Newspaper coverage of presidential primary debates.

• Author: Benoit, William L.

Televised presidential campaign debates are an important component of the modern presidential campaign. Political debates have several important advantages over other media. First, they give the electorate an opportunity to compare the major candidates "head-to-head" as they discuss the same topics at the same time. Second, Jamieson (1987) commented on the length of these discourses: "As messages running an hour or longer, debates offer a level of contact with candidates clearly unmatched in spot ads and news segments.... the debates offer the most extensive and serious view of the candidates available to the electorate" (p. 28). Third, candidates (and their advisors) prepare extensively for these events. Nevertheless, unexpected questions or unanticipated statements from an opponent are capable of provoking a more spontaneous--and perhaps more accurate--view of the candidates than is possible with highly scripted speeches or television spots with scripts and the opportunity for multiple takes.

Although most scholarly attention has been focused on the general election campaign, we should not neglect the primary campaign. Primaries and caucuses have been increasingly important in recent years (Davis, 1997; Kendall, 2000). Candidates actively develop and disseminate messages to millions of voters in primaries. Furthermore, in some years with weak candidates (e.g., Bush in 1992), in a real sense the primary campaign may essentially determine the outcome of the election by deciding who receives the nomination and is entitled to challenge the weak opponent. For example, Bill Clinton won the Democratic nomination in 1992 so he could challenge, and defeat, President George Bush in the general election. However, it is possible that one or more of the other Democrats (e.g., Paul Tsongas, Bob Kerrey, Jerry Brown) could also have defeated Bush. If so, then the Democratic primary ultimately decided the election by determining who had the chance to face Bush in the general election. Clearly, the primary phase, and the discourse produced in it, merits scholarly attention.

Unlike presidential debates in the general campaign phase, which have been studied extensively (e.g., Benoit & Wells, 1996; Bishop, Meadow, & Jackson-Beeck, 1978; Bitzer & Rueter, 1980; Carlin & McKinney, 1994; Coleman, 2000; Friedenberg, 1994, 1997; Hellweg, Pfau, M., & Brydon, 1992; Hinck, 1993; Jamieson & Birdsell, 1988; Kraus, 1962, 1979, 2000; Lanoue & Schrott, 1991; Martel, 1983; Schroeder, 2000), debates in the primary phase have been unjustly slighted. In fact, the first presidential debate, broadcast on radio, occurred in 1948 between two Republican candidates, Harold Stassen and Thomas Dewey (see Kane, 1987), twelve years before the Nixon-Kennedy debates. Furthermore, John Kennedy tuned up for the 1960 Nixon-Kennedy debates by facing fellow-Democrat Hubert Humphrey in a debate in West Virginia (Berquist, 1960). Primary debates occurred in three years without general debates (1956, 1968, 1972) and in the last four campaigns primary debates have outnumbered general debates by six to one (Benoit, Pier, Brazeal, McHale, Klyukovksi, & Airne, 2002). Clearly, much primary debate discourse is disseminated to the public.

Of course, fewer people are likely to watch a primary than a general debate. This becomes a smaller factor when one considers that the potential audience for a general debate includes all voters in the United States. However, in an important sense the audience for, say, a New Hampshire Republican primary debate consists only of the registered Republican voters in New Hampshire, so the audience for a primary debate need not be as large to potentially make a difference. Furthermore, a meta-analysis revealed that the effects of primary debates on viewers are significantly larger than the effects of general debates (Benoit, Hansen, & Verser, 2003). This is probably because the candidates who participate in primary debates (in mostly January and February) are much less well known than those who debate in the Fall campaign. Another way to look at this is that primary debates serve a particularly important need because so much less is known about the primary candidates.

