Presidential Primary Debates as a

Presidential Primary Debates as a Party-Strengthening Exercise By Mike Yawn Department of Political Science Sam Houston State University 936.294.1456 pol_rmy@shsu.edu Paper prepared for the Annual Meetings of the Southern Political Science Association Annual Meetings, January 6-9, 2005, New Orleans. 1 Presidential Primary Debates as a Party-Strengthening Exercise Numerous studies have addressed the effects of presidential debates in general elections, with most focusing on media effects (Lang & Lang 1979), audience knowledge (Abramowitz 1978, Wald & Lupper 1978, Geer 1988, Lanoue 1992; Holbrook 1999), agenda setting (Katz & Feldman 1962; Carlin 1992), candidate styles (Lang & Lang 1979, Sears & Chaffee 1979, Lanoue & Schrott 1989a; Schrott & Lanoue 1992, Yawn & Beatty 2000), voters’ issue positions (Abramowitz 1978, Lanoue & Schrott 1989b) or voters’ evaluation of candidates (Lang & Lang 1962, Kraus & Smith 1962, Hagner & Rieselbach 1978, Dennis, Chaffee, & Choe 1979, Holbrook 1994). Fewer studies have applied these same topics to the matter of presidential debates in the primary season (Lanoue & Schrott 1989, Yawn, Ellsworth, Beatty, Kahn 1998), and these studies show that primary debates have a larger effect in terms of altered vote preferences and changes in audience knowledge. But no study has examined the effects of a presidential primary debate on the party members’ attitudes toward the debating candidates’ party. This study rectifies this oversight in the literature. I examine a sample of more than 300 audience members at a presidential primary debate in Phoenix, AZ. The survey consisted of three waves: one immediately before the sample watched the debate live in Gammage Auditorium on November 22nd 1999 and the second immediately after the same debate. The third, which was conducted approximately three weeks later, followed a second Arizona Republican Primary debate. The latter wave consisted of only twenty-one respondents, and was designed to collect richer information regarding respondents’ thought processes. The first AZ debate involved Steve Forbes, Orrin Hatch, Alan Keyes, and John McCain. The second debate 2 involved all six candidates: the four mentioned above, as well as Gary Bauer and George Bush. The survey measured audience attitudes across many items: vote preference, candidate support, Republican party support, partisan identification, candidate electability, and the respondents’ likelihood of voting. This study contributes to the existing literature on campaigns and debates in three ways. Peripherally, the study examines the persistence of campaign effects in the midst of an ongoing campaign. George Bush, for example, participated in two debates in the weeks between the second and third survey waves, and while his feeling thermometer scores rose somewhat between waves, they never reached the heights he enjoyed in the first wave of the survey. Second, the party’s candidates who did participate successfully exploited this free media time simply by attending—a finding that strongly supports Woody Allen’s observation that ninety percent of life is just showing up. All four candidates improved their standing with Republican voters. Even the otherwise hapless Orrin Hatch—the loser, according to a majority of the sample—increased his feeling thermometer scores and electability ratings. Finally, the Republican party benefited from the exposure, and perhaps, the free air time to bash the Democratic party. Not only did Republicans in the audience indicate they were more likely to vote, but the party’s candidates’ standings improved among the party faithful, while Democratic candidates fell in the eyes of the Republican audience. Moreover, this sample of Republicans indicated they were more supportive of the party, reinforcing already strong party ties. Contrary to the belief that debates diminish a party’s candidates, either through divisive exchanges or merely by appearing onstage with the likes of Gary Bauer or Al Sharpton, it appears that debates may improve a party’s standing, at least among the party faithful. 