TWO EXISTING EXPLANATIONS

The most common explanation for the rise of negativity is that consultants increasingly believe that attack ads are more effective than positive ads (Iyengar Reference Iyengar2011). Although the scholarly literature does not indicate that attack ads work better than positive ads (Fridkin and Kenney Reference Fridkin and Kenney2011; Lau et al. Reference Lau, Sigelman and Rovner2007), practitioners apparently believe that these work better, and this perception helps account for this increase. Surely, this hypothesis has some merit, but it is unclear how much. Attacks have always been part of American politics. Although negativity is now more common, attacks are hardly new. Thomas Jefferson, for example, was attacked as being the “anti-Christ” in 1800.Footnote ¹ In 1864, Abraham Lincoln was referred to as “a liar, buffoon, ignoramus, swindler, and butcher” (Geer Reference Geer2006, 67). Harry Truman, during the 1948 presidential campaign, equated the Republicans and Thomas Dewey to Adolf Hitler and Nazi Germany (Karabell Reference Karabell2001). Why would attacks be more effective in 2012 than in 1948 or in 1864? Do consultants know more about campaign tactics now than they did then? Perhaps, but when we think about the political skills of people like James Madison, Abraham Lincoln, or Franklin Roosevelt, that claim becomes less appealing. These individuals were masters at their trade. Certainly candidates from the past have gone on the attack, often fiercely, and with success. To say consultants have new and better techniques to practice the “dark arts,”Footnote ² is easy, but nearly every era has witnessed new ways, such as radio in the 1920s, to communicate with voters. Surely, practioners of their day made the most of technological changes as they unfolded. I do not dismiss this hypothesis, but I urge caution before putting too much weight on it.

The second explanation for the rise of negativity involves the polarization of the parties (e.g., Geer Reference Geer2006). The idea is simple and compelling. Candidates now disagree more about policy than they did 30 years ago. These disagreements manifest themselves in attack ads. Figure 2 shows a strong statistical correlation (r = .88) between negativity and polarization. Correlation, of course, does not mean causation. Perhaps negativity is driving polarization, as argued by Ansolabehere and Iyengar (Reference Ansolabehere and Iyengar1995), but that seems unlikely for several reasons. To begin with, evidence that supports claims that ads influence the choices of voters is mixed (see Lau et al. Reference Lau, Sigelman and Rovner2007) and even if these ads do, the duration of the effect is a matter of days (Gerber et al. Reference Gerber, Gimpel, Green and Shaw2011). How could the tone of ads have the power to shape the positions parties take on issues? A structural change in the country, such as polarization, would be more likely the driving force behind negativity. In past research, I have showed that at the presidential level the increased polarization of the parties does seem to generate more negativity (Geer Reference Geer2006). Consider that personal attacks have not increased in frequency over the last 50 years, but “issue” attacks have increased, which is what we should expect as the country has polarized. Perhaps most telling is that the jump in “issue” negativity corresponds more strongly with increasing party polarization than with the overall measures of negativity—evidence that at least indirectly supports my contention (Geer Reference Geer2006, 145–53).

Figure 2

Polarization and Negativity, 1960–2008

A NEW EXPLANATION

This article offers a third explanation: that the news media bear (at least some) responsibility for this rise in negativity. The core idea is that the news media now cover negative ads so extensively that they have given candidates and their consultants extra incentive to produce and air them. Candidates want to get their message out, hoping to control the terms of the debate. They can air a positive ad and seek to influence voters with that spot. But the news media will likely ignore it. Why cover an ad in 2008 that said John McCain wants to improve K–12 education? Surely the senator from Arizona wanted better educated children. Nothing is newsworthy in such a spot. A negative ad, however, can generate controversy and conflict, drawing attention from journalists. So, when John McCain aired a spot claiming that Barack Obama favored “legislation to teach comprehensive sex education to kindergartners,” that drew substantial interest from the press.Footnote ³ This type of coverage by the news media helps candidates get their message out to the public and allows them to shape the all-important narrative of the campaign. Now the way to influence that narrative is increasingly through attack ads, making negativity an even more important tool for consultants in their battle to win votes than just a few decades ago.

The single best example of the news media's interest in negative advertising comes from the 2004 presidential campaign. As a result of the press attention paid to a set of controversial ads aired by Swift Boat Veterans for Truth (SBVT) against John Kerry, the term “swift boat” has become part of the American political vocabulary. These attacks became so well known that political commentators now refer to making nasty allegations during a campaign as being “swift boated.” The facts are that few Americans actually saw the attacks aired directly on TV. SBVT made limited ad buys for these spots in three battleground states (Iowa, Wisconsin, and Ohio).Footnote ⁴ An estimated 1 million people saw these spots on TV,Footnote ⁵ which is not even 1% of the voting public. Americans learned about these ads from the news media's coverage of them. Journalists became enamored with these attack ads about Kerry and gave them a huge amount of attention. Consider that in September, 80% of Americans had read or heard something about these advertisements.Footnote ⁶ Yet, public awareness of these spots could not have come directly from the ads. The public's awareness of the ads came from the press attention to them.

