International researchers, among them two Rochester experts, successfully tested 25 different approaches. Two proved most effective.
As a country we are deeply divided. That much we can agree upon. But there may be ways to bridge the chasm, according to a new megastudy published in Science.
The researchers discovered that some attitudes-specifically, support for undemocratic practices (such as trying to curtail free speech, spreading intentional misinformation, or attempting to curb voting in certain precincts) and partisan violence-are clearly distinct from partisan animosity (a strong dislike of or deep distrust between supporters of the opposite party). Thus, lessening animosity does not necessarily lead to a reduction in the other two attitudes.
On the flip side, successful interventions that reduced partisan animosity tended to reduce a host of other problems, such as social distrust, social distance, opposition to bipartisan cooperation, biased evaluation of politicized facts, and support for undemocratic candidates.
According to James Druckman, the Martin Brewer Anderson Professor of Political Science at the University of Rochester (one of the six original researchers to start the megastudy), two of the evaluated approaches were particularly effective across outcomes-even for reducing support for undemocratic practices and partisan violence:
- Correcting misperceptions of the other side
- Using cues to show that elites, such as political candidates or officeholders, place great importance on upholding democratic norms
"People tend to exaggerate the antidemocratic attitudes or the violent inclinations of the other side. And they exaggerate just how different the other side is from your own," says Druckman, who is an expert on democracy and political division. "When you provide information or opportunities to interact with people from the other side and learn about some commonality, that seemed to be a pretty effective approach. The focus on addressing misperceptions seems to be very, very useful."
Megastudy assesses 250-plus approaches
Back in 2019, when Druckman first got involved with the project as a faculty member at Northwestern University, he figured the task ahead would be considerable. Little did he know the gargantuan undertaking it would become.
Pockets of people-academics in various disciplines and civic organizations alike-had begun working on behavioral interventions to address the three most pertinent threats to US democracy: partisan hostility, antidemocratic attitudes, and noticeable support for political violence. But the findings remained fragmented. What was needed was one large, unified collection of ideas where the best could be tested in rigorously controlled, scientific experiments.
The resulting study spans a whopping 32,059 study participants, 252 treatment ideas (whittled down to 25 interventions), 85 coauthors, five countries, and dozens of universities, institutions, and organizations. The researchers come from the fields of political science, sociology, psychology, economics communications, and marketing, as well as from civic organizations.
One approach does not fit all
While most of the tested interventions (self-administered online modules about 8 minutes in length) worked to reduce partisan hostility, only about a quarter successfully lowered support for undemocratic practices. And just one-fifth proved useful in reducing support for political violence.
The degree to which those interventions worked was also telling.
The researchers found that it's much easier, for example, to reduce a person's animosity toward someone of the opposite political opinion than it is to reduce someone's support for political violence. According to Druckman, that's in part because support for violence is already relatively low, so there's less room for reduction. Yet, Druckman notes, the minority who supports violence is "likely more extreme in its opinions and, therefore, harder to convince to change its views."
Of the 25 tested interventions in the megastudy:
- 23 significantly reduced partisan animosity by up to 10.5 percentage points
- 6 significantly reduced support for undemocratic practices by up to 5.8 percentage points
- 5 reduced support for partisan violence by up to 2.8 percentage points
At times, the results were surprisingly mixed.
Druckman tells the story of one video that showed a montage of political violence around the globe-in Venezuela, Russia, and Zimbabwe-where pro-democracy supporters engaged in violent struggles. The video was effective in lessening partisan animosity and antidemocratic attitudes. However, it backfired when it came to study participants' condoning partisan violence, which the video increased.
"I think people may have thought that you need a kind of defensive violence to counteract the tearing down of democracy, perhaps, but it's hard to say for sure," Druckman notes.
Which megastudy interventions work best?
The answer may depend on what specific undemocratic problem you are trying to address.
Partisan animosity
Highlighting sympathetic, politically dissimilar people and emphasizing their common ground was effective in reducing partisan animosity. For example, a top-scoring intervention ("correcting division misperceptions") involved watching a four-and-a-half-minute Heineken commercial from 2017, titled "Worlds Apart."
In the video, perfect strangers of diametrically opposite views are paired: a feminist with a man who equates feminism with "man hating," a homophobic man with a transgender woman, a climate activist with a climate change denier.
