If facts alone rarely change minds, what does? Researchers and practitioners have found several approaches with real evidence behind them. They share a surprising feature: the persuader talks less and listens more.

Deep canvassing

Deep canvassing was developed by volunteers at the Los Angeles LGBT Center's Leadership LAB. Instead of reciting talking points at the door, canvassers hold a conversation of roughly ten to twenty minutes built on exchanging personal stories, nonjudgmental listening, and inviting the voter to reflect on their own experiences.

A typical conversation goes like this: ask the voter's view and rate it on a scale; ask why; share a personal story related to the issue; invite the voter to recall a time they needed understanding or were treated unfairly; then ask whether their view has changed. The canvasser does not argue. The voter does most of the reasoning.

In 2016 the political scientists David Broockman and Joshua Kalla published a field experiment in the journal Science finding that a single deep canvassing conversation reduced prejudice against transgender people, and the effect lasted for at least three months. The finding mattered partly because an earlier, widely publicized study of canvassing had been retracted over fabricated data, irregularities that Broockman and Kalla themselves helped uncover. Later studies by the same researchers found similar effects on other issues, though effects are usually modest in size and depend on skilled, well-trained canvassers.

Ideas from motivational interviewing

Motivational interviewing is a counseling method developed by the psychologists William Miller and Stephen Rollnick, first for helping people with drinking problems. Its central insight is that people resist being pushed. When a helper argues for change, the person tends to argue against it. Miller and Rollnick call the urge to correct others the righting reflex and advise resisting it.

Instead, the helper draws out the person's own reasons for change, sometimes called change talk, using open questions, affirmations, reflective listening and summaries. A useful question in civic conversations: “On a scale of one to ten, how sure are you about that?” If the answer is seven, ask “Why not a ten?” The person then explains their own doubts in their own words, which is far more persuasive to them than your objections.

Finding shared values

Level 4 noted that people weigh moral concerns differently. One practical use is moral reframing: presenting your position in terms of values the listener already holds. Someone who prizes tradition and loyalty may respond to an environmental argument framed around protecting the land that past generations handed down, rather than one framed only around harm to distant strangers. Studies by researchers such as Matthew Feinberg and Robb Willer have found modest persuasive effects from this kind of framing. Reframing must be honest; it means finding real common ground, not dressing your view in borrowed language you do not mean.

Telling stories

Stories persuade in ways statistics do not. A specific person with a name, a problem and a choice draws listeners in and makes an abstract issue concrete. Psychologists call the experience of being absorbed in a story narrative transportation, and absorbed listeners tend to counterargue less. Good civic stories are true, specific and short, and they leave room for the listener's own conclusion.

Limits and ethics

These methods are not tricks. They work best when the persuader is genuinely open to learning, and they can be misused if someone fakes curiosity to manipulate. A good rule: use only methods you would be comfortable having used on you, and be prepared to change your own mind too.

Worked example: a deep canvassing conversation

Volunteers in a county are building support for a ballot measure that would fund home care aides for elderly and disabled residents. Lena knocks on the door of a voter named Frank.

Ask and rate. “On a scale of zero to ten, how do you feel about the measure?” Frank says a three; taxes are already too high.

Listen without arguing. Lena does not reply with numbers. She asks why, listens, and reflects: “So you're worried about paying for things that might not work.” Frank nods.

Motivational interviewing move. “You said three, not zero. What makes it a three?” Frank thinks, then says his sister struggled to care for their mother alone. He has just given his own reason to support the measure.

Exchange stories. Lena shares her own story: her grandfather wanted to stay home but could not without help, and her mother nearly lost her job caring for him. Then she asks Frank to say more about his sister. He talks for several minutes about how hard it was.

Shared values. Lena reflects what she hears: “It sounds like family looking after family matters a lot to you.” Frank agrees. She does not invent a value; she names one he expressed.

Re-rate. “Thinking about all that, where are you now?” Frank says maybe a six. He still worries about cost, and Lena acknowledges that concern honestly rather than brushing it aside.

Lena talked for perhaps a third of the conversation. Frank did most of the reasoning, using his own experience. That is what makes deep canvassing different from reciting talking points, and what research suggests makes its effects last.