Nonverbal behavior and political leadership - nonverbal behavior and the dimensions of relationship

Nonverbal Behavior and Political Leadership - NONVERBAL BEHAVIOR AND THE DIMENSIONS OF RELATIONSHIP

Historian Shelby Foote describes the way a military leader's nonverbal behavior, in a particular instance, was misconstrued. During the Civil War, a Union general checked into the Willard Hotel in Washington, D.C. He struck an observer as having 'no gait, no station, no manner.' Rather, his aspect of 'rough, light-brown whiskers, a blue eye, and rather scrubby look withal as if he was out of office and on half pay' suggested someone who need not be taken seriously. The desk clerk assumed a superior air. When the general wrote his name in the regis ter, 'U.S. Grant Galena, Illinois,' things changed fast. The clerk rang the bell loudly, and the observer took a new look. On second glance, he 'perceived that there was more to him than had been apparent before . The `blue eye' became a `clear blue eye,' and the once stolid-seeming face took on 'a look of resolution, as if he could not be trifled with.' ' (Foote, 1963, pp. 3-4).

People's impressions of Ulysses S. Grant during the Civil War, or any leader at any time, are based on several kinds of information, in­cluding appearance, nonverbal behavior, and context. Leaders most often speak or write, but their words are often qualified by their non-verbal behavior, and their verbal and nonverbal behavior together are interpreted quite differently depending on contextual information. This chapter considers the role of nonverbal behavior in political leadership. Obviously, nonverbal behavior does not exist in a vac­uum. It combines with words to help create an overall impression or reaction. These impressions and reactions are key elements in lead­ing and following.

We will review briefly some basic theoretical formulations about the role of nonverbal behavior in communicating information about rela­tionships, examine anecdotally the role of nonverbal behavior in influ‑

encing viewer impressions of candidates in political debates, and dis­cuss several studies my colleagues and I have conducted at Williams College on nonverbal behavior and perceptions of leadership.

NONVERBAL BEHAVIOR AND THE DIMENSIONS
OF RELATIONSHIP

Timothy Leary's (1957) classic volume, Interpersonal Diagnosis of Personality, argued from extensive data bases that interpersonal be­havior could be usefully catalogued along two dimensions (cf. Carson, 1969). First, interpersonal behaviors vary along an affective dimen­sion, that is, in how much positive vs. negative feeling they express. Second, they vary along a status dimension, that is, in how much dom­inance vs. submission they express. Thus, interpersonal behaviors can be categorized as expressing love or friendliness on the one hand versus hate or hostility on the other. Or, they can be categorized as re­vealing neither distinctly positive nor negative affect. Also, those be­haviors can be categorized as expressing dominance, submissiveness, or neither. Since a behavior can express one pole or the other of each dimension, or neither pole, the result, very roughly, is 8 different kinds of interpersonal behavior: behavior that is simply dominant, behavior that is friendly and dominant, behavior that is simply friendly, and so forth. In theory there is a ninth kind of interpersonal behavior, one that expresses neither positive nor negative affect, nor dominance vs. sub­missiveness.

One of Leary's key contentions was that each type of interpersonal behavior invites or elicits a complementary type. Specifically, on the af­fective dimension, both friendly and unfriendly behaviors invite simi­larly friendly or unfriendly behavior in return. In converse, on the status dimension, behaviors invite their opposite or counterpart. Thus while friendly behavior invites friendly behavior in return, dominating behavior invites submission in return. So, for example, friendly-domi­nant behavior from one person invites friendly-submissive behavior from another. Furthermore, while every interpersonal behavior invites its complement, people often but not always behave in the way they are invited to (Markey, Funder, & Ozer, 2003). For example, friendly-domi­nant behavior might produce friendly-dominant behavior in return, rather than friendly-submissive behavior. The second individual is happy to have a friendly interaction, but he or she does not want to be submissive.

One can see interpersonal behavior in flux along these lines in one of President John F. Kennedy's exchanges with reporters. Kennedy had a good relationship with the press, and he enjoyed having the upper hand in that relationship. His behavior was friendly-dominant. Luck­ily for him, most reporters were willing to be essentially submissive in their exchanges with him, complementing and therefore reinforcing his interpersonal behavior. On one occasion during the 1960 cam-

paign a reporter repeatedly pressed the assertion, which Kennedy de­nied, that he had advocated reducing the federal debt early in his presidential term. Kennedy responded very firmly, but with a smile on his face: 'No, never. No ' The reporter accepted Kennedy's denial, re sponding in a friendly-submissive way, complementing Kennedy's be­havior, after a little urging. The complementary behavior from the reporter avoided a spiraling exchange of dominant behavior from both parties, and it cut off the possibility that behavior in the exchange might become hostile. Leaders frequently employ dominant interper­sonal behavior, inviting submissive behavior in return.

Leary argued that although people's actions vary with the situation, each individual's actions are marked by certain preferred categories of interpersonal behavior. These preferred categories reflect 'security operations.' People behave in ways that are comfortable for them, in large part because they are secure or comfortable when others behave in the complementary fashion that is elicited by their own behavior. The person who characteristically behaves in a friendly-submissive manner does so because he is comfortable performing that kind of be havior and equally comfortable with others behaving in a complemen­tary friendly-dominant fashion.

Roger Brown (1965) proposed an analysis of interpersonal relation-ships that highlighted the same two dimensions identified by Leary. Brown called them the dimensions of status and solidarity. Just as an interpersonal behavior, or a set of interpersonal behaviors in an inter-action, or across interactions, can be classified as friendly vs. un­friendly and dominant vs. submissive, so a relationship can be classified as being friendly or hostile, or neither one particularly, and also as one in which the two parties to the relationship have equal or differential status. The pair might have the same status level, as in the case of two corporals in an infantry unit, or one might have higher status, as in the case of a supervisor and a subordinate.

In this light, one person's interpersonal behavior not only elicits a particular kind of interpersonal behavior from another, it also ex-presses a definition of their relationship. John Kennedy's friendly-dominant behavior with reporters not only elicited friendly-submis­sive behavior from them, it signaled his definition of the relationship as one that was warm and cordial, but also one where he had a higher status. In expressing a definition of the relationship, especially along the status dimension, people express their view of themselves and oth ers. Erving Goffman's (1955) essay on face-work makes this point very clear. In social encounters, Goffman argues, a person performs 'a pat tern of verbal and nonverbal acts by which he expresses his view of the situation and through this his evaluation of the participants, especially himself. Regardless of whether a person intends to [do this], he will find that he has done so in effect' (p. 213).

