When the dust finally settled on the Diederik Stapel case in 2012, the investigating committee had reviewed all 137 of the social psychologist’s publications, finding fraud in 55 of them, indications of misconduct in 10 more, and that 10 dissertations he supervised contained fraudulent data he provided. In two of the dissertations, every chapter of the student’s work was based on data that Stapel had simply made up.
In their report, the committee noted two primary motivations for such an extensive review: correcting the scientific record, and protecting innocent scientists, especially those early in their careers.
“The scientific literature must be cleansed of everything that is fraudulent, especially if it involves the work of a leading academic,” the committee wrote. “No more time and money must be wasted on replications or meta-analyses of fabricated data. Researchers’ and especially students’ too rosy view of the discipline, caused by such publications, should be corrected.”
Stapel’s modus operandi was creating fictitious datasets or tampering with existing ones that he would then “analyze” himself, or pass along to other scientists, including graduate students, as if they were real.
“When the fraud was first discovered, limiting the harm it caused for the victims was a matter of urgency,” the committee said. “This was particularly the case for Mr. Stapel’s former Ph.D. students and postdoctoral researchers, whose publications were suddenly becoming worthless.”
Why revisit the decade-old case of Stapel now?
Because its echoes can be heard in the unfolding case of Harvard Business School Professor Francesca Gino as she faces allegations of data fraud, and her coauthors, colleagues, and the broader scientific community figure out how to respond. Listening to these echoes, especially those of the Stapel committee, helps put the Gino situation, and the efforts to remedy it, in greater perspective.
The Many Co-Authors project has the potential to bring clarity to the allegations of fraud, possible revocation of tenure, retracted papers, defamation lawsuit, and crowdfunding campaign for three scientists’ legal defense.
Particularly resonant is the Many Co-Authors project, a new initiative led by a group of Gino’s coauthors, who are coordinating a review of Gino-led studies to help address concerns surrounding the body of Gino-coauthored work. These researchers are driven by the same two concerns as the Stapel committee, correcting the scientific record and protecting early career scientists, and have developed an online platform to help do so.
The Many Co-Authors project has the potential to bring clarity to the allegations of fraud, possible revocation of tenure, retracted papers, defamation lawsuit, and crowdfunding campaign for three scientists’ legal defense. And it has the potential to do so quickly.
To appreciate the clarity that the Many Co-Authors project could bring to the situation, it is important to understand how it began, just how messy it has become, and how much clean up is required.
A brief background
The evidence of potential fraud in Gino-led studies emerged through the work of three scientists—Uri Simonsohn of the Esade Business School, Joseph Simmons of the University of Pennsylvania, and Leif Nelson of the University of California, Berkeley—who, along with a set of anonymous researchers, first alerted Harvard to the alleged fraud in 2021, and then made their analyses public this June on their blog Data Colada. Across a series of four posts, they detail alleged data fraud in four separate studies led by Gino. (For more background on the specific allegations, see my previous article here.)
“We believe that many more Gino-authored papers contain fake data,” they wrote. “Perhaps dozens.”
Gino remained largely silent after the allegations surfaced, only putting out a brief statement in June saying, “There will be more to come on all of this.” (Disclosure: Gino has served as an advisor to Behavioral Scientist in the past.)
By that time, Harvard had placed Gino on administrative leave without pay based on the university’s own investigation into possible research misconduct. And shortly after, all four publications that contained the questionable Gino-led studies had been retracted or were in the process of being retracted by their respective journals.
Then, the story moved from the lab into the courtroom.
In August, Gino put out a longer statement denying the allegations and announcing a $25 million lawsuit against Harvard, Harvard dean Srikant Datar, and Simonsohn, Simmons, and Nelson of Data Colada.
“I want to be very clear: I have never, ever falsified data or engaged in research misconduct of any kind,” she wrote on Linkedin. “I had no choice but to file a lawsuit against Harvard University and members of the Data Colada group, who worked together to destroy my career and reputation despite admitting they have no evidence proving their allegations.”
In the lawsuit, she claims the three Data Colada scientists made false and defamatory statements about her and her work, which the lawsuit refers to as a “vicious, defamatory smear campaign.” Harvard, she argues, based their findings on “speculation,” and failed to follow its own protocol for handling potential research misconduct, singling her out.
As dramatic as it all may be, ultimately, the case is a disruption to the scientific process and scientific community.
I reached out to Gino for comment, and her legal team replied with a statement from her lawyer, Andrew T. Miltenberg.
“This lawsuit is not an indictment on Data Colada’s mission, nor efforts to replicate and validate data with the intent of upholding a standard of research excellence,” his statement read.
