How to talk about politics again, according to the research

Historical setting

Americans have always argued about politics. In 1798 the argument got so bitter that Congress made some criticism of the government a crime, and only opposition editors were prosecuted archives.gov. Three years later, the side that lost the election of 1800 handed over power anyway. The story is in the election of 1800.

People were not nicer then. They disagreed deeply and still shared one system, one set of rules and one country. The shared ground was wider than the fight.

It still is. Most Republicans and most Democrats give the same answer on a long list of questions; see the starting ten. Here is what the research says about talking from that ground.

1. Start from what you share

The tip: Open with something you have in common.

The research: Erik Santoro and David Broockman paired strangers from opposing parties for video calls in two randomized experiments. Talking about their "perfect day" sharply reduced how much each side disliked the other party. Conversations that focused on disagreement, such as why each person supports their own party, had no effect pmc.ncbi.nlm.nih.gov.

The limits: The warmth faded in the long term; in the first study the effect had decayed by a 3-month follow-up. The conversations also produced little change in attitudes about democratic accountability pmc.ncbi.nlm.nih.gov. Liking each other more is a beginning.

Try it: "Before we get into it, what do you actually want for this town in ten years?" Our starting ten lists national questions where most Republicans and most Democrats already agree.

2. Ask before you tell

The tip: Ask about their experience before you share your view. Then actually listen to the answer.

The research: In a 2016 study in Science, canvassers had roughly 10-minute conversations with 501 voters in South Florida that encouraged them to actively take another person's perspective. Prejudice dropped and stayed lower for at least 3 months ocf.berkeley.edu. A later study by Joshua Kalla and Broockman separated three conversational strategies and found that "perspective-getting," asking the other person about their own experiences, durably reduced exclusionary attitudes on its own isps.yale.edu.

The limits: An earlier canvassing study by different authors, published in the same journal, was retracted, and the 2016 paper says so ocf.berkeley.edu. This result stands on its own data, and later field experiments by the same researchers point the same way (see tip 4). We read the perspective-getting study in abstract only isps.yale.edu.

Try it: "Has this ever affected you or someone close to you? What happened?"

3. Listen like you mean it

The tip: When they talk, do not prepare your reply. Listen with attention and without judgment.

The research: Guy Itzchakov, Avraham Kluger and Dotan Castro ran four laboratory experiments on how listeners "characterized by empathy and a non-judgmental approach" affect the people speaking to them cris.huji.ac.il. They predicted that high-quality listening would lower speakers' social anxiety and defensiveness, make them more aware of contradictions in their own views, and make their attitudes less extreme, and they report that all four experiments supported these predictions cris.huji.ac.il.

The limits: These were lab experiments, not political conversations in the field, and we reviewed the abstract and publisher summary only cris.huji.ac.il. Listening opens the speaker up. It does not settle the question.

Try it: Repeat back what you heard before you respond. "So for you, the real worry is the cost. Did I get that right?"

4. Trade real stories

The tip: Facts alone bounce off. A true story from your own life, exchanged for theirs, lands.

The research: Kalla and Broockman ran three pre-registered field experiments in which 230 canvassers talked with 6,869 voters in 7 U.S. locations. In the first, face-to-face conversations that used arguments alone had no effect. Otherwise identical conversations that added a "non-judgmental exchange of narratives" produced changes that lasted at least four months. Two more experiments replicated the result escholarship.org.

The limits: The effects were small: d = 0.08 in the first experiment and 0.08 and 0.04 in the replications escholarship.org. One conversation is a small step. Many of them add up.

Try it: "Can I tell you why this matters to me?" Then tell it plainly. Then ask for theirs.

5. Understand the values behind the view

The tip: When someone lands somewhere different, look for the value that got them there.

The research: Across six studies on issues including same-sex marriage, universal health care, military spending and making English the official language, Matthew Feinberg and Robb Willer found that liberals and conservatives both tend to argue from their own moral values, and that arguments framed around the values of the listener were typically more effective sdimakis.github.io. People often reach different answers from different starting values, and each side tends to speak only its own.

The limits: Feinberg and Willer's first study recruited participants online through Amazon Mechanical Turk sdimakis.github.io, and online experiments do not always carry over to real conversations.

Try it: "What matters most to you here: fairness, freedom, safety, tradition?" Then listen for it. You will often find you hold that value too.

6. Check what you think they believe

The tip: You probably think the other side is more extreme than it is. Check before you assume.