The Swift Boat case is an interesting one, but does it represent a larger systematic shift in the behavior among journalists covering presidential campaigns? The answer is yes (see figure 3).Footnote ⁷ Starting in 1988, the news media began to pay more attention to political advertising during presidential campaigns. The subject drew some attention in the press prior to 1988, but the number of stories in the New York Times and the Washington Post jumped from 88 in 1984 to 197 in 1988. The amount of attention increased again in 2004—surely reflecting the Swift Boat ads. This trend is not limited to newspapers. The network news shows (e.g., CBS Evening News) also indicate a big jump in coverage of negativity in 1988 (Geer Reference Geer2006).

Figure 3

News Media Attention to Advertising, 1960–2008

These data about the news media's increasing interest in ads seems compelling. But a skeptic might reasonably note that the press has always enjoyed conflict and ask, for example, about the so-called Daisy Spot? The Daisy ad, which Lyndon Johnson aired in 1964, is perhaps the most famous negative ad of all time. Although it aired only once, we still talk about it nearly 50 years later. The conventional wisdom is that this ad, which was aired in the early days of television and a quarter of a century before the jump reported in figure 3, generated a good deal of attention by the news media. Yet figure 3 suggests that the spot drew little coverage—at least by comparison to coverage of advertising in more recent campaigns. To probe further I took a closer look at the attention paid to the Daisy Spot by the press in 1964 and compared it to the attention given to other “controversial” ads (West Reference West2009).

Table 1 reports the number of stories in the New York Times and the Washington Post for seven well-known spots. Two things stand out in the table. First, the data dramatically confirm the amount of attention the news media paid to the Swift Boat controversy. It received more press coverage than all of these other spots combined. Second, and more relevant to the question at hand, the Daisy Spot received the least attention of the seven ads I examined. Admittedly, the difference between the attention paid to the Daisy Spot and the Tank Ad (or the Revolving Door Ad) is minor. But those two ads, while well known, are not nearly as famous as the Daisy Spot. Further, the Willie Horton ad drew four times more attention than Daisy, and the Boston Harbor ad drew three times more attention. The Willie Horton ad remains controversial today, so that difference is not so surprising. Considering the press paid so much more attention to the Boston Harbor spot than the Daisy ad is not only a bit curious, but also dramatically underscores my core claim about the surge in coverage of negative ads.

Table 1

What about the Daisy Spot?

These data suggest that we may want to reconsider the idea that the Daisy Spot drew so much attention from journalists in 1964. Perhaps the conventional wisdom that this ad received a lot of attention may speak more to current discussions about negativity than what actually unfolded in the Johnson–Goldwater campaign. That is, we think the Daisy Spot is so controversial because it has become part of the current dialogue over attack advertising. The rise of negativity may have made this ad far more famous now than it was at the time.

However one catalogs the Daisy Spot, the data in figure 3 and table 1 offer strong confirmation of what I call the “McKinnon Hypothesis;” namely that “Ads are about news coverage these days.” It is not, however, just any kind of advertising that draws the interest of journalists; it is, as suggested earlier, negative advertising that is of most interest. No positive ad aired in the last 50 years approaches the attention given to the ads in table 1. Even the famous “Morning Again in America” ads of Ronald Reagan drew little press attention by the standards of the spots mentioned above.

Evidence supporting this claim can also be found in the personal interviews I conducted with leading journalists and consultants.Footnote ⁸ As Dan Okrent, the first public editor of the New York Times, said when commenting about news coverage, “negative is where the story is.” Editors, Okrent argues, are not going to be interested in whether candidates favor world peace. Surely the nominees do. What is newsworthy is something we do not know, something that is pointed or in dispute. Dan Balz, national political correspondent for the Washington Post, agrees, noting that journalists “love conflict.” Balz goes on to point out that this love of conflict often leads journalists to “exaggerate how negative things are.” Balz comments that we often “make a big deal out of small things.” David Chalian, ABC News political director, continues this theme, talking about the need among journalists for “controversy.” And, as Chalian notes, “Negative ads get at that. They are the most base form of controversy being injected into the campaign.” Of course, the fact that all the ads listed in table 1 are negative certainly suggests that the attack ads are what have drawn interest.

