PCL: Political Communication Lab, Stanford University
PCL: Political Communication Lab, Stanford University
  • PCL

Going Negative

  • C-SPAN Video: Negative Campaign Advertising
    In Shanto Iyengar's 1995 talk at George Washington University he discusses the research methodology and results described in Going Negative: How Campaign Advertising Shrinks and Polarizes the Electorate.

Advertising Manipulations

1992 California Primary and General Elections

Excerpt from Going Negative

Winning, but losing

How negative campaigns shrink electorate, manipulate news media

By Stephen Ansolabehere and Shanto Iyengar

Once upon a time, this country divided itself neatly along party lines. Most people voted; those who did not tended to be poorer, less well educated, and more apathetic, but still party loyal. Television has changed all that. Now, we are split by a new division: between loyalists and apathetics. On the one hand, media propaganda can often shore up loyalists to vote for their traditional party; on the other hand, that same propaganda is increasingly peeling off a band of citizens who turn from independence to apathy, even antipathy, toward our political institutions.

Political advertising is everywhere. In the past, every two years, like clockwork, the American public would be bombarded for a few weeks with televised campaign advertisements; now, with advertising increasingly being brought to bear on such major legislation as trade agreements and health care, it is hard to avoid contact with paid political advertising. The amounts of money spent on political advertising are staggering: hundreds of millions of dollars are poured into what has become the main means of political communication in the United States.

Public regard for politicians has sunk to an all-time low; by wide margins, Americans believe that governmental institutions inflict more harm than good on their collective well-being. The more the campaign rages, the less we seem to respect and like any of its contestants, or even the contest itself

Since 1960, voter turnout in presidential elections has dropped 10 percentage points, from 62 percent of the voting-age population to 52 percent in 1992. The public's sense of its own effectiveness has taken an even more dramatic dive, in 1960, nearly 75 percent of the American public felt confident in the capabilities of government and their own efficacy; today only 40 percent do.

Alabama to Wyoming, New York to California–candidates have taken the "attack dog" tenor, and their bark has kept many voters away from the polls.

Our experiments, and corroborating evidence from actual Senate election results, reveal that high-tech advertising campaigns can stimulate people to vote and instill a sense of confidence in government, but only through positive campaign messages

We monitored each of the 34 U.S. Senate campaigns in 1992. Senate elections provide an especially good test of our claims, since Senate candidates rely heavily on advertising and since many of our experiments dealt with the two California Senate campaigns during the 1992 elections.

We recorded whether the tone of the overall campaign in the state was negative, mixed, or positive. A race was negative if both candidates in the general election relied heavily on attack advertisements. A race was positive if both candidates largely avoided personal or issue-based attacks and, instead, focused on reasons to vote for the candidates. A race had mixed tone if one candidate relied on positive and the other on negative messages or if both candidates used a fairly even mix of positive and negative messages.

Although 1992 was "the year of hope and change," the tenor of the Senate campaigns was overwhelmingly negative. Positive campaigns occurred in 12 states: Alaska, Hawaii, Idaho, Iowa, Kansas, Maryland, Nevada, North Dakota, South Dakota, Utah, Vermont, and Wisconsin. While these states accounted for a third of the Senate races, they contained only 13 percent of the nation's voting-age population. The rest of the electorate feasted on negative advertising. Six states, containing a quarter of the electorate, had mixed campaigns: Alabama, Arizona, Florida, Illinois, Missouri, and Oklahoma. Fifteen states, with 62 percent of the voting-age population, had full-blown negative campaigns: Arkansas, California (two Senate seats), Colorado, Connecticut, Georgia, Indiana, Kentucky, New Hampshire, New York, North Carolina, Ohio, Oregon, Pennsylvania, South Carolina, and Washington.

The best single predictor of campaign tone, it turns out, is the closeness of the race. The tighter the contest, the meaner the campaign.

The hostility of the 1992 Senate campaigns drained the electorate as much as it wore down the opposition. The positive Senate campaigns averaged high turnout rates–57 percent of the voting-age population. Turnout in the mixed tone races was almost five percentage points lower, 52.4 percent, and turnout in the negative races was down even further, to 49.7. After removing the effects of other factors, the difference in turnout between the positive and negative races was 4.5 percentage points–strikingly similar to the size of the effect produced by our experiments.