The twist? They get to know each other first and find commonalities before they are revealed to each other as holding views the other abhors. Then they are given a choice: walk out or discuss the issue over a beer (no points for guessing the brand). In the video, every single person opts for the discussion, which looks civilized and friendly.
The second-highest scoring intervention (and a close second to the Heineken video) came from a team that includes Cameron Hecht, then a postdoctoral researcher at the University of Texas at Austin and now an assistant professor in the University of Rochester's Department of Psychology. Hecht focuses on developing solutions to societal problems, such as disparities in academic motivation, mental health issues, and, yes, political polarization.
His team's approach-titled "common exhausted majority identity"-managed to reduce partisan animosity by more than 10 percentage points. How? The researchers gave study participants information that reframed polarizing content from news and social media companies as a calculated manipulative strategy. They explained that this strategy is designed to artificially deepen political division and manufacture outrage-because companies use manipulation as a tool to maximize and maintain their own audiences.
The team's intervention, which uses text, images, and voiceover narration as part of an online module, takes advantage of people's reactance, a well-established psychological concept in which any blatant attempt to change a person's behavior is likely to result in negative reactions and push back.
The researchers used the participants' natural reactance to harness it against someone else.
"We show them how they're already being manipulated by a third party, in this case news media, and explain that the way to fight back against that control and regulation is to engage in the behavior that we think is good and beneficial," says Hecht.
Undemocratic practices
Correcting misperceptions about another person's political views and instead highlighting the risk of a democratic collapse was especially effective in lowering support for undemocratic practices.
The highest scoring intervention here (titled "correcting democracy misperception") asked participants eight questions about what they thought people from the other party believed when it came to democracy-undermining actions. Participants then guessed the other side's willingness to use violence, to reduce the number of polling stations in unfriendly districts, or to accept the results of elections if they lost.
After each guess, participants received the correct answer, based on recent surveys. The answers clarified that most supporters of the other party do not, in fact, condone actions that undermine democracy and instead also support key elements of democracy.
Partisan violence
Strategies that involve political elites who model healthy democratic behavior, such as having rival candidates stress the importance of sticking to democratic principles, were especially effective in reducing support for partisan violence. For example, the megastudy researchers found that a video (referred to as a "pro-democracy bipartisan elite cue") featuring the 2020 Republican and Democratic gubernatorial candidates in Utah was one of the most successful interventions in this category.
In the video, Republican Spencer Cox (who went on to win the election) and Democrat Chris Peterson, standing opposite each other, tell the viewer that "we are currently in the final days of campaigning against each other but our common values transcend our political differences, and the strength of our nation rests on our ability to see that."
Translating the findings into the real world
"Just because the Utah governor ad worked in this study setting, we can't just play it a few times and expect it to unify the country," cautions Druckman.
But the video's success in changing hearts and minds (at least in the study's setting) offers important insights.
"If we take a more aggressive stance toward trying to get partisans from different sides to show some agreement," says Druckman, "that could have a really positive effect over time."
Besides creating a unifying research framework and providing a theoretical understanding of the mechanisms and psychological tenets that were driving antidemocratic attitudes, the researchers were also looking for effective, scalable interventions.
To that end, Hecht and his team members have already begun work on a follow-up study of their intervention. This time the team examines whether participants, after experiencing the intervention, were switching away from divisive news sources, unfollowing social media accounts that made them angry at other Americans, or removing social media apps from their phone.
Early data, Hecht says, looks promising.
Meet your experts
James Druckman
Martin Brewer Anderson Professor of Political Science
An expert in political behavior and a member of the American Academy of Arts and Sciences, Druckman studies public opinion formation, political polarization, political and scientific communication, political psychology, and experimental and survey methods. He has published approximately 200 articles and book chapters. His latest coauthored book, Partisan Hostility and American Democracy: Explaining Political Divisions and When They Matter (University of Chicago), was published in 2024.
Cameron Hecht
Assistant Professor of Psychology
Hecht's research "seeks to identify psychological processes that contribute to societal problems, develop theory-based interventions that target these processes, and identify the features of contexts that enable these interventions to be effective." His recent studies have been published in the Journal of Research on Adolescence and the Proceedings of the National Academy of Sciences.