3 Context of the Arizona Debates On October 23rd, 1999, the Republican Party held its first official primary debate. Five candidates debated; the media yawned. Indeed, only C-SPAN carried the debate nationally. Missing was George W. Bush, the Republican frontrunner, and his absence, according to the Boston Globe’s Jill Zuckman, created a “void” unfilled by the other candidates. Bush’s strategy was clear. As Darmouth’s Linda Fowler noted, “the front-runner doesn’t give free air time to his rivals” (Ratcliffe, Hines & Roth, 1999). Further, Bush announced that he would not be attending any other scheduled debates in 1999. “We intend,” said Karen Hughes, “to debate next year when people are beginning to pay attention to the presidential race” (Ratcliffe, Hines & Roth, 1999). Hughes was wrong, but, at this early stage in the primary season, Bush had every reason to feel confident. He had, according to William Mayer (2003), “perhaps the most successful invisible primary season in the modern history of presidential elections.” The numbers tell the story: by October Bush led the Republican pack, with 68% of Republicans supporting him for the Republican nomination. John McCain, the Republican senator from Arizona, was second with eleven percent (Mayer 2003). The Texas Governor was also in the midst of a fundraising feat unprecedented in electoral history. By the end of 1999, Bush had raised more than 67,000,000 dollars—more than the rest of the Republican field combined (Mayer 2003).i No candidate similarly positioned had ever lost the nomination for the Presidency. But concerns were emerging. In early October 1999, Linda Fowler noted “There are already sufficient doubts about him as a candidate in terms of his ability to handle himself in setting where he’s not scripted” (Ratcliffe, Hines & Roth 1999). Such doubts were reinforced on November 3rd, when a reporter asked Bush to name the leaders of four foreign countries. Bush 4 responded peevishly, and he was able to identify only one of the leaders. Such performances not only increased doubts about his knowledge base, but also raised concerns about his emotional intelligence.ii By December, the Manchester Union Leader published a front-page editorial referring to Bush as an “empty suit” (McQuaid 1999). The candidates, particularly Steve Forbes, were quick to jump on these concerns. After Bush missed the first debate, held in New Hampshire, Forbes referred to Bush as a “big loser” (Zuckman 1999) and called the Governor’s absence “an insult to the voters of New Hampshire” (Mehren 1999). Rival Gary Bauer said that Bush’s decision to “high-tail” it rather than debate was an “incredible insult” (Zuckman 1999). John McCain’s spokesman, Dan Schnur, characterized Bush’s absence at the second debate as the Governor “thumbing his nose at the people of New Hampshire” (Bush Plans 1999). By the second debate, Bush felt sufficient pressure to call a press conference explaining his absence, a move that was met with much derision. Steve Forbes characterized Bush’s justification for missing the debate as “blah, blah, blah” (Roth & Hines 1999). Even an audience member at the second debate got into the act, qualifying her question to the present candidates by noting that she had originally hoped to pose her question to George Bush. These doubts about Bush redounded to the benefit of Senator John McCain. By late October a New Hampshire poll showed McCain surging to 26%, up from 10% in August. The press was quick to pick up on the candidates’ shifting fortunes. According to the Seattle PostIntelligencer, “McCain’s poll numbers and public profile have been decidedly on the upswing” (GOP Candidates 1999). The Boston Herald noted that McCain was “pulling closer” (Battenfield & Miga 1999), while the Houston Chronicle noted that the senator was enjoying “momentum” (Roth & Hines 1999). By the time of the Arizona Presidential Primary Debate on 5 November 22nd, 1999, a New Hampshire poll showed a statistical dead heat between Bush and McCain. Whatever divisions existed among the party faithful, however, it was clear that after eight years of President Bill Clinton, Republicans wanted a candidate who could win. According to Jeb Bush, speaking before the 2000 general election, Republicans were sufficiently hungry to be back in power that they would back the horse that looked like the winner: “Eight years in the wilderness allows for a higher tolerance for diverse views” (Bruni 2002, 148). The exit polls of 2000 support this notion, showing that fewer Republicans than Democrats defected when voting for President. In short, Republican primary voters during the winter of 1999 were overwhelmingly backing Bush, at least nationally, but their motivations, it seemed, had more to do with his electability than with his political profile. If their confidence in him faltered, many were prepared to jump ship and back McCain—as many of them did following Bush’s second debate.iii Amid this political backdrop, Senator John McCain, Senator Orrin Hatch, Alan Keyes, and Steve Forbes attended Arizona State University’s Gammage Auditorium for the third Republican primary debate. Televised nationally, the debate “put a spotlight on Sen. John McCain of Arizona, currently Bush’s closest rival in several key states” (GOP Presidential 1999). By the time of the first Arizona debate, Bush was no longer a sure thing. Data, Survey Design, and Measures To measure the impact of the Arizona primary debate on the Republican party and its candidates, I used a pretest-posttest design, asking