Over the spring and summer of 2021, three behavioral scientists, acting on a tip, uncovered multiple instances of apparent data manipulation in the work of Francesca Gino, a celebrated Harvard Business School professor. Although the trio held regular academic posts—Joe Simmons, at Wharton; Leif Nelson, at Berkeley; and Uri Simonsohn, at Esade, in Barcelona—they occasionally moonlighted as a kind of informal internal-affairs bureau for the behavioral sciences, a discipline that had never done a particularly good job of policing its own research practices. Although they had come to enjoy some grudging respect, their probes did not earn them many friends. They saw themselves as decent guys—which, despite a measure of pugnacity, they are—and the Gino investigation afforded them an opportunity to demonstrate that their reputation for vigilantism was unfair.

In the past, their standard protocol had been to solicit comment from the subject of a critique, and then to publish their findings on their blog, Data Colada. This time, they decided, they would step aside in favor of a proper institutional process. This was something of a gamble. Universities had plenty of incentives to bury evidence of academic misconduct and allow the offender to slip quietly away. Gino’s status made this especially likely: although not quite a household name, she was a tenured, titled professor; a lively contributor to the TED-industrial complex; the author of two self-help guides for the aspiring entrepreneur and goal setter; and a consultant for companies such as Disney and Procter & Gamble. She was an ideal ambassador for the H.B.S. brand—confident, prolific, and sufficiently vague in her pronouncements that an executive could come away from one of her business-lite talks feeling affirmed in whatever previous beliefs he happened to entertain. Her extreme productivity, mostly untroubled by memorable ideas, was self-endorsing.

The nature of this investigation was unusual. Often, Data Colada’s findings were inferential—the trio could identify fishy data, even if they couldn’t quite reconstruct what had happened to it—but the results of the Gino audit seemed more difficult to question. Proof, for Harvard, would be relatively easy to secure. University administrators usually have access to researchers’ original data files, which, in some cases, can be at risk of becoming conveniently misplaced. A lot of Gino’s data had been collected with the online survey platform Qualtrics, and Harvard had only to compare the original versions with those which had been attached to the papers. In the estimation of the Data Colada team, this could be largely accomplished in the course of an afternoon. They compiled a single-spaced, eighteen-page dossier and dispatched it in confidence to an H.B.S. administrator.

On October 27, 2021, Gino received official notice of an inquiry, and was instructed to come to campus and turn over all “HBS-issued devices” by the end of the business day. Harvard’s investigation (for reasons that Data Colada could not fathom) lumbered on for the next eighteen months. I was in touch with the Data Colada members for much of this time, initially about a separate but related case, but they spoke to me only on the condition that I not pursue any additional reporting until the Gino inquiry had culminated in a final report. In the meantime, Gino continued to advise students, teach classes, and publish papers. Harvard had given her a handsome salary. (She owned a house in Cambridge worth almost four million dollars, on top of a two-and-a-half-million-dollar house next door, which she originally planned to demolish; she withdrew her application to the city’s historical commission about ten days after she was notified of the investigation.)

In June, 2023, she was placed on administrative leave, without pay or access to campus, for two years. Her named professorship got rescinded, and the H.B.S. dean sought to initiate the lengthy process required to revoke her tenure. Her colleagues were notified, and journal editors who had published her work were contacted about the various data infidelities, as a first step to retraction. In the weeks to follow, Data Colada rolled out a series of posts in which its members offered their own analysis of the papers in question. Three of the four papers were, in relatively short order, retracted. (The fourth had already been retracted following the discovery of a separate apparent fraud.) On LinkedIn, Gino posted that she was “limited into what I can say publicly,” but that “there will be more to come on all of this.”

A few weeks later, Gino filed a twenty-five-million-dollar lawsuit—for defamation, among other things—against Data Colada and Harvard. Gino’s initial complaint suggests that there were perfectly innocent explanations for all the alleged anomalies in the data sets and, furthermore, that if there were in fact no innocent explanations she was not the one holding the bag. (Gino has never wavered in her denial of any wrongdoing. In April, one of her lawyers noted that the idea that “she engaged in fraud is verifiably false.” ) As her attorney told the investigators, “In all four of the studies in question, Professor Gino had relied on the help of research assistants on any given project.”