“What the lawsuit demonstrates is how both Harvard and Data Colada deviated from their stated processes and treated Prof. Gino’s investigation differently from other misconduct investigations, and the burden of proof required to decimate a career and revoke tenure must be based on evidence, not theory. Anything less is dangerous precedent for all academic researchers.”
Unsurprisingly, Harvard’s account differs from that of Gino and her legal team’s. In a recent email to Harvard Business School (HBS) faculty, Datar explained why Harvard had placed Gino on leave and sought retractions for the work in question.
“After a comprehensive evaluation that took 18 months from start to completion, the investigation committee—comprising three senior HBS colleagues—determined that research misconduct had occurred,” his email said. “After reviewing their detailed report carefully, I could come to no other conclusion, and I accepted their findings.”
He added: “I ultimately accepted the investigation committee’s recommended sanctions, which included immediately placing Professor Gino on administrative leave and correcting the scientific record.”
While it is unclear how the lawsuit will play out, many scientists have expressed concern about the chilling effects it might have on scientists’ willingness to come forward if they suspect research misconduct.
“If the data are not fraudulent, you ought to be able to show that. If they are, but the fraud was done by someone else, name the person. Suing individual researchers for tens of millions of dollars is a brazen attempt to silence legitimate scientific criticism,” psychologist Yoel Inbar commented on Gino’s statement on Linkedin.
It is this sentiment that led 13 behavioral scientists (some of whom have coauthored with Gino) to create a GoFundMe campaign on behalf of Simonsohn, Simmons, and Nelson to help raise money for their legal defense.
According to the GoFundMe campaign, estimated legal costs for the three scientists could range from $50,000 to $600,000 depending on how the case proceeds. The GoFundMe page reports that Simonsohn’s, Simmons’s, and Nelson’s universities have only agreed to cover a portion of the expenses. As of writing, the campaign had raised over $310,000 against a target of $250,000.
A dramatic disruption
As dramatic as it all may be, ultimately, the case is a disruption to the scientific process and scientific community. But before both can move forward, scientists need to know what in the scientific record needs to be cleaned up, and perhaps, if they have worked with Gino, what cleaning up they might need to do themselves.
The scientific process “is seriously disrupted by interference from fraudulent data and findings based on questionable methodology,” observed the Stapel committee in their final report a decade ago. “The scientific researchers and institutions involved are duty bound to call a halt to this disruption.”
Harvard’s Datar would seem to agree, calling the need to correct the scientific record “a measure incumbent on every responsible academic institution.”
Yet Datar and Harvard, despite whistling the Stapel committee’s tune, haven’t fully backed up their stated commitment to the scientific record by making their findings public, and they are unlikely to do so. By contrast, the three universities where Stapel had been a scientist, including the University of Tilburg where he was a professor when his fraud was discovered, prioritized full disclosure from the start of their investigation. (A Harvard spokesperson declined to comment due to the pending litigation.)
This leaves the scientific community caught in a climate of chilling litigation and institutional hot air, no closer to resolving the questions surrounding the body of Gino’s coauthored work.
The newly formed Many Co-Authors project offers a productive way forward, and possibly a model for how the scientific community can address instances of possible fraud in the future.
Removing doubt, gaining clarity, quickly
At its most basic level, the Many Co-Authors project is a centralized database where Gino’s coauthors can disclose if Gino collected the data in the studies they collaborated on. It will also provide coauthors with a chance to report if they have access to any raw data provided by Gino, if they can replicate the findings, and whether a retraction might be needed. The database of coauthors’ reports will be online and open-access.
Uri Simonsohn, one of the members of Data Colada, a defendant in Gino’s lawsuit, and a Gino coauthor, conceived of the idea shortly before Data Colada published its first post in June.
“I knew in recent years, because I happen to know some of her junior coauthors, that she didn’t do any of the data collection, and they would be severely impacted if their work was considered to be as illegitimate as any other paper of hers,” Simonsohn told me in late July.
For Simonsohn, one of the most important aims of the project is “protecting really early career [scientists], grad students or people who just graduated, from having their career ruined by basically all of their work being cast into doubt,” he said.
The scientific community is caught in a climate of chilling litigation and institutional hot air, no closer to resolving the questions surrounding the body of Gino’s coauthored work.
Early on he wasn’t exactly sure what the platform would look like, just that something needed to be done. “I started calling people and inviting them to the project, telling them, Look, it’s going to be a work in progress, we’re going to figure it out as we go.”
Eventually, a leadership team formed, made up of a half-dozen of Gino’s coauthors, including Juliana Schroeder, a professor at the University of California, Berkeley.
