The acknowledgment that a study on changing voters’ minds about same-sex marriage may have been faked casts a shadow over a strategy that has been embraced by liberal activists nationally. (The' N. Pham/Virginian-Pilot via AP)

The study results confirmed everything veteran gay rights activist David Fleischer had seen in his years of door-knocking — that a short but heartfelt conversation at the doorstep could truly change the mind of a voter who opposes same-sex marriage.

So it was a shock this week when Fleischer learned there is evidence that the study, penned by a young PhD student at UCLA, was faked. The study’s co-author, an esteemed Columbia University political science professor, asked the journal Science to retract the groundbreaking paper, saying he was deeply embarrassed by the incident.

The acknowledgment cast a shadow over a strategy that has been embraced nationally by liberal activists seeking to flip conservative voters on a host of issues, including abortion and transgender rights. And it has undermined hopes that there may be a simple way to alter deeply held beliefs, which political scientists have long warned are difficult if not impossible to nudge.

The study’s main author, Michael J. LaCour, said in a statement that he is “gathering evidence and relevant information so I can provide a single comprehensive response” to the request for retraction. Science issued an “editorial expression of concern” to advise readers that it was investigating “serious questions” about the study’s validity.

Fleischer, who heads the Leadership Lab at the Los Angeles LGBT Center, said he was “shell-shocked” by the revelation. The longtime political strategist and gay rights advocate had innovated the canvassing approach, starting in 2009, shortly after California voters approved a ban on same-sex marriage that would later be struck down by the U.S. Supreme Court.

Fleischer had been eager to find out whether he could change conservative voters’ minds by sending trained canvassers, some of them gay, to their doorsteps simply to talk. Over the years, these voters usually had not slammed the door. Instead, they leaned in and engaged, admitted to biases, related personal stories and asked probing questions. When he went out canvassing himself, he often saw a look in their eyes that suggested their thinking had changed.

Fleischer decided to embark on a study to find out whether his suspicions could be proved empirically. At the suggestion of Columbia University professor Donald P. Green, he partnered with LaCour, who had been a student in one of Green’s summer seminars and struck the professor as bright, energetic and engaged.

LaCour developed a plan to determine the short- and long-term effects of these doorstep interactions, and his results were stunning: Before the canvass, the voters as a group resembled the state of Nebraska in terms of their views on same-sex marriage. After the canvass, they looked more like Massachusetts.

The study made headlines in The Washington Post, the New York Times, the Wall Street Journal and elsewhere and was featured prominently in an episode of public radio’s “This American Life” titled “The Incredible Rarity of Changing Your Mind.” It spawned a partner study on abortion, also by LaCour, that sent women who had undergone the procedure door-to-door in Los Angeles County to speak of their personal experiences.

Most recently, it led to a project in Miami focused on transgender rights. But LaCour was unavailable for that project. So Fleischer reached out to a different researcher. He asked David Broockman, another former student of Green’s who was then a graduate student at the University of California at Berkeley, to copy LaCour’s template.

Broockman quickly ran into difficulties. LaCour’s success in coaxing voters to participate in the study appeared to be unusually high, and Broockman was unable to replicate it in a pilot project. He reached out to the survey company LaCour said he had used, but the company had no record of working with LaCour. Nor did it have an employee using the name LaCour listed as his contact.

Broockman found other issues of concern. There was no evidence that LaCour had compensated the study participants, as he had claimed. And the data LaCour included with his article bore a striking resemblance to a previous study on voter attitudes. That was the smoking gun, said Green, who had served as LaCour’s co-author.

Green said he approached LaCour’s academic adviser and hatched a plan. In a meeting Monday morning, the adviser laid out the case against LaCour and asked him to produce his raw data. LaCour couldn’t do it, Green said. LaCour also declined to provide contact information for respondents in the survey so their participation could be verified.

“No one of the irregularities we report alone constitutes definitive evidence that the data were not collected as described,” Broockman and his colleagues wrote in a 27-page report posted to his faculty page at the Stanford Graduate School of Business, where Broockman recently took a job as an assistant professor. “However, the accumulation of many such irregularities, together with a clear alternative explanation that fits the data well, leads us to be skeptical the data were collected as described.”

Green echoed that impression in his retraction letter.

“Michael LaCour’s failure to produce the raw data coupled with the other concerns noted above undermines the credibility of the findings,” he wrote. “I am deeply embarrassed by this turn of events and apologize to the editors, reviewers and readers of Science.”

For Fleischer, an enduring question will be: Why didn’t LaCour simply collect and use the data? After all, his organization was sending trained canvassers out regularly, and LaCour had access to their work. It would have been possible to measure their impact.

Another conundrum: LaCour’s data seemed to confirm what Fleischer had long believed, that the most long-lasting effects occur when the canvassers are gay and speak of their own experiences. Now, Fleischer said, it’s hard to say whether a gay canvasser is more effective than a straight one, or whether a woman who had an abortion is more persuasive on the subject than a man.

Fleischer said he remains convinced of the fundamental effectiveness of the approach, that face-to-face interaction can build tolerance and change people’s hearts.

“We think we’re on the right track with voter persuasion,” Fleischer said. But “we’ll never know for sure . . . until we have an honest, independent evaluation.”