An even more stringent test of the demobilizing effects of negative campaigning in 1992 is ballot rolloff. Ballot rolloff occurs when people vote for offices high up on the ticket, but ignore less important elections. In the positive Senate races in 1992, 3.3 percent of those who voted for president did not vote for senator. In the negative Senate races, the rate of ballot rolloff was 6 percent.

Negative campaigning may keep people away from the polls for three different reasons. First, negative advertising may discourage supporters of the candidate who is attacked, Attack advertising might defuse partisan support for the opposition, just as advertising in general reinforces partisan preferences. For a supporter reacting to negative information, dropping out may be easier than switching to the attacker.

An alternative explanation is that negative advertising makes the public disenchanted with both candidates. The electorate may curse a "plague on both houses." By this account, candidates unintentionally depress turnout among their own supporters by using negative advertising. If this were true, the effects of negative advertising would set in equally among the ranks of both the candidates' supporters and nonpartisans.

Finally, negative campaigning may diminish the power of civic duty and may undermine the legitimacy of the entire electoral proccess. Campaigns that generate more negative than positive messages may leave voters embittered toward the candidates and the rules of the game.

The negative tenor of campaigns can he traced to the competitive nature of political advertising, to the activities of organized interests, and, last but not least, to the ways in which reporters cover the campaign. Politicians, interest groups, and journalists all act in ways that serve their own best interests. Few of these players really want to produce highly negative campaigns, but the interplay among them produces the kind of campaigns that voters have come to loathe.

"Politics," Lloyd Bentsen reflected after the 1988 election, "is a contact sport." The main event is the head-to-head competition between the candidates. This, above all else, drives candidates to assail one another with 30-second spot ads. Put bluntly, candidates attack out of fear: fear that the opposition will throw the first punch, fear that they will appear weak if they don't respond in kind.

In addition, candidates attack to expand the scope of the political conflict, to drag organized interests and the media into the fray. The more intense the conflict, the more people are drawn to it. The more a candidate attacks, the more he or she makes news; the more conflict there is, and stories about the conflict, the more likely the candidate's proponents are to join the fray.

The media are less partisan, but have an equally important effect on the tenor of campaigns. Journalists report the campaign with the verve of sportswriters covering a title fight. Their job, after all, is to sell papers and attract viewers, and elections are full of great material–the mistakes and weaknesses of the candidates, the twists and turns of public opinion, and the jabs and hooks of political debate. Campaign commercials, especially the negative ones, are ideally suited to the dictates of a good news story. They pack a sensational story with good visuals and good sound into 30 brief seconds. Nothing grabs the public's attention like the smell of a scandal or the prospect of a political upset.

The need to punch back and the temptation to get in the first blow are driven by the same electoral forces. Voters are often most receptive to attack advertisements when the candidate who is attacked responds with self-promotional advertisements rather than a counterattatck, Republicans and Democrats want to see their own candidate score points. Partisans get an extra charge when the man or woman they support blindsides the opposition with a negative ad. By the same token, partisans are averse to seeing their own candidates on the receiving end of such a blow. This sort of judgment by voters drives candidates to run negative commercials. Even candidates who would in principle like to run wholly positive campaigns are led to attack to protect themselves from the airborne assault of the opposition.

On September 7, 1964, Lyndon Johnson's presidential campaign launched "Peace, Little Girl" The advertisement shows a sweet young girl plucking the petals from a daisy and counting from one to nine. Then, a male announcer counts down from ten. With each successive number the camera jumps to a closer shot of the child's face. At zero, a nuclear blast fills the screen, and President Johnson says: "These are the stakes–to make a world in which all of God's children can live, or to go into the dark. We must either love each other, or we must die."

