subtractive method may have been used to construct the conditions in a multifactor experiment, the logic of the additive-factors method could be used to search for possible differences in processing stages should no apparent differences be obtained on the verbal or behavioral outcome variable of interest (e.g., recall, percentage of correct responses on a task, attitude change). Alternatively, if differences between two con-

Social Processes

655
cal activity resulting from the isolation of the~e ~resumably identical stages of processing are sirnilar, then convergent evidence is obtained that the same fundamental stage is a part of both processes, although these data do not prove that the stages are the same. The more peculiar the physiological profile is to a given stage within a particular experimental context; however, the greater the value of the convergent evidence conferred. Monitoring phasic changes in skin resistance alone, for example, provides weak convergent evidence for a particular psychological stage, since it is characteristic of so many divergent stages (e.g., see Schwartz & Shapiro, 1973). There are two additional concerns that should be considered when using this framework to investigate an elementary stage of a social process. First, the subtractive method contains the implicit assumption that a stage can be inserted or deleted without changing the nature of the other constituent stages. But this method has long been criticized for ignoring the possibility that manipulating a factor to insert or delete a processing stage might introduce a completely different processing structure (Kulpe, 1895). This problem can be attenuated by using multiple operationalizations to insert or delete a stage. If each operational insertion or deletion of a stage alters the processing structure of the task differently, then no distinctive features attributable to the conceptual processing stage of interest should be observed across operationalizations, and alternative experimental operations and means for isolating the stage will have to be devised. Parametric studies of a processing stage can also provide important information about the range over which a stage is manifested as a particular physiological profile, thereby improving an investigator's ability to generate appropriate comparison conditions and predictions, and to draw clear interpretations. A failure to find the same profile of physiological response across a wide range of levels of a social stimulus believed to invoke a given processing stage does not in itself indicate whether a new stage is invoked or whether the old stage is manifested differently within the organism at various levels of stimulation; it does clearly indicate an important limitation when one is interpreting the physiological profile obtained in a subsequent study of this processing stage. As Donchin (1982) suggests, "each hypothesis so tested generates predictions for its own specific range of validity. The observed relations mayor may not be universally applicable" (p. 460). Second, in order to construct the set of comparison tasks using the subtractive method, one must already have a clearly articulated hypothesis about the sequence of events that transpires between stimulus and overt response. This assumption renders the subtractive method particularly useful in testing an existing theory about the stages constituting a social process and in determining whether a given stage is

ditions emerge on the outcome variable of interest, then the logic of the subtractive method may be invoked to help interpret the significance of the accompanying psychophysiological data. Hence, these analytic frameworks serve as complementing heuristic aids.

The Comparison of n with n - 1 Stages: The Subtractive Method
At the simplest level, experimental design begins with an experimental and a control condition. The experimental condition represents the presence of some factor, and the control condition represents the absence of this factor. The experimental factor might be selected because it is theoretically believed to harbor n stages of a social process; the construction of the control condition, on the other hand, is guided theoretically to incorporate n - 1 stages of the social process, and hence to yield a different impact on social cognition (e.g., attitudes) or behavior (e.g., supportive action). The principle underlying the inclusion of physiological measures in these designs is that the physiological character, or "markers," of the various constituent stages of a social process might be deduced through the systematic application of the procedure of stage deletion. Thus, this approach is sometimes termed the "subtractive method" (Donders, 1868). If, for example, differences in performance result when individuals perform a task while observed versus unobserved, then the presence of an observer can be conceived as introducing one or more additional processing stages. If a theory including this stage or stages is posited to explain people's performances, then comparison conditions can be constructed and physiological measures obtained to determine the physiological markers of this and each of the accompanying processing stages. When, in a subsequent study, one of these stages is thought to be responsible for the differential impact of two conditions on behavior (e.g., in a study of social influence), then analyses of the concomitant physiological activity can again be informative, in one of two ways. 1. If the patterns of physiological activity resulting from the isolation of presumably identical stages are dissimilar, then the similarity of the stages is challenged, even though there may be similarities between the subsequent behavioral outcomes. The greater the evidence from multiple operationalizations that a particular stage is accompanied by a specific profile of physiological responses across the ranges of the stimuli employed in the investigations, the more challenging is the dissimilarity in obtained physiological profiles (i.e., the failure of this template matching when two possibly similar stages are isolated and compared). 2. If, on the other hand, the patterns of phvsiologi-

,.------------~-------656
among the set constituting two separate social processes. Note, however, that confirmatory evidence can still be questioned by the assertion that the addition or deletion of a particular stage results in an essentially different set of stages. Again, the inclusion of several experimental and comparison tasks [i.e., multiple operationalizations) appears to be the judicious action regarding this criticism, which can always be applied and can never be disproven (cf. Pachella, 1974). The subtractive method can also yield data that is helpful in deriving comprehensive, empirically based models of the foundations of social processes. Again, assume that the comparison is between experimental conditions with differential impacts on behavior as a function of some social stimulus. If the physiological profile that differentiates these conditions is similar to a distinctive physiological profile that has been found previously to characterize a particular processing stage, then the possibility is raised that the same processing stage has been detected in another context, although one cannot logically conclude that this stage has definitely been detected. The stronger and more distinctive the link between a physiological profile and a processing stage within the ranges of stimuli employed, however, the stronger this possibility. Finally, if other (e.g., behavioral) sources of converging evidence are also obtained in the experiment or from subsequent experimental research, then the plausibility of this inference is enhanced. Similarly, if the sequence of physiological activity differentiating two conditions can be modeled by concatenating the physiological responses found previously to characterize even more rudimentary types or stages of a social process, then a model of a set of stages distinguishing these two conditions would be suggested, though certainly not confirmed. Again, converging evidence for this empirically derived model could be marshaled from other (e.g., observational, verbal) measures obtained in the experiment and from subsequent experimental studies. Monitoring incipient somatic activity at sites that have been associated with distinctive dimensions of information processing (e.g., affective, linguistic) in a persuasion experiment, for example, does not guarantee that these types of processes are reflected in the activity at the respective recording sites (e.g., corrugator or zygomatic muscle regions, perioral muscle region), but it does potentiate a more informative assessment than is provided by relying solely upon self-reports and/or subsequent behavioral measures (Cacioppo & Petty, 1981a).
PROCESSES

Generating Candidates for Stages: The Additive-Factors Method
An analytic framework adapted from the additivefactors method for isolating stages of informational flow (Sternberg, 1969; see reviews, critiques, and sug-

gestions by Pachella, 1974; A. F. Sanders, 1980; Schweickert, 1980) is particularly helpful in conceptualizing the potential set and sequence of stages constituting a social process when one is (1) designing a study with no strong a priori notion of what the constituent stages might be, or (2) interpreting psychophysiological data from conditions that yield similar results on the verbal or behavioral measure of interest (e.g., performance, aggression, attitude change). The basic assumption of this analytic framework, like that underlying the subtractive method, is that the recording interval during which a given social process unfolds is filled with a sequence of independent processing stages. Each stage receives the input from a previous step and performs a particular transformation on this input before passing it on to the next stage of the sequence. Differences in the output of a stage are interpreted as possibly being due to differences in the input from the preceding stage, which in turn can derive from either a different input to the inaugural stage or a different set of preceding stages. Thus, a stage is assumed to perform a constant function (e.g., informational transformation), regardless of the nature or duration of the stages preceding or following it. Recall that the operational definition of a "stage" that we have adopted structures the social process such that (1) all the stages of a social process are executed in a sequence (although concurrent operations may be executed within a processing stage-cf. A. F. Sanders, 1980)2; and (2) each stage has an impact on physiological, verbal, and/or behavioral responses (otherwise, it has yet to qualify as a processing stage). The psychophysiological profile obtained during a social process is viewed simply as the aggregate of the profiles characterizing each constituent stage plus noise (e.g., measurement error, individual differences, irrelevant physiological processes). Several interesting propositions can be derived from these assumptions. First, if the manipulation of an independent variable affects a specific processing stage with physiological manifestations, then the aggregate of the physiological profiles is expected to change accordingly. Second, if two (or more) independent variables affect different processing stages, each with physiological manifestations, then the independent variables are expected to produce independent effects on physiological activity. Finally, if two (or more) independent variables modify each other's effects on a processing stage (or set of stages), then the independent variables are expected to contribute interactively to the profile of physiological activity. Thus, the effect of the experimental factors on psv2. Schweickert (1980) has discussed a method for identifying the separate paths of concurrent processes using reaction times. but application of his concept of "critical-path scheduling" is beyond the scope of the present chapter.

.-

SociaL Processes

657
. ferences about potenditions. J The strength of the In, 'on of this method tial stages drawn from the apphcatl h ed only by '1 -an to psvc h op hvsi vsiologica d ata La be en anc h lausibility of design considerations that attenuate t e Ph f in ' le t e use 0 a1 ternanve explanations. For examp , 'of ore verSions terna 1 replications. which are tWOor m f in 0 t h e same experiment ditferi ittermg on 1Y I the sequenceraerol the experimental stimuli. and the use of multlP etk li tice trials prior to the experiment de~rease th~ I ~~: hood that a difference in phYSiological profiles tween two conditions is artifactual (e.g., is due to random error. sensitization. carryover effects, or pr~ctice effects). Moreover. as noted in the precedmg section, the use of multiple operationalizations and parametric manipulations reduces the likelihood that an obtained difference in physiological responding is attributable to the experimental manipulations' changing the entire sequence of processing stages, the unintentional and unnecessary confounding of a "psychological" with a "biological" factor, or the manipulations' exceeding the range of validity for the psychophysiological manifestation of a particular processing stage. In sum, the additive-factors analytic framework provides useful heuristics for exploring what the fundamental stages of a social process might be, even though no preconceptions exist regarding the nature of these stages and/or no overt verbal or behavioral differences are obtained on the measure of interest in an investigation of the social process. If a factor or set of factors produces no effect on the verbal or behavioral measure of interest but is differentiated physiologically, then the possibility is raised that the factor initiates distinct processes with potentially unique and important verbal and behavioral consequences that have not yet been identified. The application of psychophysiological procedures, therefore, again potentiates a more powerful analysis of the foundations of social processes than is obtained using verbal or behavioral measures alone. It should be emphasized, though, that the link between physiological profiles and psychological processing stages is far from perfect. Thus, subsequent experimental research (e.g., application of the subtractive method) should follow the application of the additive-factors analytic method to test specifically the models that emerge.

chophysiological activity is assumed to be the sa regardless of the level of the others, if these factr;:;~ pertain to separate stages of a given social process or if they have an additive impact on the same processing stage(s) '. However, their effects on psychophysiologi_ cal activity are expected to be interactive if they modify each other's influence on a common processing stage or set of stages. In interpretations of psychophysiological data using this analytic framework, these propositions are reversed to yield the following heuristics: (1) If two or more experimental factors each exert a main effect on the same measure of physiological activity, then the effect of one factor would appear not to depend upon the level of the others; this suggests either that different stages are involved or that the effects of these factors on a given stage in this type of processing is strictly additive. (2) If two or more experimental factors exert main effects on distinctive aspects of physiological activity, then different processing stages, possibly reflecting different types of processing (e.g., linguistic, visual), are suggested. (3) If two or more experimental factors interact in their effects on physiological activity, then the factors may affect at least one common processing stage. Thus, when an experimental factor, or set of factors, affects the physiological responses associated with a particular social event (e.g., a persuasive communication) without altering the outcome measure (e.g., attitude change), it raises the possibility that the nature of one or more of the constituent stages of the social process has also been changed, with at present unobserved consequences for social behavior (e.g., attitudinal persistence). It would, of course, be an error to assume that the inferences derived using these heuristics are logically compelling, even should the initial assumptions of this method prove to be correct. If the manipulation of an experimental factor results in similar outcomes on the target behavior but different outcomes on physiological profiles, then it is possible that a factor influencing a processing stage (or set of stages) has been identified; if this possibility is supported in subsequent experimental research, then application of this analytic method will have contributed to the delineation of a social process by highlighting the role of a factor and identification of a stage that may otherwise have erroneously been overlooked or deemed irrelevant. It is also conceivable, however, that varying the level of the experimental factor redefines the entire sequence of the constituent stages, yielding a different social process; that the levels of the manipulated factor fall beyond the range of validity for a given stage manifesting within the organism in a particular fashion; or that physiological noise (e.g., an irrelevant but confounded biological process such as cardiac-somatic coupling, or measurement error) contributes differentially to the physiological profiles across con-

Summary
The analytic frameworks outlined above place a special value on basic studies of psychophysiological mechanisms. First, the more factors that are found to
3, The inclusion of "noise" in the foregoing expression for the aggregate physiological manifestation of a social process is intended to represent clearly that not every difference in physiological response can be assumed to represent a distinctive processing stage or to have behavioral implications.

-_-_._-_.

-~--------------------..,...--~-

658
affect specific aspects of physiological acnvirv, the better our understanding of possible links between elementary stages constituting a social process and the obtained psychophysiological profiles. For example, experiments that outline various psychophysiological relationships and their ranges of validity, research identifying physiological responses that reflect solely biological processes or individual differences, and procedures for attenuating measurement error all aid investigators interested in applying psychophysiology to the study of social process. Specifically, they help in the construction of informative comparison conditions, selection of physiological measures to record, and interpretation of the obtained data. Second, since both analytic methods conceive of a social process as being composed of a sequence of processing stages that are perhaps differentiated physiologically, it is necessary to understand the manner in which the constituent profiles of physiological responses combine to form an aggregated physiological response profile. Psychophysiological concepts, such as the law of initial value, habituation, and sensitization, are of concern. For example, although the law of initial value applies to skin resistance responses (SRRs), it does not generally apply to measures of skin conductance responses (SCRs; see Hord, Johnson, & Lubin, 1964). These kinds of observations are important, because they can point to conditions where the assumption that the aggregated physiological response profile represents the simple addition of responses characterizing each constituent stage is dubious, or, as in this example, where the extraction of particular measures (e.g., skin conductance rather than skin resistance) allows clearer interpretations to be drawn. Finally, psychophysiological research on the limits of physiological response is of help to investigators who, when faced with accepting the null hypothesis concerning the main effects of or interactions between two or more independent variables, must select among a number of options. They should be able to determine whether the failure to find significant effects is attributable to a ceiling effect for physiological activity, an absence of power in the statistical tests, too meager a sample of physiological responses to draw firm conclusions, or the true absence of main or interactive effects of the factors on a broad array of at least one class of physiological measures (e.g., somatovisceral, electrocortical).

PROCESSES

studies of social processes. Two caveats are in order. First, our review is meant to be illustrative rather than comprehensive. Second, although most of the pertinent studies reveal diligence in the construction of the comparison tasks, much weaker rationales exist for the physiological measures selected in most of these studies. Many reflect the investigators' interest in consequent verbal or behavioral data rather than the special utility of the selected physiological measure( s) for isolating the stages of a social process, and rarely has a study been designed or interpreted with either of the analytic frameworks described above in mind. Hence, this review is neither definitive nor designed to settle any pressing theoretical issues. Rather, it is meant to illustrate the potential of applying psychophysiological concepts and procedures to identify possible candidates for, and to test specific hypotheses about, the building blocks of social processes. Selected for review are two of the oldest and most vigorous areas of research in social psychology. The first, which deals with interpersonal social processes, concerns the effects on performance of being observed. The second, which deals with intrapersonal social processes, pertains to the formation, maintenance, and modification of attitudes.