The idea that interpersonal behavior expresses a definition of the re­lationship is developed in Watzlawick, Bean, and Jackson's (1967)

Pragmatics of Human Communication. They argued that every be­havior is a communication and that communication takes place on two levels simultaneously. First, at the content level behavior communi­cates about external tasks, problems to be solved, recreational activi­ties, the stock market, the National Football League, appropriate dress for a party, etc. However, at the relationship level behavior communi­cates each person's view that the relationship has high solidarity, that it is cordial or even affectionate and close, or that the relationship has low solidarity In addition it communicates each person's view of the relative status of the two parties in the relationship-they are equal or one is superior.

In sum, the works of Leary, Brown, Goffman, and Watzlawick et al. suggest that both interpersonal behavior and interpersonal relation-ships can be classified according to dimensions of friendliness and dominance or status, and that one's interpersonal behavior communi­cates one's view of oneself, the other, and their relationship, and also invites the other to complement the behavior and thereby signal accep­tance of that definition of the relationship. In our example above, John Kennedy's firm but smiling, 'No, never. No,' defined his relationship with the reporter as friendly but unequal, with Kennedy being domi­nant, or having higher status. The reporter accepted the definition.

Watzlawick et al., and to some extent Roger Brown, in his 1986 So­cial Psychology, the Second Edition, argues that nonverbal behavior plays a critical role in communicating at the relationship level. That is, while words and nonverbal behavior are completely intertwined (Henley, 1977), one can look at their separate contributions to commu­nication. Watzlawick et al. holds that most communication about rela­tionships is done nonverbally for two reasons. First, there are constraints on the candid expression of interpersonal feelings. Sec­ond, our verbal vocabulary for describing our feelings about others, especially along lines of solidarity and status, is quite limited. Brown believes that words can express feelings in addition to views about ex­ternal matters, but that nonverbal behavior is important in communi­cating about relationships because deceiving people with nonverbal behavior is more difficult than simply telling lies. Our feelings about others sometimes leak, whether we like it or not, through nonverbal channels, especially voice quality. Nonverbal communication that is consistent with words underscores the credibility of what is said. In sum, our views of relationship are expressed implicitly but credibly through various nonverbal expressions: gestures, tone of voice, pro­xemics, facial expressions, etc.

NONVERBAL BEHAVIOR AND LEADERSHIP ENACTMENTS

People express their view of themselves and their relationship to oth­ers largely through nonverbal behavior. In the case of leaders, interper­sonal nonverbal behavior will typically be assertive and express

superiority in the relationship. In Watzlawick's terms, it will be 'one-up' behavior. Others are invited to be 'one-down,' essentially follow­ers. As many leadership theorists have argued (e.g., Burns, 1978; Hol­lander, 1993), if leadership is to exist, others must accept the invitation to follow. Burns emphasizes that leadership is a relation-ship. Hollander begins with the assumption that followers 'accord and withdraw support to leaders' (p. 29). Thus leadership is negotiated and followers accord support to individual leaders depending on their own motives. But how precisely do leaders use nonverbal behavior to appeal to others to accord them support?

First, while our focus is on nonverbal behavior, as noted above non-verbal and verbal behavior are distinguishable but inseparable. They combine in a Gestalt. Roger Brown (1986) wrote: 'Certainly, good ac-tors can contribute something to the emotional impact of Romeo and Juliet, but it is generally supposed that Shakespeare's exclusively ver­bal contribution (the written transcript) is not negligible.' (p. 497). Both verbal and nonverbal elements are important. Also, recall that Erving Goffman's quote above refers to the 'pattern of verbal and non-verbal acts' that expresses a person's view of himself, others, and the situation. Finally, Howard Gardner's (1995) Leading Minds contends that leadership is primarily about the 'stories' leaders tell. In most cases the stories are told or 'related' with words. But in addition to telling stories, leaders embody their stories to varying degrees. Their behavior other than words, that is, their nonverbal behavior, may illus­trate the story, or may contradict it. Ronald Reagan's story of the im­portance of a strong American military was somewhat undermined by the fact that he spent World War II in Culver City making movies rather than fighting in Europe or the Pacific. On the other hand, Pope John XXIII preached a message of humility and openness, and embodied it in his nonverbal behavior. He smiled, bowed, and listened. His well-il­lustrated story provoked a strong negative reaction, a counterstory, from the church hierarchy. But Pope John had the advantage of telling and embodying a story that was resonant with the teachings and life of Jesus. His story was compelling, and many followers accorded him support (Gardner, 1995).

One exploration of leadership that makes plain the interaction of words and nonverbal expression is John Keegan's (1987) Mask of Command, a study of military leadership:

lLeaders] are both shown to and hidden from the mass of humankind, revealed by artifice, presented by theatre. The theatrical impulse will be strong in the successful politician, teacher, entrepreneur, athlete, or di-vine, and will be both expected and reinforced by the audiences to which they perform . at they should know of him must be what they hope and require. at they should not know of him must be concealed at all costs. The leader of men in warfare can show himself to his followers only through a mask, a mask that he must make for himself, but a mask

made in such form as will mark him to men of his time and place as the leader they want and need. (p. 11)

Keegan illustrates this perspective exceptionally well in his treat­ment of Alexander the Great. Usefully, he comments both on the theat­rical (largely nonverbal) and oratorical (largely verbal) aspects of Alexander's leadership. 'Theatricality was at the very heart of Alexan­der's style of leadership His appearances in the field of battle [were] dramatic stage entries, tellingly timed and significantly costumed' (pp. 47-48). As for oratory, Alexander combined verbal and nonverbal ele­ments in leading effectively. He had a 'forceful and collected style' and used the rhetorical devices of a prebattle speech, urging his soldiers to follow him, and to achieve victory once again as they had in the past.

Richard Brookhiser's (1996) biography of George Washington ex-plains the role of nonverbal behavior in our first president's leader-ship. He reports that an English visitor wrote during Washington's first presidential term 'Washington has something uncommonly majestic and commanding in his walk, his address, his figure, and counte­nance' (p. 52). Brookhiser writes that he 'had physical authority in its simplest form, and though he enhanced it with exercise and adorn­ment, they functioned as supplements, not substitutes' (p. 56). His fa­cial expression often showed a towering temper, but usually under firm control. Washington was not an orator, but when he spoke the nonverbal supports increased his impact and capacity to lead.

