Tell people that elections like theirs are sometimes decided by a handful of votes, and more of them show up. A field experiment during the 2014 US congressional primaries found that changing a single number in a phone script raised turnout by 1.6 percentage points.
The study is by Daniel R. Biggers of the University of California, Riverside, David J. Hendry of the Hong Kong University of Science and Technology, and Gregory A. Huber of Yale University. It was published online in American Politics Research on October 23, 2023.
Political scientists have long argued that people vote more when they expect a close race, because their vote is more likely to matter. Testing that cleanly is hard, since close races also bring more campaigning and media coverage.
Two scripts that differed by one number
The researchers ran the experiment in Massachusetts, Michigan, Minnesota, Missouri, New Hampshire, Tennessee and Wisconsin. A professional survey firm phoned randomly chosen registered voters in the four days before each state’s primary.
Both closeness scripts said that more than 7% of roughly 160 seriously contested 2012 US House primaries were decided by fewer than a certain number of votes. One group heard 350 votes; the other heard 2,500. Both statements were true, because the 12 races decided by under 350 votes were also decided by under 2,500. The scripts said nothing about the voter’s own race.
Two other groups heard a standard election reminder or a placebo call asking how often they went to the grocery store. Turnout was then checked in official voter files.
The closer-sounding message beat a plain reminder
Of 8,453 people reached with the 350-vote message, 26.0% voted. Of 8,402 who heard the 2,500-vote message, 24.5% voted. That 1.6 point gap is a 6.4% relative increase. A regression with controls put the effect at 1.2 points, or about 5%.
The 350-vote message outperformed the 2,500-vote version in six of seven states, and worked about equally well in districts with and without a competitive House primary. Compared with the placebo group, where turnout was 22.7%, the 350-vote message raised turnout by 3.0 points. The 2,500-vote message and the standard reminder raised it by 1.8 and 2.0 points.
The authors write that the perception that “one’s vote doesn’t matter” is malleable rather than fixed.
Who was reached, and what is still unknown
The sample was people reachable by phone, and their average age was 62 to 63, much older than the electorate. The authors suggest this may make the estimates conservative, because older voters already turn out more. The study covered low-profile midterm primaries, so it is unclear how the message would work in a presidential general election. The team also did not measure directly whether the calls changed beliefs about a vote’s power or something else, such as a sense of civic duty.
For campaigns, the practical point is cheap and honest: the same true fact, framed around the smaller margin, moved more people to the polls than a standard get-out-the-vote reminder.
Study Details:
- Title: Messages Designed to Increase Perceived Electoral Closeness Increase Turnout
- Authors: Daniel R. Biggers, David J. Hendry, Gregory A. Huber
- Journal: American Politics Research
- Publication Date: October 23, 2023
- DOI: 10.1177/1532673X231206139