Furthermore, not everyone watches primary debates; many learn about debates from news reports. Page (2003) reported that "the television viewing audience for the highest-rated debate this fall--the October 9 forum in Phoenix on CNN--was smaller than the audience for the lowest-rated prime-time entertainment show on network television. That debate was watched by 1.8 million Americans" (p. 5A). Note that voters can learn about primary debates held in their own state which they chose not to watch and voters who live in other states can also learn about debates from news coverage. It is even possible that those who watch a debate can have their perceptions influenced by a news story about that debate the next day. Kendall (1997) argued that "there is much evidence of the influence of the media's interpretation of the debates" (p. 1). Hellweg, Pfau, and Brydon (1992) concluded that "news commentary does influence viewers' perceptions about debates" (p. 99). This raises the question of whether that coverage accurately reflects the content of debates.

The question of whether news reports of debates accurately reflect their content is important for several reasons. First, the news plays a gate-keeping function, reporting some but not all of the information available. As Patterson (1994) explained, "The news is not a mirror held up to society. It is a selective rendition of events" (p. 60). Some voters learn from campaign messages like debates whereas others learn from news reports about the campaign. In order to understand the flow of information in a political campaign--what information is available from which sources--we need to know if there are biases in the "selective rendition" of stories about presidential primary debates.

Second, Ansolabehere and Iyengar (1995) have argued that negative advertising may lead to cynicism and smaller voter turn-out (Ansolabehere & Iyengar, 1995). Importantly, Ansolabehere and Iyengar did not content analyze television spots to measure campaign tone; they content analyzed news stories about the campaign (Finkel & Geer, 1998, point this out) and although they conclude that negative advertising decreases turn-out, what their data actually indicate is that as negativity in campaign stories increases, turn-out drops. Thus, we need to examine the negativity of these stories. It is possible that if reporters over-emphasize the attacks in primary debates, that could have pernicious effects on voter turn-out.

Thirdly, voters decide how to vote based on various criteria. Public opinion poll data (reported in Benoit, 2003) reveal that some voters say policy (issues) is the most important determinant of their vote for president whereas other voters say character is the most important determinant of their vote. This is of course why candidates discuss both policy and character in their messages. Newspaper reports that emphasized one topic (policy or character) could be shortchanging those voters who consider the other topic to be the most important determinant of their vote.

Furthermore, as we will discuss in the literature review, casting a vote is in some ways like cost-benefit analysis (Benoit, 1999). Candidates use three functions in their messages: acclaims (positive statements identifying the candidate's benefits or favorable qualities), attacks (negative statements identifying an opponent's costs or unfavorable qualities), and defenses (statements which refute attacks, attempting to dispel alleged costs). If news coverage emphasizes one function more than others, it would make it more difficult for voters to make a decision (arguably voters need to know both costs and benefits to make an informed decision).

Finally, one may assume that reporters know what is best for voters and that the topics they choose to emphasize are more important than the ones stressed by the candidates themselves. This is highly paternalistic and it is debatable whether reporters deserve this trust. While we know of no evidence that directly addresses the question of whether reporters emphasize issues that are best for voters, two studies suggest caution. First, Benoit and Hansen (2001) correlated the frequency with which reporters asked questions about issue topics (e.g., education, taxes, health care) in debates with the importance of that topic to voters. They found no significant correlations: Reporters' questions in debates did not emphasize the topics which were most important to voters. Hence, it would be risky to assume that reporters who write about debates would emphasize the issues that matter most to voters. Second, Graber (1989) discussed the results of a survey of television and newspaper editors. The three most important criteria for selecting a story are conflict, proximity, and timeliness; "conspicuously absent from their choice criteria was the story's overall significance" (p. 86).

Thus, news coverage of presidential debates influences our perceptions of the debates and accordingly merits scholarly attention. However, little research has investigated the question of whether media coverage of presidential debates are an accurate representation of the content of debates themselves. This essay begins by reviewing the literature, describes its purpose and method, and presents and discusses the results of this study.

LITERATURE REVIEW

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

Functions: Acclaims, Attacks, and Defenses

Most content analytic work identifies two functions of political campaign messages...