What the research says about campaigns
Political scientists test campaign tactics with experiments and careful studies. Each page gives the short answer, what the studies found, what they can't tell us, and every source, read in a free copy and checked by a second reader.
- Do candidate visits change votes?
Candidate visits seem to move few votes: in the closing stage of the 2012 presidential race, one study finds visits had only a small effect on how people planned to vote, and it faded fast.
- Do party conventions and running mates move the polls?
A party's national convention usually lifts its presidential nominee in the polls; in U.S. races from 1964 to 1988, the typical gain was about 6 percentage points.
- Do debates change votes?
Debates rarely change how people vote: a study of 56 TV debates in seven countries finds no detectable effect on how voters reach their final choice.
- Does knocking on doors win votes, or just bring out voters?
Knocking on doors brings out voters: across U.S. experiments, a visit that reaches someone raises the chance they vote by about 2.5 percentage points on average.
- Do endorsements change votes?
Endorsements can change some votes, but how many depends on surprise: in the 2000 and 2004 U.S. presidential races, unexpected newspaper endorsements moved readers far more than expected ones.
- How much does being the incumbent help?
Being the incumbent, the person who already holds the office, still helps, but in U.S. House races its estimated value fell to 1.2 percentage points of the vote in 2024, the lowest since the 1950s.
- Does a landslide give the winner a mandate?
Political scientist Robert Dahl argues that winning, even by a landslide, gives a president the right to try to pass policies, not a mandate: a message from voters to carry out particular plans.
- Why does the president's party lose seats in midterm elections?
From 1934 to 2022, the president's party lost seats in the U.S. House of Representatives in every midterm election except 1934, 1998 and 2002.
- Do attack ads work, and should a candidate answer them?
A 2007 review of research finds no good evidence that negative campaigning is an effective way to win votes, even though attacks tend to be more memorable.
- How accurate are election polls?
Since 1936, final national polls for U.S. president have missed the gap between the top two candidates by an average of 4.4 percentage points, in one direction or the other.
- Do yard signs win votes?
Lawn signs seem to win a few votes: combining four U.S. experiments in a way not planned in advance, precincts picked at random to get signs gave the advertised side an estimated 1.7 percentage points more of the vote.