Negativity is a powerful force in the news. This is all the more the case with political advertising. Political ads are ubiquitous in presidential campaigns and are often the largest expenditure in an election (Iyengar 2011; Fowler and Ridout 2012). Ads are also becoming increasingly negative, due in large part to outside groups like 527s and Super PACs. Furthermore, in order for campaigns and outside groups to get the most from their money, negative ads are becoming more concentrated in fewer media markets (Fowler and Ridout 2012). Thus, the majority of voters are not exposed to the ads themselves but instead become aware of the negativity through the news media (Geer 2006; 2012). The aim of this project is to understand the way negative presidential ads are covered in the press.
When it comes to political ads, specifically the negative ones, scholarly and conventional wisdom are divided on their value and effectiveness (Lau, Sigelman, and Rovner 2007; Lau and Rovner 2009). One area of agreement is that negative ads generate more news coverage compared to their positive counterparts (Geer 2012; Ridout and Smith 2008; Fowler and Ridout 2009). John Geer (2012) has persuasively claimed that negative ads now serve two important purposes: (1) attacking the opposition and (2) receiving "earned" or free coverage through the news media. In other words, negative ads are a contemporary version of a presidential campaign press release (Geer 2012; see also Ansolabehere and Iyengar 1995). In this sense, negative ads are powerful not because of the ad content itself, but because the media is publicizing the presence of negative ads--the negative ads are the story.
One limitation, however, is that previous findings are based on the quantity of coverage. For example, Geer (2012) relies solely on article counts in the New York Times and Washington Post. By focusing on the quantity of news, the literature cannot tell us about the quality of coverage. So, yes, negative political ads generate more news coverage than positive ads but we have no empirical evidence about how the news media frames this coverage. This is an important endeavor because voters rarely retain the content of the ad itself and are reliant on the news media for their perceptions of candidates and the ads (Hill et al. 2013; Druckman 2005; Aday and Devitt 2001; Fowler and Ridout 2012). (1) Additionally, scholarly and conventional wisdom tend to equate free media with positive coverage (Lynch 2014). However, free media coverage may be negative and unwanted because of the potential for creating a backlash against the presidential campaign associated with the ad. This makes more sense considering the news media's predilection for negativity and drama (Soroka and McAdams 2015; Bennett 2012; Geer 2012). Despite calls for an examination of the quality of news coverage of negative ads (Fowler and Ridout 2009), we are not aware of any extensive analysis. (2) Based on the theoretical assumptions of scholarly and conventional wisdom regarding negative ads and the news media, the empirical expectations that we test below are straightforward: Free media coverage generated from negative political ads will advantage the campaign associated with the attack. Advantages for the campaign will be controlling the news narrative through the duration of coverage by having more quoted sources and favorable headlines and news frames compared to the campaign being attacked in the negative ad.
Using the New York Times, USA Today, and Washington Post, we conduct a content analysis of national news coverage of the controversial Swift Boat Veteran's for Truth (SBVT) ads from the 2004 presidential election to determine whether scholarly and conventional wisdom correctly assumes that the free coverage generated from negative ads helps the campaign associated with the attack. We selected the SBVT ads for several reasons. We desired to find a well-known case in which the news media played a major role in informing the public about an attack ad. Also, we wanted to find an example that was representative of the contemporary campaign environment in which candidates are not directly responsible for all the ads that are aired. Outside groups now generate the majority of negative ads run during campaigns and rely upon the media to cover these ads and thereby expand the number of voters who are exposed to them. The SBVT ads meet both of these standards.
The SBVT ads are frequently cited among scholarly and conventional wisdom as among the most infamous negative ad spots in presidential campaigns, along with 1964s "Daisy" and 1988s "Willie Horton" (see, e.g., Geer 2006; Geer 2012; Lau and Rovner 2009). Furthermore, this is an appropriate case for analysis because "the messages from this controversial [Swift Boat] spot reached the public not through ad buys, but through the megaphone of the news media" (Geer 2006, 133). Furthermore, Geer (2012) notes that Swift Boat ads exceeded press coverage compared to other controversial or negative spots combined (424).