A final example is Ronald Reagan. As David Gergen (2000) writes in Eyewitness to Power, Reagan himself attributed his reputation as 'the great communicator' to what he said: 'It was the content. I wasn't a great communicator, but I communicated great things' (p. 216). Rea­gan, of course, was kidding. He had honed his speaking skills over many years, first as an actor and then as a pitchman for General Elec­tric, and later for Barry Goldwater. There are many mannerisms and nonverbal attributes that made Reagan effective. Just as one example, for the moment, many have commented on his 'honeyed voice.' Gergen wrote that 'He spoke in warm, velvety tones that enveloped lis­teners and made them feel good-about themselves and about him he talked softly, even gently at times' (pp. 218-219). But importantly as well, the voice and manner matched the words, and gave them addi­tional credibility and impact. Reagan could also speak in a stern, force­ful manner, as when he challenged the Russians to historic change in Berlin: 'Mr. Gorbachev, tear down this wall.'

There is one fascinating study of the impact of Reagan's manner and the way his nonverbal behavior drew people in (McHugo, Lanzetta, Sullivan, Masters, & Englis, 1985). Undergraduates at Dartmouth College watched videotaped excerpts from Reagan press conferences. They were asked to report their reactions to what they saw and heard, and their emotional reactions were also assessed via facial electromyo­graphy, skin resistance, and heart rate. The students' prior attitudes

toward Reagan influenced their self-reports of their emotional reaction to the videotaped excerpts. They reported more positive reactions if they liked Reagan and agreed with his policies. However, their auto­nomic and facial muscle responses were independent of their prior at titudes. The results suggest that even those who disagreed with Reagan couldn't help but like him, and weren't fully aware of how positively they were responding to him.

In sum, leaders' appeal to followers or potential followers is based on some combination of their words and their nonverbal behavior. In politics today, much of our information about candidates for office co mes through watching them on television. We hear them, but we also see them. The importance of television in conveying potentially influ­ential information about nonverbal behavior is underlined by the McHugo et al. (1985) study noted above. They conclude their discus­sion of students' reactions to Ronald Reagan stating 'the present re­sults indicate that expressive displays can cause emotional reactions that are independent of prior attitudes when emitted by a powerful political leader and presented on television. To the extent that there is a trend toward candidate style variables and away from political party and issue positions in determining vote choice, the role of nonverbal behavior in electoral politics is increasingly important.' Nonverbal behavior 'may play an important role in forming and modifying impressions of political leaders who gain extensive exposure to voters through television' (p. 1528).

One of the ways that political candidates get access to voters is through televised debates. Televised debates between John F. Kennedy and Richard M. Nixon were important in the 1960 election and they have been a staple of presidential elections every four years starting with the 1976 debates between Jimmy Carter and Gerald Ford. We wi11 describe the role of televised debates throughout their history on the American political landscape, and then discuss some of our own re-search on the role of nonverbal cues in debates.

NONVERBAL BEHAVIOR IN PRESIDENTIAL DEBATES

It has been widely believed since the publication of Marshall McLuhan's (1964) Understanding Media that televised debates were critical to John F Kennedy's slim victory over Richard Nixon in the election of 1960. McLuhan argued that voters who watched the debate on television tended to perceive Kennedy as the winner while those who listed on the radio thought that Nixon won. There are data show ing that Kennedy, but not Nixon, was helped by his appearance in the first of four joint debates. Tannenbaum, Greenberg, and Silverman (1962) found that Kennedy was perceived as much more 'experienced' after the first debate. This change undermined Nixon's campaign slo gan, 'Experience Counts,' implying that he, Nixon, had more relevant experience for executive leadership than Kennedy. Also, the changes in

voters' perceptions of Kennedy were all in the direction of what people had specified as their image of the 'ideal President.' (Weiss, 1968). Changes in voters' impressions of Nixon were unsystematic. Can we begin to specify some of the elements that made Kennedy more appeal­ing on television?

One possibility is that Kennedy was simply better looking than Nixon. We have some data that suggest that good looks are a part of the story, but only a part. In an experiment, Williams College students were shown excerpts from the first 1960 Kennedy/Nixon debate in three conditions. In an Audio Only condition, participants simply lis tened to an audio recording of debate excerpts. In the Audiovisual con dition, the excerpts were ordinary televised versions. In an Audio Still condition, an audio track was accompanied by a video tape that showed still or nonmoving pictures taken from the video tape, each still segment lasting about 20 seconds. The relative ratings of Kennedy and Nixon were least favorable to Kennedy in the Audio Only condition and most favorable to Kennedy in the Audiovisual condition, with the Audio Still condition falling between the other two. Although extremely preliminary, these results suggest there was more going on than people simply judging that Kennedy was better looking.

A close look at a videotape of the debate suggests some of the visual factors that might have made a difference. The camera shows the de-bate moderator, Howard K. Smith, seated behind a small table. On ei ther side of him is a chair. Kennedy sits on the left of the screen, Smith is in the middle, and Nixon is on the right. Farther to the left and right are podiums behind which the candidates are to speak. The procedure that is followed is that each man rises from his chair next to Smith's ta ble, and walks to his respective podium to speak. In this setting, there are several telling differences in the candidates' appearance and non verbal behavior. First, when Howard K. Smith introduces the two can didates, Kennedy nods in a relaxed self-assured manner, while Nixon fidgets, moving his arms awkwardly, and nods in a jerky manner. Sec­ond, at several times during the candidates' eight-minute opening statements, first by Kennedy, then Nixon, the other candidate is shown, Nixon while Kennedy is speaking and Kennedy while Nixon is speaking. Kennedy is shown taking notes while Nixon speaks, writing in a very rapid, focused and confident manner. Nixon sometimes seems drawn and haggard, watching Kennedy when Kennedy speaks. The fact that Kennedy appears more in command of himself and the occasion when Nixon speaks rather than vice-versa led some observ ers to joke 'The cameraman was a Democrat.' Third, at the end of each of their opening statements, the camera follows the candidates back to their chairs. Kennedy walks back deliberately, sits down, folds his hands, crosses his legs, and looks self-satisfied. In contrast, Nixon moves in one direction and then the other, seemingly unsure of where he should be going, and then sits down, looking somewhat awkward and confused.