Observed versus Unobserved Performance
In 1898, Triplett reported that people performed a well-practiced motor task (e.g., bicycle riding) more quickly when others were engaged in the same task than when alone. Several years later, Meumann (1904; cited by Cottrell, 1972) reported that a subject's performance on a simple motor task improved when observed by the experimenter. In a follow-up study, Meumann found that the mere presence of a passive observer in the subject's room was sufficient to improve the subject's performance. These early observations prefigured what is now a fundamental if not surprising principle in social psychology; An individual's performance in an endeavor can be altered dramatically simply by moving the task from a nonsocial to a social context. The next six decades of research on the effects of the presence of others produced an inconsistent picture. On some occasions, the mere presence of an observer or coactor enhanced an individual's performance on a task, whereas on other occasions the presence of others interfered with the individual's performance. Zajonc (1965) organized much of this research by proposing that the presence of members of the same species facilitates the single most likely response from an organism at that point in time, rather than the correct response. The mere presence of another member of the species, therefore, was viewed as facilitating the performance of easy tasks and impairing the performance of difficult tasks (i.e.,

PSYCHOPHYSIOLOGICAL CONTRIBUTIONS TO THE DELINEATION OF SOCIAL PROCESSES
In this section, the guidelines outlined above are used to interpret psychophysiological data obtained in

Social Processes

659
physiological measure has been palmar sweating, obtained using either the palmar sweat index (which provides a rough indication of the number of active sweat glands at the finger tips; see Dabbs, Johnson, & Leventhal, 1968) or the sweat bottle technique (which involves measuring the conductivity of distilled water following the immersion of the hand in it for a few seconds; see Strahan, Todd, & Inglis, 1974). These measures are somewhat crude, but they have revealed increased palmar sweating immediately before or following the task in just over half of the published studies in which it has been obtained (cf. Geen, 1980; Moore & Baron, 1983). Geen (1977) provides an illustrative experiment. Eighty female subjects performed a set of eight difficult anagrams. Subjects performed the task either alone or while being observed by the experimenter. Of those being observed, some were told that the experimenter was watching "in order to evaluate the subject's task performance" (p. 715); others were told that the observations were being made "so that the experimenter would have a basis for advising the subject on ways to improve performance on a subsequent task" (p. 715); and still others were not told anything about the reason for the experimenter's observations. Time to solve the anagrams and change in palmar sweating from before to after the task served as the dependent measures. Geen found that subjects performed the task fastest when working unobserved, next fastest when observed by a "helpful" experirnenter, and slowest when observed by an evaluative or unexplained experimenter. Palmar sweating was affected similarly, with the highest levels found when an evaluative observer was present and the lowest levels exhibited when no observer was present. The physical versus concealed (but acknowledged) presence of an observer has seldom been manipulated in the same experiment. Instead, in the majority of the experiments documenting increased palmar sweating prior to the task, observers have been in the test situation with the subject (e.g., Carver & Scheier, 1981; Droppleman & McNair, 1971; Martens, 1969a, 1969b, 1969c); conversely, in the majority of the studies failing to find greater palmar sweating, observers have been concealed (Bargh & Cohen, 1978; Cohen, 1979; Cohen & Davis, 1973). In the only study in which palmar sweating was monitored before and during the performance of a task (copying German prose), the presence of an observer (the experimenter), who was seated clearly in view several feet in front of the subject, resulted in more palmar sweating prior to the onset of the task, but slightly less palmar sweating during the performance of the task (Carver & Scheier, 1981). In a study in which several measures of autonomic activity were monitored, Geen (1979) gave subjects success or failure feedback on a preliminary task, which they performed either alone or while being

tasks whose correct response is not dominant for the performer). The term "social facilitation" came to be used generically to refer to the facilitatory and inhibitory effects of conspecifics on task performance.

Mere Presence and Arousal
Zajonc ( 1965) based his analysis upon the notion that the mere presence of conspecifics increases an organism's level of general drive, which in the HullSpence formulation (Spence, 1956) increases the likelihood of the performer's emitting the dominant response to a task. Zajonc suggested that the mechanism by which the presence of others increases a person's general level of drive is physiological arousal. Zajonc ( 1965) acknowledged, however, that his postulate about the underlying physiological mechanism was a speculation based primarily on a few studies of the effects of population density on the secretion of adrenal corticosteroids in animals. More recently, Zajonc (l980a) has suggested that the presence of conspecifics introduces more uncertainty into the situation, and that increased physiological arousal and drive are innate responses to uncertainty. Zajonc has maintained throughout that it is "the mere presence of others that is the sufficient condition bringing these effects about" (Zajonc, 1980a, p. 38). This formulation led to a spate of research (e.g., see recent reviews by Borden, 1980; Geen & Gange, 1977; Moore & Baron, 1983; G. S. Sanders, 1981; Zajonc, 1980a). In a demonstrative study, Zajonc and Sales (1966) presented lists of words to subjects, whose task it was to identify each word as it was presented. Some of the words were presented repeatedly, whereas others were infrequently presented. This was done to manipulate the habit strength associated with each word. Zajonc and Sales ( 1966), who were working within the Hullian model of "response potential drive X habit," then asked subjects to identify which word was presented tachistoscopically; some subjects were observed and others were not. In fact, no word was actually presented to subjects on the tachistoscope, but Zajonc and Sales found, as they had expected, that subjects who were observed during this task reported seeing words with high habit strength (i.e., those seen frequently before the pseudo recognition task) more frequently and words with low habit strength less frequently than did subjects working alone. Other behavioral investigations have found that the physical presence of an observer ( 1) interferes with short-term recall in paired-associate learning but aids long-term recall; (2) improves performance on simple tasks and impairs the initial performance on complex tasks; and (3) facilitates rather than inhibits task performance following practice and mastery of these tasks (Geen & Gange, 1977). Physiological responses have been monitored in only a few of these experiments, and the most popular

=

660
observed by the experimenter, who was standing several feet behind them. Next, subjects performed a difficult paired-associates learning task. Subjects who were observed during the preliminary task were again observed, and those who completed the preliminary task alone performed the paired-associates task unobserved. Geen found that subjects who were with an observer performed more poorly on the difficult paired-associates task and displayed more spontaneous skin resistance responses (SSRRs) following failure feedback than did subjects who received failure feedback but performed the tasks alone; no differences were observed on the measures of HR or maximum skin conductance level (SCL), however. Whether subjects were observed or alone had no effect on performance or on either physiological measure when subjects had received success feedback following the preliminary task .. SCL and HR have been monitored simultaneously in similar experiments employing tasks ranging from listening to a persuasive communication (Borden, Hendrick, & Walker, 1976) to a pseudorecognition task (Henchv & Glass, 1968) and a reaction time task (Moore, 1977). In each of these investigations, the presence of observers failed to influence SCL or tonic HR. Kissel (1965), who only monitored SCL, found that it was actually lower when subjects worked on four insoluble tasks in the presence of a friend (who was working on an unrelated task in the same room) than when subjects worked on the insoluble tasks alone or in the presence of a stranger. The SCLs of subjects in these latter conditions did not differ significantly; this finding replicates those of Geen (1979), Borden et al. (1976), Henchv and Glass (1968), and Moore ( 1977), all of whom employed strangers rather than friends (see also Shapiro & Leiderman, 1967; Shapiro, Leiderman, & Morningstar, 1964).4 The influence of observers on forehead (frontalis region) electromyogram (EMG) activity has been examined in at least three other experiments (cf. Moore
4. A few additional studies are tangentially relevant to the present discussion. In an unpublished study, Shapiro (cited in Schwartz &. Shapiro, 1973) periodically recorded the skin potential level of subjects during a 15·min rest period, which they knew would be followed by a simple task. Schwartz and Shapiro ( 1973) note that "these subjects were known to each other, not friends but costudents in a nursing school" (p. 400). Some subjects were seated alone, and others were seated in groups of three. Subjects in triads were asked to refrain from interacting, though they were in visual contact with one another. Shapiro found larger decrements in skin potential level in subjects who anticipated the task in the presence of others rather than alone-a result reminiscent of Kissel's (1965; see also Back &. Bogdonoff, 1964). Finally, Latane and Cappell ( 1972), who monitored only HR, found that solitary rats exhibited higher HRs when a conspecific was temporarily placed in the openfield situation with them. However, if the placement of another rat In the situation increased the somatic activity of the subject, then the increased HR could be attributable to the increased somatic activity of the subject, rather than the physical presence of a conspecific (see Obrist, 1976, 1981).

PROCESSES

in a supine position and were instructed tokeep their eyes open while listening to a recording of a humorous (Chapman, 1972; cf. Chapman, 1973) or mystery (Chapman, 1974) story. Wallerstein ( 1954) had previously found that attentive listening to recorded stories caused forehead tension to increase, and Chapman replicated this effect. Furthermore, Chapman ( 1974) varied whether or not the subject was observed, and in the latter condition whether the observer was physically present or concealed. He found that subjects exhibited steeper forehead EMG gradients when observers were either present or concealed than when subjects listened to a story unobserved. There was also a nonsignificant tendency for the physically present observer to evoke higher levels of forehead EMG activity than did the concealed observer. In another study, Musante and Anker (1972) instructed subjects that their task was to inhibit all movements when three tones were sounded. Subjects were then exposed to the tones either while the experimenter watched or while alone. Under these conditions, Musante and Anker (1972) observed lower levels of forehead EMG activity following the tones when in the presence of another than when alone. These physiological data have been somewhat difficult to characterize, for the following reasons: (1) No study has been designed specifically to isolate the elementary stages of the process underlying the effects on performance of being observed; (2) few studies have simultaneously recorded multiple measures of physiological activity; (3) no study has reported the intermeasure correlations between physiological and performance variables'': and (4) most studies have been conducted under the assumption that one physiological measure is as good as another (or several) when indexing physiological activity. Nevertheless, these studies do not support the conception that the behavioral effects of being observed are mediated by general physiological or autonomic arousal (cf, Vanderwolf & Robinson, 1981). Furthermore, no study has ever explicitly tested Zajonc's ( 1965) original speculation that the mere presence of others influences behavior through its effects on adrenocortical activity. Indeed, the adrenomedullary complex, which is innervated by and affects the sympathetic branch of the autonomic nervous system, appears a more likely candidate for the underlying physiological mechanism of
5. In one of the few related studies in which a physiological index and performance were correlated, Church (1962) tested 92 subjects in a simple reaction time task under competitive and noncompetitive testing conditions. Church found that subjects responded more quickly, reported being more alert, and exhibited higher SCLs when tested under competitive conditions. These measures were not correlated, however. "Put simply, the physiology does not seem to 'mediate' the overt performance, while both are separately. affected by the nature of the social setting" (Schwartz &. Shapiro, 1973, p. 399).

& Baron, 1983). In two studies, subjects were placed

.-

'P'

s
SociaL Processes

7

661
hibition of overt practice (Berger, Carli, Garcia, & Brady, 1982; Berger, Hamptom, Carli, Grandmaison, Sadow, Donath, & Herschlag, 1981). Moreover, the notion that the mere presence of conspecifics is su~i~ cient to produce these behavioral effects by enhancing drive remains a fundamental postulate in several of these formulations (Markus, 1981; Zajonc, 1980). Do the accumulated physiological data do anything more than question Zajonc's (1965) initial speculations regarding the physiological mechanism underlying the facilitatory and inhibitory effects of observed performances? We believe so. Since most of these models do not make or differ in their psychophysiological predictions, we do not begin this analysis with any preconceptions regarding the component processing stages that may emerge from a reconsideration of the accumulated physiological data in light of the heuristics outlined above (see the discussion of analytic frameworks ). To begin with, recall that although the presence of an observer was not found to affect HR, it was found that palmar sweating and time to perform a difficult task were greatest when an observer was described as evaluative, moderate in levels when the observer was described as watching in order to offer helpful suggestions later, and least when there was no observer present during task performance (Geen, 1977). Consider also the conceptually similar finding that SSRRs and errors on a difficult task were more frequent when the subject was observed performing a task associated with failure than when the subject was alone or observed performing a task associated with success, and that measures from these latter two conditions did not differ (Geen, 1979). These experimental conditions, when viewed as converging operations yielding similar physiological results, suggest a fundamental stage in social facilitation. In adducing the characteristics of this stage, it is of interest to note that "EDA responds to threatening stimuli, that this response is not attributable to somatic activity, and that-at least in many cases-HR does not respond to these stimuli" (Fowles, 1980, p. 93). Katkin and his colleagues (Hirschman & Katkin, 1971; Katkin, 1965, 1966, 1975; Rappaport & Katkin, 1972), for example, report finding that "the creation of psychologically stressful situations results in a substantial increase in the frequency of spontaneous electrodermal responses" (Masling, Price, Goldband, & Katkin, 1981, p. 396). The link between the EDRs and the punitive consequences of a stimulus is unlikely to be invariant (e.g., novel stimuli also elicit EDRs; d. Venables & Christie, 1980), and the appearance of the consequence (e.g., SCRs, SSRRs) does not prove the presence of the antecedent (threats of punishment). The link between EDRs under similar conditions (e.g., Block, 1981; Masling et al., 1981; Szpiler & Epstein, 1976) simply raises the possibility that the processing stage invoked by the introduction of an

!

"social facilitation effects" than the hypothalamicpituitary-adrenal cortex axis posited by Zajonc ( 1965). Zajonc based his speculations upon evidence that crowded conditions in animal colonies lead to increased secretion of adrenocortical steroids. But the actions of the hypothalamic-pituitary-adrenal cortex mechanism are tonic rather than phasic: The physiological and behavioral effects of adrenocortical steroids are slow to develop, are slow to dissipate, and operate primarily to increase an organism's readiness for energy expenditure ("arousability") rather than present activity at sympathetically innervated effectors (Zillmann, 1984). The release of catecholamines (I.e., epinephrine) from the adrenomedullary cornplex, on the other hand, occurs more quickly in re~ sponse to a stimulus, has been associated with psychological stressors (in contrast to norepinephrine, which has been associated more with physical stressors), contributes to elevated activity in the sympathetic nervous system, and dissipates within a short period of time. As Zillmann (1984) has noted, , , . the level of adrenomedullarv activity may vary quasi-tonically throughout the day (rising during work periods and declining during recumbency ... ), the energizing hyperglycemic effect of medullary secretion is essentially phasic. Generally speaking, the energy supplied is absorbed by one episode of vigorous, strenuous action-or by just a few such episodes at the most.
(p. 112)

No published study has explicitly addressed the notion that the presence of others affects the phasic actions of the adrenomedullary system or of the hvpothalamic-pituitary-adrenal cortex mechanism. Nevertheless, what is known at present about the behavioral effects of these endocrinological mechanisms points to the adrenomedullary system as more likely than the hypothalamic-pituitary-adrenal cortex system to be involved in producing the short-latency behavioral effects observed in social facilitation research. Consistent with this reasoning, Singer, Lundberg, and Frankenhaeuser (1978) found that perceptions of control were related to catecholamine response to crowding on commuter trains.

II

Reconsideration
The behavioral data that have accrued since Zajonc's ( 1965) conceptualization of "social facilitation" have also led to a spate of theoretical work. A number of alternative explanations have now been proposed to account for the inhibitory and excitatory effects of observers, including learned drive (Cottrell, 1972), a modified conception of learned drive (Geen, 1980; Weiss & Miller, 1971), objective self-awareness (Duval & Wicklund, 1972), distraction/conflict (Baron, Moore, & Sanders, 1978; G. S. Sanders, 1981), self-presentation (Bond, 1982; Borden, 1980), self-consciousness (Carver & Scheier, 1981), and in-

----------

662
observer gauges the potential punitive consequences aroused by this intrusion upon the organism's environment. The characterization of this processing stage as threat assessment gains plausibility only when consideration is given to the nature of the experimental conditions that led to these and related physiological results (e.g., Kissel, 1965, concerning SCL and the presence of friends vs. strangers; Fowles, 1982, concerning HR and positive incentives). It seems more plausible, for example, that the presence of an evaluative experimenter is more threatening but no more novel to a subject than the presence of an experimenter who is observing a subject's performance in order to be helpful. Moreover, the finding that an experimenter who simply watched subjects perform had the same physiological and behavioral effects as an observer who was explicitly described as being evaluative suggests that observers in an experimental context are assumed to be evaluative (Geen, 1977). The palmar sweating data associated with the salience of the observer. on the other hand, may simply illustrate the limitation of inferring processing stages from insensitive physiological data. The notion of an early threat assessment stage in social facilitation leads to an expectation regarding the effects of concealed versus unconcealed observers: The more threatening the observer. the clearer the physiological effects should be, and the stronger the behavioral effects should be. Thus. the presence of the observer may serve as a weaker manipulation affecting the same processing stage as the description of the observer as evaluative versus helpful. The data are not in complete accord with this expectation. Palmar sweating was generally higher and facilitatory-inhibitory behavioral effects were obtained when a subject performed a task while observed by someone present rather than while alone. However. no differences in palmar sweating were generally obtained (even though the behavioral effects emerged) when subjects performed the task while being watched by a concealed observer rather than while alone. There are two obvious but very different interpretations for these results: (1) Given that the patterns of physiological activity resulting from the isolation of what are presumably identical processing stages appear dissimilar, one might conclude that the processing stages are, in fact, dissimilar. (2) Alternatively, the experimental conditions may have isolated different levels of the same processing stage. but the physiological "index" for gradations in the processing stage may have been insensitive or invalid across this range (see pp. 655-656).0
6. The same argument applies to the inconsistent results regarding SCLs. Nonspecific fluctuations and electrodermal levels are viewed as being relatively independent (Katkin, 1965; Szpiler & Epstein. 1976). In reviewing these data. however. Fowles (1980) suggested that SCL may simply be less sensitive: "[I]t seems more likely that electrodermal level is inconsistently related to anticipation of punishment because of peripheral physiological limitations" (p, 95).