people attending the debate to fill out short questionnaires immediately before the debate. Immediately after the debate, these same individuals completed a second questionnaire before leaving the auditorium. The pretest 6 questionnaire was one page, consisting of seventeen items. The posttest questionnaire was two pages, consisting of thirty-two items.iv Unlike most other debate research, this study measures the impact of a debate witnessed in person, rather than on television. According to social cognition researchers, direct observation makes social phenomena more vivid and sensorially proximate to the viewer (Nisbett & Ross 1980, 5). In practice, however, social scientists have been unable to demonstrate empirically that more vivid events are more capable of changing attitudes than less vivid events (Taylor and Thompson 1082; Fiske & Taylor 1991). According to Fiske and Taylor (1991), direct experience “does not necessarily change attitudes more effectively than does second-hand contact” (255). Thus, “vividness does not work well empirically, although intuitively it seems as if it should” (Fiske and Taylor 1991, 255).v This design also differs significantly from the two designs most typically employed by previous research in other ways. The first method commonly used to study debate effects involves interviewing people (either face-to-face or by phone) who have watched the debate live or (more likely) on television (Geer 1988; Lanoue 1992; Lemert 1993). This design has the advantage of reaching a wide range of people exposed to the debate. In other respects, it is less desirable. Often, because of the time necessary to complete all of the interviews, subjects are interviewed a few days or even a week after the debate. Very likely, these subjects have been exposed in the interim to other campaign stimuli, either media coverage of the debate or another campaign event. As Alan Schroeder has noted “it is virtually impossible to isolate debates from other influences on voters’ decisions” (213). Moreover, with a telephone survey, it is difficult to determine whether the individual was actually exposed to the debate or whether they only indicated so in response to some social desirability concerns. 7 This design circumvents the problems cited above. Posttests were administered immediately after the debate, before the audience is exposed to other stimuli. Furthermore, there is no question that the respondents were actually at the debate. Finally, interaction among these respondents was likely more limited than interaction would have been among television viewers. Thus, our respondents may be much more individuated than is typical of those who watch debates on television and less affected by group pressure (Asch 1951; Dittes & Kelly 1956). The second design commonly found in the literature (Zhu, Milavsky, & Biswas 1994; McIntosh 1989; McIntosh 1993) involves recruiting participants to view a recorded version of the debate in an experimental setting. The main limitation of this design is the artificiality of the setting. When people watch a debate on television, they normally watch at home, in a relaxed atmosphere with friends and family. In the experimental setting, subjects watch the debate with a large number of strangers—aware they are being observed. Such a setting may encourage subjects to pay closer attention to the debate, heightening the impact of the experimental stimulus and limiting the generalizability of the findings. Furthermore, the subjects of such studies are typically undergraduate students who may respond differently to debates than a more representative sample. These problems are also avoided in the present design. First, since the study took place during the actual debate, there was nothing artificial about the “experimental” setting. Second, our respondents were a varied sample of people, not just students, who came to observe the debate. Demographically, they reflect the party faithful in Arizona. The respondents were audience members of the debate. Tickets were sold to the general public, although a percentage of the tickets were reserved for interested students. Thus, the audience was a self-selected group, more interested than the “average” voter. Although it is not 8 altogether clear how the self-selected nature of the sample affects our results, prior research indicates that attitude change is least likely among those with the most information. Bauer, for example, argues The audience selects what it will attend to. Since people generally listen to and read things they are interested in, these usually are topics on which they have a good deal of information and fixed opinions. Hence the very people most likely to attend to a message are those most difficult to change (1969, 401; see also Milburn, 1991, 110). Thus, the debate audience may have been