The project made sense to Schroeder. “Four papers have been retracted. There was fraudulent data found in those papers. They share the common coauthor of Francesca Gino. Therefore, I think it’s very reasonable to go ahead and audit the rest of the Gino coauthored papers,” Schroeder told me in early August.
In addition to Simonsohn and Schroeder, Max Bazerman and Julia Minson of Harvard, Don Moore of Berkeley, and Maurice Schweitzer of the University of Pennsylvania, joined the leadership team. (Though Simonsohn is involved in both this project and Data Colada, he explained that this is not a Data Colada project.)
The initiative took shape over the course of only a few weeks in June and July; lightspeed in academic time. Simonsohn built the online reporting platform in about two weeks. Then, the leadership team along with other coauthors spent a week and a half planning, drafting, negotiating, and finalizing what information all 150 of Gino’s coauthors would be invited to report. It was a process that unfolded over long email threads and zoom calls.
What emerged was a platform that has the chance to bring clarity and transparency to the case.
What will coauthors report?
Over a Zoom call in July, Simonsohn walked me through the backend of the Many Co-Authors platform and explained what information coauthors can submit.
A coauthor’s report consists of three parts: a data review, a question about the possibility of retraction, and a “living document” where they have an open text box to provide any additional thoughts about the specific study that they want to communicate to readers.
“What we’re hoping to do through this initiative is identify which findings are reliable, whether there’s reason to believe that any of the data might be shaky, and basically give people more confidence when they’re thinking about citing a Gino coauthored paper,” Schroeder said.
In July, all 150 coauthors were invited to participate in the project, and the results of the coauthor-led reviews are set to be made public on the Many Co-Authors website, as early as mid-October*. Update 9/30/23: The Many Co-Authors leadership team has updated their timeline from going public from September 1, then October 1, to mid-October.
The coauthors’ data review involves only studies where Gino played a role in the data collection, and coauthors can complete a report for each paper they coauthored with Gino. So for a multi-study paper, only those in which Gino provided the data are up for review. The reason for this was to avoid undue scrutiny on coauthors work, which could be perceived as a sort of double victimization on top of an association to the alleged fraud, both Simonsohn and Schroeder told me.
When they begin the data review, coauthors are first asked to report whether Gino was involved in the data collection. They can answer: Yes, No, or Don’t Know. If the answer is No, then that’s the end of the review. If the answer is Yes, then they are also asked to report the approximate date that the study occurred, which could help identify the study’s raw data if further investigation is needed.
Next, coauthors report whether they ever had access to the raw data file, to which they can respond: Never, Don’t know if ever had, Yes but not anymore, and Yes and still do.
The final question of the data review is: Do you have the data necessary to reproduce the published results? Here, coauthors can answer: Not sure, No, Yes (but can’t be posted online for privacy reasons), and Yes (can be posted online).
A coauthor’s report consists of three parts: a data review, a question about the possibility of retraction, and a “living document” where they can provide any additional thoughts about the specific study that they want to communicate to readers.
After completing the data review, coauthors are asked about the possibility of retracting the article, to which they can enter: Have not decided yet, Do not plan to retract the article, Will request/have already requested to retract the article, or It has already been retracted.
Finally, coauthors have the opportunity to provide additional context for their responses in a living document, which is designed to provide authors with a space to communicate to readers anything they want that’s not captured in the multiple choice questions.
“Lots of co authors are going to have different ways that they want to deal with this and different things they want to say about the papers,” Schroeder said. The living document is the space where they can do that.
Taken together, the Many Co-Authors team believes this information will provide greater clarity and transparency around Gino-coauthored publications. Update 9/30/23: The Many Co-Authors project posted screenshots of the authors’ questionnaire and can be viewed here.
“Readers will have more immediate information about what to believe about that paper, and they can immediately update their beliefs,” Schroeder said. “Whether they think, Okay, some of these studies may or may not be trustworthy, or there’s no reason not to trust this paper.”
Both Simonsohn and Schroeder told me that the Many Co-Authors team considered capturing additional information, but balancing so many different coauthor perspectives, the need for high quality information, and the time constraint, they felt this was a worthwhile start.
When the site goes public, they don’t expect to have all the coauthors’ reviews submitted but hope to have the majority, with additional reviews coming in over the course of the fall.
“I think this is something that all of the coauthors probably want to do, because everybody wants to feel confident in the finding,” Schroeder said, “both the people who wrote the papers and the people who are reading them.”
What will the Many Co-Authors’ report reveal?
The Many Co-Authors project has the potential to answer a number of critical questions: How frequently did Gino alone provide the data for her coauthored work? How many coauthors report finding no issues versus indications of fraud? Will any additional publications be retracted? Will any established findings be called into question?
