Top Ten Myths About Presidential Campaigns

(1) Campaign messages do not influence voters (minimal effects)

Early research on presidential campaigns in the 1940s and later concluded that campaign messages had “minimal effects.” However, campaigns probably have more effects in more recent years. It is also difficult to detect any net effects when two campaigns (such as Kerry and Bush) are campaigning hard for the same votes. If one candidate were to cease campaigning in a state and the other continued, the effects would be more obvious. We found that voters who lived in battleground states in 2000, and who were exposed to ads and campaign appearances from both candidates, had more accurate issue knowledge than voters who lived in non-battleground states, where the candidates did not campaign.

(2) Campaigns are mostly negative

The messages that are most negative are usually from surrogates (e.g., keynote speakers like Senator Zell Miller at the Republican National Convention, and non-candidate groups' ads, such as PAC ads or ads from 527 groups). News coverage, on the other hand, is more negative than the candidates themselves. For example, our analysis of the New York Times coverage of the general campaign from 1952-2000 found that it was mostly negative in tone: 59% negative, 41% positive. However, application of the same procedures to the candidates' own messages revealed that candidates were significantly more positive. General TV spots from 1952-2000 were positive in 61% of statements (39% attacks). Statements in general debates (1960, 1976-2000) were also mainly positive (63%, 37% attacks in debates). If we believe candidates run mainly negative campaigns, it is very likely due in large part to the negative coverage of the news – and to some extent from attacks by surrogates.

(3) Campaigns have little substance/mostly image

Most statements in candidate messages are about policy or the issues. In general TV spots (1952-2000), 61% of statements concern policy (39% character). In general debates (1960, 1976-2000) 75% of statements discussed policy and only 25% character. However, news coverage can emphasize character more than policy. For example, our analysis of New York Times coverage of the general campaign from 1952-2000 found that more statements discussed character than policy (32% character, 26% policy; and 41% horse race).

Of course, candidates do not go into great detail about policy. For example, in 1996 Senator Bob Dole proposed a 15% across-the-board tax cut; President Clinton urged targeted tax cuts. There was no reason for either candidate to discuss 12% or 17% across the board tax cuts, because that was not a “real” choice. Nor was there much reason to talk about other potential specific tax cuts that Clinton did not advocate: Again, this was not a choice. It does not require details to grasp the option facing voters: 15% across-the-board tax cuts versus Clinton's package of specific tax cuts.

(4) Voters do not learn from TV spots

TV spots may seem annoying. They certainly “spin” like whirling dervishes. However, much research has demonstrated that political advertising enhances issue knowledge, influences perceptions of the candidates, and can change vote preference. Of course ads do not change the minds of every voter. In September of 2004 about 12% of the electorate reported that they were undecided and we know that in several years, most recently in 2000, elections have been decided by far, far less than 12% of voters. Ads can inform and persuade voters.

(5) Attack advertising decreases turnout (“demobilization”)

Some research reports that negative ads depress turnout. Other studies find that negative ads increase turnout. Probably, some voters get disgusted with the process and don't vote, whereas other voters get mad (at attacks on “their” candidate) and are more likely to vote. The best evidence comes from “meta-analysis” which can statistically ad together research from different studies. That approach concludes that negative ads do not substantially decrease turnout. Furthermore, our analysis of negativity in ads and other messages found no significant relationship with turnout from 1952-2000. Although some scholars believe that negative ads decrease turnout, the weight of the evidence suggests otherwise.

(6) News coverage of campaigns is good at informing voters

Candidates are biased, of course, so the news must be good at informing voters, right? Research has consistently found that the news – newspapers, television, news magazines – focuses most on the “horse race.” The news wants to tell what is “new” to attract readers, viewers, and listeners. The polls that come out constantly are an easy topic for a story. Voters are better at telling who is ahead than at knowing the candidates' issue positions. The horse race also adds an element of competition and, as long as one side is not running away with the election, some suspense. Horse race coverage also includes stories analyzing the campaigns' strategies. Research spanning 1988-2004 (Democratic primary campaign), but mostly from 1968 on, has consistently found that horse race coverage is more common than coverage of the candidates' character (qualifications for office) or policy positions. Our study of the 2004 Democratic primary coverage (local and national newspapers, national television) found that 68% of the coverage was about horse race, 18% about character, and only 14% about policy.

(7) News has a liberal/conservative bias

Some new outlets (this newspaper, that television network) show a liberal bias; other news outlets (that newspaper, this television network) reveal a conservative bias. Overall, there is no significant slant in news in one direction or the other. Of course, a specific voter does not read all newspapers or watch all television networks. So, there is bias (both directions) and that can affect individual voters. However, considering all voters (and all news outlets), the bias in one direction is “canceled out” by bias in the other direction.

(8) Every vote counts

As much as I HATE to disagree with this statement, the Electoral College rules mean that every vote does not count. Electoral College rules (two states may be exceptions) specify that the winner takes all. A candidate who wins a state by one vote receives all of the states' Electoral College votes. It does not get a candidate one iota closer to the White House to win by a landslide rather than by one vote. The extra popular votes, after a majority plus one, do not help and therefore do not really count. Similarly, if you voted for the candidate who lost in a state, your candidate receives nothing from the Electoral College. This is why the candidates in 2000 and 2004 have ignored the non-battleground states. They focus their campaign appearances and ads in battleground states, the ones were the outcome is uncertain.

The Electoral College winner takes all rule means that voters are short-changed. We found that voters who lived in non-battleground states in 2000, where the candidates did not campaign, had less accurate issue knowledge than voters who lived in the battleground states where the candidates actively campaigned. Furthermore, it is possible that as voters realize that their votes might not really count, that decreases turnout. This is why we should abolish the Electoral College and use direct election, the method used for other elective offices in America.

(9) All attacks are despised by voters

Voters routinely say that they hate mudslinging. And, of course, false and misleading attacks (as well as false and misleading positive statements) are wrong. However, looking deeper into the nature of attacks reveals that most voters believe that attacks on policy are fair, whereas attacks on character are considered unfair. Laboratory studies have confirmed that attacks on policy have larger effects than attacks on character. Furthermore, content analysis of presidential campaign messages from 1948-2000 reveals that winners are more likely to attack on policy than losers; losers are more likely to attack on character than winners. There is no guarantee that every policy attack will succeed or every character attack will fail, but clearly there are differences from different topics of attack.

(10) Debates only reinforce existing beliefs

Debates do reinforce voters' existing beliefs, but that is not all they do. Some voters change their attitudes and which candidate the prefer after watching a debate. Keep in mind that in close elections, such as 2000, debates do not have to sway the votes of millions of voters to affect the outcome. (Debates also have other effects, such as informing voters on the issues and influencing perceptions of the candidates' character)