PROCESSES

Which interpretation is to be favored? Unlike the first, the second interpretation hinges on the assumption that because of the gross nature of the palmar sweating index. small differences induced by the (less threatening) presence of a concealed (in contrast to unconcealed) observer may not have been noticeable. That is, palmar sweating, which reflects surface sweating on the palms, may not have been a sufficiently sensitive index to gauge small but behaviorally significant changes in perceived threat. Two additional predictions follow uniquely from the second explanation: (1) More sensitive physiological measures should differentiate the conditions where subjects were observed by an unconcealed versus concealed observer; and (2) social facilitation effects should be stronger when the observer is unconcealed than concealed. In an experiment pertinent to the former hypothesis, Chapman ( 1974) monitored forehead EMG activity as subjects listened to a story while unobserved, watched by a concealed observer, or watched by an unconcealed observer. As would be expected if palmar sweating were simply insensitive to subtle differences in the potential punitive consequences aroused by the observer, Chapman found forehead EMG activity to be higher during the story when the subject was observed than when unobserved, and slightly though not significantly higher when the observer was present than when concealed. Whether an observer is concealed or not has not generally peen viewed as informative, since the facilitatorv or inhibitory effects of being observed have emerged in both instances (Markus. 1978; cf. Geen, 1980). Unfortunately, no performance data were provided by Chapman (1974). and it is uncertain from the existing psychophysiological studies whether the behavioral effects on performance would be strengthened if a physically present observer were contrasted with a concealed observer. Several experiments, however. provide tentative evidence for a related prediction: that the "attentiveness" of an observer influences the behavioral effects of observed performances. Strube. Miles, and Finch (1981) recorded the running speed of joggers who were in the presence of an attentive or an inattentive observer. They found that the joggers with an attentive observer ran around the track more quickly than did the joggers with an inattentive observer. In another study, subjects performed a pseudorecognition task alone, in the presence of two inattentive and blindfolded persons, or in the presence of two observers (Cottrell, Wack, Sekerak, & Rittle, 1968). Subjects in the presence of the blindfolded observers performed similarly to the subjects who were alone, whereas subjects in the presence of observers reported seeing stimuli with higher habit strength than those in the other conditions. Finally, Cohen and Davis (1973) taught subjects to solve hidden-word problems in a particular manner and then shifted the test items so that a more efficient problem-solving strategy became available.

........... ---------------Social Processes

663

Subjects in some of the conditions in their experiment performed the task unobserved or while being watched through a one-way mirror by nonevaluative or evaluative observers. Cohen and Davis (1973) found that the concealed observer described as being evaluative inhibited subjects most from finding this new problem-solving strategy, that the concealed but less evaluative observer inhibited subjects less often, and that unobserved subjects were the most likely to find the more efficient problem-solving strategy (see also Henchv & Glass, 1968; Sasfy & Okun, 1974). At present, therefore, parsimony would appear to favor the adduction from these data of one stage, which functions to gauge the potential punitive consequences aroused by the introduction of a conspecific. The remaining physiological data suggest a second processing stage bearing upon energy expenditure (e.g., efforts to emit socially valued behaviors for personal advancement within the group). The existence of this second processing stage is suggested by the experiments in which somatic activity was monitared. Chapman's (1973, 1974) experiments, for instance, demonstrated that a somatic response pre~ viouslv found to covary with the performance of a task (listening to a story) was found to be of greater magnitude when the subject was observed than when not observed. Moreover, Musante and Anker ( 1972) provided evidence that when trying harder to perform a task meant relaxing a specific muscle region, then the one somatic measure obtained (frontalis EMG) was lower when the subject was observed than when not observed. Finally, the somatic changes obtained by Chapman (1973, 1974) and Musante and Anker ( 1972) were exhibited at the onset of the task rather than in anticipation of the task; this result further suggests that the presence of the observer affects how hard subjects strive to perform a specific behavior, rather than how tense or generally aroused they become. Thus, this processing stage may not reflect a change in the generalized drive (see criticisms of this concept by Bond, 1982; Zillmann, 1984), but rather a stage of effortful striving, based in heightened svmpathetic activity and catecholamine circulation, that fosters short-term demands for vigorous activity (e.g., fight, flight, or task performance). In other words, it is possible that the presence of observers heightens sympathetic tonus and efforts to deal with sometimes conflicting environmental demands." Evidence that the underlying processing stage may be better characterized as "effortful striving" than as generalized drive is provided by several recent experiments (e.g., Bond, 1982; Groff, Baron, & Moore, 1983). For example, Bond (1982) found that subjects who were observed performed better on difficult tasks embedded within the context of simple tasks when observed than when unobserved, whereas subjects performed simple tasks
7. This processing stage may correspond to the excitatory component in Zillmann's (1983, 1984) three-factor theory of emotion.

embedded in a pool of difficult tasks more poorly when observed than when unobserved. Note, too, that given both the timing of the physiological changes (see also Carver & Scheier, 1981) and the nature of threat assessment and effortful striving, it seems reasonable to suggest that (1) these two stages have a special propensity for combining, and (2) the decision regarding the potential punitive consequence (or threat) aroused by the presence of the observer precedes and influences the mobilization of temporary energy resources (i.e., effortful striving)." Finally, although these stages have been derived from research that typically appears under the rubric of "social facilitation," the ultimate value of the application of psychophysiological research to social processes lies in the identification of the finite set of fundamental stages that exist in various combinations to yield a wide range of social processes. Thus, the present stages are assumed to be among the total set of elementary building blocks that make up other social processes, such as social inhibition. It is of interest, therefore, that Latane, Williams, and Harkins (1979) found that subjects expended more physical effort (i.e., clapped and shouted more loudly) when they believed that they were personally accountable for their performance; if the presence of others made it possible to diffuse the responsibility for their performance, then each individual's personal efforts declined. Petty, Harkins, Williams, and Latane (1977) found that this effect applied to cognitive as well as physical efforts. Are the preceding stages involved in producing this social inhibition? There are no experiments on social inhibition to our knowledge that have included psychophysiological measures, but if social inhibition (i.e., individuals expending less effort to perform a task) emerges at least in part as a result of the presence of conspecifics' lowering the potential punitive consequences for a performer, psychophysiological research on this issue may prove interesting (e.g., diffusion of responsibility would be expected to be associated with lower levels of SSRRs).

Summary
The proposed stages of the social process described above are undeniably crude, but their derivation should serve to illustrate the application of psychophysiological procedures to delineate the building blocks of social processes. Moreover, these stages provide a starting point from which to construct corn8. The wording used here is not meant to suggest that subjects are aware of either stage. That is, subjects are not viewed as necessarily being consciously involved in, or able to report on, their decision regarding the threat value of an observer or their allocation of effort to emit socially valued behaviors. Indeed, it seems more probable to us that human subjects will be able to provide verbal data that are products of rather than first-hand reports on these processing stages, whereas these processes in subhuman species may be evident in a more primitive form (cf. Uetz et at., 1982).

664
paris on conditions (e.g., concealed vs. unconcealed observers) and select physiological measures (e.g., task-relevant and task-irrelevant EMG, SCRs, SCl, and/ or SSRRs) in future research. It is somewhat encouraging, nevertheless, that behavioral predictions can be derived from the analysis of the existing psychophysiological data (e.g., regarding concealed vs. unconcealed observers), and that the various competing models of social facilitation are compatible with the existence of the stages that have been derived. Some of the current formulations suggest that the presence of conspecifics has an innate threat value (Markus, 1981; Zajonc, 1980); others suggest that this threat value is learned (Cottrell, 1972); and still others suggest that what has been learned is that the presence of others in specific conditions poses a threat (e.g., Geen, 1981). There are existing formulations that emphasize the role played by the stage of threat assessment (e.g., evaluation apprehension), others that emphasize the stage of effortful striving (e.g., drive conceptions), and still others that straddle the two (e.g., self-presentation). Indeed, in light of recent controversies in the area (cf. Geen, 1981; Markus, 1981; G. S. Sanders, 1981), psychophysiological studies may help distinguish between operational and conceptual differences among theories. For example, Zajonc, Heingarner, and Herman ( 1969) demonstrated that the photophobic cockroach (Blatta orienta lis ) runs a simple maze more quickly and a complex maze more slowly when conspecifics are present and the context is punitive (i.e., when the maze is illuminated). They suggested that the presence of conspecifics is innately threatening. Of course, this study, while interesting, cannot be taken as evidence that the presence of conspecifics is always threatening, arousing, or energizing. Ethological research on territorial instincts clearly suggests that the presence of conspecifics is innately threatening only within specific environmental contexts (e.g., during feeding). The observations by Uetz, Kane, and Stratton ( 1982) of the normally inhospitable spider (Metepeirc spinipes) provide a case in point. Spiders share the part of the web that forms the suprastructure (i.e., the "space web") for their individual retreats and preycatching spirals. These spiders mdividually reconstruct their sticky prey-catching spirals daily, whereas they work cooperatively, though not daily, to maintain the communal space web. Uetz et al. (1982) found that these spiders adjusted group size and territorial distances to match the availability of valued resources (i.e., foodstuff). For example, they found that the distance to the nearest conspecific within a Colony was correlated (r -.97) with the availability of prey. When they manipulated the availability of prey (e.g., by moving the colony, or by attracting prey to the area of the colony), the spiders adjusted this nearest-neighbor distance. Uetz er al. (1982) concluded,

PROCESSES

The combination of solitary and communal behavior exhibited by this species suggests that it represents an intermediate stage in the evolution of social behavior in spiders .... the advantages of group living-exploitation of habitats free of competing species. increased prey capture efficiency. architectual stability of webs. and so on-would usually outweigh the advantages gained by maintaining maximum distances from conspecifics at the cost of aggressive behavior. (p. 542)

Finally, the various demonstrations that more cogitative threats can be posed by humans (e.g., ernbarrassment, evaluation apprehension) and can produce facilitatory or inhibitory effects on performance should come as no surprise, given humans' greater capacity to learn, employ symbols, and make controlled assessments of impending threats.

Attitude Formation, Maintenance, and Change
The concept of "attitude" has long been central to disciplines within the social sciences, including social psychology (e.g., Allport, 1935; McGuire, 1969), clinical and counseling psychology (Corrigan, Dell, lewis, & Schmidt. 1980; Strong, 1978), political science (e.g., Oskamp, 1977; Tursky & [amner, 1983), communications (e.g., Roloff & Miller, 1980; M. J. Smith, 1982), and consumer behavior and marketing (Kassarjian & Robertson, 1981). There are perhaps more social-psychophysiological studies dealing with attitudes and attitude change than any other topic. Unlike the situation in social facilitation research, verbal rather than behavioral measures have been the more common outcome variable in attitude studies. An exhaustive survey of this research is not attempted here, although interested readers may wish to consult reviews of this literature in Cacioppo and Petty (1983b) and Waid (1984). Our aim in this section is to provide descriptions of the psychophysiological data pertaining to a few of the distinctive processing stages in attitude change. An "attitude" represents a general and enduring favorable or unfavorable feeling about a person, object, or issue. Ever since Thurstone ( 1928) developed the first attitude scale, attitude measurement techniques have tended to focus on the assessment of a general evaluation of, or affective reaction to, a stimulus. For example, the respected scaling techniques of Thurstone (1928), Likert (1932), and Guttman ( 1944), as well as the more popular and contemporary semantic differential (Osgood, Suci, & Tannenbaum, 1957) and single-item rating scales, are all designed to gauge how much one likes or dislikes, promotes or detracts from, or feels generally favorable or unfavorable toward a stimulus (cf. Cacioppo, Harkins, & Petty, 1981). Different theories of attitude formation, maintenance, and change postulate different intervening pro-

=

Social

Processes

665
One important implication of this distinction is that psychophysiological assessments designed to tap the affective "roots" of an attitude may generally be insensitive (cf. Cacioppo & Sandman, 1981; Petty & Cacioppo, 1983), unless a physiological index of an evaluative belief is formed artificially-for instance, through semantic conditioning (cf, Tursky & [amner, 1983). Second, attitude development and change may involve a stage representing an affective reaction, but, depending upon the antecedent conditions, this stage may be preceded by others, such as the instigation and reduction of cognitive dissonance (Fazio er aI., 1977; Festinger, 1957) or fear (Janis & Feshbach, 1953; Leventhal, 1970); by attention, comprehension, and cognitive elaboration (Cacioppo & Petty, 1981a; Petty & Cacioppo, 1981); or by surprisingly little else (Wilson, 1979). As in the discussion of observed versus unobserved performance, it is assumed that the various theories of attitude formation, maintenance, and change can ultimately be viewed as unique subsets and sequences of a finite number of elementary stages in which these particular stages may be included.

cesses between intitial exposure to some attitude-relevant stimulus (e.g., a persuasive message) and the subsequent attitudinal response (e.g., resistance, yielding; see Petty & Cacioppo, 1981, for a review). Only recently, however, have psychophysiological procedures been applied to the study of the underlying attitudinal processes. Psychophysiological investigations designed to measure attitudes are far more common than those designed to explore attitudinal processes. The rationale for searching for a psychophysiological index of attitudes is the understandable concern that people may not report their true attitudes, either because they are not aware of how they really feel or because they are not willing to disclose these feelings (Cook & Sellriz, 1964). Although these studies in and of themselves are not particularly informative about the rudimentary processes underlying attitude formation, maintenance, and change, they do furnish data about the physiological response profiles characterizing an elementary component of these processes-namely, affective (l.e., favorable-unfavorable) reactions. Initially, one may think that the affective reaction is the attitude or the fundamental process underlying attitude formation and change. After all, attitude development and change are conceived of as involving the arousal of new positive or negative feelings about the attitude stimulus. This conceptualization, however, simply suggests that an affective reaction is a fundamental stage in attitude development and change; it does not mean that other stages are not involved. That is, the notion of a general feeling of liking or disliking accompanied by bodily responses (i.e., an affective reaction) should not be viewed as synonymous with the notion of an attitude. Consider a person whose attitude toward some object was developed long ago and has not undergone change recently (e.g., an attitude toward one's favorite Bogart movie or one's first love). It is conceivable that should this person be asked to indicate his or her attitude toward the object, the person would not have to re-experience the feelings once associated with it; recalling the evaluation of the object should be sufficient. The individual might even recall that he or she had once felt a strong affection for the object, but this memory could be accessed without also accessing (or re-experiencing) these feelings. In these instances, the attitude is akin to a bit of knowledge regarding one's evaluation of the object, rather than to a nonspecific feeling of liking or disliking accompanied by bodily responses. In sum, then, it seems reasonable that most attitudes are maintained as general beliefs or memories of one's evaluations of and affective reactions to stimuli, rather than in the form of affective reactions per se. In this way, people can hold many attitudes simultaneously (which they must in this complex world if attitudes endure), even though they cannot possibly maintain as many feelings (affects) simultaneously as they retain attitudes.

Affective Reactions
In one of the earliest studies relevant to our search for physiological profiles of affective reactions, Dysinger ( 1931) monitored SRRs as subjects were exposed to words that varied in the pleasantness ratings they elicited (cf, McCurdy, 1950). Dysinger found that the SRRs were correlated with the extremity of the ratings, but were unrelated to the direction of the reaction: The more pleasant or unpleasant the word, the larger the SRR it elicited. A slightly different finding emerges when complete attitude statements (e.g., Dickson & McGinnies, 1966), attitude objects (e.g., J. B. Cooper & Singer, 1956), or people (e.g., Rankin & Campbell, 1955) are presented, rather than individual words. Generally, the more discrepant the attitude stimulus from an individual's own position, the larger the SRR. For example, recall that C. E. Smith (1936) found that students showed the largest SRRs when they disagreed with the statement presented and they believed their own position was in conflict with group norms. Two decades later, J. B. Cooper and Singer (1956) demonstrated a similar effect. They asked 126 students to rate and rank 20 ethnic and national groups. Of these 126 subjects, 20 who had extreme attitudes toward their most liked and disliked groups subsequently served in a laboratory study in which skin resistance was monitored while the experimenter complimented the most disliked nationality, derogated the most liked nationality, complimented a neutral nationality, and derogated a neutral nationality. Subjects emitted the largest SRRs to derogatory statements about the most liked nationalities and to complimentary statements about the most disliked nationalities, and the SRRs to these two types of statements were equivalent.