Finally, there is a fascinating sequence after the opening statements. The first question is directed to Kennedy. He begins answering the question from his chair, right next to Smith. Kennedy is violating the rules for the debate, which specify that when speaking, the candidates should rise from their chairs and walk to their podiums. Nixon looks agitated and gets Smith's attention, and gestures that Kennedy should go to his podium. Perhaps the rules are very much on Nixon's mind, since he was unsure of where he should go the minute before, when he concluded his opening statement. Smith also seems unsure of what ac­tion to take. He begins to bang his gavel, his only prop at the table, but then stops, whispers to Kennedy, and points to the podium. With al-most no interruption in his response, Kennedy gets up and walks to the podium and continues answering the question. There is a large dif­ference between the ways this moment plays out on television vs. the radio. On the radio, there would be a very brief, but perhaps notice-able, pause in Kennedy's reply. This might be taken to signal that Ken­nedy is unsure of how to respond. On television, an interesting nonverbal mini-drama plays out. 1Wo men, Nixon and Smith, look slightly perplexed and agitated. One, Kennedy, responds to the flow of events calmly and gracefully. To be sure, the account above is simply this author's guess of how the moment might have come across to ra­dio vs. television audiences. However, we do know that something be­yond the candidates' appearances as captured in still frames affected people's reactions. This moment is a strong candidate for one of the somethings that was important.

Following the 1960 election, there were no presidential debates for sixteen years. Lyndon Johnson wanted no part of televised debates when he ran against Barry Goldwater in 1964, and Richard Nixon, once burned, did not choose to debate Hubert Humphrey in 1968 or George McGovern in 1972. However, a series of debates did take place between the presidential and vice-presidential candidates in 1976. Jimmy Carter and Gerald Ford had three debates, and Walter Mondale and Bob Dole participated in the first vice-presidential debate.

Again, a fascinating combination of visual and verbal information was important in shaping people's perceptions, and among the visual elements, a large part was played by candidates' nonverbal behavior. One immediate impression from the outset of the first Carter-Ford de-bate was that Ford was favored by the stage setting. The candidates stood behind large wooden podiums, unlike the simple orchestra di­rector-style podiums used by Nixon and Kennedy. Gerald Ford, a large former football player at the University of Michigan, leaned over his po­dium, and seemed to dominate it. Ca_ ter, a smaller man, stood behind his podium, and looked more like a choirboy than an athlete. The ini­tial visual impression favored Ford. He appeared more commanding than Carter. A more decisive moment took place in the vice-presiden­tial debate that year between Dole and Mondale. Dole began talking about the number of soldiers who had been killed or wounded during

'Democrat wars' of the 20th century. Viewers and commentators were shocked that Dole would refer to the two world wars, the Korean con­flict, and the war in Viet Nam as 'Democrat wars.' Moreover, as he spoke, Dole leaned on his elbow against the podium, looking very much like a gunslinger in a saloon in an old west cowboy movie. Dole's dark hair and eyes and dry manner of speaking contributed to a very negative image. Mondale calmly criticized the idea that any reasonable person would think that the war against Nazi Germany, and other wars, were partisan. Many years later Mondale (1999) said that he was prepared for Dole's comment, because he had been making it on the campaign trail. However, Mondale had told his advisers that they were 'nuts' to think that Dole would say such a thing in a nationally tele­vised debate: 'He wouldn't be that stupid.' Many political commenta­tors believed that Dole's performance in his debate with Mondale seriously hurt the Ford campaign.

The next two elections, in 1980 and 1984, turned out to be landslide victories for Ronald Reagan, first over Jimmy Carter and then over Walter Mondale. In the single Carter-Reagan debate in 1980, Reagan's slow, calm, and gentle manner of speaking convinced uncertain voters that he was not too bellicose and that he could be entrusted with the power to make war or peace. Reagan's paralinguistic abilities served him extremely well. When he stated at the outset that 'our first priority must be world peace,' his voice sounded sincere, calm, and most of all, reassuring. James David Barber's (1992) discussion of varying 'climates of expectation' implies that the electorate wanted reassur­ance in 1980, and Reagan delivered it. Reagan's 'velvety tones' and 'honeyed voice,' described by David Gergen (2000), were on full display.

In 1984 Reagan's acting abilities saved him from the embarrass­ment of losing two debates to Walter Mondale, and paved the way for a forty-nine state landslide. Reagan performed very poorly in his first de-bate with Mondale, so poorly, in fact, that political pundits began to ask whether Reagan, then 73 years old, was still mentally and physi­cally up to the job of being president. It was perhaps inevitable that during the second debate someone would ask Reagan about his rocky performance and his age. As is well-known, Henry Trewhitt of The Bal­timore Sun asked Ronald Reagan: 'Mr. President, I want to raise an is-sue that I think has been lurking out there for two or three weeks, and cast it specifically in national security terms. You already are the oldest President in history, and some of your staff say you were tired after your most recent encounter with Mr. Mondale. I recall yet that Presi­dent Kennedy had to go for days on end with very little sleep during the Cuba missile crisis. Is there any doubt in your mind that you would be able to function in such circumstances?' The words of Reagan's re­sponse have been reported often. They worked very well for the Presi­dent: 'Not at all, Mr. Trewhitt. And I want you to know that also I will not make age an issue in this campaign. I am not going to exploit, for

political purposes, my opponent's youth and inexperience.' But the impact of what Reagan said, and the entire moment, were greatly af­fected by the way he said it. As in the Kennedy-Nixon case described above, there unfolded a nonverbal mini-drama, or at least a nonverbal episode.

It plays out as follows. First, when Reagan realizes that the question wi11 be about age, he moves his body and nods in a very serious and confident-looking way. He appears ready for the question. Clearly, he had anticipated the question and planned a response. Second, he speaks in what can be described as a mock-serious manner. That is, he looks and sounds serious, and firm, but the words are not serious, so it becomes clear that the manner is not serious either. Third, the im­mediate reaction, which Reagan would have anticipated to some de­gree, works wonderfully in his favor. The audience laughs, Trewhitt laughs, and Walter Mondale laughs. Mondale (1999) said he knew that Reagan was scoring big points, but that he could only do the natural thing, which was to laugh at a very clever and funny remark. Fourth, in the melee, Reagan immediately reaches for a glass of water, calmly and confidently sipping from it, thereby prolonging the moment of laugh-ter, and triumph. He has a confident and self-satisfied look on his face. Fifth, Reagan displays suppressed laughter as Trewhitt tries to stop laughing.

