The Washington PostDemocracy Dies in Darkness

Here’s what persuades Americans to support democracy over party

Our new study tested 25 different approaches with both Republicans and Democrats. Here are three that made a difference.

Sisters Lori Ediger, on left, of Aurora, Neb., and Diana Johnson of Henderson, Neb., listen to presentations during the Nebraska Election Integrity Forum on Aug. 27 in Omaha. (Rebecca S. Gratz/AP)

American democracy is facing arguably its greatest stress test since the Civil War. Less than two years after an attack on the peaceful transfer of power, many prominent figures are undermining U.S. democratic ideals. More than 120 Republican nominees in the midterm elections deny that Joe Biden won the 2020 presidential election, including prominent Senate candidates like Herschel Walker in Georgia and Mehmet Oz in Pennsylvania, although not a single credible audit or court hearing has found evidence of election fraud.

Yet there is little sign that U.S. voters will punish candidates who are breaking democratic norms — such as acknowledging that their preferred candidate lost the last election. Research finds that the majority of voters of both major political parties report that they would continue to support their party’s candidate even if the candidate did such undemocratic things as reducing the number of polling stations in areas supporting the rival party or ignoring court decisions by judges appointed by someone in the other party.

But why do voters support undemocratic candidates? And what might motivate Americans to elevate democratic principles over partisan gain?

Don't miss any of TMC's smart analysis! Sign up for our newsletter.

How we did our research

To answer these questions, we designed a massive experimental study called the “Strengthening Democracy Challenge.” Together with collaborators from Stanford University, MIT, Northwestern University and Columbia University, we used crowdsourcing to recruit more than 250 testable ideas for how to improve Americans’ democratic attitudes, designed by researchers, activists and practitioners from around the world. We received a wide range of submissions and selected 25 to evaluate in a tournament-style experiment.

In April and May, we tested these approaches in a massive online survey with more than 32,000 participants recruited by Bovitz from opt-in panels. The sample was designed to approximate the national U.S. population of Democrats and Republicans with respect to gender, age, race, education and strength of party identification. We randomly assigned participants to one of the 25 approaches, or to be in one of two control groups. We used weighting to protect against bias that could result from participants dropping out of the study at different rates, depending on which group they were assigned to.

Participants in the study had different experiences depending on which group they were assigned to, and these different experiences reflected different strategies submitters thought would improve Americans’ democratic attitudes. For example, some participants were assigned to a group that all viewed a short video of the Republican and Democratic nominees in the 2020 Utah gubernatorial race — Spencer Cox and Chris Peterson — speaking together in support of the legitimacy of the 2020 election. Other groups had different experiences — some viewed videos, others interacted with a chatbot or read short essays — with the content varying based on the submitting team’s strategy. Before completing the study, all participants answered survey questions related to democracy and polarization, allowing us to examine how responses might vary depending on which group the participants were randomly assigned to.

For example, we asked participants how likely they would be to vote for four hypothetical candidates representing their own party on a scale ranging from 0, if they definitely would vote for the other party’s candidate, to 100, if they definitely would vote for their own party’s candidate. In each question, the candidate from the participant’s party was described as having done something anti-democratic that helped their own party: reducing the number of polling stations in districts likely to support the opposition, ignoring unfavorable court rulings, prosecuting journalists and saying they would deny the results of elections their party lost.

We then measured which of the 25 approaches most reduced support for these undemocratic candidates, compared with responses from the two control groups.

Republican candidates are increasingly sharing misinformation, new research finds

What worked to increase Americans’ commitment to democracy?

One of the most effective approaches showed respondents vivid images of societal instability and violence after democratic collapse in several countries, including Venezuela and Zimbabwe, before culminating in footage of the Jan. 6, 2021, U.S. Capitol riot, with narration highlighting the potential for democratic failure in the United States. The success of this approach — which on average reduced support for undemocratic candidates by 4.5 points on a 100-point scale — suggests that Americans simply aren’t imagining what might happen to their own society if democracy were to fail.

Another top-performing approach built on the fact that Democratic and Republican voters greatly overestimate how much voters from the rival party support subverting democratic norms; partisans’ estimates of how much their rivals report supporting anti-democratic actions are more than twice as high as their rivals’ actual reported support for these actions. We found that this approach, which simply provided respondents with the accurate polling data, led to a 4.2 point reduction in support for undemocratic candidates.

These strategies influenced people’s commitments to democratic rules either by correcting their sense of how willing their rivals are to break the rules or by making them take seriously the threat of backsliding into social chaos. However, our results revealed another, less obvious strategy: reducing partisan animosity.

For example, one approach had respondents watch a recent viral U.K. video, produced by Heineken, showing conversations between people with different political orientations connecting with one another and finding mutual respect over pints of beer. Though the video focused on ideological differences in the United Kingdom, it strongly reduced American respondents’ animosity toward the other side more than any of the other approaches we tested. It also reduced partisans’ willingness to support candidates endorsing undemocratic moves — probably because by reducing animosity for rival partisans, it made respondents less determined to defeat the rival party no matter what.

U.S. democracy is in crisis. Civics education can help.

These findings fit with other research on Americans’ voting. According to one analysis, American partisans’ voting is more influenced by their animosity for their rival partisans than by their liking for their fellow partisans, a tendency that has grown steadily since 1980. Our results show that partisan animosity also encourages people to tolerate unethical moves by their own party’s leaders — and thus can erode democracy.

Understanding why many voters are willing to consider voting for undemocratic candidates can help policymakers, activists and others to overcome these motivations — knowledge that may be useful for those who want to motivate Americans to defend fair play in the U.S. political system.

Professors, check out TMC’s newly indexed and improved classroom topic guides.

Robb Willer (@RobbWiller) is a professor of sociology, psychology and organizational behavior and the director of the Polarization and Social Change Lab at Stanford University, where he studies the social and psychological forces shaping Americans' political attitudes.

Jan Voelkel (@JGVoelkel) is a PhD candidate in sociology at Stanford University, where he studies social change, group conflict and morality.