666
It has been suggested that SRRs reflect the extremity but not the direction of an attitude (e.g., Lemon, 1973). This inference is based upon the notion that the SRRs are reflecting the emotional arousal elicited by the stimulus. Although this inference has been appealing to some because of the utility of such a measure (cf. ]. B. Cooper, 1959), the support is equivocal. Specifically, the presentation of pleasant versus unpleasant words and of attitude statements that are consistent versus inconsistent with subjects' own attitudes might initially be thought to isolate the same fundamental stage-namely, an affective reaction. The effects of these manipulations on SRRs, however, are different. The former manipulation leads to similar increases in SRRs, whereas the latter does not. These different effects are sometimes interpreted as signaling that exposure to consistent attitude statements does not elicit an affective reaction, but this interpretation lacks independent support. Hence, alternative interpretations should also be considered. For instance, the experimental conditions eliciting SRRs in these studies resemble the conditions necessary to elicit an orienting response (e.g., novelty, signal value-cf. Cacioppo & Sandman, 1981; Petty & Cacioppo, 1983; Shapiro & Crider, 1969). Shapiro and Crider (1969) noted over a decade ago that statements conflicting with a person's attitude may elicit SRRs because "this inconsistency may be seen as a source of novelty, incongruiry, or violation of expectancy and thereby a determinant of orienting reactions in the individual exposed to the communication" (p. 22). Alternatively, if subjects are stressed or frustrated when exposed to a subset of the attitudinal stimuli in the experimental setting (e.g., by being exposed to taboo words or by having others disagree with them), then the SRRs may not reflect the affective reaction to the attitude stimulus per se, but rather the reaction to the context in which it appears. Thus, although SRRs have been one of the most commonly employed physiological indices of attitudes (cf. Schwartz & Shapiro, 1973), these responses are elicited by such a variety of physical and psychological stimuli that they have provided an ambiguous marker for the processing stage representing an affective reaction. A very different physiological assessment of this processing stage was advanced by Hess (1965), who argued that the pupillary response to a stimulus that has been repeatedly presented covaries with the person's attitude. According to Hess, the more liked the stimulus, the more pupillary dilation; the more disliked the stimulus, the more pupillary constriction. This bidirectional physiological index of attitudes is not necessarily evident when the stimulus is first presented, presumably because of other influences on pupil size (e.g., interest, novelry, mental load); rather, in Hess's view, this relationship emerges after multiple exposures to the stimulus. A recent experiment by Metalis and Hess (1982)

PROCESSES

illustrates this relationship. Male and female undergraduates were shown a series of pictures depicting either normal people in neutral settings, people with grotesque skin diseases, or nude men and lor women in erotic postures. Each photograph was preceded by a control slide of comparable overall luminance. The pupillary response was defined as the average change in pupil size to the picture slide following the control slide; these slides were paired and presented three times (but sampled only twice-the first and third presentations). They reported finding pupillary dilation by men and women to the erotic photographs, and the largest dilations were evoked by the photographs of men and women in erotic poses. Subjects also displayed pupillary constriction to photographs of skin diseases, and pupillary responses to the neutral photographs fell between those evoked by the erotic and disease pictures. The association between pupillary responses and affective reactions to stimuli has also been assailed, however, and definitive responses to the concerns expressed by critics have not been forthcoming. One concern expressed has been methodological and notes that many nonpsychological factors can affect pupil size (e.g., the wavelength of the visual stimulus-see Goldwater, 1972; Woodmansee, 1970). For example, ]anisse and Peavler (1974) suggested that different visual fixation points may account for the effect in Hess's original research, wherein females were found to show larger pupillary responses to male pin-ups than males. The same criticism can be leveled against Metalis and Hess ( 1982). For example, Metalis and Hess obtained subjects' evaluations of each photograph in addition to assessing their pupillary reactions, and found that these measures were not correlated except when diseases were depicted. Unfortunately, it is uncertain whether pupillary constriction evoked by pictures of skin diseases reflected a negative affective reaction or a focusing on the lighter area of the photograph. In particular, the more a person disliked a given photograph in this set, the more likely he or she may have been to fixate on an area of normal rather than diseased skin. It is noteworthy, therefore, that no investigator has documented a significant relationship between attitudes and pupillary responses when auditory, olfactory, or tactile stimuli have been employed (Goldwater, 1972). Another concern that has been expressed is that pupillary responses may reflect a processing. stage under certain test conditions, but that this stage IS not necessarily related to affective or evaluative reactions. Beatrv (1982), for example, reviews research. demonstrating that pupil size varies systematically With overall cognitive load, a point to which we ~e~rn. below (see pp. 669-670). Hence, although pupil slz7 IS occasionally cited as a bidirectional measure of attltudes or as a marker for affective reactions, the accumulated

1

..
Social Processes

667
Salt, Mandel, & Klerman, 1976a, 1976b), for instance, reasoned that different affective reactions should be identifiable in changes in the EMG activity of facial muscles when subjects are overtly inactive. In an early experiment, Schwartz et al. (1976b) asked people to imagine positive or negative events in their lives. In their observations of changes in the mean amplitude of integrated EMG (lEMG) activity over facial muscle regions, Schwartz et al. found that peopie generally responded with higher IEMG activity over the depressor annuli oris and zygomatic major muscle regions and lower IEMG activity over the corrugator supercilii muscle region when imagining happy, in contrast to sad, events. In another study, changes in mean-amplitude IEMG activity were rnonitored over the corrugator supercilii and zygomatic major muscle regions of the face while subjects engaged in emotional imagery (e.g., "Picture the last situation in which you laughed," "Picture the last funeral you went to"). Again, the changes in meanamplitude IEMG activity indicated elevated activity over the zygomatic major region and lowered activity over the corrugator supercilii region when subjects engaged in pleasant, in contrast to unpleasant, emotional imagery (see S. L. Brown & Schwartz, 1980; Fridlund & Izard, 1983). We (Cacioppo & Petty, 1979a, Experiment 2) reasoned that patterns of low-amplitude EMG activity over facial muscles might distinguish positive from negative affective reactions to a persuasive communication. Changes in the mean-amplitude IEMG activity were recorded over the corrugator supercilii, depressor anguli oris, and zygomatic major muscle regions using miniature surface electrodes while undergraduate students anticipated and then listened to either an involving proattitudinal communication (e.g., recommending more lenient visitation hours in dormitory rooms) or an involving counterattitudinal communication (e.g., recommending stricter visitation hours in dormitory rooms). A control group was forewarned only that they would hear a message and subsequently heard a news story about an archeological expedition in Asia. The observed patterns of subtle facial IEMG activity and reported evaluations of the messages were consistent. Subjects who were exposed to the proattitudinal message showed slightly higher rnean-amplitude IEMG activity over the zygomatic major and depressor anguli oris muscle regions and lower meanamplitude IEMG activity over the corrugator super~ cilii muscle region than subjects who were exposed to the counterattitudinal message. The neutral message was rated by subjects as being relatively pleasant, and the pattern of IEMG activity over these muscle regions during the message more closely matched that found during the proattitudinal than during the counterattitudinal communication. In another study (Cacioppo, Petty, & MarshallGoodell, 1984), IEMG activity was recorded over the

evidence is at present more in accord with the conclusion reached by Woodmansee (1970): "[Tjhe pupil does not measure attitude or qualitatively different affective states. There is ample evidence, however, that the pupil, in its reflex dilation reaction may be used to indicate ... attentiveness, interest, and perceptual orienting" (p. 532). A more recent attempt to obtain, under circumscribed testing conditions, a physiological marker of the affective reactions to attitudinal stimuli relies upon one of the oldest sets of scientific observations (and one of the most commonly employed naruralistic assessment procedures) of bodily responses and emotions: subtle changes in facial expression. Darwin (1872/1904) linked affective reactions directly to facial expressions and argued that different facial expressions were biologically adaptive. Both Izard ( 1971) and Ekman (1972) have provided evidence that different emotions can be linked to unique facial expressions across a wide variety of situations and cultures, and they too suggest that these distinctive facial displays have functional significance (e.g., as a form of communication). Clearly, facial expressions can be used in deceptive as well as veridical communication in social settings (Kraut, 1982; Zuckerman, Larrance, Spiegel, & Klorman, 1981). Ekman and Friesen (1975) have argued that cultural differences in the facial expression of emotion are due to cultural differences in "display rules." These observations limit the conditions under which changes in facial expression can be used to mark an affective reaction during attitudinal processing (i.e., subjects should not be aware that facial ex~ pressions are being monitored). Moreover, betweensubject comparisons of facial expressiveness to gauge the intensity of affective reactions may be misleading, since several studies suggest that intense facial displays may actually help relieve some of the potency of an emotional reaction (cf. Buck, 1980; Notarious, Wem~ pie, Ingraham, Burns, & Kollar, 1982). Particularly informative for studying the affective processing stage, however, may be spontaneous facial expressions obtained in a within-subjects design; through the use of EMG techniques, these "expressions" include (or concern primarily) those that are so subtle that neither the person nor an observer can detect them in a videotape replav.? Schwartz and his colleagues (e.g., S. L. Brown & Schwartz, 1980; Schwartz, 1975; Schwartz, Fair, Mandel, Salt, Mieske, & Klerman, 1978; Schwartz, Fair,

9. We are not suggesting. however. that covert facial expressions always reflect veridical affective reactions. The presentation of a stimulus that has been associated repeatedly with deceptive facial expressions might elicit covertly deceptive expressions. and covert expressions of emotion that are evinced while an individual is being scrutinized by an observer might be subject to the same display rules as are overt expressions of emotion in these conditions.

- -----668
zygomatic major, corrugator supercilii, levator labii superioris, masseter, orbicularis oris, and superficial forearm flexors muscle regions while subjects imagined or performed simple physical or attitudinal tasks. In the physical tasks, subjects either imagined or actually lifted a l o-kg or a 35-kg weight. In the attitudinal tasks, subjects adopted an attitudinal set of either agreement or disagreement as they silently read neutral text or imagined reading an editorial. The tasks were ordered randomly, and subjects were focused on the distinction between imagining and performing tasks rather than the distinction between physical and attitudinal tasks. Multivariate analyses of the topographical features of the IEMG responses over each muscle region were performed to determine the effects of the tasks on changes in IEMG response waveforms, and univariate analyses were conducted to isolate the features of the waveform (e.g., mean amplitude, variance time) that were altered by tasks (cf. Cacioppo, Marshall-Goodell, & Dorfman, 1983). Results for the physical tasks showed that the form of the IEMG response over the forearm was dramatically affected by the parameters of the tasks, whereas the IEMG waveforms obtained from facial muscles were not. Results for the attitudinal tasks showed that the parameters of the tasks influenced the form of the responses over facial muscles previously associated with affective processing in similar paradigms (e.g., corrugator supercilii, zygomatic major), whereas the form of the IEMG responses over the forearm and facial muscles not implicated previously in low-level affective processing (e.g., orbicularis oris) were not differentiated by the attitudinal tasks. These data support the notion of temporal and spatial specificity in incipient skeletal-motor activity during low-level affective processing, given that subjects are relaxed, unobtrusively observed, and involved in the tasks. Finally, subsequent analyses of covariance suggested that the analyses of the topography of phasic IEMG responses provided significantly better differentiation of the antecedent conditions than did analyses of mean amplitude alone. Lanzetta and his colleagues have also found that subtle changes in the mean amplitude of IEMG activity over facial muscles accompany affective reactions in social settings. In one study, subjects viewed a videotape of a confederate who was ostensibly receiving an electric shock (Vaughan & Lanzetta, 1980). With each delivery of an electric shock, the confederate grimaced in pain. EMG recordings over the medial frontalis, masseter, and orbicularis oculi muscle regions, which were selected because of their possible involvement in expressions of pain, revealed that higher mean-amplitude IEMG activity was present over the latter two sites in subjects immediately after the confederate's grimace as compared to preceding levels.

_._'

_L-

_

PROCESSES

In a follow-up study, Englis, Vaughan, and Lanzetta ( 1982) constructed a stock market game that subjects ostensibly played with another person. Subjects tried to guess which market indices would rise or decline and were rewarded with money for correct guesses but punished with mild finger shock for incorrect guesses. During the game, subjects viewed another "subject" on a video monitor. (All subjects actually were exposed to a videotape of a confederate.) Some subjects found that when the confederate smiled, their own guess was correct and they were rewarded, and when the confederate expressed pain, their own guess was incorrect and they were punished. Other subjects observed the opposite: When the confederate smiled, their own guess was incorrect and they were punished, but when the confederate expressed pain, their own guess was correct and they received a reward. After completing the game, subjects simply watched a videotape of the confederate. Surface EMG over the corrugator supercilii, masseter, and orbicularis oculi muscle regions, as well as skin conductance and HR, were monitored continuously during the game and postgame periods. Consistent with the previous findings using surface EMG recordings, Englis et al. ( 1982) reported that when the confederate and subject shared outcomes, mean amplitude of the IEMG activity over the monitored regions was higher in both phases of the experiment when the confederate exhibited a pained than when he or she exhibited a happy facial expression. On the other hand, when the outcomes of the confederate and subject were asymmetrical, this profile tended to be reversed. Thus, the incipient changes in facial muscle activity appeared to mark subjects' affective reaction to, rather than mimicry of, the confederate's overt facial expression. In sum, preliminary evidence suggests that subtle facial muscle changes may reflect the general nature of subjects' affective reactions to attitudinal stimuli as long as the subject is generally relaxed, inactive, involved in the task, and unobtrusively observed so that display rules or deceptive expressions are not invoked and the small muscular signals of interest are not masked by background somatic activity. It has not yet been determined, however, whether changes in the mean amplitude of IEMG activity over facial muscles reflect the specific nature of simple positive or negative evaluations (e.g., mild anger vs. sadness), whether they are sensitive to the intensity as well as to the direction (i.e., positive or negative) of affective reactions (cf. Schwartz et al., 1976a), or whether additional features of the IEMG response (e.g., variance amplitude, mean time) will improve the value of subtle facial IEMG profiles as markers of this processing stage (cf. Cacioppo, Marshall-Goodell, & Dorfman, 1983). Finally, the ease with which subjects can control their facial expressions heightens concerns about demand charcacteristics in this type of research. Inter-

Socia! Processes

669
region and over the trapezius muscle region, HR, breathing rate, and cephalic pulse amplitude were cal, culated for the 60 sec immediately preceding and following the forewarning and instruction to subjects to collect their thoughts on the issue. Persuasive messages were not presented. in order to avoid confound, ing the reception of messagearguments with the issue, relevant thinking that subjects produced in response to the forewarning and experimental instruction. Resuits indicated that the mean amplitude of perioral IEMG activity, HR, and breathing rate were height, ened during the "collect thoughts" interval. More, over, the mean amplitude of non-oral IEMG activiry and cephalic pulse amplitude were unchanged. suggesting that subjects were not simply more generally aroused during this interval. The discrepancy between subjects' initial attitudes and the impending recommendation was varied in this study to be low. moderate, or high (e.g., advocating that tuition be raised from 5.5% to 100% per quarter when subjects wanted no tuition increase). This rna' nipulation had profound effects on the number of favorable and unfavorable thoughts listed and the agreement with the recommendation that they reported, but it had no effect on any of the physiological measures obtained. Correlational analyses also indicated that neither the autonomic measures monitored nor somatic measures of activity over the perioral or trapezius muscle regions reflected the affective tone of the issue-relevant thinking (as assessed using the thought, listing technique; see Cacioppo & Petty, 1981c, for a review of this procedure). In combination with the studies of physiological markers de, scribed in the preceding section, these data suggest that the processing stages of issue-relevant thinking and affective reactions have distinctive physiological consequences in the persuasion paradigm (Cacioppo & Petty, 1981a). Further evidence that silent linguistic processing results in changes in EMG activiry localized over the perioral muscle region was obtained in two experiments using the well-researched orienting task para, digm developed by cognitive psychologists to study encoding operations (see Cermak & Craik, 1978; Craik & Lockhart. 1972). In the first experiment. subjects were asked in half of the trials to determine whether or not a trait adjective was self-descriptive, and in the other half to determine whether or not the trait adjective was printed in upper-case letters (Cacioppo & Petty, 1979b). HR and the mean amplitude of the IEMG activiry over the perioral region and over the non preferred forearm (superficial forearm flex, ors) were recorded during the first few seconds following the presentation of the trait adjective. Finally. subjects were asked to recall as many of the trait adjectives as possible at the completion of the study. Results revealed that subjects could recall more of the

ested readers may wish to consult Fridlund and Izard ( 1983), who review methodological concerns and offer suggestions for securing interpretable data. Message Processing Research on persuasive communications has recently moved from focusing on how well people learn and subsequently recall externally provided messagearguments (e.g., Eaglv, 1974; N. Miller & Campbell, 1959) to what people understand to be the meaning, personal implications, and merits of these message arguments (cf. Cialdini, Petty, & Cacioppo, 1981; Eagly & Himmelfarb, 1978). Thus, the emphasis in much of the recent research on persuasion has moved from the preliminary stages of message reception to succeeding stages, such as issue-relevant thinking (e.g., message elaboration). In this section, we exam, ine the psychophysiological markers of one stage of this sequence, issue-relevant thinking. Psvchophysiological applications to the study of attentional pro' cesses and learning are reviewed elsewhere in this volume. Before proceeding, we should clarify that not all instances of resistance or susceptibility to persuasive communications are predicated on issue-relevant thinking, or what we have characterized as the "central route" to persuasion (Cacioppo & Petty, 1982b; Petty & Cacioppo, 1981). A recent estimate places the averageperson at the receiving end of hundreds of persuasive messages a day from advertisers alone (Will, 1982), and when the virtual deluge of appeals, prompts, suggestions, and counsels that an individual receives daily from personal contacts (e.g., friends, sales clerks, physicians) is considered, it is obvious that people do not have the luxury of either consider' ing or rejecting every recommendation to which they are exposed. Recommendations that have important personal consequences, for instance, are more likely to elicit extensive issue,relevant thinking that is then instrumental in determining their susceptibiliry or re' sistance to an appeal (Petty & Cacioppo, 1979), whereas acceptance of recommendations that have no personal consequences and/or on topics about which individuals know little is likely to be determined by peripheral cues (cf. Cacioppo & Petty, 1982b; Petty & Cacioppo, 1981). Extensive issue,relevant think, ing, therefore, is likely to be a fundamental stage of some, but certainly not all, attitudinal processes. In an early study to determine the physiological profiles of issue,relevant thinking, subjects were told the topic and position of an audio communication that they were to hear later in the experimental session, and subjects were instructed to take the next 60 sec to "collect their thoughts" on the issue (Cacioppo & Petty, 1979a). Mean-amplitude IEMG activitv over the perioral (e.g., orbicularis oris) muscle

