Inaccuracies in media coverage of the 1996 and 2000 Presidential Debates.

• Author: Benoit, William L.

**********

Heather Currie (*)

In 2000, at least 46.6 million people watched the first presidential debate (Rutenberg, 2000, p. A30). The huge size of the audience for televised debates means that their potential for influence on voters and the election is considerable.

This potential for influence is particularly important given the fact that several elections have had relatively small margins of victory. Zakahi and Hacker report that "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" (1995, p. 100). This was never more true than in the 2000 election which turned on the outcome in Florida, and was certified by the Secretary of State as a Bush win by a mere 537 votes. Clearly, messages like presidential debates need not sway millions to have the potential to sway the outcome of presidential elections.

Furthermore, evidence suggests that presidential debates have in fact changed attitudes and voting intentions (see Benoit & Wells, 1996; Hellweg, Pfau, & Brydon, 1992). Middleton (1962) found that the 1960 Nixon-Kennedy debates were "extremely important" for the voting decision of one out of every eight voters. Wayne argued that "Kennedy and Carter might not have won without the debates" (1992, p. 229). Geer (1988) argued that "a sizable minority of the public altered their preference for president" from the 1976 and 1984 debates (p. 498). Kelley (1983) reported that about twenty percent of the electorate said they had decided how to vote after watching the 1980 Carter-Reagan debate (see also Davis, 1982). Thus, presidential debates have the potential to influence election outcomes and deserve study.

In fact, research has suggested that debates may have become the single most important influence on voters. Paul Kirk, cochair of the Commission on Presidential Debates, reported that "focus groups and exit polls told us that more people based their decision in 1992 on the debates than any other single means of information throughout the course of the campaign" (1995). Furthermore, Owen (1995) suggested that debates have increased in importance over the years. Debates are an important source of information for voters.

However, despite the fact that millions of voters watch the debates, many others only learn about the debates from news coverage. Even those who do watch the debates are exposed to media coverage of them. Thus, media coverage of debates, as well as the debates themselves, is important. Kendall (1997) argued that "there is much evidence of the influence of the media's interpretation of the debates" (p. 1). Lang and Lang's (1984) study of the first 1976 presidential debate revealed that, of the viewers who were surveyed immediately after the debate (without exposure to media commentary), about twice as many thought that Carter had done a better job. Those who were surveyed after seeing post-debate media commentary thought, again by approximately a two-to-one ratio, that Ford had done the better job. Steeper (1980) reported that over half of those surveyed changed their minds about who did better in the second Carter-Ford debate following the post-debate media commentary (see also Patterson, 1980). Hellweg, Pfa u, and Brydon (1992) concluded that "television news commentary does influence viewers' perceptions about debates" (p. 99). Thus, news coverage of presidential debates influences our perceptions of the debates and accordingly merits scholarly attention.

Scant research has addressed the question of whether media coverage of presidential debates are an accurate representation of the debates themselves. This essay begins by reviewing the literature on political messages generally and presidential debates in particular. Then we report a study on accuracy of media coverage of the 1996 and 2000 debates.

LITERATURE REVIEW

Research on political messages often analyzes this discourse using two key dimensions: functions (positive and negative or attack messages) and topic (issues or policy along with image or character). We will review the literature on the functions of presidential campaign messages.

Functions: Acclaims (Positive), Attacks (Negative), and Defenses

Research on television spots routinely examines the functions of these messages (see Benoit, 1999; Kaid & Johnston, 2001; West, 1997). However, the positive/negative dimension is much less common in studies of presidential debates (Carlin & McKinney, 1994; Coleman, 2000; Friedenberg, 1994, 1997; Hellweg Pfau, & Brydon, 1992; Hinck, 1993; Jamieson & Birdsell, 1988; Kraus, 1962, 1979, 2000; Lanoue & Schrott, 1991). Bitzer and Rueter (1980) counted attacks in the Carter-Ford debates, but did not compare attacks with positive comments.