Is there a chance these reviews absolve Gino? Or, that a pattern of fraud emerges that points to someone other than her? What impact will this new information have on the viability of the lawsuit?
And, with fresh eyes on old projects, will other problems come to light? Might coauthors find revisiting previous work valuable in its own right?
In the absence of Harvard making its investigation public, the Many Co-Authors project offers the most viable path toward understanding the trustworthiness of Gino-coauthored papers and answering these broader questions. But even if Harvard did release its report, the Many Co-Authors project would provide a critical resource for good scientific hygiene going forward.
For instance, although the Stapel committee was exceptionally thorough, some of Stapel’s work still lives in a sort of scientific no-man’s-land.
“It bothers me that … I’ll find a paper out there by Stapel, and I don’t know, why is it not retracted?” Simonsohn said. “Did somebody decide this was legit? Did they never check? Or did they check, but the journal refused to retract? Did an author get in the way? There’s so many scenarios.”
He added: “That seems like a really bad outcome. You go through all this effort, and, at the end, you end with this stuff in limbo.”
The Many Co-Authors platform could help clear up uncertainties like these and give more confidence to those considering building off of the research, as it provides a live, centralized, open-access resource to better understand the context of specific publications, from the people closest to the work. Any time someone wants to better understand a Gino-coauthored paper, they can review the coauthors’ report on the Many Co-Authors site.
The Many Co-Authors platform provides a live, centralized, open-access resource to better understand the context of specific publications, from the people closest to the work.
In Simonsohn’s own review of the 2013 paper he and Gino wrote together, he came to the conclusion that data were legitimate, but wrote in his living document that he still planned to do several additional analyses.
In Schroeder’s case, she coauthored seven papers with Gino. “I will be posting my very detailed audits in the living documents,” she said. Based on those audits, she and her coauthors had already come up with a plan for each paper if further action needs to be taken.
If coauthors do discover anomalies or fraud, it is up to them to decide what steps to take. The Many Co-Authors project serves only as the platform. There are so many possible scenarios and moving parts for every paper, and every journal and editor has its own policies, that standardizing a response would neither make sense nor be possible. However, the Many Co-Authors team might help develop guidelines for authors on the possible steps they could take to correct the scientific record, if it is needed.
Looking ahead, Simonsohn and the Many Co-Authors team hopes to summarize the coauthors’ findings in a paper, along with lessons learned from the process of the project.
“To me, the end goal is making social science more trustworthy, long term,” Simonsohn said.
Restoring trust
“Is science’s purported self-cleansing ability really up to the challenge of dealing with such a serious infringement of scientific integrity?” the Stapel committee asked of themselves.
A similar question must now be asked of those who have the power to correct what might need to be corrected in this case.
Suspecting a colleague of misconduct is likely one of the worst things a scientist can experience, and confronting the possibility of fraud will not be easy for Gino’s coauthors. They know her as a colleague, a mentor, a friend.
It was Stapel’s colleague and friend, Marcel Zeelenberg, whom he had known for nearly 20 years, who had to confront Stapel with the fraud allegations. A conversation they had in the living room of Zeelenberg’s home.
“When this is all over, I would like to talk to him,” Zeelenberg told The New York Times in 2013. “Then I’ll find out if he and I are capable of having a friendship.”
Like friendship, science is built on trust, and that trust is damaged. Many people have separated themselves into the Gino camp or the Data Colada camp, and it is hard to imagine what could move them back together, then forward, except for a thorough analysis of Gino-led work by those who are closest to it. The Many Co-Authors project will not answer every question, but it has the potential to offer everyone paying attention a clearer view of the situation than they have now.
The Many Co-Authors project offers a productive way forward, and possibly a model for how the scientific community can address instances of possible fraud in the future.
“Ultimately, trust forms the basis of all scientific collaboration,” the Stapel committee wrote. “If, as in the case of Mr Stapel, there is a serious breach of that trust, the very foundations of science are undermined.”
Very human scientists—with motives, emotions, reputations, and relationships—are the ones who can violate this trust and the ones who can restore it.
Two years ago, after Data Colada discovered fraudulent data in one of her coauthor’s studies, Gino wrote: “The work they do takes talent and courage and vastly improves our research field. I am grateful for their efforts.”
Don Moore, a Gino coauthor and member of the Many Co-Authors leadership team, would agree. Alongside his donation to Data Colada’s legal defense, he wrote:
“Science is self-correcting when scientists correct it.”
*Update 9/30/23: The Many Co-Authors leadership team has updated their timeline from going public from September 1, then October 1, to mid-October.