more resistant to change than prospective voters who did not attend. Unfortunately, there is no way to test this possibility. A dozen undergraduate and graduates students were stationed at the entrances to the auditorium and distributed surveys. In all, three hundred thirty one Republican respondents completed both the pretest and the posttest. The demographics of our respondents reflect the nature of the settings. Fifty-seven percent of the sample was male, with a mean age of 42. More than half the sample had a college degree, with more than a fifth of the sample holding at least a Master’s degree. Caucasians accounted for more than ninety percent of the sample and almost half (44%) of the sample had a household income of at least $75,000. The surveys were designed to identify the respondents’ favorability ratings of President Bill Clinton, plus the leading presidential candidates of the two parties, as well as the respondents’ support for the Republican Party. In addition, the survey measured the electability of each Republican candidate vis-à-vis both Bill Bradley and Al Gore. Finally, various questions were asked assessing the respondents’ likelihood of voting, respondents’ responses to the 9 candidates’ styles, and respondents’ responses to the candidates’ absence or presence at the debate. Findings Primary debates are attended by partisans who seek information, a forum to support their chosen candidate, or simply for the excitement of the political process. They are often held early in the primary season, with many taking place during the “invisible primary,” and they typically contain a surprising number of obscure candidates. In the last two presidential election cycles, for example, the likes of Dennis Kucinich, Gary Bauer, and Alan Keyes have participated in primary debates. Given the obscure nature of many of the candidates, I expect that many audience members will become more knowledgeable about the candidates. To test this, I rely on a low—but reasonable—threshold: Does the respondent have sufficient knowledge to evaluate the candidate? Respondents were asked to rate the Republican Party and various political figures on a scale from 0-10, with “Don’t Know” as an explicit response category. If, as expected, the primary debate arms audience members with knowledge about the candidates, the percentage of “Don’t Know” responses should be lower on the posttest than on the pretest. The results, presented below, are consistent with these expectations: Table One Here Although largely an interested and informed lot, more than one in five respondents listed “Don’t Know” when asked on the pretest to rate Orrin Hatch and Alan Keyes. On the posttest, however, fewer than one percent had trouble rating any of the attending candidates. Clearly, the debate offered candidates the chance for exposure and partisans the opportunity to learn more about their party’s slate of candidates. 10 Of course, a party only benefits from this exposure if the audience responds favorably to that which it has been exposed. Testing this is a simple matter of tracking feeling thermometer scores from the pretest to the posttest for each candidate. Although individual debate performances vary from contest to contest, it is expected that candidates who attend the debate and do not make fools of themselves will see their feeling thermometer scores rise. For those candidates who fail to appear, their feeling thermometer scores will fall, perhaps because of the zingers the other candidates will direct at their absent rivals, the perception that the absent candidate is scared, or because of some sort of relative evaluation system used by the audience. In addition to asking audience members to rate each of the Republican candidates for their party’s nomination, the survey included Democratic candidates Al Gore and Bill Bradley, as well as President Bill Clinton. Given the partisan nature of these events, it is expected that the audience will evaluate members of the opposition party lower on the posttest than on the pretest. The results, presented below, are almost uniformly in line with expectations: Table Two Here As expected each Republican candidate who was present at the debate benefited from increased audience evaluations. Indeed, the findings were all statistically significant at the .01 level. This is perhaps best exemplified by the case of Orrin Hatch. Sixty-four percent of the audience identified him as the loser of the debate, yet his feeling thermometer scores increased almost a half of a point on an eleven-point scale. For Alan Keyes, the perceived winner of the debate, the debate was particularly rewarding. Coming in with low name recognition, a silver tongue, and a potent, partisan message, Keyes increased his feeling thermometer scores from 7.14 to an eye-popping 8.57, the highest