[Last updated (5/28/2015, 7:58 p.m.) with new information from an Emory University professor about Mr. LaCour’s work.]
It looks as if another paper by Michael J. LaCour may be bogus.
On Thursday afternoon Gregory J. Martin, an assistant professor of political science at Emory University, published a paper on the university’s website that identifies multiple serious problems with a paper by Mr. LaCour on the effect of bias in the news media.
In that paper, Mr. LaCour, the UCLA graduate student who is under fire for apparently faking the data for another paper, published in Science, examined whether people watch or listen to only programming that aligns with their political leanings. Mr. LaCour came to a contrarian conclusion: The “vast majority of partisans consume predominately centrist newscasts.”
Mr. Martin was intrigued by the paper, titled “The Echo Chambers Are Empty: Direct Evidence of Balanced, Not Biased, Exposure to Mass Media,” when he learned of it last summer. He set about trying to replicate it. Mr. Martin emailed Mr. LaCour, who he says provided him with the information he needed. But when Mr. Martin tried to run the numbers, “it didn’t work.”
He assumed that the data set Mr. LaCour was using must be different, perhaps better, than the one he had used. The thought that Mr. LaCour had perpetrated a fraud did cross Mr. Martin’s mind, though. “I couldn’t match what he had, and I chalked it up to the difference in the data set,” he said. “I was kind of skeptical of it, but I didn’t take the time to check it out.”
But when news of the apparently faked Science paper broke, Mr. Martin gained access to the same database Mr. LaCour used for his study. He ran the numbers again. Again, he didn’t get the same results. “I don’t know how he could have gotten those figures,” Mr. Martin said. “I don’t think it’s possible to do, using the data and the method he describes.”
He noticed other problems too. Mr. LaCour lists a slew of television and radio shows that he examined for evidence of political bias. But some of the shows don’t appear in the database he said he had used, according to Mr. Martin. Some of the radio shows don’t appear to have available transcripts anywhere.
One example: Mr. LaCour writes that he looked at transcripts of shows from 2006. But the transcripts for the show Real Time With Bill Maher, which is included in Mr. LaCour’s data, don’t start appearing in the database until 2008.
Mr. Martin’s paper is the latest bad news for Mr. LaCour, a scholar whose reputation has taken a severe beating. In little more than a week, Mr. LaCour has gone from being celebrated as a co-author of the widely touted Science paper to becoming notorious for apparently fabricating the data that the paper was based on.
He has promised to explain himself by Friday, though at this point the evidence against him — grants that don’t exist, a survey that was apparently never performed — seems hard to refute.
Even before Mr. Martin’s analysis, other evidence had raised the question: Was there a broader pattern of deceit here?
New York magazine’s The Science of Us blog caught Mr. LaCour trying to scrub his CV of a teaching award that appears not to exist. Mr. LaCour recently uploaded a new version of his CV that eliminated information about hundreds of thousands of dollars in grants he had previously claimed.
There were other small changes, too. The older version of his CV said that he graduated magna cum laude from the University of Texas at Austin in 2009. The new version excised that honor. (Mr. LaCour did not respond to an email asking about the change. A spokesman for the University of Texas said that such information could be released only to the student.)
The LaCour case is shaping up as one that will be talked about for years to come — unless his forthcoming statement somehow clears things up. But as the blogger Neuroskeptic put it: “He’s going to need a bloody good rebuttal.”
The journal Science on Thursday retracted the paper that set off the firestorm, citing three factors: the misrepresentation of survey incentives, the false sponsorship statement, and Mr. LaCour’s inability to produce original data. The journal noted that Mr. LaCour “does not agree to this retraction.”
Mr. Martin said he had hesitated to publish his analysis of Mr. LaCour’s work for fear of “beating a dead horse.” He decided to go ahead because he wanted it on the record that this paper was “probably not legitimate” either. “A lot of people were talking about the LaCour data,” Mr. Martin said. “He was certainly seen as a star in the field.”