In a time of bitter polarization, Berkeley researchers find a promising solution
The concept, developed by political scientists Alia Braley, Gabriel Lenz and colleagues, focuses on correcting our negative misconceptions about political opponents. A new megastudy in the journal Science finds it’s more effective than other interventions tested.
October 18, 2024
Just days before the presidential election, U.S. political divisions are bitter and deep, and polarization seems like a permanent condition. But a new approach developed by UC Berkeley researchers shows strong promise as a way to ease suspicion and rebuild trust among voters on all sides, according to a mega-study released today in the journal Science.
The concept, developed by political scientists Alia Braley and Gabriel Lenz and colleagues at MIT, appears to create a powerful impact with a relatively simple, data-based message: Voters on the other side of the divide are not as extreme as you think they are — and in fact, most of them value democracy just as much as you do.
The Science study matched the Berkeley approach against 24 other ambitious potential solutions and found it had an overall impact that surpassed all the others. It was top-ranked in reducing voter support for undemocratic practices and also highly effective in reducing support for partisan violence and easing partisan animosity.
At the root of growing anti-democratic sentiment in the U.S. is mutual fear among both Republicans and Democrats that their political opponents intend to undermine democracy, Braley said. “If you want to stop democratic backsliding, you have to be sensitive to the fear and try to turn down the temperature. That could happen through policy decisions, or just through decisions on how you talk to the other side.”
Added Lenz: “What was so surprising about our study was just how eager both sides were to learn that their opponents value democracy. I expected people to be resistant to the idea — for example, I expected Democrats to resist the idea that Republican voters actually support democracy. But they seemed to be reassured.”
A grim overview of American democracy
The new Science megastudy opens with a survey of recent research that, taken together, presents a grim picture of the health of democracy in the U.S.
“The health of American democracy is under threat,” it says. “Although the U.S. clearly remains a democracy, featuring free and fair elections, many experts warn of the potential for the future decline of U.S. democracy.”
While voters profess to support democratic values and practices, the new study says, data shows that anti-democratic sentiments are widespread, especially among Republicans and conservatives. Distrust across the divides has grown toxic. A majority of voters in both parties are willing to vote for candidates who have violated democratic principles, including those who have lied about the real outcome of the 2020 election, clearly won by Democrat Joe Biden. Violent threats against elected officials are surging.
Although partisans on both sides share many attitudes, data reveals some significant differences: Democrats report higher levels of support for partisan violence and for segregating themselves from Republicans. Republicans report higher levels of support for undemocratic candidates and practices, social distrust and opposition to bipartisan cooperation.
The study traces links between dysfunctional politics and a range of harmful impacts: poor mental health, strains in family relationships, avoidance of romance, workplace retaliation and weak collective response to national crises such as the COVID pandemic, which left more than a million Americans dead.
An unusual, but fascinating, political experiment
Against such a backdrop, the megastudy’s two primary authors — sociologists Jan G. Voelkel of Cornell University and Robb Willer of Stanford University — designed research to assess what approaches might help to improve the nation’s political health. They issued a broad call for ideas, and received more than 250 proposed treatments from 400-plus scholars across the social sciences, and from political practitioners, as well.
They assembled a panel of experts to evaluate the treatments, and that process narrowed the field to the 25 most promising contenders, including the Berkeley concept developed by Lenz, a professor of political science, Braley, a Ph.D. student in political science, and MIT computer scientists. Then they engaged more than 32,000 participants online to see what worked.
The models spanned a wide range: Some stressed a message that all voters share common economic interests or common moral values. Some showed participants a video of two people bonding despite their political differences, or an animation about how democracy allows politically diverse people to work together. In one treatment, participants played a trivia game that encouraged bipartisan teams to work together. Another featured “befriending meditations.”
The researchers then tested how these interventions changed attitudes about three primary areas of concern: partisan animosity, support for undemocratic practices and support for political violence.
The results were encouraging. Of the 25 treatments, 23 significantly reduced partisan animosity. But only six treatments significantly reduced support for undemocratic practices, and only five significantly cut support for partisan violence.
The Berkeley-MIT approach, focused on correcting misperceptions about opponents’ democratic values, stood out for its strong impact across those three concerns. It was the top-ranked intervention for reducing anti-democratic attitudes. It ranked third in reducing support for political violence, and seventh in easing attitudes of partisan animosity. It also was among the most effective in reducing participants’ support for un-democratic candidates.
The other top interventions were sophisticated videos. But the Berkeley approach used a simple question and answer format, with text and images only. Data was central to the approach: Participants were asked to what extent most people on the other side of the divide support undemocratic actions. Many people believed that their opponents were strongly hostile to democracy. But then they were given recent public opinion data that showed opponents’ support for democratic values and practices actually was much higher.
Even with such a low-tech approach, the data were persuasive.
In addition to Braley and Lenz, MIT computer scientists Hossein Rahnama, Dhaval Adjodah and Alex Pentland also helped to develop the intervention.
The psychological dynamics that drive Americans apart
Lenz’s past research has focused on a deceptively simple truth about political behavior: Most people don’t follow politics closely. Instead, they fall in behind the direction set by the leaders of their preferred political party.
People on all sides fundamentally support democracy, the researchers said, but political leaders have tremendous power to undermine those values.
“We started with a theory: One of the reasons that democracy is backsliding is because a polarized electorate is afraid that the other side is trying to undermine democracy,” Braley explained. “And they’re afraid because, usually, there’s an aspiring autocrat who creates that fear. That leader says: ‘OK, the other side is stealing the election.’ Then people feel they have to do whatever it takes to save democracy. And so you have people voting in aspiring autocrats.”
In Turkey, Hungary and Venezuela, “people elected autocrats under the guise of democracy,” Braley continued. “And you see this happening in the U.S. Trump says Democrats are trying to undermine democracy, and that creates fear in Republicans who otherwise want to live in a democracy. But they’re giving leeway to Trump to overstep democratic norms. Likewise, that’s creating fear in Democrats who then could become less democratic, in kind.”
The Berkeley-MIT team found a way to reduce the fear that the other side poses an existential threat. In effect, their model short-circuits the feedback loop, slowing or stopping the anti-democratic spiral. When people see data that reassures them about the underlying good faith of their opponents, Lenz said, their own trust is reinforced.
How would the intervention work in real-life, day-to-day politics?
At every level of the political system — from national politics to community organizations — leaders can reinforce trust by constantly stressing that they support democracy and by recognizing that this is a value shared across the political spectrum.
That can also be an important consideration in the practice of bridging divides, a system of trust- and consensus-building that is being pioneered at UC Berkeley. The leaders of national and campus bridging initiatives have approached Braley and Lenz for advice on how to improve such efforts to reduce polarization.
“People should be very happy to hear that voters on all sides really do support democracy,” Lenz said. “Instead of support for democracy being narrow, it’s very broad-based. People love democracy. They think it’s important. And that’s really reassuring.”
Learn more:
Learn more about UC Berkeley initiatives to ease polarization through projects designed to bridge even difficult political and cultural divides.
Read about the new Berkeley Center for American Democracy.