After the laughter subsides, Reagan continues. Looking as if he were reaching back into long-term memory from Eureka College, he adds 'If I still have time, I might add, Mr. Trewhitt, I might add, that it was Sen­eca or it was Cicero, I don't know which, that said, if it was not for the elders correcting the mistakes of the young there would be no state.' Trewhitt then ratifies Reagan's knockout punch by stating 'Mr. Presi­dent, I'd like to head for the fence and try to catch that one before it goes over but I'll go on to another question.'

We have data showing clearly that more than the words of Reagan's age comment carries the moment (Fein, Goethals, & Kassin, 1999). In two experiments Williams College students were presented with three different televised versions of the second Reagan-Mondale debate dis­cussed above. In the Intact version the complete exchange between Rea­gan and Trewhitt described above was included. Also included was another memorable Reagan one-liner. In discussing national defense Reagan says that Mondale 'has a commercial out where he's appearing on the deck of the Nimitz and watching the F -14s take off and that's an image of strength, except that if he had had his way when the Nimitz was being planned, he would have been deep in the water out there, because there wouldn't have been any Nimitz to stand on. He was against it.' Rea­gan's comment was followed by hoots and laughter. In a Soundbites De­leted condition the Nimitz exchange and the age exchange were both edited out. In a Reaction Deleted condition the initial comments were in­cluded, but the audience reactions immediately after the remarks ('he was against it'; 'my opponents youth and inexperience') were deleted.

In the case of the age exchange Reagan's drinking water and suppressing laughter were among the elements specifically deleted.

In an initial experiment participants viewed tapes that were about 40 minutes long. Thus slightly less than half of the entire 90-minute debate was shown. Participants then answered a questionnaire begin­ning with ratings of each candidate's performance on 0-100 point scales. The most telling way to look at the data is to consider the differ­ence between Reagan's ratings and Mondale's. In the Intact condition, Reagan's ratings were 7.50 points higher, 66.25 to 58.75. In the Soundbite Deleted condition Reagan's ratings were 9 points lower (64.17 to 73.03), and in the Reaction Deleted condition Reagan's rat­ings were nearly 24 points lower (49.29 to 72.86). Consistent with these data, participants were asked at end of a questionnaire Who Won? Reagan, Mondale, or Neither? In the Intact condition, many more participants chose Reagan as the winner over Mondale, 60% to 17%. In the Soundbite Deleted condition, the perceived winner was re­versed-Mondale was named the winner more frequently, 43% to 18%. However, in the Reaction Deleted condition, Reagan was completely swamped. Seventy-six percent named Mondale the winner and only six percent named Reagan.

These data suggest a number of things. First, Reagan's one-liners made an immense difference in the debate. It changed him from being perceived as a loser to being perceived as a winner. But, even more im­portant, it is not the words themselves that are important. It is dy­namic of the whole situation. When Reagan makes a witty remark, it invites a response. When that response is not heard, the comment it-self seems out of place. Although we cannot be sure from these data, Reagan's follow through on the age question seems particularly power­ful. Not only did he deliver a witty remark, he anticipated the reaction and then managed, with body movements, physical action, and facial expressions, to extend the moment and to bring it to a successful con­clusion with his remark about elders correcting the mistakes of the young. at is impressive is the total orchestration of the moment, us­ing both words and nonverbal behavior to move the audience to his side. No wonder Reagan once commented that he couldn't imagine being a successful president without having been an actor.

A second study was conducted to replicate and extend the study de-scribed above. This time the entire 90-minute debate was used, with segments described above deleted in the Soundbite Deleted and Reac­tion Deleted conditions. In this study, both candidates are given per­formance ratings of about 60 in the Intact condition. In the Soundbite Deleted and Reaction Deleted Conditions, Mondale's ratings are about 12 points higher. Clearly the soundbites are important, but they only work when the unfolding dynamic of the total situation is seen by view­ers. Consistent with our findings, suggesting that what people see is important, a study by Patterson, Churchill, Burger, and Powell (1992) considered the impact of recorded segments from the same 1984 Rea-

gan/Mondale debate. There were four segments showing each candi­date, averaging 44 seconds each. Participants were exposed to the candidate's statements in four modalities: audio only, visual only, au­diovisual and text. Reagan was rated more favorably in all four varia­tions, but the greatest difference was in the visual only condition. Relative to Mondale, Reagan appealed most to voters who simply watched him. Similarly, studies by Masters and his colleagues of facial displays by Reagan and Mondale in the 1984 campaign suggest that Mondale's facial expressions failed to communicate warmth and reassurance (Masters, Sullivan, Feola, & McHugo, 1987; Sullivan & Masters, 1988).

The story continues to unfold in later debates. In 1988 George Bush had a great height advantage over Michael Dukakis. When they shook hands after their first debate, the difference was striking. During their first debate Bush suggested, without specifically pointing to Dukakis that Dukakis was the 'iceman who never makes a mistake.' This fram­ing attempted to convert Dukakis' considerable advantage in articula­tion over Bush into a liability-Bush was imperfect but warm; Dukakis was a cold, Northern intellectual. Dukakis' manner in the second de-bate between the two candidates seemed to give proof to the Republi­can's implication. When the first question asked whether he would favor an irrevocable death penalty for the killer who raped and mur­dered his wife, Dukakis looked unfazed by the image of his dead wife, and said, no, and then changed the subject. The image of Dukakis as the iceman was frozen. In 1992 Clinton's 'I feel your pain' voice and di­rect physical approach to citizens asking questions in the town meet­ing format helped him look caring and engaged, especially in contrast to George Bush, who was looking at his watch.

Overall, several studies as well as anecdotal accounts of presiden­tial debates suggest that more than words are important. Both visual information and paralingual cues play a role in perceptions of political candidates. We would like to report some preliminary findings from a study that looked at the role of several different kinds of candidate in-formation in a mayoralty election in Seattle in 1997. We find some tan­talizing evidence that candidate and voter ethnicity interact with verbal vs. nonverbal behavior to influence people's reactions.