Figure 4 confirms the idea that negative ads draw the lion's share of attention from journalists. When taking a close look at the stories on the nightly network news, about 75% are about negative ads. In fact, since 2000, the share of attention paid to attacks is about 80%.Footnote ⁹ The notable exception is 1992, which reflects the interest by journalists in the Perot ads, which were almost all positive. Journalists do indeed like conflict and have given attack ads substantially more air time than positive ads.

Figure 4

Percentage of TV News Stories on Negative Ads

These data actually call for a modification to the McKinnon Hypothesis: “Negative ads are all about news coverage these days.”

WHY DID THE NEWS MEDIA CHANGE?

Why did the news media start paying much more attention to political advertising during the 1988 presidential campaign? Thomas Patterson (Reference Patterson1993) has ably pointed out that journalists started shifting their coverage of news in the 1970s from description to interpretation of events. By the 1990s, his data show that 80% of stories were “interpretative” compared to only 10% in 1960. Negative advertising provides a wonderful opportunity for journalists to interpret strategy and to assess the conflict between the candidates. So, there is a fit between the changes in the news media and negativity.

It also appears that Lee Atwater, campaign manager for vice president George H. W. Bush, was a central figure behind this surge of attention. Joe Klein made this point about Atwater, as did Ed Goeas, Dan Balz, and Adam Nagourney. When I asked journalists and consultants what happened in 1988 to produce this change, each of these individuals raised the name of the mercurial operative from South Carolina, without prompting on my part. My question was open ended, but Atwater was at the top of their list. Klein's account is perhaps most telling. Klein recalls observing focus groups organized by Atwater in Patterson, New Jersey, in June 1988. Atwater was showing these elite journalists the power of attacks on the soon-to-be Democratic nominee, Michael Dukakis. According to Klein, “We were all skeptical.” The public does not care about “flag factories.” Yet these attacks “moved the dials,” he said. Klein contends this experience gave journalists a “new appreciation for the impact of the dark arts of consultants.”

During the Bush campaign, some of the most controversial and discussed negative ads were aired (recall table 1), and as we know, Bush went on to win the 1988 election. A quick look at the campaign suggests that Dukakis was ahead in the summer and then after the attack ads began to fly from the Bush campaign, Bush took command. From this perspective, it seemed like Atwater was able to transform the race. In retrospect, however, we now know that Bush regained his lead in the Gallup poll before he aired any of his attack ads against Dukakis (Geer Reference Geer2006). Bush's victory in 1988 was mostly likely due to a strong economy and the popularity of President Reagan (Vavreck Reference Vavreck2009).

The perception that these attack ads were decisive lingered, fueling a sense of unhappiness among journalists concerning the 1988 campaign. This unhappiness, however, gave the press reason to continue their newfound interest in advertising. Most notable here is David Broder's call to action. Reflecting on the problems during the 1988 campaign, Broder made the following suggestion on January 14, 1990, in the Washington Post:

Broder, in effect, recommended that journalists focus on ads and treat them like speeches. Ads should be assessed, judged, and measured. Clearly, Broder was much more concerned about negative ads than positive ads. As the dean of political reporters, his call to arms carried weight. His colleagues now had even more reason to continue the so-called Ad Watches. By 1992, the major newspaper in every state had instituted some form of an ad watch (Kahn and Kenney Reference Kahn and Kenney1999). While Broder's intent was good, the effects of his recommendation are less clear.

IMPLICATIONS

The usual assumption is that the news media are simply reflecting what is happening in campaigns. They are covering negative ads because candidates and their consultants produce them. That is only part of the story. The news media's recent surge in attention to negativity has altered the incentives of candidates to produce and air negative ads. Ads today are about news coverage—they are “not about persuading voters,” observes David Chalian. Peter Fenn, a Democratic consultant, expands the point that “ads are for the media these days,” contending that spots “are often video press releases designed to play into the 24-7 news coverage.” Sam Fiest of CNN concurs, noting the ads are often “video press releases.” Matt Erickson, another Democratic consultant, was even more pointed, noting that “negative ads at the presidential level have taken the function of press releases.” This connection between ads and press releases at the presidential level underscores one of the core arguments of this article.

Mark McKinnon has already made the point that presidential ads are no longer for voters, but for the news media. It seems to be a near consensus among the people I talked to. In fact, when I asked GOP consultant Alex Castellanos to name a few ads that were created in presidential campaigns for news coverage, he responded, “Well, about all of them.” He mentioned the much discussed “wolves” ad from 2004. The spot sought to generate discussion in the country about security and terrorism. “We wanted the media,” contends Castellanos, “asking the question: was Kerry able to protect the country?” These comments all underscore Chalian's observation that “consultants know they can drive news coverage with ads.”

