LGBT activists say you can change voters' minds on issues like gay marriage simply by sending out teams to have personal, two-way conversations with them.
A new report in the journal Science shows that this tactic works. But it took a research scandal in 2014 and a two-year redo to get the proof.
The original idea was simple: organizers with the Los Angeles LGBT Center would knock on doors of people who might be opposed to same-sex marriage. Canvassers would introduce themselves, share their own story and chat for about 10 to 15 minutes.
The results from a 2014 study showed that the strategy – called deep canvassing – works, but it was especially effective if the canvasser was LGBT themselves. In that case, what would have taken years to make people more accepting of same-sex marriage shrunk to just months.
"LGBT canvassers had an impact that's never really been seen in social science like this," canvasser Jackson Darling said to Take Two shortly after the report's release.
It turned out the results, however, weren't true because Michael LaCour, the researcher the Center worked with, faked his survey data of voters before and after.
"The scandal," says organizer Laura Gardiner, "it absolutely rocked our team emotionally. It felt like a big punch to our collective gut."
But the Los Angeles LGBT Center wanted to prove that their work was not in vain.
The Center teamed up with David Broockman and Joshua Kalla, the researchers who debunked LaCour's report.
They tried the study again in early 2015, this time on people's attitudes toward transgender people.
"Transgender and gender non-conforming people are the most targeted part of the LGBT community," says Center organizer Ella Barrett. "We needed a method to reduce prejudice against transgender people.
Canvassers walked through neighborhoods in Miami to knock on people's doors. They wouldn't ask them to sign or pledge anything: just to talk.
In this case, some canvassers were transgender while others only shared the story of a friend or loved one who is trans.
"What we found really blew me away," says the study's co-author David Broockman.
One in 10 significantly changed the way they view transgender people, and that impact lasted for at least the three months studied so far.
"If you look at existing research, what you tend to find is that personal conversations – like by mail or television – have impacts that evanesce and disappear," he says, "so people will be persuaded one day but forget it the next."
One important difference is that it did not matter whether the canvasser identified as part of the affected group in question. As long as the conversation was a dialogue where both sides could talk personally, it was effective.
"A lot of these conversations involved voters not just hearing what a canvasser has to say, but talking about their own lives," says Broockman.
The conclusion could dramatically shift activists' strategy the next time a fierce debate is on the horizon, like laws similar to North Carolina's controversial measure preventing protections for LGBT people from discrimination.
"When we humanize transgender people face-to-face, we can reduce prejudice," says Ella Barrett. "Communities can begin to do deep-canvassing conversations before there is a big political flashpoint."