The Functional Theory of Political Campaign Discourse has been applied to each of the presidential debates. Benoit and Harth cock's (1999) analysis of the Nixon-Kennedy debates found that positive comments (acclaims) were most common at 49% of the utterances, attacks constituted 39% of the remarks, and defenses were 12%. Wells (2000) found that in 1976, there were 52% acclaims, 42% attacks, and 7% defenses; in 1980 he reported 50% acclaims, 39% attacks, and 10% defenses; in 1984 there were 52% acclaims, 37% attacks, and 11% defenses. Benoit and Brazeal (in press) reported that in the Bush-Dukakis debates, acclaims were 59% of the utterances, attacks were 33% and defenses comprised 8% of the utterances. Benoit, Blaney, and Pier (1998) studied the Clinton-Dole debates, indicating that 59% of their comments were acclaims, 33% were attacks, and 7% were defenses. Benoit, Pier, McHale, Hansen, and McGuire, (in progress) reported that in the most recent campaign, the debate consisted of 74% acclaims, 24% attacks, a nd 2% defenses. Thus, in each of the presidential debates, acclaims outnumbered attacks and defenses were the least common utterance form. However, research has yet to address the question of whether media coverage accurately reflects the tenor of these messages.

Benoit, Pier, Brazeal, McHale, Klyukovksi, and Airne (in press) analyzed twentyfive presidential primary debates from 1948 (a radio debate between Dewey and Stassen) to 2000. They found that 63% of the utterances in these debates were acclaims, 32% were attacks, and 4% were defenses, a pattern similar to that found in general debates. Reber and Benoit (in press) compared content analyses of a Republican and a Democratic primary debate from 2000 with newspaper reports of those debates. They found that newspaper coverage over-represented attacks, reporting 45% attacks when 31% of the utterances in these two debates were attacks. Acclaims were under-represented (40% in stories but 58% of these debates). Defenses were also over-represented (16% of the newspaper accounts, 12% of these debates). So, what research exists suggests that newspaper coverage stresses attacks and defense more than acclaims.

TOPICS: POLICY VERSUS CHARACTER ADVERTISING

Once again, research on television spots frequently contrasts policy (issue) comments with character (image) remarks (Benoit, 1999; Joslyn, 1980; Kaid &Johnston, 2001; West, 1997). However, most works on presidential debates (Carlin & McKinney, 1994; Coleman, 2000; Friedenberg, 1994, 1997; Hellweg, Pfau, & Brydon, 1992; Hinck, 1993; Jamieson & Birdsell, 1988; Kraus, 1962, 1979, 2000; Lanoue & Schrott, 1991) do not content analyze them for policy versus character. Bitzer and Rueter (1980) discuss issues raised by the candidates, but do not contrast this with discussion of image or character.

The Functional Theory has analyzed the topics of past presidential debates. Benoit and Harthcock (1999) reported that in the Nixon-Kennedy debates 78% of their comments addressed policy matters, whereas 22% concerned character. Wells (2000) found that in 1976, policy accounted for 86% of remarks and character for 140/0; in 1980, policy constituted 77% of utterances and character 24%; in 1984, policy comprised 81% of comments and character 19%. Benoit and Brazeal (in press) found that Bush and Dukakis devoted 66% of their utterances to policy and 34% to character. Benoit, Blaney, and Pier (1998) found that Clinton and Dole addressed 72% of their remarks to policy and 28% to character. Benoit, Pier, McHale, Hansen, and McGuire, (in progress) reported that in 2000, 76% of utterances concerned policy and 24% pertained to character. Clearly, the debates focus on policy more than character. However, once again we have no evidence about the accuracy of media reports of these debates.

Benoit, Pier, Brazeal, McHale, Klyukovksi, and Airne (in press) found that primary debates focused on policy in 630/n of utterances and on character in 37% of remarks. Thus, these debates also emphasized policy over character. Reber and Benoit (in press) found no significant difference in the proportion of policy and character utterances between the two 2000 primary debates and the newspaper coverage of those debates.

Of course these two topics (policy, character) in political messages are not as discrete as one might assume. Benoit and Wells (1996) argue that a candidate's stance on issues shapes that candidate's image, and that a candidate's image probably influences perceptions about his or her issue stances. This relationship between...