Given that candidates want to influence voters, the free media offers a powerful way to get out their message. For presidential elections, Chalian said, “The free media narrative is the single most important thing to control.” As shown earlier, however, candidates must have a message that sparks controversy or otherwise it gets no coverage. The best way to do that is through attacks—that is, negative ads. For candidates to get their message out and to control this all-important narrative, the news media have unintentionally given candidates even more reason to create and to air attack ads. This argument recasts how we think about the rise of negativity because the news media have given candidates more reason to be negative.

During the elite interviews I conducted during the fall of 2009, I asked about the merits of this hypothesis that the news media have unintentionally given campaigns more reasons to produce and air negative ads. Here are some of the reactions. John Harris of Politico said, “I absolutely believe that.” “That is probably correct,” commented Dan Balz. Tom Fiedler, who had been editor at the Miami Herald, had some reservations, but was “willing to believe” that coverage of ads had altered the incentives of consultants. Adam Nagourney added, “I think that is right.” Mark McKinnon responded immediately, saying that it was “exactly right.” Alex Castellanos simply said “yes.” Nick Ayers “completely agrees.” Sam Fiest was a bit more cautious, but still supportive, observing “there is something to it.” Finally, David Chalian noted: “Hard to argue it is not true.” He then went on to say that the news media “have crafted a marketplace for negative ads.” This observation highlights the core claim of this article.

In short, we need to revise how we think about the causes of this rise in negativity. The news media are not just reflecting the goings on of campaigns. Instead, their coverage has altered the conduct of campaigns. They do more than cover the process; they shape it. That is, the increase in attacks in presidential campaigns is partly the result of the news media's extensive coverage of advertising, in general, and negative advertising, in particular. As Nick Ayers comments, journalists have “incentivized the process” for attack ads. Negative ads are now being run and produced not so much with an eye toward influencing voters directly, but with the hope of altering the news media's narrative in the campaign. It is that narrative, then, that can provide candidates a chance to win over voters and secure a victory in the election.

A long-standing assumption, certainly in the academic world, is that the purpose of ads is to influence voters. Specifically, social scientists have almost exclusively focused on sorting out the direct impact of ads on voters through television. We need to adjust our research designs to consider changes in the media and how we can best study political behavior under these new conditions (e.g., Clinton and Geer Reference Clinton and Geer2012; Geer, Lau, Nickerson, and Vavreck Reference Geer, Lau, Nickerson and Vavreck2012; Prior Reference Prior2007).

My argument has special relevance for the upcoming 2012 presidential campaign. With the rise of Super Pacs and their ability to air nearly unlimited ads, we should see lots of spots that offer journalists an “interesting” story. Without accountability keeping these ads in check (Geer Reference Geer2006), we could see a spate of Swift Boat like spots. The ads themselves need not be a problem, but journalists' excessive coverage of these ads could be a problem and might yield many Swift Boat scenarios. If journalists do pay attention to the extreme messages that will likely appear in these ads, not only will the news media be giving more attention to these questionable claims, but it will also further encourage additional outrageous ads by Super Pacs (and candidates as well). Journalists need to be sensitive to these issues and make decisions about what claims in ads are, and are not, newsworthy.

The changes described here are, of course, part of larger changes that have unfolded in the mass media over the last two decades. I focus only on one very small piece of the pie. It is, nonetheless, an important piece. We tend to assume that most attack ads we see on television (and the Internet) reflects the increasing belief by consultants that negativity works. Negativity is a fundamental part of campaigns and have been since our country's the first elections. But it is not clear that negativity is on the rise because suddenly politicians and their aides think negative ads work better. When Harry Truman attacked Thomas Dewey in 1948, I am sure it was on a belief that it would secure votes. We can quibble over whether consultants are more enamored with attack politics than in the past. Those quibbles certainly should not lead us to forget that polarization (recall figure 2) is likely a key to the story. And even more importantly, that discussion misses the central objective of this article: The news media need to be part of this conversation over the rise of attack politics that has unfolded over the last 20 years. Perhaps for these reasons Alex Castellanos claimed that the “news media are the most negative force in American politics.” I do want to go as far as Castellanos, but it is important to shine a bright light on this recent and potentially troubling development. By doing this, we can forge a more complete understanding of attack politics in the twenty-first century.

ACKNOWLEDGMENTS

I want to thank the Shorenstein Center at Harvard University for the chance to spend the fall of 2009 in residence. It was a great experience, providing me both the chance to conduct many of the interviews reported here and to have time away from teaching to develop some of the ideas contained in these pages. I also thank Vanderbilt University for support that made this project possible. Finally, I would like to express appreciation to Marc Hetherington who collected some of the data on polarization presented here (see figure 2).