---', ---------~

I

r
,

670
traits whose meaning and self-reference had to be determined than whose orthographic appearance was judged. In addition, the former task led to higher mean-amplitude IEMG activity over the perioral region than the latter task, but similar levels of HR and mean-amplitude IEMG activity over the non-oral muscle region. This finding has been conceptually replicated in a follow-up study using a wider variety of orienting tasks and a different presentation modality (Cacioppo & Petty, 1981b). Finally, although changes in mean-amplitude IEMG activity over the perioral region can mark major differences in the extent of such types of linguistic processing as issue-relevant thinking, they can also reflect reading, recalling information, and problem solving (see reviews by Garrity, 1977; McGuigan, 1978) and listening to persuasive communications (Cacioppo & Petty, 1979a, Experiment 2). Recall that in this latter experiment, subjects were told that they would hear a communication in 1 min, and following this period all subjects were exposed to a taperecorded message. Some subjects anticipated and heard a proattitudinal message; others anticipated and heard a counter attitudinal message; and still others anticipated only hearing some unspecified message that, as it turned out, was judged to be an enjoyable news story. Changes in HR and changes in meanamplitude IEMG activity over the perioral and over selected facial muscle regions were monitored. (The results of the latter measures have been described in the preceding section and are not repeated here.) Previous experimental work had indicated that subjects engage in extensive topic-relevant thinking when anticipating an involving counterattitudinal appeal (Petty & Cacioppo, 1977); as expected, higher meanamplitude IEMG activity over the perioral region was observed when subjects anticipated the counterattitudmal message, but not the proattitudinal or undescribed communication. Moreover, comparable elevations in mean-amplitude IEMG activity over the perioral region were observed during the presentation of these messages. Thus, although substantial differences in the extent of issue-relevant thinking can lead to localized differences in IEMG activity over the perioral muscle region, so too may message comprehension or extensive issue-irrelevant thinking."?
10. HR in this study was not found to change during the anticipatory interval; this result contrasts with the increased HR observed following the forewarning of a counterattitudinal appeal when subjects were explicitly instructed to "collect their thoughts" on the issue (Cacioppo & Petty, 1979a). As outlined above (see pp. 655656). this result can be interpreted in several ways. First, differences in processing stages may have existed across the two studies. For example, the increases in HR observed when subjects were instructed to "collect their thoughts" perhaps reflect an additional processing stage concerning the presence of a social incentive (d. Fowles, 1982). Alternatively, changes in HR may have simply been insensitive to the relatively small differences in covert verbalizations (Johnson & Campos, 1967; Lynch, Thomas. Long. Malinow, &

PROCESSES

It should be noted that although changes in the IEMG activity localized over the perioral muscle region appear to vary with substantial differences in the extent of linguistic processing in the persuasion paradigm, these changes may not reflect rudimentary differences in imagery or in general cognitive load if linguistic processing is constant across conditions (B. B. Brown, 1968; Totten, 1935; cf. McGuigan, 1978). Beatty (1982) has contended that task-evoked pupillary responses provide a sensitive measure of the overall information-processing load on an individual. Although to date no experiment has documented the course of the pupillary response during the presentation of persuasive communications or documented its utility in gauging the cognitive load on a recipient of a persuasive message, the pupillary response may prove to be a sensitive marker of this attitudinal processing component. According to Beatty ( 1982; see also Chapter 3, this volume), the pupillary response
provides a reliable and sensitive indication of within-task variations in processing load. It generates a reasonable and orderly index of between-task variations in processing load. It reflects differences in processing load between individuals who differ in psychometric ability when performing the same objective task. (1982. p. 291)

Thus, if an attempt at social influence causes individuals to think about the verbal arguments with which they can respond, parallel increases in pupillary and perioral EMG activity can reasonably be expected; however, if the influence attempt primarily evokes images rather than verbal consideration, then divergences in pupillary and perioral EMG responses can occur. When investigations of pupillary responses in a persuasion context are conducted, it is of course important to control for potentially confounding features of the persuasive message and context, such as novelty, the luminosity of the material in the setting, and the person's focal point during the presentation (e.g., by using audiotaped messages). Cognitive Dissonance "Attitude-discrepant behavior" refers to actions people take that they know are inconsistent with their attitudes toward the target of their actions. For instance, if a person is opposed to being assertive but finds himself or herself voluntarily being extraordinarily assertive in a social setting, then the person is engaging in an attitude-discrepant action. One may initially think that people would seldom do such a thing, but a voluminous literature in social psychology demonstrates that people are often "induced" to
Katcher. 1981) or mental concentration (cf .. Col~s & DuncanthdcdantlclJohnson • 1975' , Lacey et al.. 1963) normally elicited. by A mona I . pation of or exposure to a persuasive communication. .I research is required to determine the reliability and pSycho!oglca significance of this difference in HR response.

SociaL ProcesSe5

671
equally liked (the stimuli the subject had ranked third and eighth). Deciding between two similarly attractive alternatives, in contrast to one attractive and one unattractive alternative, is thought to create more cognitive dissonance, since when two alternatives are relatively equal in their attractiveness the individual foregoes more attractive features in the unchosen alternative and/or accepts more unattractive features in the chosen alternative than when the alternatives are very different in their attractiveness. Indeed, Gerard (1967) observed a constriction of the blood vessels in the fingers of 10 of the 12 subjects who had chosen between the two attractive paintings, whereas he observed digital vasoconstriction in only 1of the 12 subjects following the choice between an attractive and an unattractive painting. Unfortunately, the normal attitudinal effects of choosing between two attractive, in contrast to attractive and unattractive, alternatives were not observed. Moreover, this same paradigm has been found to cause subjects who have made a difficult rather than a simple choice to attend more carefully to the chosen alternative. Since digital vasoconstriction is a component of the orienting response, there is some question whether the vasoconsttiction observed by Gerard (1967) is attributable to the arousal of cognitive dissonance or to possibly more pronounced orientations to the paintings by subjects in the "high-dissonance" conditions. Fazio and Cooper ( 1983) review several conceptually similar but unpublished studies. Buck ( 1970), for example, in a doctoral dissertation, reported a study in which subjects were induced, under conditions of high choice, to deliver a series of audio tones, mild electric shocks, or painful shocks to another individual. Since harming an innocent victim is typically counterattitudinal for people, the delivery of intense electric shocks by subjects to their partners was assumed to constitute an attitude-discrepant action for which they were personally responsible and that had clear and foreseeable negative consequences. Buck, who monitored HR and SCRs, found that subjects who delivered intense shocks showed the largest SCRs and subjects who delivered harmless tones exhibited the smallest SCRs. No differences in HR were found, however. 11

perform an attitude-discrepant behavior when interacting with others. The theory of cognitive dissonance, which has stimulated more research in social psychology than any other formulation, addresses the consequences of this kind of social behavior. According to Festinger (1957), two elements of information (cognitions) are dissonant for an individual when the individual knows that one suggests the opposite of the other. Exceptions to this general postulate were soon found, and the theory of cognitive dissonance underwent a series of changes and refinements (Aronson, 1969; Brehm & Cohen, 1962; Wicklund & Brehm, 1976; cf. Greenwald & Ronis, 1978). At present, cognitive dissonance is thought to develop when an individual accepts personal responsibility for an action associated with negative consequences that were foreseeable or that should have been foreseeable (cf. Cialdini et al., 1981). According to dissonance theory, the reduction of cognitive dissonance can occur in one of several ways, including (1) denying responsibility for, control over, or negative consequences from the behavior; (2) changing the attitude to make it more consistent with the behavior; or (3) misattributing the dissonant state to a cause that is either temporary and self-correcting or beyond one's control (e.g., a pill whose "side effects" are described as making one feel unpleasant). Finally, the mode selected for dissonance reduction is thought to follow the path of least resistance. Hence, denial of responsibility for one's actions or misattributing the cause of a dissonant state to an external agent is usually preferable to changing one's attitude, since the latter requires more cognitive reorganization and effort. The spate of research generated by cognitive dissonance theory has led to the emergence of a number of alternative theories, such as Bern's (1965,1972) selfperception theory and Tedeschi, Schlenker, and Bonoma's (1971)' theory of impression management. The distinctive feature of dissonance theory is the postulated internal state of dissonance. Unfortunately, Festinger (1957) was not very explicit about this internal state. He described the presence of cognitive dissonance as physiologically discomforting (p. 2), for which "one can substitute other notions similar in nature, such as 'hunger,' 'frustration,' or 'disequilibrium,'" (p. 3); he also described dissonance as a drive-like state (p. 18). Despite the proliferation of theories, published psychophysiological studies of dissonance arousal are uncommon (see Fazio & Cooper, 1983). The first and only published study for over a decade on the physiological responses to dissonance arousal was reported by Gerard (1967). Subjects were exposed to several paintings and were asked to rank the paintings from their most to least liked. Subsequently, subjects were given a choice between two paintings they liked about equally well (the stimuli the subject had ranked third and fourth) or between two that were not nearly

11. Fazio and Cooper ( 1983) review a second study in which HR and SCRs were recorded. by Gleason and Katkin (1978). Subjects were instructed to spend 5 min thinking of arguments either supporting or attacking the recommendation that average grades at their university be lowered (a counterattitudinal appeal). Gleason and Karkin found that subjects who were trying to think of arguments favoring the counterattitudinal appeal exhibited higher HR and more SSCRs than subjects trying to think of arguments attacking the appeal. As Fazio and Cooper ( 1983) note. however. there is no reason to believe that Gleason and Karkin's manipulations created different levels of dissonance arousal. since it is unclear whether subjects had any choice in this activirv or anticipated negative consequences from preparing their speeches.

. ---- ----

.. ... '-~!~"

.

.

.._ .,.,..-----~---

---

672
Croyle and Cooper (1983) provide an interesting and informative set of observations in this area. They note that in previous dissonance research in which physiological measures were obtained, the typical dissonance effects on attitudes were not observed (e.g., Buck, 1970; Gerard, 1967). They suggest that subjects may have been misattributing the dissonance to the electronic gadgetry associated with physiological recordings, thereby rendering attitude change unnecessary. Croyle and Cooper dealt with this problem by conducting rwo experiments. The first was designed to validate that a given set of procedures induced cognitive dissonance. Subjects were induced under the perception of high choice to write an essay either supporting or attacking the recommendation, "Alcohol use should be totally banned from the university campus and eating clubs"; another group of subjects were induced under the perception of low choice to write an essay favoring this recommendation. Only the subjects who perceived that they had high choice and who wrote the essay favoring this counterarritudinal recommendation were expected to experience dissonance as a result of their decision. No physiological measures were obtained; rather, subjects were simply asked their attitude toward the recommendation after they had written the essay. As would be expected, subjects in the high-choice/ counter attitudinal-essay condition showed more attitude change than subjects in the remaining conditions. In a follow-up study, different groups of subjects were exposed to the same experimental manipulations while spontaneous SCRs (SSCRs) were measured "to examine the impact of simple mental and physical tasks on the electrical activity of the skin." Following the composition of their essays, subjects were given a 3-min rest period. An analysis of covariance was conducted using the number of SSCRs emitted during this rest period as the criterion and the number of SSCRs displayed during the pre-essay period as the covariate. As expected, Croyle and Cooper found higher levels of SSCR activity in the high-dissonance than in the low-dissonance conditions. Even though the sparse data on the physiological characteristics of dissonance have not consistently indicated that dissonance arousal initiates a diffuse and unitary increase in physiological arousal (e.g., Buck, 1970), the physiological data are consistent with the view that the instigation of dissonance leads to heightened activity in the sympathetic nervous system (cf. ]. Cooper ee al., 1978; Fazio & Cooper, 1983), and they provide an interesting picture of dissonance as an attitudinal processing stage. Specifically, the observation of increased EDRs in the high-dissonance condition while HR remains constant across high- and low-dissonance conditions is the same pattern found to characterize the threat assessment stage involved in observed performances (see the earlier discussion of observed vs. unobserved performance). It is particularly interesting to note,

PROCESSES

therefore, that cognitive dissonance is not aroused if individuals believe ( 1) that there is no choice but to act as they do (Linder, Cooper, & [ohes, 1967; Sherman, 1970); (2) that no one can identify them with the discrepant behavior, so they can deny that they seriously mean what they say or do (Brehm & Cohen, 1962; Riess & Schlenker, 1977); or (3) that the actions for which they are personally responsible cause no harm to themselves or significant others (Calder, Ross, & Insko, 1973; Collins & Hoyt, 1972; ]. Cooper & Worchel, 1970; Goethals & Cooper, 1975). Dissonance is aroused only when a person is forced to conclude that he or she is the willing causal agent of some discrepant decision or action that leads predictably to negative consequences that are (or can be) recognized by others. In other words, a fundamental feature of cognitive dissonance is social dissonance (cf. Fazio & Cooper, 1983; Schlenker, 1980). The logic of the experimental conditions that give rise to cognitive dissonance, therefore, is in accord with the "stage" suggested by perusing the physiological data: A criterial, and probably early, processing stage involves determining the potential punitive consequences for oneself as a result of one's decision (see footnote 8). At present, there is insufficient psychophysiological data on dissonance arousal and reduction using various operationalizations and measures to permit us to derive candidates for other processing stages. Another finding from behavioral studies of dissonance, however, is pertinent here, given the commonality berween the threat assessment stage that possibly underlies both social facilitation and cognitive dissonance phenomena. It will be recalled that the presence of an observer was found to facilitate the performance of sets of simple tasks and to impair the performance of sets of difficult tasks. The arousal of cognitive dissonance apparently has the same facilitatory-inhibitory effects on performance (Kiesler & Pallak, 1976). In an illustrative study, Pallak and Pittman ( 1972) examined the performance of subjects under conditions of high versus low dissonance on simple and difficult Stroop tests. Subjects performed a boring task that involved pronouncing words as they appeared on a memory drum. After 5 min of this boring task, subjects were told that they were to perform the same task using the same words for an additional 30 min. The experimenter subtly induced half of the subjects to agree to perform this task while maintaining the impression that subjects were personally responsible for choosing to perform the task (high-dissonance group), whereas the other subjects were given no choice but to perform ~he ta~k yowdissonance group). Immediately following this decision" by subjects, but prior to the contin~ation of work on the boring pronunciation task, subjects performed either the simple (i.e., low-response-competition) or difficult (i.e., high-response-competition) Stroop test. Pallak and Pittman (1972) found that

J

Social Processes

673
Lazarsfeld (1949), we recently presented individuals with several statements about social behavior (e.g., "A bystander is more likely to offer help when witnessing an emergency with others than when alone"). Our respondents assured us that many of these claims were so apparent as to warrant neither empirical documentation nor theoretical investigation. Despite the apparently self-evident nature of the principles of social processes when unveiled, people are in fact quite poor at specifying these principles correctly. The interesting aspect of the propositions that we gave to the participants in our inquiry was that in every instance we presented a general "principle" that had been disconfirmed in empirical research. That is, social processes appear more obvious than they are. In this chapter. we have outlined (1) the history of the application of psychophysiological procedures to study social processes; (2) the means by which psychophysiological procedures can contribute to the delineation of the found.a.tions of social processes; and (3) selected candidates for the elementary processing stages underlying selected interpersonal and intrapersonal processes. Specifically, we have advanced the position that social processes are constituted by the concatenation of unique combinations of a finite number of elementary and generally sequential processing stages. In applying psychophysiology to isolate and identify these elementary stages, we have argued against the general search for universal physiological descriptions (or "correlates") of social processes; rather, we have emphasized how these procedures can be employed within limits to delineate the rudimentary building blocks of social processes. Our review suggests that ( 1) the application of psychophysiological concepts and procedures to investigate the stages of social processes need not be reductionistic or incompatible with the use of social paradigms and methodologies; (2) the selection of psychophysiological measures and procedures and the construction of comparison conditions should be specifically linked to the theoretical question being posed; (3) at least some elementary processing stages, like some chemical elements, show an affinity for one another; and (4) differences among very different social processes may indeed ultimately be understood in terms of the differences in the concatenation of a subset of a finite number of processing stages. though a great deal needs to be accomplished before these fundamental stages are identified and their positions in various social processes are understood.
ACKNOWLEDGMENTS We wish to thank Barbara L. Andersen, Robert S. Baron, and Marcia Ward for comments on an earlier draft of this chapter. Preparation of this chapter was supported by National Science Foundation Grant Nos. BNS 80-23589, BNS 82-17096 & BNS 8444909.

subjects from the high-dissonance, in contrast to the low-dissonance, group increased the number of correct responses they made on the simple test and increased the number of errors they made on the difficult test across the trials of the test. This suggests that effortful striving may be a second processing stage underlying dissonance as well as social facilitation processes. Further research is needed to determine the similarities and differences in the processing stages underlying these two very different social processes, of course; however, psychophysiological recordings may contribute importantly to the delineation of their respective processing stages and to the specification of common elementary stages. Summary We have reviewed a variety of ways in which psychophysiological procedures have been related to attitudes and attitude change. Studies have typically measured only one bodily response (e.g., SRR or pupillary dilation) to assess attitudes, and these studies have been vulnerable to the criticisms that attention, interest, distress, or cognitive load was being measured, rather than positive or negative feelings about the stimulus per se. Assessments of facial EMG activity may prove to be a valid, though possibly insensitive, physiological index of attitudes, but the benefits of such a measure may not outweigh the costs in most instances. The application of psychophysiological procedures, including the assessment of skin conductance and facial EMG, to investigate the underlying processing stages appears to us to be a more useful and promising endeavor. Psychophysiological studies of attitudinal processes are still uncommon, and it is clear even from a casual perusal of this literature that many more processing stages may exist than the ones that have been studied. Moreover, the physiological data pertaining to these stages are themselves so sparse that the conceptualizations of these stages are likely to undergo considerable revision and refinement as more data accrue. The commonality in a processing stage suggested by psychophysiological studies of social facilitation and cognitive dissonance, however, is encouraging, given the thesis that psychophysiological investigations of social processes should uncover commonalities between and distinctions among the constituent stages underlying social processes.

