The more we learn about Francesca Gino’s lawsuit, the more problems that arise
Last summer, I wrote about Harvard Business School professor Francesca Gino, a famous “dishonesty researcher” who was suspended from Harvard after it was revealed that someone had manipulated the data in four papers she co-authored over the course of a decade. In response, Gino — who has maintained her innocence — sued Harvard and the bloggers who first published the allegations, claiming she’d been defamed.
I’ve written about how her lawsuit will have negative consequences for the scientific community, stifling critics and making it even less likely that research fraud, when it happens, will come to light. But there’s one respect in which the lawsuit has been enormously valuable for casting light on the allegations against Gino.
Harvard’s lengthy internal investigation of the allegations was not originally released to the public. Last week, a Massachusetts judge ordered it unsealed, granting the world a look at Harvard’s process as they investigated the possibility that one of their star researchers had fabricated her data — and a look at Gino’s defense.
Overall, the document makes the allegations of Gino’s misconduct look more warranted than ever.
But that still may not be enough to prevent an expensive legal battle. Given that it is increasingly evident that scientific fraud is nowhere near as rare as anyone hoped, we desperately need better processes for identifying it. We also need a better means of protecting the people who stick their necks out to bring it to light.
What Harvard found
The Gino saga turns on four papers — published in 2012, 2014, 2015, and 2020. In each of them, independent data detectives found telltale signs of manipulation: rows that had been altered or inserted in the data, and which conveniently produced dramatic effects supporting her hypotheses. The data detectives alerted Harvard of their concerns.
Harvard commissioned an independent investigation into the allegations. The resulting report is almost 1,300 pages — longer than Infinite Jest — though that doesn’t mean it’s dull. It details how the research team systematically determined that, for each of the four papers, the data was indeed manipulated. And it includes Gino’s theories of who did it and why.
One explanation Gino offered Harvard for the manipulated data stands out: her theory that the data was manipulated by an academic rival of Gino’s seeking to take revenge over a disagreement.
If that sounds far-fetched, well, the report’s authors agreed. The major problem is that while there are many people who could have manipulated the data for any one of the studies, the only common denominator across all of them — over eight years — was Gino.
“In order to falsify data across all four studies’ records,” the report observes, “actors with malicious intentions would have needed the following: First, they would have needed access to both Professor Gino’s Qualtrics accounts and her computer’s hard drive, as two allegations (1 and 2) involve discrepancies in Qualtrics data and one allegation (3) involves discrepancies in the computer’s data.”
Then they’d also have needed a co-conspirator in order to falsify the data associated with allegation 4. Then they’d also have needed deep familiarity with how Gino stored and labeled data on her computer and her planned timetable for each study, just to maliciously alter the study to make her look bad.
Then, having carefully engineered this evidence of data misconduct, they would need to sit still for years before revealing the evidence of falsification.
It’s a stretch, to put it mildly.
Could this possibly be defamation?
“Defamation,” defamation lawyer Ken White told me last summer, “is an unprivileged false factual statement about someone which causes harm. Where a lot of the action comes in is determining what is a provably false factual statement.”
The Harvard report makes it abundantly clear that Gino will have an uphill battle convincing any court that the Data Colada bloggers made a provably false factual statement. For one thing, the data was clearly manipulated.
So far, her primary argument for why she didn’t commit fraud is, basically, that maybe she was carefully and elaborately framed for it by someone with access to her computer and her logins over the course of eight years.
That’s a fundamentally unserious theory of this situation, and Harvard’s analysis commission was correct to conclude it was highly unlikely. The only reason we’re obliged to take it seriously is that Gino’s lawsuit is a serious matter, whether the underlying claims are reasonable or unreasonable.
“The system is so broken that being sued for defamation in a case like this will cost hundreds of thousands of dollars and go on for years,” White told me earlier. “Realistically, you could wind up going to trial. Even if you’re going to win at trial, eventually you’re going to be ruined doing it.”
Gino doesn’t need to win her lawsuit to have a devastatingly chilling effect on independent experts searching for fraud. She doesn’t even need to propose a credible theory of how the data manipulation could have happened without her involvement. It doesn’t matter if her explanation strains credulity. “The process is the punishment,” as White put it.
That’s a huge problem because scientific fraud is a huge problem. Between the dishonesty researchers who have one by one turned out to be dishonest and the cancer research that turned out to be reusing Photoshopped versions of the same test result pictures, the last few years have been full of discomfiting reminders that, yes, some people will cheat to get ahead in science, and we lack a robust process for catching them.
Scientific integrity currently depends on the willingness of individuals to speak out when they see fraud, and it’s precisely that willingness Gino’s lawsuit targets.
A version of this story originally appeared in the Future Perfect newsletter. Sign up here!