ETHNICITY, NONVERBAL BEHAVIOR, AND PERCEPTIONS OF LEADERSHIP

Past studies have identified several different kinds of nonverbal infor­mation that might be relevant to voters' perceptions of leaders, partic­ularly leaders who appear in televised debates. In a study of viewers' reactions to Bob Dole and Walter Mondale in their 1976 vice-presiden­tial debate, Krauss, Apple, Morency, Wenzel, and Winton (1981) iden­tify several kinds of paralinguistic cues that might be relevant once semantic content has been removed from speech, including 'pitch,

amplitude, rate, voice quality, contour, etc.' (Krauss et al., 1981, p.312). They also note the different kinds of information that is con­veyed visually through posture, gestures, movement, facial expres­sion, and looks. A particularly interesting study by Warnecke, Masters, and Kempter (1992) found that American adults who were shown si­lent videotaped images of American and European (French and Ger­man) leaders 'feel more negatively when seeing the foreigners and judge them more negatively' than American leaders (p. 267). This ef­fect disappeared when sound was included so that nationality was made known. Warnecke et al. suggested that the negative emotions and judgments of foreigners are based on the 'preconscious monitoring of nonverbal cues' (p. 267). Perhaps people feel comfortable when they observe someone who moves in familiar ways, and have a more posi­tive reaction to those individuals. The impact of these differences in comfort level and judgment is lessened when information in addition to information about the way individuals move is presented, perhaps due to a dilution effect (Nisbett, Zukier, & Lemley, 1981). Warnecke specifically suggests that 'dynamic nonverbal cues,' quite possibly head movements, account for their data. They conclude: 'bottom line, our responses to leaders-and to others more generally-are often governed more by feelings or `gut reactions' than by conscious verbal judgments' (p. 280). These feeling or reactions, in turn, are based on responses to nonverbal cues.

Pursuing the issues raised by Warnecke et al. we conducted a study whose major focus was the extent to which ingroup favoritism in judg ing two candidates in a political debate would vary depending on what verbal and nonverbal information participants had about the two can didates (Farmer, 1998). It considered especially the role of movement cues. Do people respond to the motions of others differently depend­ing on whether those others are similar or different in their ethnicity? Warnecke et al. suggest people respond differently to the motions of others depending on their nationality. What about responding differ­ently to the motions of leaders from different ethnic groups within a nation, specifically, the United States?

Let us begin by identifying four different kinds of information that are conveyed in televised debates. These are words, voice quality, looks, and movement. It is possible to look at the impact of each of these sources of information on perceivers separately. Words can be presented in a written transcript or can be read in a bland voice, thereby neutralizing any effect of differences in candidates' voice qual­ity. Voice quality can be isolated by presenting audio tapes of candi­dates' content-filtered speech (cf. Krauss et al, 1981). Looks can be presented with still shots from photographs or videotapes. And move ment can be presented with silent videotapes. Furthermore, much in-formation about looks can be removed from silent recordings by blurring the face, as is often done to hide the identity of persons speak ing on a videotape. Just as each of the four kinds of information can be

Our study employed a debate that took place during the 1997 Seat­tle mayoral election between Charley Chong, an Asian-American, and Paul Schell, a white male. Information was presented to participants in six different variations. There were five different audio-visual combi­nations: audio only (Audio Only), still pictures only (Still Only), video with no sound (Video Only), audio with still pictures (Audio Still), and full channel audiovisual (Audiovisual). Also, we included a second au dio only variation (Audio Ethnicity) in which we told the participants the names and ethnicities of the two candidates. In addition to the four

Table 5.1
Possible Combinations of Information to be Presented About Candidates in Debates

Information

Presentation

A Transcript (Tr)

B Content filtered speech (CF)

C Stills only (St)

D Blurred video (BV)

AB Audio only (Au)

AC Tr + St

AD Tr + BV BC CF + St BD CF + BV

CD Video only (V)

ABC Au + St ABD Au + BV ACD Tr + V BCD V + CF ABCD Au + V

(full channel videotape)

Words Voice Looks Motion

Yes No No No

No Yes No No

No No Yes No

No No No Yes

Yes Yes No No

Yes No Yes No

Yes No No Yes

No Yes Yes No

No Yes No Yes

No No Yes Yes

Yes Yes Yes No

Yes Yes No Yes

Yes No Yes Yes

No Yes Yes Yes

Yes Yes Yes Yes

kinds of information discussed above-looks, movement, words, and voice quality-information about race is included in all the conditions except Audio Only. In the Audio Ethnicity condition this information is provided by the experimenter's instructions. In the Still Only, Video Only, Audio Still, and Audiovisual conditions that information is pro­vided by the candidate's appearance and by their names, which were shown from time to time on the videotaped broadcast. Table 5.2 shows the kinds of information that are presented in each of these variations.

We were interested in how white and non-white students would react to the candidates in the different audio-visual variations. First, we felt that Charley Chong would do better with more visual information and less auditory information. Charley moves and gestures more than Paul. We felt that he would be perceived as more active and dominant in a video only variation. On the other hand, Paul speaks more smoothly and has a more soothing voice than Charley, though pretesting showed that neither man's voice suggested his ethnicity. Second, we expected that there would be some in-group bias such that overall nonwhite students would have a more favorable reaction to Charley, the Asian-American candidate, relative to Paul, the white candidate, in comparison to the relative ratings that white students gave to the two candidates.

Again, however, the major focus was on the way ingroup favoritism varied according to whether or not participants saw the candidates move. There are two pairs of conditions in which the only difference is that one variation shows the candidates move while the other variation does not. First, the Video Only variation shows the candidates move while the Still Only variation simply shows still pictures. There is no sound in either one. The only difference is the presence or absence of motion. Similarly, the Audiovisual condition shows the candidates move while the Audio Still condition does not. There is sound in both variations, but the difference is motion. We predicted that ingroup pref­erences, as revealed in white vs. nonwhite students' preferences for Paul vs. Charlie would be stronger in the Video Only variation than in the Still Only variation, and stronger in the Audiovisual variation than in the Au‑

Table 5.2
Information Presented in Seattle Mayoralty Race Study

dio Still variation. Furthermore, in line with Warnecke et al. findings, we predicted that the greatest ingroup favoritism would occur in the Video Only condition. Favoritism in that condition would be stronger than in the Audiovisual condition because there would be no audio information to dilute the effect of the nonverbal motion information.

