Can a friendly chat with a stranger help fight prejudice? Two years ago, a seemingly groundbreaking study said it could — until it was debunked as a piece of brazen scientific fraud. But the researchers who uncovered its problems, Stanford assistant professor David Broockman and UC Berkeley graduate student Joshua Kalla, still thought the underlying idea could have merit. Now, they're releasing their own study to support it, suggesting that engaging someone in a face-to-face conversation can help increase acceptance of transgender people.
Broockman and Kalla started work on their paper, published today in Science, because of researcher Michael LaCour. In a widely reported 2014 study, LaCour had claimed that door-to-door canvassing organized by the Los Angeles LGBT Center had dramatically shifted some voters' opinion of gay marriage. Looking to build on his findings, Broockman and Kalla began working with the LGBT Center themselves to study a new effort aimed at reducing transphobia in Florida.
As the project moved forward, though, red flags piled up. LaCour had gotten an almost unbelievably high response rate from participants, the data was suspiciously neat, and the survey company he'd supposedly worked with denied any connection. Though LaCour continued to defend the research, his co-author agreed to retract the paper, and its results are now almost completely discredited.
"The essential thing to understand here is that Mike LaCour never actually measured anything at all."
Still, Broockman and Kalla chose to continue the project. "We already sort of laid the groundwork for [the paper] before we knew the issues with the original. But we pressed ahead," says Broockman. The retraction didn't disprove LaCour's idea; it just meant there was no meaningful research on it. "The essential thing to understand here is that Mike LaCour never actually measured anything at all," says Stephen Deline, who helped lead the anti-transphobia canvassing project. "As best anyone can tell, he instead made up all of his data and never actually surveyed a single voter."
With the help of the LGBT Center and Florida LGBT advocacy group SAVE, Broockman and Kalla recruited registered voters via survey and sent trained canvassers to their homes. Of the roughly 500 people who came to the door, half got a talk about recycling, just like participants in LaCour's study. The other half went through a 10-minute conversation about discrimination against transgender people. Unlike a standard canvassing pitch, the session was aimed at actively engaging participants, asking about a time they had faced prejudice and asking them to think about how this might connect to the experience of transgender people.
The team followed up with long surveys that included a handful of questions about gender identity and expression, including whether participants supported including trans people in anti-discrimination laws and whether they would support a close friend who decided to transition. The survey also included a "feeling thermometer" for their overall opinion of transgender people. Over the next three months, the group that had talked about transphobia was consistently more accepting on average than the group that had discussed recycling, even after being shown an anti-trans attack ad.
The group that discussed transphobia remained more accepting, even after three months
These are encouraging results, especially during a heated debate over extending anti-discrimination protections to transgender people. The rhetoric around "bathroom bills," which ban people from public restrooms that don't match the gender on their birth certificate, relies on portraying trans people as predatory. Alongside legal discrimination, trans people also face elevated levels of violence, poverty, and job loss — and understanding how to change people's minds is one small part of the solution. But the paper raises one of the same concerns LaCour's study did before it was debunked. LaCour's research was so impressive because it contradicted evidence that trying to change people's beliefs just entrenches them more. As Dylan Matthews of Vox put it, permanently reducing homophobia with a short conversation turned out "too miraculous" to be true. Is reducing transphobia different?
Broockman insists that their new study turned up more convincing results. LaCour, for example, found that people were much more effectively convinced by a gay canvasser than a straight one. In a This American Life segment about the study, a gay man wins over a voter by describing the impact of discrimination on his own life. In the results published today, a transgender volunteer wasn't notably more convincing than a cisgender one. "We don't see this pattern where you have to be a member of a certain group to reduce prejudice against that group," says Broockman. Instead, he compares their process to therapy. "The canvasser is not drawing attention to themselves or their story. They're trying to be there to kind of guide the voter through their own thought process."
"Most people don't change their minds; it's not like a silver bullet."
The new study is also not promising dramatic increases in tolerance. "Most people don't change their minds; it's not like a silver bullet," says Broockman. They estimate that about one in 10 people who were canvassed became less prejudiced. On the 1-100 feeling thermometer scale, their attitude toward transgender people rose 10 points after the conversation. But Broockman notes that people tend to give rough and inconsistent numbers on feeling thermometers, and it's just one factor in their overall measure of acceptance.
At this point, Broockman says he's primarily interested in proving that their canvassing method can produce solid — and non-fabricated — data. "I think scientists love saying 'Oh, well, future research can consider this... blah blah blah...' as lip service. I really do believe that in this case, what we're offering here in part is a model for experimental design," he says. "I see this much less as the epilogue to last year's event, and much more a really exciting prologue to a lot of research to come."