CONCLUSION
Social scientists are chagrined when people see as "obvious" scientific propositions regarding the role of situational factors in social behavior or regarding the effects of human association on the behavior of individuals (d. Holden, 1982; Lazarsfeld, 1949). For example, in a variation of an analysis provided by

674
REFERENCES Allport. G. W. Attitudes. In C. Murchison (Ed.), Handhook of so<:wl psychology ( Vol. 2). Worcester. Mass.: Clark University Press. 1935. Aronson. E. The theory of cognitive dissonance: A current perspective. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 4). New York: Academic Press. 1969. Back. K. W.. & Bogdonoff, M. D. Plasma lipid responses to leadership. conformity. and deviation. In P. H. Leiderman & D. Shapiro (Eds.), Psychobiolol(lcal approaches to social behavior. Stanford. Calif.: Stanford University Press. 1964. Bargh, ). A.. & Cohen. ). L. Mediating factors in the arousalperformance relationship. Motivation and Emonon, 1978. 2. 243-257. Baron. R. S.. Moore. D.. & Sanders. G. S. Distraction as a source of drive in social facilitation research. Journal of Personalityand Social Psychology. 1978.36.816-824. Beatty. ). Task-evoked pupillary responses. processing load. and the structure of processing resources. Psycholol(lcal Bullelln. 1982. 91. 276-292. Bern. D. ). An example analysis of self-persuasion. Journal of Expenmental Social Psychology. 1965. 1. 199-218. Bern. D. J. Self-perception theory. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 6). New York: Academic Press. 1972. Berger. S. M .. Carli. L. L.. Garcia. R.• & Bradv.}. ) .. Jr. Audience affects and anticipatory learning: A comparison of drive and practice-inhibition analyses. Journal of Personalityand Social Psychology. 1982. 42. 478-486. Berger. S. M.. Harnptorn, K. L.. Carli. L. L.. Grandmaison. P. S.. Sadow, ). S.. Donath. C. H .• & Herschlag, L. R. Audienceinduced inhibition of overt practice during learning. Journal of Personalityand SocUlI Psychology. 1981. 40. 479-491. Blascovtch, J .. & Katkin, E. S. Visceral perception and social behavior. In). T. Cacioppo & R. E. Petty (Eds.), Social psychophysiology: A sourcebook. New York: Guilford Press. 1983. Block. A. R. An investigation of the response of the spouse to chronic pain behavior. PsychosomaticMedicine. 1981.43.415422. Bond. C. F.. Jr. Social facilitation: A self-presentational view. Journal of Personalityand Social Psychology. 1982. 42. 1042-1050. Borden. R. ). Audience influence. In P. B. Paulis (Ed.), Psychology of group influence. Hillsdale. N.).: Erlbaum, 1980. Borden. R. ) .• Hendrick. C .• & Walker.]. W. Affective physiological. and attitudinal consequences of audience presence. Bulletin of the Psychonomic Society. 1976. 7. 33-36. Boyd. R. W.. & Dilvlascio, A. Social behavior and autonomic physiology: A sociophysiologic seudv, Journal of Nervous and Mental Disease, 1954. 120. 207-212. Brehm. ). W .• & Cohen. A. R. Explorations in cognitivedissonance. New York: Wiley. 1962. Brener. ). Sensory and perceptual determinants of voluntary visceral control. In G. E. Schwartz &). Beatty (Eds.), Biofeedback: Theory and research. New York: Academic Press. 1977. Brown. B. B. Visual recall ability and eye movements. Psychophysiology. 1968. 4. 300-306. Brown. S. L.. & Schwartz. G. E. Relationships between facial electromyography and subjective experience during affective imagery. Biolol(icalPsychology. 1980. 11. 49-62. Buck. R. Nonverbal behavior and the theory of emotion: The facial feedback hypothesis. Journal of Personality and Social Psychology. 1980. 38. 811-824. Buck. R. W .. Jr. Rela[ionshiPs between dissonance-reducing behavior and tension measures following aggression. Unpublished doctoral dissertation. University of Pittsburgh. 1970. Cacioppo. ). T. Effects of exogenous changes in heart rate on facilitation of thought and resistance to persuasion. Journal of Personality and Social Psychology. 1979.37.489-498. Cacioppo,]. T. Social psychophysiology: A classic perspective and contemporary approach. Psychophysiology, 1982. 19,241-251.

PROCESSES

Cacioppo.}. T.. Harkins. S. G .. & Petty. R. E. The nature of attitudes and cognitive responses and their relationships to behavior. In R. Petty. T. Ostrom. & T. Brock (Eds.), Cognitive responsesin persuasion. Hillsdale. N.).: Erlbaurn, 1981. Cacioppo. ). T.. Marshall-Goodell. B.. & Dorfman. D. Skeletomuscular patterning: Topographical analysis of the integrated electromyogram. PsychophysIOlogy. 1983. 20. 269-283. Cacioppo.). T.• Marshall-Goodell. B.. & Gorrnezano, I. Social psychophysiology: Bioelectrical measurement. experimental control. and analog/digital data acquisition. In ). T. Cacioppo & R. E. Petty (Eds.), Social psychophySIology: A sourcebook. New York: Guilford Press. 1983. Cacioppo. J. T.. & Petty. R. E. Attitudes and cognitive response: An electrophvsiological approach. Journal of Personalityand Social Psychology, 1979.37.2181-2199. (a) Cacioppo.). T.• & Petty. R. E. Lip and nonpreferred forearm EMG activity as a function of orienting task. Journal of Biological Psychology. 1979. 9. 103-113. (b) Cacioppo. ). T.. & Petty. R. E. Electrornvograms as measures of extent affectivity of information processing. American Psychologist. 1981.36.441-456. (a) Cactoppo.}. T.• & Petty. R. E. Electromyographic specificirv during covert information processing. Psychophysiology. 1981. 18. 518-523. (b) Cactoppo.}. T.• & Petty. R. E. Social psychological procedures for cognitive response assessment: The thought-listing technique. In T. V. Merluzzi. C. R. Glass. & M. Genest (Eds.), Cognitive assessment.New York: Guilford Press. 1981. (c) Cacioppo. ). T.• & Petty. R. E. A biosocial model of attitude change: Signs. symptoms. and undetected physiological responses. In). T. Cacioppo & R. E. Petty (Eds.}, Perspectives in cardiovascular psychophysiolol(Y. New York: Guilford Press. 1982. (a) Cacioppo. ). T.. & Petty. R. E. Language variables. attitudes. and persuasion. In E. B. Ryan & H. Giles (Eds.), Attitudes towards language vanallon; SOCIal nd applied contexts. London: Edward a Arnold. 1982. (b) Cacioppo.). T.. & Petty. R. E. Foundations of social psychophysiology. In). T. Cacioppo & R. E. Petty (Eds.), Social psychophysi0101(Y: A sourcebook. New York: Guilford Press. 1983. (a) Cacioppo. ). T.• & Petty. R. E. (Eds.). Social psychophysiology: A sourcebook. New York: Guilford Press. 1983. (b) Cacioppo.). T.. & Petty. R. E. Soviet contributions to social psychophysiology. In ). T. Cacioppo & R. E. Petry (Eds.), Social psychophyslOlol(Y: A sourcebook. New York: Guilford Press. 1983. (c) Cacioppo.}. T.• & Petty. R. E. Social psychophysiology. In L. Stegagno (Ed.), Psychophysiology. Turin. Italy: Boringhieri Publishers. in press. Cacioppo. ). T.. Petty. R. E.. & Marshall-Goodell. B. Electrornvographic specificity during simple physical and attitudinal tasks: Location and topographical features of integrated EMG responses. Biological Psychology. 1984. 18. 85-121. Cacioppo. ). T.• & Sandman. C. A. Psychophysiological functioning. cognitive responding. and attitudes. In R. E. Petty. T. M. Ostrom. & T. C. Brock (Eds.), CognitIVe responses in persuasion_ Hillsdale. N. ).: Erlbaurn, 1981. Cacioppo. ). T.. Sandman. C. A.• & Walker. B. B. The effects of operant heart rate conditioning on cognitive elaboration and attitude change. Psychophysiology. 1978. 15. 330-338. Calder. B. ) .• Ross. M.• & Insko. C. A. Attitude change and attitude attribution: Effects of incentive. choice. and consequences. Journal of PersoTUllity Social Psychology. 1973. 25. 84-99. and Campbell. D. T.• & Stanley.). C. Experimental and quasi-experimental designs for research. Chicago: Rand McNally '. 1963. .. Cannon. W. B. The James-Lange theory of emotions: A critical examination and an alternative theory. American Journal of Psychology. 1927. 39. 106-124. .. Carver. C. S., & Scheier. M. F. The self-attention-mduced feedback loop and social facilitation. Journal of Experimental Social Psychology. 1981. 17. 545-568.

Social Processes

675
Dabbs, J. M., Jr., Johnson. J. E .• &. Leventhal. H. Palmar sweating: A quick and simple measure. Journal of Experimental Prychology, Darwin. C. The expression of the emotions in man and animals. london: Murray. 1904. (Originally published, 1872) Dickson, H. W., &. McGinnies. E. Affectivity and the arousal of attitudes as measured by galvanic skin response. American Journal of Ps-ychology, 1966,79, 584-587. DiMascio. A., Boyd, R. W., &. Greenblatt, M. Physiological correlates of tension and antagonism during psychotherapy: A study of "interpersonal physiology." Ps-ychosomanc Medicine. 1957. and the irrelevance of dissociationism: A reply to Schwartz and Pritchard. Ps-ychophysiology. 1982, 19,457-463. Donders, E C. Die schnelligkeit psvchischer Prozesse. Archit/e fur Anatomie und Physiologie, 1868. 657-681. Droppleman, L. E, &. McNair, D. M. An experimental analog of public speaking. Journal of Consulting and Clinical Prychology, 1971, 36,91-96. Duffy, E. The concept of energy mobilization. Ps-ychological Review, 1951, 58, 30-40. Duffy, E. The psychological significance of the concept of "arousal" or "activation." Ps-ychological Review, 1957.64,265275. Duffy, E. Activation. In N. S. Greenfield &. R. A. Sternbach (Eds.), Handbook of psychophysiology. New York: Holt, Rinehart &. Winston. 1972. Duval. 5., &. Wicklund, R. A. A theory of objectitJe self-awareness. New York: Academic Press. 1972. Dysinger, D. W. A comparative study of affective responses by means of the impressive and expressive methods. Prychological Monographs, 1931, 41(No. 187}. Eagly, A. H. Comprehensibility of persuasive arguments as a determinant of opinion change. Journal of PeTsonaliry and Social Psychology, 1974,29,758-773. Eaglv, A. H., &. Himmelfarb, S. Attitudes and opinions. In M. Rosenzweig &. L. Porter (Eds.), Annual r~iew of prychology (Vol. 29). Palo Alto, Calif.: Annual Reviews. 1978. Ekman, P. Universal and cultural differences in facial expression of emotion. In J. R. Cole (Ed.), Nebraska Symposium on Motivanon (Vol. 19). Lincoln: University of Nebraska Press. 1972. Ekman, P., &. Friesen. W. V. Unmasking the face. Englewood Cliffs, N.J.: Prentice-Hall, 1975. Ekman, P., &. Friesen, W. V. Facial action coding system. Palo Alto, Calif.: Consulting Psychologists Corporation, 1978. Englis, B. G., Vaughan, K. B., &. Lanzetta, J. T. Conditioning of counterernpathic emotional responses. Journal of Expenmenral Social Psychology, 1982, 18, 375-391. Ericcson, K. A., &. Simon, H. A. Verbal reports as data. Ps-ycholigical Ret/iew, 1980,87.215-251. Fazio, R. H., &. Cooper, J. Arousal in the dissonance process. In J. T. Cacioppo &. R. E. Petty (Eds.), Social psychoph;mology: A sourcebook. New York: Guilford Press. 1983. Fazio, R. H .• Zanna, M. P.. &. Cooper, J. Dissonance and self-perception: An integrative view of each theory's proper domain of application. Journal of ExpeTimental Social Prychology. 1977, 13, 464-479. Fesringer, L. A theory of cognmve dissonance. Stanford. Calif.: Stanford University Press, 1957. Fowles. D. C. The three arousal model: Implications of Gray's twofactor learning theory for heart rate. electrodermal activity. and psychopathy. PsychophYSIOlogy, 1980, 17, 87-104. Fowles, D. C. Heart rate as an index of anxiety: Failure of a hypothesis. In J. T. Cacioppo &. R. E. Petry (Eds.), Perspecnves in cardiooascular psychophysiology. New York: Guilford Press, 1982. Fridlund, A. J., Cottam, G. L., &. Fowler, S. C. In search of the general tension factor: Tensional patterning during auditory stimulation. Psychophysiology, 1982, 19, 136-,·145. Fridlund, A. J., &. Izard, C. E. Electromyographic studies of facial expressions of emotions and patterns of emotion. In J. T. Caci-

L. S.• &. Craik, F. 1. M. (Eds.), Levels of pTOcessing in human memory. Hillsdale. N.J.: Erlbaum, 1978. Chapman. A. J. Some aspects of the social facilitation of "humorous laughter" in children. Unpublished doctoral dissertation. UniverCermak. sity of Leicester. 1972. Chapman. A. J. An elecrromvographic study of apprehension about evaluation. Prychological Reports. 1973.33.811-814. Chapman. A. J. An electromyographic study of social facilitation: A test of the "mere presence" hypothesis. British Journal of Prychology. 1974.65. 123-128. Christie. M. J.• &. Mellett, P. G. Foundatioo.s of ps-ychosomatics. Chichester. England: Wiley. 1981. Christie. M. J .• &. Todd. J. L. Experimenter-subject-situational interactions. In P. H. Venables &. M. J. Christie (Eds.), Research in prychophysiology. London: Wiley. 1975. Church. R. M. The effect of competition on reaction time- and palmar skin conductance. Journal of Almcmnal and Social Prychology. 1962. 65. 32-40. Cialdini, R. B.• Petry. R. E .• &. Cacioppo. J. T. Attitudes and attitude change. Annual R~iew of Ps-ychology. 1981. 32. 357-404. Cohen. J. L. Social facilitation increased evaluation apprehension through permanency of record. Monocnon and Emotion. 1979. 3. 19-33. Cohen. J. L.. &. Davis. J. H. Affects of audience status. evaluation. and time of action on performance with hidden-word problems. Journal of PeTsonaliry and Social Prychology. 1973, 27. 74-85. Coles. M. G. H .• &. Duncan-Johnson. C. C. Cardiac activity and information processing: The effects of stimulus significance. detection. and response requirements. Journal of Experimental Prychology: Human PeTception and PeTformance, 1975. 1,418-428. Coles. M. G. H .• Jennings. J. R .• &. Stern. J. A. (Eds.). Prychophysiological perspecnees: Festschrift for BeatTice and John Lacey. New York: Van Nostrand Reinhold. 1984. Collins. B. E.• &. Hoyt. M. G. Personal responsibility for consequences: An integration and extension of the "forced compliance" literature. Journal of Experimental Social Psychology. 1972.8. 558-593. Cook. S. W .• &. Sellnz, C. A multiple-indicator approach to attitude measurement. Ps-ychological Bulletin. 1964. 62. 36-55. Cooper. J .• &. Worchel. S. Role of undesired consequences in arousing cognitive dissonance. Journal of PeTsonaliry and Social Prychology. 1970, 16. 199-206. Cooper. J .• Zanna, M. P.• &. Taves. P. A. Arousal as a necessary condition for attitude change following induced compliance. Journal of PeTsonaliry and Social Prychology. 1978. 36. 1101-

1968,78.347-350.