Participants arrived for the study and were told that it was investi­gating the way different media presentations affect people's percep­tions of candidates in political debates, and that different groups of participants were watching different presentations of the debate (a rare instance of truth-telling, or partial truth-telling, in the social psy chology lab). They were told, again truthfully, that they would see or hear brief portions of a debate between Paul and Charlie who were can­didates for mayor of Seattle. Then each candidate was described in two bland sentences (e.g., Paul 'has always wanted to make a differ­ence'). In the Audio Ethnicity variation the experimenter added that Paul is Caucasian-American and that Charlie is Asian-American. The tapes participants saw or heard all 3 minutes and 11 seconds. They contained ten different segments, each lasting about 19 seconds. The candidate's remarks dealt with local transportation and economic de­velopment issues, so as to avoid national political issues. In the still variations, a representative still shot during the segment was shown for the entire 19 seconds.

The participants were 132 Williams College students who received either cash or course credit. The nonwhite students were African-American, Asian American, and/or Latino. Ideally, we would have used only white and Asian-American students, but it would not have been possible to recruit as many Asian American students as we needed. The results showed that the reactions of the three nonwhite groups to the two candidates were similar.

After watching or listening to the tape, participants rated each can didate's overall performance in the debate on a nine-point scale. Par­ticipants also indicated their perceptions of the two candidates on several bipolar nine-point scales including likeable-unlikeable, leader-not a leader, active-passive, confident-unsure, etc. The orthogo nal solution to a factor analysis revealed two factors. Four measures correlated with the first factor but not the second: dominant-submis sive, leader-not a leader, shy-outgoing, and active-passive. These four measures were combined to create an Overall Leadership measure. Five other measures correlated highly with the second factor but not the first: calm-nervous, appealing-not appealing, rough-smooth, like­able-unlikeable, attractive-unattractive. These five measures were combined to create an Overall Likeability measure. All analyses em­ployed a three-factor ANOVA with two between-subject factors (Me­dium and Participant Ethnicity) and one within-subjects factor (Candidate).

The results showed two significant effects on the overall perfor­mance measure. First, there was a Medium by Candidate interaction

showing, as predicted, that Paul was perceived as the better performer in the four variations with sound (Audio Only, Audio Ethnicity, Audio Still, and Audiovisual) while Charlie was perceived as the better per-former in the two variations without sound (Still Only, Video Only; p < .02). These results are shown in Table 3. Second, while there was no predicted Candidate by Participant Ethnicity interaction, showing overall ingroup preference, there was a three-way interaction showing, as predicted, that there were varying degrees of ingroup preference in the different media variations (p < .05).

The degree of ingroup preference in each media variation can be il­lustrated by adding the white participants' rating of Paul and the non-white participants' rating of Charlie and subtracting from that sum the white participants' rating of Charlie plus the nonwhite participants rating of Paul. The figures for this measure are also shown in Table 5.3. In descending order ingroup preference was strongest in the Video Only variation (1.44), as predicted, and decreased in the remaining variations as follows: Audiovisual (1.10), Audio Ethnicity (1.00), Still Only (0.95), Audio Still (-0.52), and Audio Only (-1.16). In line with specific predictions, ingroup preference was stronger in the Video Only variation than the Still Only variation (1.44 vs. 0.95, p < .02) and stronger in the Audiovisual variation than the Audio Still variation (1.10 vs.-0.52,p < .01).

In addition to the confirmed predictions, there are a number of other results of interest. First, the participants strongly preferred Paul in the Audio Only condition. This is especially true for nonwhite stu­dents. The basis for the nonwhite students' preference for Paul in this condition is not at all clear. It is the only condition in which partici­pants do not know the candidates' ethnicity, and therefore the only condition in which issues of ingroup favoritism would not arise. Com­paring the Audio Only and Audio Ethnicity conditions, both white stu­dents and nonwhite students evaluate Charlie more favorably in the Audio Ethnicity variation, where they know that he is Asian-American. This effect is much larger for nonwhite students. Nonwhite students

Table 5.3
Overall Performance Ratings

evaluate Paul less favorably in the Audio Ethnicity condition, whereas white students evaluate him about the same. Thus nonwhite students give very different evaluations of the two candidates in the two varia­tions, suggesting that their ingroup preferences are strongly shaping their evaluations in the Audio Ethnicity variation, where ethnicity is known. The white students change their evaluations less. They do be-come more positive about Charlie, perhaps reflecting a tendency to be politically correct, but their changes are less than those of the nonwhite students. Second the Medium by Candidate interaction is in­teresting in its own right. Both white and nonwhite students respond more positively to Paul when heard and Charlie when seen. Those re-actions are shared across the two groups of participants.

One interesting and perhaps reassuring finding is that participants' evaluations of the candidate's performance is not simply a matter of which one they like better. We noted above that a factor analysis yielded an Overall Leadership measure and an Overall Likeability measure. The results on the five-item Overall Leadership measure were very similar to the single overall performance measure discussed above, which was our main measure. The same significant three-way interac­tion of the Medium, Candidate and Participant Ethnicity variables ob­tained, with greater ingroup preference in the Video Only (0.90) and Audiovisual (0.95) variations than in the Still Only (0.28) and Audio Still (0.83) variations. In contrast, the Overall Likeability measure simply showed a main effect for Candidate, with both groups of partici­pants liking Paul more than Charlie in all the Medium variations. Thus while participants consistently liked Paul more, they did not always consider him the best leader. The latter judgments were apparently influenced by ingroup favoritism, while the former were not.

The study reported above is an initial attempt to consider the inter-action of verbal and nonverbal information in influencing people's re-actions to political candidates. We considered whether white students would show some degree of preference for a white candidate and non-white students a preference for a nonwhite candidate, and found that it depended on the kind of nonverbal information that was presented about the two candidates. Although the results could be stronger and need to be replicated, they do suggest that people show in-group pref­erences more when they see political candidates from different groups moving rather than simply seeing their still pictures. The results also suggest that the response to motion is stronger when the motion is not accompanied by sound. Candidates' words seem to take some atten­tion away from the candidates' movements and weaken whatever ap­peal may arise from positive reactions to the nonverbal behavior of candidates who are ethnically more similar. This may essentially be a dilution effect (Nisbett, Zukier, & Lemley, 1981).