We content analyzed three months of New York Times, USA Today, and Washington Post coverage of the SBVT advertising campaign. The period of analysis was between the first day a Swift Boat aired (August 5, 2004) and Election Day (November 2, 2004). The news items were accessed from LexisNexis using the search term "Swift Boat." The papers were selected because they are major dailies that devote significant resources to national politics and campaigns. The three have dominated the daily American newspaper market for more than 30 years and were among the largest and prestigious newspapers in America in 2004. Due to the reach and significance of these three national dailies, the topics and tone of coverage that they adopt influence other newspapers throughout the country. These prestige papers also influence other mediums like local television and the blogosphere.
Letters to the editor, items outside of Section A, and advertisements were excluded. News items with passing references to the Swift Boat ads were also excluded from analysis. For example, an item in the New York Times covering the Republican National Convention in New York City informally tested whether GOP delegates could tell the difference between pastrami and corned beef quoted Benjamin L. Ginsberg, the lawyer that resigned from the Bush campaign because of his connection to the Swift Boat ads controversy (Fairfield 2004). For news items to qualify it had to have Swift Boat in the headline, or at least two paragraphs referring to the Swift Boat ads, or it spurred discussion of a related topic such as the value of political ads. (3) This raises the issue of selection bias because we are not sampling from the totality of the 2004 presidential election coverage. However, our focus is on the framing of the SBVT ads and its relationship to free media coverage so we find this targeted search more suitable.
The SBVT ads were intended to discredit John Kerry's Vietnam War record and call into question his ability to be a competent commander in chief. The question guiding our coding scheme was the following: Is there a reasonable expectation that the Kerry campaign would perceive coverage as positive or negative? Thus, we placed each indicator into one of two dichotomous categories: evidence that news coverage was "Good for the Kerry campaign" or evidence that news coverage was "Bad for the Kerry campaign." It follows that what was good for the Kerry campaign was bad for the Bush campaign and what was bad for Kerry was positive for Bush. Any paragraphs that did not contain information falling into one of these two categories we coded as neutral.
The majority of our analysis focused on framing. Frames serve as information organizers as they have the potential to identify and evaluate a problem as well as suggest a remedy (Entman 2004). News frames highlight relevant aspects of an issue while also ignoring or omitting other equally relevant aspects of the issue (Entman 2004). The seven "Good for Kerry" frames included (1) denunciations of the SBVT ads or criticisms and refutations of the evidence and arguments in the SBVT ads, (4) (2) support for John Kerry and/or his military record, (5) (3) links between SBVT and the Bush campaign, (6) (4) President Bush refusing to denounce the SBVT ads, (7) (5) charges of Federal Election Commission (FEC) violations committed by the Bush campaign, (8) (6) fear that the SBVT ads may cause a backlash against President Bush, (9) and (7) questions raised about President Bush's National Guard service. (10)
The seven "Bad for Kerry" frames included (1) uncritical descriptions of the SBVT ads, (11) (2) unqualified questioning of Kerry's military record and his potential as president and commander in chief, (12) (3) no link between SBVT and the Bush campaign, (13) (4) charges of FEC violations committed by the Kerry campaign, (14) (5) commentary on the power or effectiveness of the SBVT ads, (15) (6) the Kerry campaign responding too slowly to the SBVT ads, (16) and (7) President Bush served honorably in the Air National Guard. (17)
Many of the paragraphs were embedded with conflicting frames. If a paragraph contained multiple frames, both positive and negative, then it was coded as such. (18) This is the only way to capture the complexity and magnitude of the frames in each news item. However, in accordance with conventional and scholarly wisdom, we anticipate that negative frames of Senator Kerry should outnumber positive frames.
We also coded the headlines as positive, (19) neutral, (20) or negative for the Kerry campaign. (21) The headline must make an explicit reference to Kerry or Bush for it to qualify as positive or negative. Headlines without their names were coded as neutral. (22) Like news frames, our hypothesis expects that negative headlines will be associated with Kerry whereas the Bush campaign will receive more positive headlines.