The commercial aired just once, but it caused an instant furor. NBC, the network on which the commercial aired, was flooded with phone calls and letters. Sen. Barry Goldwater requested equal time from NBC to clarify his positions on the tactical use of nuclear warheads. Out of shock and voyeurism, television news programs on all three networks featured the commercial and the resulting controversy. Bill Moyers, President Johnson's press secretary, reported to the president a week after the spot aired:

[W]hile we paid for the ad only once on NBC last Monday night, ABC and CBS both ran it on their news shows Friday. So we got it shown on all three networks for the price of one. This particular ad was designed to run only one time.

Three decades later, campaign commercials remain very much in the media spotlight. Newspapers and television news programs that give even the most superficial attention to the campaign cannot avoid reporting on the candidates' paid messages.

Political advertisements also make the nightly news because they are perfect news stories. Reporters need low-cost, high-impact stories that can be compressed into a half-hour newscast. Most aspects of political campaigns are time-consuming to cover and don't make for good TV. In contrast to the standard campaign fare, 30-second spot ads contain great sound bites, arresting visuals, and sensational attacks, all in a package that firs easily into the two-minute format of television news or the fast format that many newspapers have copied from USA Today. Campaign advertisements are made for TV. They're made for TV news.

Negative advertisements make particularly tasty morsels for the media. For journalists, it is a no-lose situation when candidates attack one another. Allegations of dishonesty and incompetence lay the seeds of controversy and scandal. Even if the charges prove to be false, reporters can always rail against the candidate who aired the attack for slandering his or her opponent and engaging in sleazy campaigning. The fight itself often becomes the story.

Political consultants cater to the incentives facing newsrooms. Campaign managers produce materials that journalists can easily use. Copies of advertisements are regularly sent to television news stations and newspaper op-ed page editors even before the spots air. Sometimes campaigns do not even bother to purchase air time; they simply rely on local television stations to show the commercials for them.

Consultants also dovetail their advertisements with the candidates' speeches and staged events. One technique is "the message of the day." Running a spot on crime and visiting a ward of AIDS babies on the same day only produces confusion. Running a spot on crime and giving a speech in front of 100 uniformed policemen whose union has endorsed the candidate enables the campaign to set the public agenda. The advertisement can ride the wave of favorable news about the endorsement. What is more, by focusing on a single issue, the campaign can lead the press to focus on that issue as well. A story about the endorsement makes a natural lead-in for a story about the concerns raised in the advertisement.

All of this works to the advantage of the candidates. Turning advertisements into news stories allows candidates to set the campaign agenda and magnifies the effects of the ads considerably. Coverage of particular issues boosts the candidates and parties who are already associated with those issues in the voters' minds. Of course, news coverage of the advertisements recycles the sponsoring candidates' messages, providing them with valuable prime time, free of charge.

Not surprisingly, voters increasingly give election coverage poor marks. In 1994, for example, the Times Mirror survey of the electorate gave the media a C grade. The problem is not lack of substance, but the negativity of the message that the media convey. Seventy percent of the voters in the 1994 survey said they had learned enough to make a reasoned decision, but a majority also complained about the negativity of the news. In states where the campaigns were highly negative, people rated the media especially badly.

Following the 1988 election, reporters and editors were deeply critical of the machinations of the Bush campaign and their own complicity in the "Willie Horton" saga. How could they be so readily manipulated by one campaign? What perverse incentives led sensible, experienced reporters to cave into television's single-minded quest for good visuals and good leads?

A number of journalists decided to fight back. All came to the same conclusion: the media should act as referees. A new genre of campaign journalism, devoted to monitoring campaign advertising, was born,

By 1992, "truth-box" or "ad-watch" journalism had come into its own. The Los Angeles Times, for instance, published more than 20 ad-watch stories focusing on the two races for US Senate in California. Ad-watches appeared regularly on national and local newscasts. The producers of CNN's Inside Politics program assigned a senior correspondent, Brooks Jackson, to the task of inspecting and analyzing advertisements aired by the presidential candidates. Today, ad-watches are standard fare in print and broadcast outlets at both national and local levels.

Ad-watches represent an important shift in campaign journalism. They give reporters and editors a tool with which to assert their independence and to rebuild the integrity of their media. Now, journalists subject the candidates' messages to critical analysis and, thereby, alter the incentives facing the candidates.