It is not entirely clear how to account for the finding that observing the motion of two potential leaders increases in-group favoritism. The effect may be automatic, whereby the candidates' similarities or differences to

the perceiver become clearer as more of their nonverbal behavior is ob­served. Perhaps there is a certain comfort level with people from one's group, or discomfort level with people from other groups, especially groups that are perceived in opposition or in competition in some way. Warnecke et al. (1992) suggest that their results reflect 'preconscious monitoring of nonverbal cues' and such a process may account for our results. A somewhat different explanation emphasizes information pro­cessing, but again a nonconscious variety. Possibly confirmation biases are operating, whereby subjects have the hypothesis that the ingroup candidate is the better candidate and movement information is inter­preted in a biased way so as to support and strengthen the original bias (Snyder, 1984). Additional research is needed to explore these issues further.

CONCLUSION

Clearly, our perceptions of political leaders are affected by their non-verbal behavior. Both words and music, if we may think of nonverbal behavior as music, are important. The nonverbal information may re­inforce, qualify, or contradict the words. Highly skilled political actors, such as Ronald Reagan, have a masterful command of nonverbal be­havior and know how to use it to enhance their appeal and influence. David Gergen (2000) suggests that nonverbal behavior reveals whether political leaders are comfortable with themselves and their audience. Listeners reciprocate leaders' comfort and relax when the leader is relaxed. Thus nonverbal behavior affects the relationship be­tween aspiring leaders and potential followers, and affects both the meaning and the credibility of the leader's message. We have a long way to go to understand the subtleties of the way the nonverbal behavior of leaders influence the reactions and perceptions of various followers, or nonfollowers. But we already understand the enduring importance of nonverbal behavior in affecting the dynamic between leaders and followers.

REFERENCES

Barber, J. D. (1992). The presidential character: Predicting performance in the white house. Englewood Cliffs, NJ: Prentice Ha11.

Brookhiser, R. (1996). A man on horseback. Atlantic Monthly, 227, 50-64. Brown, R. (1965). Social psychology. New York: Free Press.

Brown, R. (1986). Social psychology, The second edition. New York: Free Press. Burns, J. M. (1978). Leadership. New York: Harper & Row.

Carson, R. C. (1969). Interaction concepts of personality. Chicago: Aldine. Farmer, G. W. (1998). The effects of nonverbal cues and ethnicity in triggering

ingroup leader preference. Unpublished Honors Thesis, Williams College. Fein, S. Goethals, G. R., & Kassin, S. M. (1999). Group influence on political

judgments: The case of presidential debates. Unpublished manuscript, Wil‑

liams College

Foote, S. (1963). The civil war: A narrative. New York: Random House.

5. NONVERBAL BEHAVIOR AND POLITICAL LEADERSHIP 1 15

Gardner, H. (1995). Leading minds: An anatomy of leadership. New York: Ba­sic Books.

Gergen, D. (2000). Eyewitness to power: The essence of leadership, Nixon to Clinton New York: Simon and Schuster.

Goffman, E. (1955). On face-work: An analysis of ritual elements in social in­teraction. Psychiatry: Journal for the Study of Interpersonal Processes, 18, 213-231.

Henley, N. (1977). Body politics: Power, sex, and nonverbal communication. Englewood Cliffs, NJ: Prentice Ha11.

Hollander, E. P. (1993). Legitimacy, power, and influence: A perspective on rela tional features of leadership. In M. M. Chemers & R. Ayman (Eds.), Leader-ship Theory and Research (pp. 29-48). San Diego, CA: Academic Press.

Keegan, J. (1986). The mask of command. New York: Viking Penguin.

Krauss, R. M., Apple, W, Morency, N., Wenzel, C., & Winton, W. (1981). Verbal, vocal, and visible factors in judgments of another's affect. Journal of Per­sonality and Social Psychology, 40, 312-320.

Leary T. (1957). The interpersonal diagnosis of personality. New York: Ronald.

Markey, P M., Funder, D. C., &Ozer, D. J. (2003). Complementarityofinterper­sonal behaviors in dyadic interactions. Personality and Social Psychology Bulletin, 29, 1082-1090.

Masters, R. D., Sullivan, D. G., Feola, A., & McHugo, G. J. (1987). Television coverage of candidates' display of behavior during the 1984 primaries in the United States. International Political Science Review, 8, 121-130.

McHugo, G. J., Lanzetta, J. T., Sullivan, D. G., Masters, R. D., & Englis, B. (1985). Emotional reactions to expressive displays of a political leader. Journal of Personality and Social Psychology, 49, 1513-1529.

McLuhan, M. (1964). Understanding media. New York: McGraw-Hill.

Mondale, W. F. (1999). Personal Communication. October 21, 1999.

Nisbett, R. E., Zukier, H., & Lemley, R. E. (1981). The dilution effect: Nondiag­nostic information weakens the implications of diagnostic information. Cognitive Psychology, 13, 248-277.

Patterson, M. L., Churchill, M. E., Burger, G. K., & Powell, J. L. (1992). Verbal and nonverbal modality effects on impressions of political candidates: anal ysis from the 1984 presidential debates. Communication Monographs, 59, 231-242.

Snyder, M. (1984). When belief creates reality. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology, 18, 248-306. New York: Academic Press.

Sullivan, D. G., & Masters, R. D. (1988). 'Happy warriors': Leaders' facial dis­plays, viewers' emotions, and political support. American Journal of Politi­cal Science, 32, 345-368.

Tannenbaum, P. H., Greenberg, B. S., & Silverman, F R. (1962). Candidate im­ages. In S. Kraus (Ed.), The Great Debates (pp. 271-288). Bloomington: Indiana University Press.

Warnecke, A. M., Masters, R. D., & Kempter, G. (1992). The roots of nationalism:

Nonverbal behavior and xenophobia. Ethology and Sociobiology, 13, 267-282. Watzlawick, P, Beavin, J., & Jackson, D. D. (1967). Pragmatics of human

communication. New York: Norton.

Weiss, W. (1968). Effects of the mass media of communication. In G. Lindzey & E. Aronson (Eds.), The handbook of social psychology (Vol. 5, pp. 77-195). Reading, MA: Addison-Wesley.