SSN Key Findings

What Election Campaigns Need to Learn to Persuade Voters about Candidates and Ballot Measures

Policy field

Connect with the author

University of California-Berkeley
Yale University

The 2016 presidential campaigns supporting Donald Trump, Hillary Clinton, and other candidates spent billions on TV ads, mail, and door-to-door canvassing to try to change voters’ minds. This was no surprise, because campaigns spend much of their time and effort trying to persuade people who already plan to vote to support their candidate.

Do their efforts pay off? Thankfully for campaign managers and interested observers, political scientists have now used experimental approaches to amass a body of evidence about how campaign contacts can change voters’ minds and which voters they can persuade. For the past two decades, political scientists have been partnering with political campaigns to conduct rigorous, real-world experiments to learn what persuades voters. Much like studies in which doctors randomly assign patients to receive either a drug or a sugar pill to learn if the drug works to improve health outcomes, political scientists have worked with campaigns to randomize their persuasive contacts – by mail, door knocks, and so forth – to learn how various kinds of efforts influence voters’ choices.

In a new article, we examine results from 49 of these field experiments conducted with real-world political campaigns. We ourselves conducted nine of these experiments, working over the last two years with Working America, the community affiliate of a major labor federation, the American Federation of Labor and Congress of Industrial Organizations (AFL-CIO). The approaches we tested in this partnership are much more precise than earlier efforts.

Most Persuasion Efforts Fall Short – But Experiments Can Help Campaigns Find Persuadable Voters

What do these experiments show? Across the 40 existing studies and our nine new studies, we find that in general elections where a Republican is running against a Democrat, campaign contacts almost always fail to persuade meaningful numbers of voters. Our best estimate of the direct effects of campaign contact on Americans’ candidate choices in general elections is essentially zero. Our findings throw cold water on the notion that it is easy, overall, for campaigns to persuade voters.

But our experiments have uncovered some exceptions to this general finding. Our findings show that when campaigns conduct experiments to identify pockets of persuadable voters, their persuasion can make a bigger difference.

  • Early experiments in Working America’s campaigns for presidential, Senate, and governor’s races found some elections with pockets of persuadable voters. Follow-up studies revealed that those pockets of voters kept being persuadable all the way until Election Day. Thus by conducting early rigorous field experiments to guide their persuasive efforts, Working America was able to dramatically improve the efficacy of their persuasive efforts. In one state, using the experimental results to guide their targeting allowed them to generate an additional 5.6 votes per 100 conversations in the presidential race, compared to a blind targeting informed only by polling, past experience, and intuition. In another state, the targeting allowed them to generate 9.3 votes per 100 conversations in a U.S. Senate race.
  • In another example, Todd Rogers and David Nickerson worked with a pro-choice organization ahead of the 2008 U.S. Senate election in Oregon to identify voters who identified as pro-choice in a survey. The group then sent them a mailer that informed them the incumbent was pro-life, which was effective at changing Senate vote choice. But it is not easy to replicate this sort of process – having a pro-life incumbent many of whose supporters were pro-choice and could be individually identified by the opposing campaign.

Cautionary Notes

Our data also shows that campaigns should not be fooled by data collected early on in an election that suggests they can persuade meaningful numbers of voters. Campaigns might think they are persuading voters early in an election cycle. However, our research shows that, by the time Election Day arrives, voters often forget early persuasion efforts. What is more, the tactics that seem to work early on in an election cycle usually stop working as the election draws near.

We also find that campaigns can more readily change minds when voters don’t have a partisan label to guide their candidate choice – as can happen in primaries and for ballot measures. Advocacy groups, this finding suggests, may get more bang for the buck by investing more in primaries and ballot measures campaigns.

Last, campaigns can do more than persuade. Rigorous evidence makes clear that campaigns can increase turnout, even in highly competitive elections, by registering new voters and turning out those who otherwise wouldn’t vote. But when it comes to winning elections, how much to invest in registration, turnout, or persuasion still is not obvious. For example, to secure a victory, there might not be enough potential voters to register or mobilize – leaving efforts at persuasion as the only option. Moreover, while increasing turnout nets at most one vote for a campaign, persuasion can net two votes, because it subtracts a vote from the opponent while adding a vote for one’s own side, presuming that the effect lasts through Election Day.

The Need for More Field Experiments

Overall, campaigns should not expect any of their efforts – for registration, turnout, or persuasion – to have overwhelming effects. Voters are difficult to influence, at least with the tactics campaigns are accustomed to using today.

With this said, so far we know the least about the effects of TV advertising, which captures the lion’s share of campaign money. Likewise, experiments have not yet tested or precisely determined the impact of candidates’ qualities, issue positions, or overall messages. In principle, rigorous methods to test the importance of these factors exist – and future campaigns and political organizations would do well to demand that the funds they expend on these various approaches be evaluated using real-world experimental tests.

Future campaigns on behalf of candidates or issues may well learn to devise new tactics and messages that are more effective than those used in the past – perhaps by applying ideas from political psychology or leveraging personal relationships in innovative ways. But new strategies need to go well beyond theory to actual field testing. Campaigns cannot take for granted that even the most promising or strongly touted new ideas will work. They must conduct rigorous field experiments to verify their impact – and hone ever-improved approaches to voters. 


Read more in Joshua Kalla and David Broockman, “The Minimal Persuasive Effects of Campaign Contact in General Elections: Evidence from 49 Field Experiments,” American Political Science Review (2017).