A functional analysis of the 1988 Bush-Dukakis presidential debates.

• Author: Benoit, William L.
• Position: George Bush - Michael Dukakis

In 1960, Nixon and Kennedy inaugurated the practice of televised presidential debates in the general campaign (there were earlier primary debates; Davis, 1997). There followed a sixteen year hiatus in the custom of giving voters an opportunity to see the leading candidates for the highest office in the land, side-by-side, discussing the same topics, in a relatively spontaneous encounter. Beginning in 1976 with the Carter-Ford clashes, presidential debates have emerged as an expected component of presidential campaigns (Friedenberg, 1994).

Clearly, presidential debates merit scholarly attention. First, they offer voters a chance to observe the principal candidates together discussing the same issues (Benoit & Wells, 1996; Carlin, 1994; Hellweg, Pfau, & Brydon, 1992). Second, debates are unique in terms of message length: "As messages running an hour or longer, debates offer a level of contact with candidates clearly unmatched in spot ads and news segments" (Jamieson, 1987, p. 28). Debates offer a more candid view of the candidates because of their spontaneity (candidates cannot use a script). Debates also attract a larger audience (more potential influence) than other message forms:

Nielson (1993) reported that the second presidential debate in 1992 attracted 43.1 million television households or 69.9 million viewers.... Those numbers contrast sharply to the 4.1 million homes or 20.5 million viewers who tuned in for each of the major party conventions.... In 1980, nearly 81 million people watched Ronald Reagan and Jimmy Carter in their only debate encounter... Miller and MacKuen (1979) noted that 90% of the adult population watched at least one of the Kennedy-Nixon debates, and 83% watched at least one Ford-Carter match up. These numbers compared favorably to 73% who read about the campaigns in the paper, 4% who read magazines, and 45% who listened to radio reports. (Carlin, 1994, pp. 6-7)

Furthermore, debates can increase voter knowledge (Benoit, Webber, & Berman, 1998; Lemert, 1993; Pfau, 1988). Hellweg, Pfau, and Brydon (1992) concluded that "debate viewing contributes to considerable learning about the candidates and their positions" (pp. 106-107). Some have suggested that debates do not affect election outcomes: Jamison and Birdsell (1988) asserted that "debates don't very often convert partisans on one side to the other" (p. 161). However, debates can influence elections by persuading undecided voters (Carlin, 1994; Pfau & Kang, 1991). Zakahi and Hacker (1995) described the slender margin in several elections:

In 1960, John Kennedy beat Richard Nixon by about 100,000 popular votes. This is a fraction of a percentage (0.2%) of the total vote. In 1968, Nixon defeated Hubert Humphrey by 500,000 votes (0.7%). In 1976, Jimmy Carter won by less than 2% of the popular vote. Polls in late September of 1976 showed an unusually large number of undecided voters (Reinhold, 1976). In 1980, Ronald Reagan beat Carter by less than 10% of the popular vote, yet two weeks before the election, 25% of the voters were still undecided. (p. 100)

Of course, 2000 was another extremely close contest. The number of voters who are neither Republicans nor Democrats has increased substantially: The proportion of independents has risen from 22.6% in 1952 to 38.0% in 1992 (Weisberg & Kimball, 1993). Neither political party enjoys a majority of citizens, so it is not possible to win the presidency without persuading millions of these voters. Additionally, the percentage of voters who belong to one party but vote for the candidate of the other party ranges from 14-27% (Nie, Verba, & Petrocik, 1999). Furthermore, there is clear evidence that debates have influenced voting behavior (Kelley, 1983; Kirk, 1995; Roper, 1960; Wayne, 1992). Thus, presidential debates clearly merit scholarly attention.

We focus on the 1988 presidential debate for this functional analysis of campaign discourse. This election was an important clash. Vice President Bush, along with Reagan, was part of a team that had won two terms in the White House--and Bush ran hard on the Reagan/Bush record. However, all factors did not point to a certain Republican victory. The last two sitting Vice Presidents who tried to move from number two to number one (Nixon in 1960 and Humphrey in 1968) had both lost. Furthermore, Bush faced Governor Dukakis, and the last two Governors to run for President had both won (Carter in 1976 and Reagan in 1980). Indeed, the next Governor to run, Clinton, would defeat Bush in 1992. Looking back at Bush's term in office, this election may well have had decisive effects on our country. For example, Dukakis may not have commanded Desert Storm in the same manner as Bush, and Dukakis might have bandied the economy differently from Bush. So, this clash deserves scholarly attention.

