Back in June I wrote about a Harvard researcher named Dr. Francesca Gino who had been put on leave after an investigation found evidence she may have manipulated data in several published papers. The possible fraud was made even more surprising by the fact that the papers in question were about the topic of honesty. Saturday the NY Times published a follow-up report based partly on a lengthy interview with Dr. Gino.

In conversation, Dr. Gino can come across as formal. The slight stiltedness of her nonnative English merges with the circumlocution of business-school lingo to produce phrases like “the most important aspect is to embrace a learning mind-set” and “I believe we’re going to move forward in a positive way.”

But she also exhibits a certain steeliness. “I am a well-organized person — I get things done,” she told me at one point. She added: “It can take forever to publish papers. What’s in my control, I execute at my pace, my rigor.”

The article makes it clear that Dr. Gino was part of a broader trend. Two decades ago there was a lot of interest in behavioral science and one figure won a Nobel Prize for his work. This was followed by several bestselling books about the field. People became academic stars and had opportunities to give speeches and TED talks. One of Dr. Gino’s collaborators was a leading figure in this field:

Dr. Gino and Dr. Ariely became frequent co-authors, writing more than 10 papers together over the next six years. The particular academic interest they shared was a relatively new one for Dr. Gino: dishonesty.

While the papers she wrote with Dr. Ariely were only a portion of her prodigious output, many made a splash. One found that people tend to emulate cheating by other members of their social group — that cheating can, in effect, be contagious — and another posited that creative people tend to be more dishonest. In all, four of her six most cited papers were written with Dr. Ariely, out of more than 100.

Papers like these helped Dr. Gino become a rising star and she moved to Harvard more than a decade ago. In 2018 she published her own mass market book (“Rebel Talent: Why It Pays to Break the Rules at Work and in Life”) but even as her own profile was rising, the view of her field was under scrutiny.

It’s often difficult to identify the moment when an intellectual movement jumps the shark and becomes an intellectual fad — or, worse, self-parody.

But in behavioral science, many scholars point to an article published in a mainstream psychology journal in 2011 claiming evidence of precognition — that is, the ability to sense the future. In one experiment, the paper’s author, an emeritus professor at Cornell, found that more than half the time participants correctly guessed where an erotic picture would show up on a computer screen before it appeared. He referred to the approach as “time-reversing” certain psychological effects.

It was papers like that one that got three academics to team up and start a blog in which they looked at data problems within the field. They called their blog Data Colada and it wasn’t long before they were uncovering evidence of data manipulation in some of the papers they looked at. When they first uncovered problems in some of Dr. Gino’s papers, the bloggers notified Harvard which asked them not to publish anything until the university could investigate the situation internally.

Harvard did investigate and even created a new, more aggressive investigation process which was used for the first time in Dr. Gino’s case. She was locked out of her email and had to turn in her school laptop. She has not been allowed back on campus while the investigation is ongoing. Meanwhile, once it became clear there was a potential problem, the bloggers published a series of posts with their findings, all of which suggest data in at least four studies was manipulated.

Dr. Gino has responded by suing Harvard and the bloggers who accused her for $25 million, claiming they defamed her. For their part, the bloggers published a new post just two weeks ago highlighting new information revealed in the lawsuit itself which (they claim) provides even more evidence of fraud.

Gino’s lawsuit (.htm), like many lawsuits, contains a number of Exhibits that present information relevant to the case. For example, the lawsuit contains some Exhibits that present the original four posts that we wrote earlier this summer. The current post concerns three other Exhibits from her lawsuit (Exhibits 3, 4, and 5), which are the retraction requests that Harvard sent to the journals. Those requests were extremely detailed, and report specifics of what Harvard’s investigators found.

Specifically, those Exhibits contain excerpts of a report by an “independent forensic firm” hired by Harvard. The firm compared data used to produce the results in the published papers to earlier versions of data files for the same studies. According to the Exhibits, those earlier data files were either (1) “original Qualtrics datasets”, (2) “earlier versions of the data”, or (3) “provided by [a] research assistant (RA) aiding with the study”. We have never had access to those earlier files, nor have we ever had access to the report that Harvard wrote summarizing the results of their investigation.

For all four studies, the forensic firm found consequential differences between the earlier and final versions of the data files, such that the final versions exhibited stronger effects in the hypothesized direction than did the earlier versions. The Exhibits contain some of the firm’s detailed evidence of the specific discrepancies between earlier and published versions of the datasets.

Dr. Gino still has supporters including people who feel the investigation by Harvard has been unfairly aggressive. As the NY Times points out, there are lots of other academics in the field who have published papers with similarly surprising findings but, for the most part, they aren’t being investigated with the same amount of focus. There may be many more papers that need to be retracted that haven’t been yet.

Update: The top comment on this story is a pretty good summary of the case against Dr. Gino.

I work in Gino’s field and have looked at the evidence in great depth. Let’s be clear about what happened:
1) The paper with Gino, Ariely and three other coauthors was retracted years ago. The other coauthors all agreed it had serious, possibly fraudulent problems – in psychology, a paper often involves a number of studies and an individual co-author does not monitor each one.
2) Two years ago, Data Colada found very suspicious patterns in four separate papers by Gino, including the previous one, specifically on studies she and her students were involved with. The evidence is circumstantial but in my view incredibly suspicious – in each of these papers there is odd looking data that just happens to massively support the main hypothesis.
3) Rather than punish, Data Colada informed Harvard. To protect Gino, Harvard asked DC not to write anything until they could do an investigation.
4) Harvard, with access to Gino’s computer, was able to see the earliest versions of data files that were eventually used in the published papers. The exact data points that DC thought were suspicious turn out to have different values in the earlier data files, and almost always values that do not support the paper’s hypothesis. Recall: 4 separate papers.
5) All coauthors agree the suspicious studies were ones Gino oversaw. The research assistants are different in each case.
6) Harvard sent evidence to the journals and the papers are either retracted or about to be. Only then was Gino put on leave.

And one more:

I do data analytics for a living and if you read the Data Colada articles, Professor Gino is either the unluckiest researcher in the history of the universe or she chronically fabricated data. I read her lawsuit and while I’m sure there’s truth in her claims about sexism and bias when HBS changed their investigative policy based on her case, it’s not lost on me that her priors about people’s tendency to cheat when (a) they have an incentive and (b) believe they can get away with it appears to be driven more from Gino’s personal psyche than how people actually behave. The other little problem with her lawsuit is she makes no claims that the actual statistical analysis is valid.

My only quip is the NYT’s seemed to go overboard presenting the case she may be innocent. When five (I’m assuming) tenured professors say they believe she’s innocent but none will go on the record, the once in the history of the universe chance she’s actually innocent goes to zero.