Several months ago, despite the objections of Gino and her attorneys, Judge Myong J. Joun, of the Massachusetts district court, ruled that a redacted version of Harvard’s full investigation report be unsealed. (The New Yorker had filed a motion in support.) This was the result of something of an unforced error on Gino’s part. In various public forums, she had criticized the report’s procedures and conclusions. The university’s lawyers argued that her repeated public references to select portions of the report’s contents had given the court no choice but to make the entirety of it available; the judge agreed.

Perhaps the least interesting aspect of the report is that a protracted process—of interviews with Gino and her collaborators, and forensic analyses of dozens of files found in her work e-mails, on her hard drive, and in her Qualtrics account—not only vindicated but expanded on the claims made by Data Colada. None of the data interventions were subtle. On the first allegation: Harvard found that twenty-eight per cent of survey responses had been manipulated manually to support the hypothesis. On the second: dummy survey responses, absent from the original files, seemed to have spontaneously materialized in the final data set. In perhaps the most comically inept example of alleged misconduct, the apparently falsified data in a file had been deliberately highlighted in gray. In one case, the description of a study had been revised, over several successive drafts and in plain sight of the co-authors, seemingly to compensate for fatal flaws that had been pointed out in the original design. Gino’s defense against this allegation is that the research team hadn’t kept careful track of which version of the paper they were working on—in other words, that an entire group of élite business-school academics was, when it came to basic project management, functionally incompetent.

The report itself concedes that the investigation was, if anything, underpowered: after having found Gino guilty of “multiple instances of research misconduct” in four papers that spanned eight years and a passel of co-authors, “the Committee is concerned about other possible instances of research misconduct in Professor Gino’s studies.” The identification of additional examples would not have required heroic exertion. In April, Science magazine published an exposé of what appeared to be repeated instances of plagiarism in Gino’s work; the Montreal-based postdoc Erinn Acland, who, out of idle curiosity, had looked into the matter, found a case of uncredited verbatim reproduction in the first sentence of Gino’s that she happened to examine. In the introduction to Gino’s book “Rebel Talent: Why It Pays to Break the Rules at Work and in Life,” the author writes, “Rule breaking does not have to get us into trouble, if done correctly and in the right doses—in fact, it can help us get ahead.” About fifteen pages later, Gino seems to have helped herself along by lifting a banal description of Milan’s fashion district from an Anglophone travel Web site. (One of Gino’s lawyers responded that she is “steadfast in her commitment to uncovering the truth in each instance, responding decisively and correcting the record if necessary,” and that it isn’t fair to litigate such accusations in “the volatile domain of public opinion.”)

What’s most remarkable about the report is the emotional tenor. The interviews with Gino’s co-authors, who are rendered speechless by the evidence of betrayal, are painful to read. In the transcripts of her appearances before the committee, its members seem desperate to solicit from Gino some excuse or alibi that will make the whole sorry episode go away. They repeatedly ask her to produce any explanation of the anomalies that would allow the business school to send her back to her post with instructions to exercise a little more oversight in the future. Gino, for her part, seems unaware that the members of the committee will not remain her colleagues for long. She is consistently ingratiating, thanking them profusely for their diligence—“I know this is not part of the job, and so I am just very grateful that you paid so close attention to everything”—and swearing that from this point on she’ll “continue to evolve in my lab practices as necessary over time to ensure accuracy in my work.”

The questioning of Gino is polite and collegial, but the final report gives vent to an exasperation with her refusal either to admit wrongdoing or to mount a defense the committee can take seriously. Many of her responses—that, say, “she wasn’t placing a high priority” on the publication of a particular paper, or that she had abandoned work that didn’t pan out—are wholly irrelevant. Others, including that she never had access to the analyses provided by an outside forensic firm hired by Harvard, are inaccurate.

