In an era of intense polarization, it can seem nearly impossible to change anyone’s mind. But a new paper by the political scientists Joshua Kalla and David Broockman offers cause for optimism. In randomized experiments involving long, nonjudgmental two-way conversations between 230 canvassers and 6,869 voters on two hot-button issues — transgender rights and undocumented immigrants — they show that persuasion is possible. Voters became more supportive of transgender rights and undocumented immigrants, and these opinion shifts persisted for at least several months.
The following interview with Kalla has been edited and condensed.
Q: Why, in general, is it so hard to change people’s minds?
A: In the realm of politics, polarization has meant that people are increasingly sorted into camps of Democrats and Republicans. People increasingly feel in-group loyalty towards their party and out-group animosity towards the other party, the parties are becoming more cohesive when it comes to issue positions, and voters are taking cues from that. More broadly, changing your mind requires you to admit that you were wrong, which requires a certain level of psychological security that is often hard — especially in politics, when you are shamed for holding particular positions.
Q: In your paper, you explain that there are more and less effective methods of persuasion.
A: We show that just having a one-way conversation — in which you give a voter talking points and make arguments — isn’t effective at changing people’s minds. And that’s how a lot of political communications work. What works are longer-form conversations in which both sides engage in a discussion about a story or narrative about a particular issue that humanizes the issue, teaches the voter about that issue and builds empathy for an out-group.
Q: How do you show that?
A: We worked with nonprofit organizations that were interested in having these kinds of conversations and learning if they can be effective. We focused on two issues — reducing discrimination against transgender people and increasing support for undocumented immigrants. First, we sent registered voters in certain areas a piece of mail recruiting them to take an online survey, which had some questions about the issue we were interested in and others that were unrelated. Among the people who took that survey, we then randomly assigned them to different experimental groups. The placebo group got a conversation totally unrelated to the subject at hand — about recycling or wearing a seat belt or donating blood. Another group was randomly assigned to a short, one-way conversation. And a final group was randomly assigned to the intervention we were interested in — that full, long conversation involving the personal exchange of narratives.
Then the group we were working with would send canvassers to go door-to-door and have those conversations. Afterwards, we did several rounds of follow-up surveys and compared people’s attitudes across different groups to see whether they became more supportive of undocumented immigrants or transgender people. For every 100 conversations, roughly five people became more supportive.
Think racial segregation is over? Here’s how the police still enforce it.
Q: What do these two-way conversations look like in practice?
A: The conversations all begin with someone knocking at your door and asking you your opinion on the issue. The canvasser might say: “Hey, in the future we might be voting on nondiscrimination laws. Do you think transgender people should be included in these laws? Zero means you think they absolutely shouldn’t be included, and 10 means you absolutely think they should.” The canvasser will ask that in a neutral, nonjudgmental way, almost like a pollster. Many voters will say they’re in the middle, just because it’s an issue they may be less familiar with. The canvasser will follow up and ask, again nonjudgmentally: “Why is that the right number for you? What’s on either side of the issue for you?” The canvasser will really try to understand where the voter is coming from and build rapport.
The second part of the script involves exchanging narratives. In the transgender case, it was narratives about a time when you were treated differently for who you were. The canvassers, if they were trans, might tell a story about a time they were discriminated against. Other canvassers might tell a story about a friend or loved one who is trans and faced discrimination. Then they’d ask the voter, “Has anything like this happened to you?” Because of the rapport built early on, many of the voters would actually open up and share a deep personal story. The canvassers then ask them to reflect on that story and how it might be similar to the experience of transgender people. Often at this point, the voter starts to talk themselves through the connection between their personal experiences and the experiences of transgender people.
The final part of the script is to return to that initial zero-to-10 rating scale to see if the voter has become more supportive. The canvasser would then ask what made you change your mind, with the hope that verbalizing it will increase the durability of the persuasion.
Q: How many voters stick around for the whole conversation?
A: The hardest part is finding someone when they’re at home and getting them to open the door to a stranger. But once they open the door, almost three-quarters of them go through and have the full conversation. And these conversations are long — on average, they lasted 11 minutes.
Q: What is it about the two-way nature of the conversations that makes them so successful?
A: One factor is that it builds trust and rapport. When we ask voters why they feel the way they feel and are genuinely curious, voters seem to enjoy that. When someone you’ve never met is interested in you, that’s a unique experience for a lot of people. Two-way conversations also force people to deeply engage and reflect. It’s very easy to tune out a TV ad, where you’re just passively watching, compared to these conversations, where someone is asking you how you feel and why.
Don’t miss our smart analysis. Sign up for the TMC newsletter!
Nikita Lalwani is a 2020 graduate of Yale Law School, where she was an editor of the Yale Law Journal.