19.99-104. Donchin, E. The relevance of dissociations

1106.
J. B.• &. Singer. D. N. The role of emotion and prejudice. Journal of Social Prychology. 1956,44,241-247. Cooper, J. B. Emotion and prejudice. Science, 1959. 130.314-318. Corrigan, J. D., Dell. D. M .• Lewis. K. N .• &. Schmidt. L. D. Counseling as a social influence process: A review. Journal of Counseling Prychology Monograph. 1980.27.395-441. Cotton. J. L. A review of research on Schachter's theory of emotion and the misarrribution of arousal. European Journal of Social Prychology. 1981. 11. 365-397. Cooper.

Cornell, N. B. Social facilitation. In C. G. McClintock (Ed.), ExpeTimencal social prychology. New York: Holt, Rinehart &. Winston. 1972. Cornell. N. B.• Wack, D. L.. Sekerak. G. J .• &. Rittle. R. H. Social facilitation of dominant responses by the presence of an audience and the mere presence of others. Journal of PeTsonaliry and

Social Ps-ychology. 1968. 9, 245-250.
Craik, F. M., &. Lockhart. R. S. Levels of processing: A framework for memory research. Journal of VeTbal Learning and VeTbal Behat/ior, 1972. 11,671-684. Crider, A. The promise of social psychophysiology. In J. T. Cacioppo &. R. E. Petty (Eds.), Social ps-ychophysiology: A sourcebook. New York: Guilford Press. 1983. Croyle, R. T., &. Cooper, J. Dissonance arousal: Physiological evidence. Journal of PeTsonaliry and Social Ps-ychology. 1983, 45, 782791.

I

\
676
oppo & R. E. Petty (Eds.), New York: Guilford Press. Gale. A The psychophysiology extraversion and EEG. In PROCESSES

Social psychophysiology: A sourcebook.

tion: Its nature, persistence,

and reinstatement.

Journal of Experi-

1983. of individual differences: Studies in P. Kline (Ed.), New approaches to psychological measurements. London: Wiley. 1973. Gale. A., & Baker, S.In VlOO or in vitro?: Some effects of laboratory environments, with particular reference to the psychophysiological experiment. In M. J. Christie & P. G. Meller (Eds.), Foundations of psychosomatics. Chichester, England: Wiley, 198 J. Gale. A .• & Smith. D. On setting up a psvchophvsiologtcal laboratory. In I. Martin & P. Venables (Eds.), Techniques in psychophysiology. Chichester, England: Wiley, 1980. Gantt, W. H., Newton, J. E. 0.. Royer, F. L., & Stephens, J. H. Affect of person. Conditional Reflex, 1966. I, 18-35. Garrity, L. I. Electromyography: A review of the current status of subvocal speech research. Memory and Cognition. 1977,5.615622. Geen, R. G. Affects of anticipation of positive and negative outcomes on audience anxiety. Journal of Consulting and Clinical Psychology, 1977,45,715-716. Geen, R. G. Affects of being observed on learning following success and failure experiences. Motivation and Emotion, 1979,3,355371. Geen, R. G. The effects of being observed on performance. In P. B. Paulis (Ed.), Psychology of gTOUpinfluence. Hillsdale, N.J.: ErIbaum, 1980. Geen, R. G. Evaluation apprehension and social facilitation: A reply to Sanders. Journal of Experimental Social Psychology. 1981, 17. 252-256. Geen, R. G .• & Gange, J.]. Drive theory of social facilitation: Twelve years of theory and research. Psychological Bulletin, 1977,84, 1267-1288. Gerard. H. B. Physiological measurement in social psychological research. In P. H. Leiderman & D. Shapiro (Eds.), Psychobiological approaches to social behavior. Stanford, Calif.: Stanford University Press. 1964. Gerard, H. B. Choice difficulty, dissonance, and the decision sequence. Journal of Personality, 1967.35,91-108. Gleason. ]. M .• & Katkin, E. S. The efiect5 of cognitive dissonance on heart rate and electrodermal response. Paper presented at the Society for Psychophysiological Research. Madison, Wisconsin, 1978. Goethals, G. R .• & Cooper, ]. When dissonance is reduced: The timing of self-justificatory attitude change. Journal of Personality and Social Psychology, 1975.32.361-387. Goldwater. B. C. Psychological significance of pupillary movements. Psychological Bulletin, 1972,77.340-355. Gould, S. J., & Vrba, E. S. Exaptation-a missing term in the science of form. Paleobiology. 1982. 8, 4-15. Greenfield. N. S., & Sternbach, R. A. (Eds.), Handbook of psychophysiology. New York: Holt. Rinehart & Winston. 1972. Greenwald. A. G .• & Ronis, D. L. Twenty years of cognitive dissonance: Case study of tile evolution of a theory. Psychological Review, 1978.85,53-57. Groff. B. D .• Baron, R. S., & Moore, D. L. Disrraction, attentional conflict. and drivelike behavior. Journal of Experimental Social Psychology, 1983, 19.359-380. Guttman, L. A basis for scaling qualitative data. American Sociological Review, 1944,9.139-150. Hebb, D. O. Drives and the C.N.S. (conceptual nervous system). Psychological Review, 1955.62,243-254. Henchy, T., & Glass. D. C. Evaluation apprehension and the social facilitation of dominant and subordinate responses. Journal of Personality and Social Psycholog" 1968, 10, 446-454. Henry, J. P., & Stephens, P. M. Stress, help, and the social environment: A sociobiologic approach to medicine. New York: SpringerVerlag, 1977. Hess, E. H. Attitude and pupil size. Scientific American, 1965.212. 46-54. Higgins. E. T., Rhodewalt, F., & Zanna, M. P. Dissonance rnotiva-

mental Social Psychology, 1979, 15, 16-34.
Hirschman, R. D., & Katkin, E. S. Relationships among attention, GSR activity, and perceived similarity of self and others. Journal of Personality, 1971, 39, 277-288. Holden, C. Academy boosts social sciences. Science, 1982, 217,

133.
Hord, D. J., Johnson, L. G., & Lubin, A Differential affect of the law of initial value (LlV) on autonomic variables. Psychophysiology, 1964, 1,79-87. Izard, C. E. The face of emotion. New York: Appleton-CenturyCrofts. 1971. Janis, 1. L., & Feshbach, S. Effects of fear-arousing communications. Journal of Abnormal and SocUlI Psychology, 1953, 48. 7892. [anisse, M. P.• & Peavler, W. S. Pupillary research today: Emotion in the eye. Psychology Today. February, 1974. pp. 60-63. Johnson, H. J., & Campos, J. J. The effect of cognitive tasks and verbalization instruction on heart rate and skin conductance. PsychophysiOlogy, 1967. 4, 143-150. Kaplan, H. B. Studies in sociophysiology. In E. Gartlv [aco (Ed.), Patient5, physicians, and illness. New York: Free Press, 1972. Kaplan, H. B. Social psychology of disease. In H. E. Freeman. S. Levine, & L. G. Reeder (Eds.), Handbook of medical sociology (3rd ed.). Englewood Cliffs, N.J.: Prentice-Hall, 1979. Kaplan, H. B., & Bloom, S. W. The use of sociological and socialpsychological concepts in physiological research: A review of selected experimental studies. Journal of Neroous and Mental Disease, 1960, 131, 128-134. Kassarjian, H. H., & Robertson, T. S. Perspectives in consumer behavior. Glenview, III.: Scott-Foresman, 1981. Katkin, E. S. Relationship between manifest anxiety and two indices of autonomic responses to srress. Journal of Personality and Social Psychology, 1965, 2, 324-333. Katkin, E. S. The relationship between a measure of rransitory anxiety and spontaneous autonomic activity. Journal of Abnormal Psychology, 1966,71, 142-146. Katkin, E. S. Electrodermal lability: A psychophysiological analysis of individual differences in response to stress, In 1. G. Sarason & D. Spielberger (Eds.), Stress and anxiety (Vol. 2). Washington, D.C.: Hemisphere, 1975. Kiesler, C. A, & Pallak, M. S. Arousal properties of dissonance manipulations. Psychological Bulletin. 1976,83, 1014-1025. Kissel, S. Stress-reducing properties of social stimuli. Journal of Personality and Social Psychology, 1965,2,378-384. Kraut, R. E. Social presence, facial feedback, and emotion. Journal of Personality and Social Psychology. 1982, 42, 853-863. Krugman, H. E. The efiectitle use of physiological measurement in adtlertising research. Paper presented at the 12th Annual Attitude Research Conference of the American Marketing Association, Hot Springs, Virginia, 1981. Kulpe, O. Outlines of psychology. New York: Macmillan, 1895. Lacey, J. I. Psychophysiological approaches to the evaluation of psychotherapeutic process and outcome. In E. A. Rubinstein & M. B. Parloff (Eds.), Research in psychotherapy (Vol. 1). Washington, D.C.: American Psychological Association, 1959. Lacey, J. I. Somatic response patterning and srress: Some revisions of activation theory. In M. H. Appley & R. Trumbull (Eds.), Psychological stress: Issues in research. New York: Appleton-Century -Crofts, 1967. Lacey, J. I., Kagan, J., Lacey, B. c.. & Moss. M. A. H. The visceral level: Situational determinants and behavioral correlates of autonomic response patterns. In P. N. Knapp (Ed.), Expression of the emotions in man. New York: International Universities Press, 1963. Laird, J. D. Self-attribution of emotion: The effects of expressive behavior on the quality of emotional experience. Journal of Personality and Social Psychology, 1974,29.475-486. Lang, P. J. The application of psychophysiological methods to the study of psychotherapy and behavior modification. In A. E.

c.

pm

Social Proce.sses

677
McGuigan. F. J. Cognitive psychophysiolo/D: Principles of covert behavior. Englewood Cliffs. N.J.: Prentice-Hall. 1978. McGuire. W. J. The nature of attitudes and attitude change. In G. Lindzey &. E. Aronson (Eds.), The handbook of sociaL psychoLOlD (2nd ed.• Vol. 3). Reading. Mass.: Addison-Wesley. 1969. Mesularn, M.• &. Perry. J. The diagnosis of lovesickness: Experimental psychophysiology without the polygraph. PsychophysioLOlD. 1972.9. 546-551. Meralis, S. A.• &. Hess. E. H.• Pupillary response/semantic differential scale relationships. JaurnaL of Research in.Personaliry. 1982. 16. 201-216. Mewborn. C. R.. &. Rogers. R. W. Affects of threatening and reassuring components of fear appeals on physiological and verbal measures of emotion and attitudes. JaurnaL of ExpenmentaL SociaL Psycholo/D. 1979. 15. 242-253. Miller. L. C .• Murphy. R.• &. Buss. A. H. Consciousness of body: Private and public. JournaL of Personaliry and SociaLPsychololf.!. 1981. 41. 397-406. Miller. N.• &. Campbell. D. T. Recency and primacy in persuasion as a function of the timing of speeches and measurements. Jaurnal of AbnormaL and SociaLPsychololf.!. 1959.59. 1-9. Moore. D. L. Are audiences distracting?: BehavioraL and physiologicaL data. Unpublished master's thesis, University of Iowa. 1977. Moore. D. L.. &. Baron. R. S. Social facilitation: A psychophysiological analysis. In J. T. Cacioppo &. R. E. Petty (Eds.). SociaL psychophysioLogy: A sourcebook. New York: Guilford Press. 1983. Moruzzi, G.. &. Magoun, H. W. Brainstem reticular formation and activation of the EEG. Electroencephalography and CLinicaL Neusophysiololf.!. 1949. 1.455-473. Murray. H. A. Explorcuionsin personaliry: A clinicaLand experimentaL study of fifty men of coLlege age. New York: Oxford University Press. 1938. Musante. G.• &. Anker. J. M. Experimenter's presence: Effects on subject's performance. PsychologicaL Reports. 1972. 30. 903904. Nisbett. R. E.. & Wilson. T. D. Telling more than we can know: Verbal reports on mental processes. PsychologicaL Review. 1977. 84. 231-259. Notarious, C. 1.. Wemple. C .• Ingraham. L. J .• Burns. T. J .• &. Kollar. E. Multichannel responses to an interpersonal stressor: Interrelationships among facial display. heart rate. self-report of emotion. and threat appraisal. JaurnaL of PersonaLiry and SociaL Psychololf.!. 1982. 43. 400-408. Obnst, P. A. The cardiovascular-behavioral interaction-as it appears today. Psychophysiololf.!. 1976. 13.95-107. Obnsr, P. A. CardiOliascuiar psychophysiololf.!: A perspective. New York: Plenum Press. 1981. Osgood. C. E.. Suci, G. J .• &. Tannenbaum. P. H. The measurement of meaning. Urbana: University of Illinois Press. 1957. Oskarnp, S. Attitudes and opinions. Englewood Cliffs. N.J.: PrenticeHall. 1977. Pachella, R. G. The interpretation of reaction time in informationprocessing research. In B. Y. Kantowitz (Ed.), Human information processes: Tutorials in performance and cognition. Hillsdale. N.J.: Erlbaum, 1974. Pallak, M. S.• &. Pittman. E. S. General motivation effects of dissonance arousal. JaurnaL of PersonaLiry and SociaL Psychololf.!. 1972. 21.349-358. Pennebaker. J. W. Accuracy of symptom perception. In A. Baum, J. E. Singer. &. S. E. Taylor (Eds.), Handbook of psychololf.! in health (Vol. 4). Hillsdale. N.J.: Erlbaurn, 1982. Pennebaker. J. W. Physical symptoms and sensations: Psychological caruses and correlates. In J. T. Cacioppo &. R. E. Petty (Eds.), SociaL psychophysioLolf.!: A sourcebook. New York: Guilford Press. 1983. Pennebaker. J. W .• & Lightner. J. M. Competition of internal and external information in an exercise setting. JournaL of PersonaLity and SociaL PsychoLolf.!. 1980. 39. 165-174. Pennebaker. J. W.• &. Skelton. J. A. Selective monitoring of physi-

Bergin &. S. L. Garfield (Eds.), Handbook of prychocherap, and behavior change. New York: Wiley • 1971. Latane, B.• &. Cappello H. The effects of togetherness on heart rate in rats. PrychonomicScience. 1972.29. 177-179. Latane, B.• Williams. K. D.• &. Harkins. S. G. Many hands make light the work: The causes and consequences of social loafing. JournaL of Personaliry and Social Psycholo/D. 1979.37.822832. Lazarsfeld, P. F. The American soldier-an expository review. Public Opinion Quarterly. 1949. 13.377-404. Leiderman, P. H.• &. Shapiro. D. (Eds.). Psychobiological approaches to sociaL behavior. Stanford. Calif.: Stanford University Press.

1964.