mean of any candidate, pretest or posttest. Overall, the attending Republican candidates enjoyed an increase of .75 points in their feeling thermometer 11 scores. By contrast, the absent candidates suffered a 1.17 point drop in their feeling thermometer scores. On average, then, the candidates could expect, a 1.92 point swing in their feeling thermometer scores simply by deciding whether to attend the debate. For those present, the debate was the best of times. For those absent, it was the worst of times. Gary Bauer, for example, was a two-time loser. Not only did he have the lowest name recognition of the bunch, he also saw his approval ratings fall nearly a full point, from 5.51 to 4.57. But it was George W. Bush who, to echo Steve Forbes assessment, was a “big loser.” His approval ratings slipped from 6.32 to 4.92—the only Republican other than Gary Bauer to slip below 5.0. Bush’s drop of 1.4 points was the largest of any candidate, and only Alan Keyes saw a greater change in terms of absolute value, although his change, of course, was in a positive direction. As expected the Democratic candidates fell in the estimation of the audience, although low pretest evaluations precluded a precipitous drop. Still, the drop of both Gore and Bradley, the two Democratic candidates for the party nomination, were statistically significant at the .01 level. Of all the Democrats, only Bill Clinton did not see a statistically significant drop in his feeling thermometer scores. No doubt, after eight years of partisan wrangling, attitudes about the President had crystallized to a point that any movement must be considered significant— particularly given his predictably low pretest rankings. Inductively, it makes sense to believe that these changes benefit the Republican Party. A majority of its candidates for the party nomination gained in the audience’s estimation, while the Democratic candidates dropped from already low levels of esteem to Hooverian depths. But did these changes extend to the Republican party? Or does the era of candidate-centered politics 12 limit the reciprocal benefit to the party? To test this, an item assessing respondents’ support for the Republican party was asked on the pretest and posttest. The results are presented below: Table Three Here Apparently, the improved standing of many of the party’s candidates—and, perhaps, the decrease in standing of Democratic candidates—also extended to the Republican Party. Feeling thermometer scores for the party increased by .39 points, from 7.61 to 8.00. Despite the predictably high mean on the pretest, this change was statistically significant at the .01 level. In addition to increased feeling thermometer scores, the Republican party appeared to benefit from a sense of optimism among the party faithful. Enthusiasm, of course, is an important component to getting out the vote. A muted base is a losing base. This debate, it seemed, had the desired effect from the perspective of the Republican Party. After viewing the debate, each of the present candidates was perceived as more electable after the debate than they were before the debate. The results are presented below: Table Four Here Table Five Here Among the present candidates, only the well-known John McCain failed to improve in terms of electability. Even with McCain, however, approximately seventy-five percent of the sample thought that he could beat either Gore or Bradley. Moreover, even including McCain in the analysis, the attending candidates enjoyed more than a 15% jump in the percentage of the sample who thought they could beat Al Gore or Bill Bradley. For the losing candidates, by contrast, about 6.5% fewer of the audience members thought they could beat Al Gore or Bill Bradley on the posttest—and this number would have likely been larger if any reasonable person could have thought Gary Bauer could win in the first place. 13 Of course, increased feeling thermometer scores and greater levels of enthusiasm are ultimately important insofar as it motivates citizens to get out and vote—and vote for a Republican candidate. This test is perhaps the stiffest of all presented here. The sample, almost by definition, rank among the most partisan of the electorate, and basic research methodology suggests that extreme scores on a pretest sample may lead to “regression to the mean,” the tendency of outliers to return to a “normal state” over time (Campbell & Stanley 1962). In this case, however, the sample regressed from the mean. On the pretest, for example, eighty-seven percent of the sample indicated they “will vote” in the Republican primary. After watching the debate, ninety-one and a half percent indicated they would vote in the Republican primary—a rate quadruple that of the national average. Conclusion and Discussion Presidential primary debates are party organized