The research: Joseph Mernyk and colleagues found that Democrats and Republicans overestimated rival partisans' support for political violence by 245% to 442%. A short, factual correction reduced people's own support for violence by 34% in one study and their willingness to engage in it by 44% in another, with effects still visible about a month later pmc.ncbi.nlm.nih.gov. In a 2018 survey of 2,100 Americans, More in Common found Republicans estimated that 54% of Democrats are proud to be American, though they acknowledge the country's flaws. The real figure was 82% perceptiongap.us.

The limits: The violence findings are survey self-reports, not observed behavior, as the authors note pmc.ncbi.nlm.nih.gov. The More in Common survey was fielded in November 2018 perceptiongap.us.

Try it: "What do you think most people on my side actually believe about this?" Then answer honestly.

Evidence

Each tip rests on a named study. Full text was reviewed for archives.gov, pmc.ncbi.nlm.nih.gov, ocf.berkeley.edu, escholarship.org, sdimakis.github.io, pmc.ncbi.nlm.nih.gov and perceptiongap.us. Abstracts or publisher summaries only were reviewed for isps.yale.edu and cris.huji.ac.il, and this page claims only what those abstracts state.

What the evidence does not show

• Not that these techniques change minds on cue. Measured effects range from none to modest, and some fade.
• Not that the methods belong to one side. Several studies were run with advocacy groups on specific issues; we cite them for the method.
• Not that every conversation ends in agreement. A good conversation can end with both of you clearer on why you see one point differently, and on how much else you share.

Sources

1. National Archives, "Alien and Sedition Acts (1798)," Milestone Documents. archives.gov (locator: summary). Accessed 2026-10-28.
2. Erik Santoro and David E. Broockman, "The promise and pitfalls of cross-partisan conversations for reducing affective polarization: Evidence from randomized experiments," Science Advances 8(25), 2022. pmc.ncbi.nlm.nih.gov (locator: abstract; Discussion). Accessed 2026-10-28.
3. David Broockman and Joshua Kalla, "Durably reducing transphobia: A field experiment on door-to-door canvassing," Science 352(6282): 220 to 224, April 8, 2016. ocf.berkeley.edu (locator: abstract; closing discussion of the retracted LaCour and Green article). Accessed 2026-10-28.
4. Joshua L. Kalla and David E. Broockman, "Which Narrative Strategies Durably Reduce Prejudice? Evidence from Field and Survey Experiments Supporting the Efficacy of Perspective-Getting," American Journal of Political Science, first published October 31, 2021. isps.yale.edu (locator: abstract) (abstract only). Accessed 2026-10-28.
5. Guy Itzchakov, Avraham N. Kluger and Dotan R. Castro, "I Am Aware of My Inconsistencies but Can Tolerate Them: The Effect of High Quality Listening on Speakers' Attitude Ambivalence," Personality and Social Psychology Bulletin 43(1): 105 to 120, 2017. cris.huji.ac.il (locator: abstract) (abstract only). Accessed 2026-10-28.
6. Joshua L. Kalla and David E. Broockman, "Reducing exclusionary attitudes through interpersonal conversation: evidence from three field experiments," American Political Science Review 114(2): 410 to 425, 2020. escholarship.org (locator: abstract, accepted manuscript dated November 17, 2019). Accessed 2026-10-28.
7. Matthew Feinberg and Robb Willer, "From Gulf to Bridge: When Do Moral Arguments Facilitate Political Influence?" Personality and Social Psychology Bulletin 41(12): 1665 to 1681, 2015. sdimakis.github.io (locator: abstract; Study 1 Participants). Accessed 2026-10-28.
8. Joseph S. Mernyk, Sophia L. Pink, James N. Druckman and Robb Willer, "Correcting inaccurate metaperceptions reduces Americans' support for partisan violence," PNAS 119(16): e2116851119, 2022. pmc.ncbi.nlm.nih.gov (locator: abstract; Discussion, limitations). Accessed 2026-10-28.
9. Daniel Yudkin, Stephen Hawkins and Tim Dixon, The Perception Gap: How False Impressions Are Pulling Americans Apart, More in Common, June 2019 (YouGov survey, November 7 to 10, 2018, N=2,100). perceptiongap.us (locator: pp. 10, 17 to 18, Table 2). Accessed 2026-10-28.

Last checked against its sources on October 28, 2026. Report a correction.