Lemon. N. Attitudes and their measurement. New York: Wiley. 1973. Leventhal. H. Findings and theory in the study of fear communications. In L. Berkowitz (Ed.), Advances in experimentaL sociaL prycholo/D (Vol. 5). New York: Academic Press. 1970. Leventhal. H. Toward a comprehensive theory of emotion. In L. Berkowitz [Ed.), Admnce.s in experimentaL sociaL psycholo/D (Vol. 13). New York: Academic Press. 1980. Leventhal. H.• &. Mosbach. P. The perceptual-motor theory of emotion. In]. T. Cacioppo &. R. E. Petty (Eds.), SociaL psychophysiolo/D: A sourcebook. New York: Guilford Press. 1983. Leventhal. H.. &. Nerem. D. R. lLIne.ss cognitionsas a source of distress in treatment. Paper presented at the annual meeting of the American Psychological Association. Los Angeles. August 1981. Likert. R. A technique for the measurement of attitudes. Archivesof Prycholo/D. 1932. 140. 1-55(Whole). Linder. D. E.• Cooper. J .• &. [ohes, E. E. Decision freedom as a determinant of the role of incentive magnitude in attitude change. JaurnaL of PersonaLiry and SociaLPsycholo/D. 1967. 6. 245254. Lindsley. D. B. Emotion. In S. S. Stevens (Ed.), Handbook of experimental prycholo/D. New York: Wiley. 1951. Lynch. J. J .• Long. J. M.• Thomas. S. A.• Malinow, K. L.. &. Katcher. A. H. The effects of talking on blood pressure of hypertensive and normotensive individuals. Prychosomatic MediCine. 1981. 43. 25-33. Lynch. J. J .• &. McCarthy. J. F. Social responding in dogs: Heart rate changes to a person. Prychophysiolo/D. 1969. 5. 389-393. Mackay. C. J. The measurement of mood and psychophysiological activity using self-report techniques. In 1. Martin &. P. H. Venables (Eds.), Techniques in prychophysiolo/D. Chichester. England: Wiley. 1980. Malmo. R. B. Measurement of drive: An unresolved problem in psychology. In M. R. Jones (Ed.), Nebraska Symposium on Monvation (Vol. 6). Lincoln: University of Nebraska Press. 1958. Manstead, A. S. R.. &. Wagner. H. L. Arousal-cognition and emotion: An appraisal of two-factor theorv, Current PrychologicaL Rel/iews. 1981. 1. 35-54. Markus. H.The effects of mere presence on social facilitation: An unobtrusive test. JournaL of Experimental SociaL Prycholo/D. 1978. 14.389-397. Markus. H. The drive for integration: Some comments. JaurnaL of ExperimentaL Social Prycholo/D. 1981. 17.257-261. Martens. R. Effect of an audience on learning and performance of a complex motor skill. JaurnaL of Personaliry and SociaL Prycholo/D. 1969. 12. 252-260. (a) Martens. R. Affect on performance of learning a complex motor task in the presence of spectators. Research Quarterly. 1969. 40. 733-737. (b) Martens. R. Palmar sweating in the presence of an audience. JournaL of ExperimentaL SociaLPsycholo/D. 1969.5.371-374. (c) Maslach, C. Negative emotional biasing of unexplained arousal. JaurnaL of Personaliry and Social Psycholo/D. 1979. 37. 953-969. Masling, ] .. Price. J .• Goldband. S.. &. Karkin, E. S. Oral imagery and autonomic arousal in social isolation. JournaL of Personaliry and SociaL Prycholo/D. 1981. 40. 395-400. McCurdy. H. G. Consciousness and the galvanometer. PrychologicaL Rel/iew. 1950.57.322-327.

678
cal sensations. Journal of Personalityand Social Psychology, 1981, 41. 213-223. Petty, R. E., & Cacioppo, J. T. Forewarning, cognitive responding and resistance to persuasion. Journal of Personality and Social PS:1Chology, 1977, 35, 645-655. Petty, R. E., & Cacioppo, J. T. Issue-involvement can increase or decrease persuasion by enhancing message-relevant cognitive responses. Journal of Personalityand SOCIal Psychology, 1979,37, 1915-1926. Petty, R. E., & Cacioppo, J. T. Attitudes and persuasion: Classic and contemporary approaches. Dubuque, la.: Wm. C. Brown, 1981. Petty, R. E., & Cacioppo, J. T. The role of bodily responses in attitude measurement and change. In J. T. Cacioppo & R. E. Petty (Eds.), Social psychophysiology: A sourcebook. New York: Guilford Press, 1983. Petty, R. E., Harkins, S. G., Williams, K. D., & Latane, B. The effects of group size on cognitive effort and evaluation. Personality and Social PS:1Chology Bulletin, 1977, 3, 579-582. Petty, R. E., Wells, G. L., Heesacker, M., Brock, T. c.. & Cacioppo, J. T. The effects of recipient posture on persuasion: A cognitive response analysis. Personality and Social Psychology Bul· letin, 1983,9,209-222. Pribrarn, K. H. Interrelations of psychology and the neurological disciplines. In S. Koch (Ed.), Psychology: A study of a science (Vol. 4). New York: McGraw-Hill, 1962. Prokasy, W. F., & Raskin, D. C. (Eds.). ElectTOdermal activiry in pS:iChological research. New York: Academic Press. 1973. Rankin. R. E.. & Campbell, D. T. Galvanic skin responses to Negro and white experimenters. Journal of Abnormal and Social PS:iChology. 1955. 51, 30-33. Rappaport. H.• & Katkin, E. S. Relationships among manifest anxiety. responses to stress. and perception of autonomic activity. Journal of Consulting and Clinical Psychology, 1972, 38. 219-224. Riddle. E. M. Aggressive behavior in a small social group. Archives of PS:1Chology. 1925 (No. 78). Riess. M.• & Schlenker. B. R. Attirude change and responsibility avoidance as modes of dilemma resolution in forced-compliance situations. Journal of Personality and Social PS:1Chology, 1977.35, 21-30. Roloff. M. E., & Miller. G. R. Persuasion: New directions in theory and research. Beverly Hills. Calif.: Sage, 1980. Rourtenberg, A. The two-arousal hypothesis: Reticular formation and limbic systems. PS:1Chological Review, 1968, 75, 51-80. Sabini, j., & Silver, M. Introspection and causal accounts. Journal of PersonaLit"y Social PS:iChology, 1981,40,171-179. and Sanders, A. F. Stage analysis of reaction processes. In G. E. Stelmach & J. Requin (Eds.), Tutorials in motor behavior. Amsterdam: North-Holland, 1980. Sanders. G. S. Driven by distraction: An integrative review of social facilitation theory and research. Journal of Experimental Social PS:iChology. 1981. 17,227-251. Sasfy. J .• & Okun, M. Form evaluation and audience expertness as joint determinants of audience effects. Journal of Experimental Social PS:iCholo". 1974. 10.461-467. Schachter. S. The interaction of cognitive and physiological determinants of emotion. In P. H. Leiderman & D. Shapiro (Eds.), PS:iChobiological approaches to social behavior. Stanford. Calif.: Stanford University Press. 1964. Schachter. S. Second thoughts on biological and physiological explanations of behavior. In L. Berkowitz (Ed.), Cognitive theories in social pS:iChology. New York: Academic Press, 1978. Schachter. S.• & Singer. J. E. Cognitive, social, and physiological determinants of emotional state. PS:iChological Review. 1962.69, 379-399. Scheier. M. F.. Carver, C. S.• & Matthews. K. Attentional factors in the perception of bodily states. In J. T. Cacioppo & R. E. Petty [Eds.), Social pS:iChophysiology: A sourcebook. New York: Guilford Press. 1983. Schlenker. B. R. Impression management: The self-concept, social identiry, and interpersonal relations. Monterey, Calif.: Brooks/Cole. 1980.

PROCESSES

Schwartz, G. E. Biofeedback, self-regulation, and the patterning of physiological processes. American Scientist, 1975,63,314-324. Schwartz. G. E. Cardiovascular psychophysiology: A systems perspective. In J. T. Cacioppo & R. E. Petty (Eds.), Perspectivesin cardiovascular psychophysiology. New York: Guilford Press. 1982. Schwartz, G. E. Social psychophysiology and behavioral medicine: A systems perspective. In J. T. Cacioppo & R. E. Petty (Eds.), Social psychophysiology: A sourcebook. New York: Guilford Press. 1983. Schwam, G. E.. Fair. P. L.. Mandel, M. R., Salt. P.. Mieske, M., & Klerman, G. L. Facial electromyography in the assessment of improvement in depression. Psychosomatic Medicine, 1978. 40, 355-360. Schwartz. G. E., Fair. P. L.. Salt, P., Mandel, M. R., & Klerman, G. L. Facial expressions and imagery in depression: An electromyographic study. Psychosomatic Medicine, 1976, 38, 337347. (a) Schwartz. G. E., Fair, P. L., Salt. P., Mandel. M. R., & Klerman, G. L. Facial muscle patterning to affective imagery in depressed and nondepressed subjects. Science, 1976, 192,489-491. (b) Schwam, G. E., & Shapiro, D. Social psychophysiology. In W. F. Prokasy & D. C. Raskin (Eds.), Electrodermal activity in pS:1Chological research. New York: Academic Press, 1973. Schweickert, R. Critical-path scheduling of mental processes in a dual task. Science, 1980, 209, 704-706. Shaffer, D. R. Some effects of consonant and dissonant attitudinal advocacy on initial attitude salience and attitude change. Journal of Personality and Social Psychology, 1975, 32, 160-168. Shapiro, D., & Crider, A. Psychophysiological approaches to social psychology. In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (2nd ed., Vol. 3). Reading, Mass.: AddisonWesley, 1969. Shapiro, D.. & Leiderman, P. H. Arousal correlates of task role and group setting. Journal of Personality and Social PS:1Chology. 1967, 5, 103-107. Shapiro, D., Leiderman, P. H., & Morningstar, M. E. Social isolation and social interaction: A behavioral and physiological comparison. In P. H. Leiderman & D. Shapiro (Eds.), Psychobiological approaches to social behavior. Stanford, Calif.: Stanford University Press, 1964. Shapiro, D., & Schwartz. G. E. Psychophysiological contributions to social psychology. Annual RevietlJ Psychology, 1970, 21, 87of 112. Sherman, S. J. Effects of choice and incentive on attitude change in a discrepant behavior situation. Journal of Personality and Social Psychology, 1970, IS, 245-252. Singer, J. E.• Lundberg, U .. & Frankenhaeuser, M. Stress on the train: A study of urban commuting. In A. Baum,]. E. Singer, & S. Valins (Eds.), Advances in environmental pS:1Chology (Vol. I). Hillsdale. N.J.: Erlbaum, 1978. Smith, C. E. The autonomic excitation resulting from the interaction of individual opinion and group opinion. Joumal of Abnormal and Social Psychology. 1936,30, 138-164. Smith, E. R., & Miller, F. D. Limits on perception of cognitive processes: A reply to Nisbett and Wilson. Psychological Review, 1978,85,355-362. Smith, M. J. Persuasion and human action. Belmont, Calif.: Wadsworth, 1982. Spence. K. W. Behavior theory and conditioning.New Haven, Conn.: Yale University Press. 1956. Spitzer, L., & Rodin, J. Arousal-induced eating: Conventional wisdom or empirical finding? In j. T. Cacioppo & R. E. Petty (Eds.), Social psychophysiology: A sourcebook. New York: Guilford Press, 1983. Sternbach, R. A. Principles of psychophysiology. New York: Academic Press, 1966. Sternberg, S. The discovery of processing stages: Extensions of Donders' method. Acta PS:iChologia, 1969,30,276-315. Strahan, R. F., Todd,]. B., & Inglis, G. B. A palmar sweat measure particularly suited for naturalistic research. PS:1Chophysiology, 1974,11,715-720.

Social Processes

679
Wallerstein. H. An electromyographic study of attention listening.

Strong, S. R. Social psychological approach to psychotherapy research. In S. L. Garfield &.. A. E. Bergin (Eds.), Handbook of psychothel'a{ry and behavior change (2nd ed.). New York: Wiley, 1978. . Strube, N. G., Miles, M. E .• &.. Finch. W. H. The social facilitation of a simple task: Field tests of alternative explanations. Personality and Social Psycho log, Bulletin, 1981. 7. 701-707. Svz, H. C. Observations on the unreliability of subjective reports of emotional reactions. British Journal of Psycholog,. 1926, 17, 119126. Szpiler,]. A .• &.. Epstein. S. Availability of an avoidance response as related to autonomic arousal. Journal of Abnonnal Psycholog,. 1976,85, 72-82. Tedeschi.]. T .• Schlenker. B. R .• &.. Bonoma, T. V. Cognitive dissonance: Private ratiocination or public spectacle? American Psychologist. 1971, 26, 685-695. Thayer. R. E. Activation states as assessed by verbal report and four psychophysiological variables. Psychophysiolog,. 1970. 7, 86-94. Thurstone, L. L. Attitudes can be measured. American Journal of Sociolog" 1928, 33. 529-544. Totten. E. Eye movement during visual imagery. Compcranoe Psycholog, Monographs. 1935. 11, 1-46. Triplett, N. The dynamogenic factors in pacemaking and competition. American Journal of Psycholog" 1898.9, 507-533. Turskv, B., &.. ]amner, L. Evaluation of social and political beliefs: A psychophysiological approach. In ]. T. Cacioppo &.. R. E. Petry (Eds.}, Social psychophysiolog,: A sourcebook. New York: Guilford Press, 1983. Uetz, G. W .• Kane, T. C .• &.. Stratton, G. E. Variation in the social grouping tendency of a communical web-building spider. Science, 1982,217,547-549. Valins, S. Cognitive effects of false heart rate feedback. Journal of Personality and Social Psycholog,. 1966.4.400-408. Vanderwolf, C. H .• &.. Robinson. T. E. Renculo-corncal activity and behavior: A critique of the arousal theory and a new synthesis. Behavioral and Brain Sciences. 1981.4.459-514. Vaughan. K. D .• &.. Lanzetta,]. T. Vicarious instigation and conditioning of facial expression in autonomic responses to a model's display of pain. Journal of Personality and Social Psycholog,. 1980, 38. 909-923. Vaughan, K. D., &.. Lanzetta, ]. T. The effect of modification of expressive displays on vicarious emotional arousal. Journal of Experimental Social Psycholog,. 1981, 17, 16-30. Venables, P. H .• &.. Christie. M. ]. Electrodermal activity. In 1. Martin &.. P. H. Venables (Eds.), Techniqua in psychophysiolog,. Chichester, England: Wiley, 1980. Waid. W. SOClophyslOlog,. New York: Springer-Verlag, 1984.

Canadian Journal of Psycholog" 1954, 8. 228-238. Weiss, R. F.. &.. Miller, F. G. The drive theory of social facilitation. Psychological R~~. 1971,78,44-57. White, P. Limitations on verbal reports of internal events: A refutation of Nisbett and Wilson and of Bern. Psychological Retli~,
1980,87,105-112. Wicklund. R. A.. &.. Brehm, ]. W. Perspectives on cognitive dissonance. Hillsdale, N.].: Erlbaum, 1976. Will. G. F. But first. a message from ... N~sweek, May 10, 1982, p.98. Wilson. W. R. Feeling more than we can know: Exposure effects without learning. Journal of Personality and Social Psycholog" 1979.37,811-821. Woodmansee.].]. The pupil response as a measure of social attitudes. In G. F. Summers (Ed.), Attitude measurement. Chicago: Rand McNally. 1970. Zajonc, R. B. Social facilitation. Science, 1965. 149, 269-274. Zajonc, R. B. Compresence. In P. B. Paulis (Ed.), Psycholog, of group in/1uence. Hillsdale, N.].: Erlbaum, 1980. (a) Zajonc, R. B. Feeling and thinking: Preferences need no inferences. American Psychologist. 1980,35,151-175. (b) Zajonc, R. B.• Heingarner, A., &.. Herman, E. M. Social enhancement and impairment in performance in the cockroach. Journal of Personality and Social Psycholog,. 1969, 13, 83-92. Zajonc, R. B., &.. Sales, S. M. Social facilitation of dominant and subordinate responses. Journal of Experimental Social Psychology. 1966.2, 160-168. Zanna, M. P., &.. Cooper,]. Dissonance and the pill: An attribution approach to studying the arousal properties of dissonance. Journal of Personality and Social Psycholog,. 1974, 29,703-709. Zanna, M. P.• &.. Cooper,]. Dissonance and the attribution process. In ]. H. Harvey, W. ]. Ickes. s, R. F. Kidd (Eds.). N~ directions in attribution research. Hillsdale. N.].: Erlbaurn, 1976. Zillmann, D. Attribution and the misattribution of excitatory reaction. In J. H. Harvey, W. ]. Ickes. &.. R. F. Kidd (Eds.). N~ directlons in attribution research (Vol. 2). Hillsdale, N.].: Erlbaurn, 1978. Zillmann, D. Hostility and aggression. Hillsdale, N.].: Erlbaum, 1979. Zillmann. D. Transfer of excitation in emotional behavior. In J. T. Cacioppo &.. R. E. Petry (Eds.), Social psychophysiolog,: A sourcebook. New York: Guilford Press, 1983. Zillmann, D. Connections between sex and aggression. Hillsdale, N.J.: Erlbaum, 1984. Zuckerman. M., Larrance, D. T .• Spiegel, N. A., &.. Klorman, R. Controlling nonverbal cues: Facial expressions and tone of voice. Journal of Experimental Social Psycholog" 1981, 17. 506524.