The media's foray into truth in political advertising also presents an important rest of the role that journalists play in campaigns. Are journalists inadvertent boosters for whomever they happen to cover, or can they serve as the referees in increasingly hostile campaigns? Can they defuse the effects of unfair advertising or do they only magnify the effects?

Since ad-watch journalism is relatively new, little evidence has been collected concerning the effects of these reports on public opinion. We conducted three experiments using actual ad-watch stories from the 1992 presidential election campaign. Each experiment involved two videotapes of a newscast. The presentations were identical except that one of the newscasts contained an ad-watch story; the other contained a nonpolitical (personal interest) filler story. Two of the ad-watch stories analyzed advertisements run by the Clinton campaign; one examined an advertisement run by the Bush campaign.

The ad-watches were actual news stories done by CNN's Brooks Jackson. Each story first replayed the particular commercial, then questioned its facts and assumptions, and offered contrary information. Whenever a false or misleading statement was encountered, the label MISLEADING or FALSE, in bold, red, capital letters, was slapped on the advertisement, Each story concluded by rating the advertisement as either inaccurate or misleading. In sum, these were actual ad-watch stories; they were representative of the genre; and they accented the misrepresentations and falsehoods in the advertisements.

If ad-watches have their intended effect, then viewers should be less favorably disposed toward and less likely to vote for the candidate whose advertisement is criticized. Specifically, we considered the margin (electoral lead) of the candidate who sponsored the ad shown in the ad-watch. To examine the effects of the ad-watches we compared those viewers who saw the story against those who saw no ad-watch (the control group).

The ad-watch stories clearly backfired. In addition, the ad-watches aimed at the negative commercials seem to benefit the candidates somewhat more. Clinton got a slightly bigger boost from the ad-watch about his negative advertisement than from the ad-watch about his positive advertisement, and the message of the Bush attack clearly got through the truth-box filter. In short, just as with the ads themselves, negativity wins in campaign reporting, and the referees can't stop it.

The stories in our experiments produced no significant changes in the intentions to vote and the feelings of political efficacy of Republican and Democratic viewers. Thirty-eight percent of the Independents in the control group reported high levels of confidence in government; only 29 percent of the Independents who saw the ad-watch did.

There are a variety of reasons that these ad-watch stories failed. First, by repeating

the advertisement itself, the ad-watch may strengthen recall of the advertisement, thus making favorable information about the candidate and unfavorable information about the opponent more accessible in memory.

Second, the ad-watch frames the issues in the terms used by the candidates, further reinforcing the sponsoring candidate's messages.

Finally, the audience may consider the ad-watch an unfair attack by the media and side with the candidate whose advertisement is scrutinized. By singling out a particular candidate for criticism, ad-watch reports run the risk of violating the basic journalistic norm of fairness.

Our findings are by no means the last word on ad-watch journalism. This form of campaign coverage is still in its infancy, and our experiments may have detected the flaws with current practices as much as with the genre itself. The challenge facing reporters and editors is to develop methods of covering advertising that do not inadvertently benefit one candidate or the other. To facilitate their roles as monitors, journalists might change the ad-watch story format along the following lines:

  1. Avoid repetition of the advertisement. While the ad's visuals make the story more attractive, refraining from showing the advertisements will minimize the danger of recycling the candidate's message.
  2. Develop two-sided ad-watches. Almost all ad-watches single out just one candidate for scrutiny. As a result, the candidate's agenda becomes the media's agenda. To prevent this, reporters should produce ad-watches that focus on a common theme on which both candidates in a race have advertised.
  3. Use nonpartisan sources. Ad-watches often include sound bites from representatives of the campaigns. This provides the candidates additional free air time and an opportunity to spin the news.

Whether these suggestions will actually improve matters is an open question. Our results counsel caution. Attempts to police the airwaves can backfire just as surely as they can succeed.

Each player in the campaign game acts, reasonably enough, to defend his or her Own interests. Candidates, groups, and journalists all have strong incentives to emphasize the negatives.

This excerpt from Going Negative: How Political Advertisements Shrink and Polarize the Electorate was originally published in Quill Magazine: