So far, more than 3 percent of the voters we’ve engaged have switched from planning to vote for Donald Trump to planning to vote for Joe Biden. Even more significant, 8.5 percent of women independents—and 4.9 percent of women overall—said they planned to change their vote to Biden. David Brockman and Josh Kalla, academics who measure what helps shift hearts and minds, told me this persuasion rate is 102 times more effective than traditional electioneering efforts aimed at persuading people to change their vote. On average, these exchanges took 15 minutes. Although spending that kind of time with voters nationwide may not be possible, and would challenge much about how candidates campaign, the lessons we learned—of listening and showing empathy—are important to recognize in this divisive election…

Research has shown time and again that people vote from an emotional place. It’s not so much that facts don’t matter. It’s that facts and talking points do not change minds. And arguing opinions at the start of a conversation about politics causes the interview subject to keep his defensive, partisan walls up and prevents him from connecting with the canvasser.

We don’t try to directly persuade people to change their minds on a candidate or an issue. Rather, we create intimacy, in the faith that people have an ability to reexamine their politics, and their long-term worldview, if given the right context. We’ve found that when people start to see the dissonance between what they believe and what they actually want, their views change—many of them come around to a more progressive perspective. For example, if a woman says she believes that immigrants are the main problem in our society, but reveals that her top personal concern is health care, then we talk about whether immigrants have anything to do with that worry. When a man says he wants to feel safe, we ask questions about what, in particular, makes him feel unsafe. If he answers COVID-19, then we talk about which candidate might be better suited to handle the pandemic.