events, designed to showcase the party’s slate of candidates, drum up media interest, and facilitate informed voting. The few studies that have addressed these debates, however, have neglected the party-oriented aspect of the event. This study redresses this oversight by examining how a primary debate can shape public perceptions of a political party, the party’s candidates—and even the candidates of the “loyal opposition.” The results suggest a different perspective than the conventional wisdom. Presidential debates, at least during primary season, are not “zero-sum” games, with one candidate winning only at another’s expense. In this study, every candidate who attended the first debate emerged a winner. Every attending candidate saw his electability ratings rise and his mean feeling thermometer scores increase; similarly, respondents’ attitudes toward the Republican Party were more favorable following the debate. The party also stood to benefit through voter mobilization, 14 inasmuch as voters indicated that they were more likely to vote after the debate than before—and this from a group already at the extreme in terms of mobilization. The only “losers” of the debate, at least from the audience perspective, were Bill Clinton, Al Gore, George W. Bush, and Gary Bauer, the latter two because of their absence, and the former two, no doubt, because of the criticisms directed at them by the attending candidates. The effects of this debate, however, were not permanent. Three weeks later, when a subset of the original sample was resurveyed, both Bush and Bauer had recovered some of their lost ground in terms of feeling thermometer scores. But both remained statistically below their original showings and all the candidates from the first Arizona debate retained at least a residue of the good will they had earned by their presence in a party-sponsored event. Of course, these results are likely context dependent and may not extend beyond this particular debate. As George W. Bush noted in February, 2000, “each primary season, every four years, there’s a different stamp, a new character to the primaries, a new, interesting twist” (Bruni 2002, 183). The same, of course, could be said for primary debates. But the context and character of this debate were such that its impact on the Republican party would be muted. It took place without its star attraction: George W. Bush. Moreover, thirty-eight percent of the audience believed that at least one of the candidates engaged in “negative tactics,” so the debate was far from a love fest.vi Even so, the audience left the debate with a more favorable opinion of the Republican party and its debating candidates. Another, less quantifiable, manner in which primary debates may help parties is by allowing candidates to serve an electoral apprenticeship—before the electoral lights reach their full wattage and heat. There is little doubt, for example, that the ten or so primary debates that George W. Bush participated in helped him in his subsequent debates against Al Gore. Frank 15 Bruni, the shrewd political writer for the New York Times, describes Bush’s appearance on Oprah during the general election campaign: Oprah wasn’t the kind of thing he was given to doing. It wasn’t his bag at all. The show was about heart-baring and soul-baring and the kind of psycho-probing that Bush detested, but several of the women around him—Laura Bush, Karen Hughes, Mindy Tucker—urged it on him, telling him it would be a good thing. And so he relented, and went, and managed to seem for all the world, as though he enjoyed being there, a feat he had not accomplished during those first, dreadful debates of the Republican primaries, which he had also resisted. The long and grueling campaign was indeed seasoning Bush (2002, 160). In this manner, the primary debates—the whole primary season—serves as a training ground for the general election. The direct beneficiary of this process is the candidate, but the benefits extend to the party as well—in the form of an increased standing in the eyes of the party faithful, increased voter mobilization, and a better prepared candidate. 16 Works Cited Abramowitz, A. I. (1978). The impact of a presidential debate on voter rationality. American Journal of Political Science 22: 680-690. Asch, S. E. (1951). Effects of group pressure upon the modification and distortion of judgment. In H. Guetzkow (ed.) Groups, Leadership, and Men, pp. 177190. Pittsburgh: Carnegie Press. Bartels, Larry M. (1988). Presidential primaries and dynamics of public choice. Princeton, NJ: Princeton University Press. Bruni, Frank. 2002. Ambling into History. New York: HarperCollins Press. Campbell, Donald T., & Stanley, J. (1966). Experimental and quasi-experimental designs for research. Chicago: Rand McNally. Carlin, D. P. (1992). Presidential debates as focal points for campaign arguments. Political Communications 9: 251-265. Davis, M. (1982). Voting intentions and the 1980 Carter-Reagan Debate. Journal of Applied Social Psychology 12: 481-492. Davis, M., & Runge, T. (1981). Beliefs and attitudes in a gubernatorial primary: Some limitations on the Fishbein model. Journal of Applied Social Psychology 11: 93-113. Davis, J. A. (1985). The logic of causal order. Sage University Paper series on Quantitative Applications in the Social Sciences, series no. 07-055. Beverly Hills, CA: Sage. 17 Dennis, J., Chaffee, S. H., & Choe, S. Y. (1979). Impact on partisan, image, and issue voting. In S. Kraus (ed.), The great debates, Carter v. Ford, 1976, (pp. 314-330). Bloomington, IN: Indiana University Press. Dielman, T. E. (1991). Applied regression analysis for business and economics. The Duxbury Series in Statistics and Decision Sciences. Boston: PWS-Kent Publishing Company. Dittes, J. E., and H. H. Kelly. (1956). Effects of different conditions of acceptance upon conformity to group norms. Journal of Abnormal and Social Psychology 53: 100-107. Fiorina, M. P. (1981). Retrospective voting in American national elections. New Haven, Conn.: Yale University Press. Fiske, Susan T., and Shelley E. Taylor. (1991). Social Cognition. New York: McGraw-Hill. Geer, J. (1988). The effects of presidential debates on the electoral preferences for candidates. American Politics Quarterly 16:486-501. Hagner, P. R., & Reiselbach, L. N. (1978). The impact of the 1976 presidential debates: conversion or reinforcement? In G. Bishop, R. G. Meadow, M. Jackson-Beeck. (eds.), The presidential debates, (pp. 57-178). New York: Praeger Hellweg, Susan A., Michael Pfau, and Steven R. Brydon (1992). Televised Presidential Debates: Advocacy in Contemporary America. New York: Praeger. 18 Holbrook, T. M. (1994). The behavioral consequences of vice-presidential debates: Does the undercard have any punch? American Politics Quarterly 22:469-482. Holbrook, T. M. (1996). Do campaigns matter? Thousand Oaks, CA: Sage Publications. Katz, Elihu, & Feldman J. J.. (1962). The debates in light of research: A survey of surveys. In Sidney Kraus (ed.), The Great debates: Kennedy v. Nixon, 1960, (pp.173-223). Bloomington, IN: Indiana University Press. Kraus, S., and Smith, R. G. (1962). Issues and image. In S. Kraus (ed.), The Great Debates, Kennedy v. Nixon, 1960, (pp. 289-312). Bloomington, IN: Indiana University Press. Lang, K., and Lang, G. E. (1962). Reactions of viewers. In S. Kraus (ed.), The Great Debates, Kennedy v. Nixon, 1960, (pp. 313-330). Bloomington, IN: Indiana University Press. Lanoue, David J. (1991). The ‘turning point’: viewers’ reactions to the second 1988 presidential debate. American Politics Quarterly 19:80-95. Lanoue, D. J. (1992). One that made a difference. Public Opinion Quarterly 56: 168-184. Lanoue, D. J., & Schrott, P. R. (1989a). The effects of primary season debates on public opinion. Political Behavior 11: 289-306. Lanoue, D. J., & Schrott, P. R. (1989b). Voters’ reactions to televised Presidential Debates: Measurement of the Source and Magnitude of Opinion Change. Political Psychology 10: 275-285. 19 Lemert, James B. (1993). Do televised presidential debates help inform voters? Journal of Broadcasting & Electronic Media 37:83-94. McIntosh, Everton G. (1989). Perceived bias in presidential debates. Perceptual and Motor Skills 68: 462. McIntosh, Everton G. (1993). Do presidential debates contribute to a change in voters’ attitudes? Perceptual and Motor Skills 77: 545-546. Markus, G. B. (1979). Analyzing panel data. Sage University Paper series on Quantitative Applications in the Social Sciences, series no. 07-018. Beverly Hills, CA: Sage. Markus, G. B. (1982). Political attitudes during an election year: A report on the 1980 NES panel study. American Political Science Review 76: 538-560. Markus, G. B. & Converse, P. E. (1979). A dynamic simultaneous equation model of electoral choice. American Political Science Review 73: 1055-1070. Martel, M. (1983). Political campaign debates: Images, strategies, and tactics. New York: Longman. Miller, Arthur H., and Michael MacKuen (1979). Informing the electorate: A national study. In Sidney Kraus (ed.), The Great Debates, Carter v. Ford, 1976, pp. 269- 297. Bloomington: Indiana University Press. Miller, A., Wattenberg, M. P., & Malanchuk, O. (1986). Schematic assessments of presidential candidates. American Political Science Review 80: 521-540. Nimmo, D. & Savage, R. L. (1976). Candidates and their images: Concepts, methods, and findings. Pacific Palisades, CA: Goodyear Publishing Company, Inc. 20 Patterson, T. E. (1980). The mass media election. New York: Praeger. Payne, J. Gregory, James L. Golden, John Marlier, and Scott C. Ratzan (1989). Perceptions of the 1988 presidential and vice-presidential debates. American Behavioral Scientist 32: 425-435. Popkin, S. L. (1994). The Reasoning voter. Chicago: The University of Chicago Press. Schroeder, Alan. 2000. Presidential Debates: Forty Years of High Risk TV. New York: Columbia University Press. Schrott, Peter R. (1990). Electoral consequences of ’winning’ campaign debates. Public Opinion Quarterly 54:567-585. Sears, David, and Steven Chaffee (1979). Uses and effects of the 1976 debates: An overview of empirical studies. In Sidney Kraus (ed.) The Great Debates: Carter vs. Ford, 1976, pp. 223-261. Bloomington: Indiana University Press. Sigelman, L., & Sigelman, C. K. (1984). Judgments of the Carter-Reagan Debate: The eyes of the beholders. Public Opinion Quarterly 48: 624-628. Steeper, Frederic T. (1978). Public response to Gerald Ford’s statements on Eastern Europe in the second debate. In G.F. Bishop et al. (eds.), Presidential Debates, pp. 81-101. New York: Praeger. Swanson, L. L., and Swanson, D. L. (1978). The agenda-setting function of the first Ford-Carter debate. Communication Monographs 45: 347-353. Taylor, Shelley E., and S. C. Thompson. (1982). Stalking the elusive ‘vividness’ effect. Psychological Review 89: 155-181. 21 Wall, V. J. Golden, and H. James (1988). Perceptions of the 1984 presidential debates and a select 1988 presidential primary debate. Presidential Studies Quarterly 18: 541-563. Yawn, M., K. Ellsworth, B. Beatty, & K. F. Kahn. (1998). How a presidential primary debate changed attitudes of audience members. Political Behavior 20: 155-181. Yawn, M. & B. Beatty. (2000). Debate-Induced Opinion Change: What Matters? American Politics Quarterly 28: 270-285. Zaller, J. R. (1992). The nature and origins of mass opinion. New York: Cambridge University Press. 22 Table One: Percentage of Audience Responding “Don’t Know” to Candidate Evaluation Candidate Percentage of Pretest Respondents Selecting “Don’t Know” Gary Bauer George W. Bush Steve Forbes Orrin Hatch Alan Keyes John McCain 39.6 2.4 1.5 20.5 21.8 1.2 Percentage of Posttest Respondents Selecting “Don’t Know” 39.9 .3 0 .6 .3 .3 23 Table Two: Feeling Thermometer Scores: Pre and Posttest Candidate Gary Bauer George W. Bush Steve Forbes Orrin Hatch Alan Keyes John McCain Bill Bradley Al Gore Bill Clinton Pretest 5.51 6.32 7.03 5.32 7.14 6.69 2.77 1.4 1.76 Post Test 4.57 4.92 7.87 5.77 8.57 6.96 2.24 1.17 1.56 Difference -.94*** -1.4*** +.84*** +.45*** +1.43*** +.27*** -.53*** -.33*** -.21 N 138 278 306 241 242 303 246 277 34 24 Table Three: Changes in Feeling Thermometer Scores For the Republican Party Pretest Republican Party 7.61 Post Test 8.0 Difference +.39*** N 260 25 Table Four: Candidate’s Ability to Beat Al Gore: Pre and Posttest Candidate Gary Bauer George W. Bush Steve Forbes Orrin Hatch Alan Keyes John McCain Pre % 9.7 86.8 48.7 13.5 24.8 75.5 Post % 8.9 74.8 65.3 26.5 56 75.2 Difference -.8 -12 +16.6 +13 +21.2 -.3% N 318 318 318 318 318 318 26 Table Five: Candidate’s Ability to Beat Bill Bradley: Pre and Posttest Candidate Gary Bauer George W. Bush Steve Forbes Orrin Hatch Alan Keyes John McCain Pre % 8.2 85.3 50.5 14.2 22.6 73.3 Post % 6.2 72.3 62.8 23.7 54.1 73.5 Difference -2 -13 +12.3 +9.5 +31.5 +.2 N 319 319 319 319 319 319 His total was also more than the total for both Bill Bradley and Al Gore combined. By contrast, in 1980, Ronald Reagan was asked a similar question about the identity of a foreign leader. Reagan responded: “No, but I guarantee you he will know my name when I am elected.” iii In December, 1999, the McCain campaign held press conferences unveiling public officials who, having previously backed Bush, were switching their support to John McCain. iv Willing audience members were given a much longer survey, tapping a number of issues not addressed in this article. The longer survey included all of the questions on the smaller survey, but also included items ranking each candidate according to one of eight candidate qualities. Of the three hundred or so members in the total sample analyzed here, approximately fifty filled out the larger survey. To ensure that the survey form did not influence the results, a comparison of the results from the two forms was conducted. No significant difference were detected. v In their classic comparison of television viewers and on-street observers of General McArthur’s parade in Chicago, the authors found that direct observers tended to be underwhelmed and bored. Television, in this case, actually “enhanced the dramatic impact of the event” (76). vi Half of those who said that one of the candidates engaged in negative tactics identified Steve Forbes as the candidate who was negative. Interestingly, however, these respondents actually rated Forbes more favorably after the debate than before. The difference (6.8 to 7.0) was not statistically significant, but it did actually increase. ii i 27