Her most robust explanation for the malfeasance is that she may have been set up by a conspiracy of former research assistants, resentful co-authors, and Data Colada. Eight pages of the report are devoted to this “malicious actor” theory. Any such party would have required access to both her Qualtrics account and her hard drive. The same party, or perhaps an additional one, would also have needed access to either the personal computer of a former research assistant or to the research assistant herself, who long ago left academia. In the latter case, the report continues, “they would have needed the ability to convince [the research assistant] to collude with them in falsifying data, and the ability to either instruct her in how to falsify the data or obtain the data from her, falsify it, and then return it to her before she forwarded it to us in May 2022 (accomplishing all of this in the relatively short timeframe—one week—between our request for [the research assistant’s] records from this study and her submission of those records).” It further notes that this would have required not just “great expertise” but the kind of perfect timing found only in “Ocean’s Eleven.” Gino’s insistence that she was framed, the report concludes with some bitterness, “leads us to doubt the credibility of her written and oral statements to this Committee more generally.”

Aside from an interview published in the Times last September, Gino has not spoken publicly to the Anglophone media. In April, however, she sat down with a reporter, whom she had known since high school, from the Torinese daily La Stampa. In this interview, she no longer flogs the malicious-actor theory; instead, she argues that the anomalous data in one of the four papers were due to spammy survey responses, and that the report was compromised by the inability of the forensic team to examine the correct files on her hard drive. These explanations are difficult to square with the evidence, and it seems more or less inconceivable that anyone could come away from the report with the notion that Gino had been subjected to punitive sham proceedings. Yet, according to Gino, it was in part at the recommendation of an unnamed Harvard colleague that she decided to file the twenty-five-million-dollar defamation lawsuit. The Italian reporter asks, “On the 26th of April there’s the first court hearing. The judge will decide which parts of the suit will be dismissed and which will go to trial. Why did you decide to sue?”

Gino replies, “The evening of June 13th, 2023, a Harvard professor was given instructions to tell me, ‘Hey, step aside, resign, it’s the best choice for you and your family.’ But she did the opposite. She bothered to read the twelve hundred and eighty-one pages and to ask herself some questions. She had her doubts, contacted me, and told me, ‘It seems like an injustice, you should resist, file a suit.’ That’s how it happened.” Gino doesn’t explain how this colleague obtained the report, which at the time almost no one had seen, although the idea that some emissary had been tasked with a diplomatic approach isn’t wholly far-fetched. (Harvard declined to comment on the facts of the case.) Still, the broader claim follows the pattern of her responses to the committee: the lawsuit, like the data manipulation, was apparently someone else’s fault.

At a Boston courthouse, on the day of the hearing, I squeezed into an elevator alongside a large group of suited figures, all of them preternaturally dour. I turned to one of them and asked, “You’re Harvard’s lawyers, right?” He didn’t look at me or smile. “We’re Harvard’s lawyers,” he said. The courtroom, a windowless box laced with a vaguely acorn-like pattern in an autumnal palette, was empty, aside from Data Colada and the trio’s lawyer, Jeffrey Pyle. The Data Colada guys aren’t accustomed to formality; both Joe Simmons and Uri Simonsohn looked as though they’d briefly glanced at a YouTube tutorial before giving up and just looping their ties decoratively around their necks. Leif Nelson, who seemed more familiar with neckwear, wore a blazer over weathered black jeans and cap-toe leather sneakers. Pyle, who appeared not only professionally but personally incensed by the fact of the lawsuit, wore cufflinks.

Gino filed in with a phalanx of counsel and sat alone, in the second row of the gallery. She looked wan in a loose cardigan over a greenish blouse, and took notes solemnly on a legal pad; her hair seemed slightly damp, as it tends to in her various TED-style appearances. Pyle’s oral arguments were short and to the point. When presented with evidence of purported fraud, he said, Gino “could have responded to this circumstance in the Marketplace of Ideas. . . . Instead, she has sued my clients for $25 million dollars, and for what? For identifying data anomalies to Harvard and writing publicly about those anomalies after, after Harvard confirmed that everything they said was right. The chilling effect of a lawsuit on science is obvious.” He continued, “A scientific disagreement isn’t a proper subject for a defamation claim. Courts and lawyers and juries . . . are ill-equipped to referee scientific controversies.”

For Data Colada’s purposes, one of the major questions to be resolved was Gino’s position as a public figure, in which case her lawyers would have to prove that Data Colada acted with malice or reckless disregard for the truth. Gino’s original complaint, unfortunately for her, went out of its way to establish that her reputation precedes her: “Plaintiff is an internationally renowned behavioral scientist, author, and teacher. She has written over 140 academic articles, both as an author and as a co-author, exploring the psychology of people’s decision-making.” Now, however, Gino’s lawyer, Julie A. Sacks, described all this as merely a description of a bog-standard academic. Judge Joun, who spent most of the hearing trying to stifle a yawn, seemed indisposed to this argument, interrupting to say, “We have a little bit more than that here. The complaint itself describes her as an internationally renowned scientist.” As she continued, Sacks did her client no favors. In her attempt to establish that Data Colada had published its blog posts with a reckless disregard for the truth, Sacks misquoted something that Simonsohn had said in a Webinar the previous summer. Pyle rose to correct the record.

In closing, Pyle referred to a recent case in the Southern District of New York, in which a group of scientists stood accused of defamation for the claim that there was something amiss in the F.D.A. trial data for an Alzheimer’s drug. The judge in that case had thrown out the defamation claim on the ground that scientific disagreements were scientific disagreements. When Sacks led an ashen Gino out of the courtroom, she didn’t turn, as is common, to shake the hands of any of the other lawyers. (Gino declined to comment on the facts of the case while it was ongoing.) But one of her lawyers has noted that the team was “pleased that Professor Gino’s story was shared in a venue where process is upheld and evaluated in a full and fair manner,” adding, “Professor Gino’s case is strong and the facts support her innocence against these unfounded claims. Professor Gino’s reputation has been tarnished by this unfair process and we hope this case can help set the record straight.”)

Last fall, I attended the high-profile fraud trial of Sam Bankman-Fried. Although that case, in the end, was legally simple—the jury found that Bankman-Fried had misappropriated customer funds and then repeatedly lied about it—it was, and remains, intellectually and morally complex. What role, for example, did Bankman-Fried’s commitment to utilitarianism play in his calculations? The Gino hearing—that of an extremely successful dishonesty researcher accused of having conducted her research dishonestly—augured something similarly invigorating. But, as it was, there appeared little to be learned aside from the fact that even once you take away someone’s shovel they will look down at their hole and use their bare hands to keep digging. The proceedings seemed terminally petty, an embarrassing dissipation of everyone’s time and resources. It was hard not to take the whole thing as evidence that a peerage of researchers had somehow been overcome by a cartel of litigants, a scientific culture easily derailed by the vanity and entitlement of the wealthy. Administrative bloat already threatens to turn the professoriate into a university rump; a case like this, in which one aggrieved actor plays to the refs of not merely the managerial bureaucracy but of the state itself, only further undermines the necessary defense of academic autonomy.

About two weeks after the hearing, the Data Colada team posted an update to their blog. They used, as an illustration, an A.I.-generated image in the sepia-toned style of a lurid mid-century courtroom drama: the three of them and their lawyer, now with fitted suits and ties knotted properly, solemnly enjoying their namesake fruity cocktails in front of a packed gallery. They hoped that the case would be dismissed. But, they continued, “if the case is not dismissed, then we go to discovery.” A lot of new data would emerge. In that eventuality, they concluded, with barely suppressed gaiety, “it is possible we would discover things that merit an additional blog post.”

The court, it turned out, was operating on an academic calendar, and the summer passed with no news. In the early afternoon of September 11th, the judge at last issued his decision. Several of the counts against Harvard itself, which related to the terms of Gino’s employment contract, would be allowed to proceed, and seem unlikely to be resolved anytime soon. (After the decision came out, one of Gino’s lawyers said, “Today’s decision clearly demonstrates Harvard treated Professor Gino differently from other misconduct investigations and their own stated policies. . . . We are pleased with the court’s decision to allow this litigation to continue and that Harvard will have to answer for how they have destroyed her career and put every member of the Harvard faculty at risk.” Harvard did not respond to a request for comment on the ruling.) The judge agreed, however, with Pyle’s arguments on behalf of Data Colada; he suggests that even a casual reading of their blog posts, which were never presented as anything but their reasonable opinion of the integrity of Gino’s work, rendered the defamation claims spurious. As the judge put it, “the sum effect of the blog posts makes plain that they represent the Data Colada Defendants’ subjective interpretation of the disclosed facts.” Science, in other words, could remain in the trust of scientists. All of the counts against them were dismissed. New cases, one can unfortunately presume, awaited them. ♦