The 1988 debates in particular merit scholarly attention. Kane argued in 1987 that "The 1988 campaign, without an incumbent candidate, will be perceived as winnable by either candidate, and neither will risk failure to debate as an issue to be used by the opponent" (p. 249). After the debates, Drew and Weaver (1991) found that they had increased voter learning about the issues in the campaign. Pfau and Kang (1991), studying the first 1988 presidential debates, found that debates tend to reinforce attitudes for partisans, but they "exert significant influence on nonpartisan viewers" (p. 123), as suggested above. Given that neither party can control the outcome of a presidential election, the effects of the 1988 debates on nonpartisan viewers (and on party vote defectors) acquire particular significance. Hinck examined the effects of the two clashes between Bush and Dukakis on their standing in the polls:

At the time of the first debate, George Bush had a substantial lead over Michael Dukakis, but after the first debate, Dukakis closed to within a few percentage points of Bush. One reason for Dukakis' gain can be seen in the way that the first debate shaped doubts about Bush's ability to present his ideas in a manner consistently dear enough to suggest that he was satisfactorily qualified to assume office. The analysis of the debates also explains why Dukakis, having narrowed Bush's lead to a small margin, was unable to overcome Bush's lead in the polls during the final weeks of the campaign. In the second debate, Bush was able to project a more coherent vision of leadership and appear more personable than Dukakis. (1993, p. 151)

Holbrook also commented on the effects of the second debate of 1988, noting that "Bush's perceived performance was strong enough to increase his standing by almost 3 percentage points (1996, p. 108). He also noted the effects of Bush's strong performance in the second debate on supporters of Dukakis: "Dukakis supporters ... who thought Bush had won the debate were almost certain to change their candidate preference" (p. 119). Thus, the 1988 debates informed voters and affected viewers, justifying a closer analysis of this encounter. Holbrook even argues for effects of the 1988 debates on partisan viewers (Dukakis supporters who thought Bush won).

We employ the Functional Theory of Political Campaign Discourse (Benoit, 1999; Benoit, Blaney, & Pier, 1998; Benoit, Wells, Pier, & Blaney, 1999) to analyze the 1988 debates. This approach has already been employed to study the 1960 (Benoit & Harthcock, 1999b), 1976, 1980, and 1984 (Wells, 2000), and the 1996 presidential debates (Benoit, Blaney, & Pier, 1998). The 1992 presidential debates have been analyzed with a preliminary version of the functional theory (Benoit & Wells, 1996). First, we review the literature on the 1988 contest and describe our purpose and hypotheses. Then, we explain our method. Next, we report the results of this investigation and end by discussing the implications of this study for political campaign communication generally.

PREVIOUS RESEARCH ON THE 1988 BUSH-DUKAJUS DEBATES

Researchers have examined various aspects of the 1988 debates, and the bulk of prior work falls into two main categories: voter response studies and rhetorical analyses. Several studies (e.g. Payne, Golden, Marlier, & Ratzan, 1989; Lanoue, 1991) examined responses from the general voting public. Clayman (1992) examined instances of disaffliation from the live audience response during the debates. Lemert, Elliott, Bernstein, Rosenberg, and Nestvold (1991) studied the influence of post-debate analysis on viewer perceptions of the candidates' performances in both debates. As mentioned above, Drew and Weaver (1991) found that the 1988 debates contributed to learning about the issues of the campaign. None of this work was designed to analyze the discourse of the debates.

Other research analyzed the debates rhetorically (Condit, 1989; Lucaites, 1989). Hinck (1993) elaborated the main problems encountered by each candidate--Bush's lack of coherence in the first debate, and Dukakis' passionless style in the second. Ryan (1990) outlined the major issues of the debates and evaluated the effectiveness of each candidate's handling of those issues.

Three studies examined the content of the 1988 debates. Calm, Howard, Stanfield, and Reynolds (1991) indicated that the candidates engaged in a roughly equal amount of clash (44% of the units coded). Statements of candidate self-praise were not specifically coded. Similarly, as their interest did not lie with the particular issues being discussed, they did not distinguish policy- and character-related statements. Hershey (1989) reported that Dukakis tended to stress domestic issues, while Bush emphasized foreign policy and defense. Additionally, Dukakis tended to mention symbolic ideas (such as family, leadership, and opportunity for the working class) more than Bush. This study is useful for understanding the content areas covered, but it does not analyze the functions of comments. Morello (1992) found that Dukakis had issued nearly twice as many attacks as Bush in...