What we do right and wrong when it comes to scientific fraud
Harvard is in the news right now for its war with the Trump administration, which has sent a series of escalating demands to the university, canceled billions of dollars in federal grants and contracts, and is now blocking the university from enrolling international students, all in an apparent attempt to force the university to conform to MAGA's ideological demands.
Stripping a celebrity professor of tenure might not seem like the best look at a moment when Harvard is in an existential struggle for its right to exist as an independent academic institution. But the Gino situation, which long predates the conflict with Trump, shouldn't be interpreted solely through the lens of that fight.
Scientific fraud is a real problem, one that is chillingly common across academia. But far from putting the university in a bad light, Harvard's handling of the Gino case has actually been unusually good, even though it still underscores just how much further academia has to go to ensure scientific fraud becomes rare and is reliably caught and punished.
There are two parts to fraud response: catching it and punishing it.
Academia clearly isn't very good at the first part. The peer-review process that all meaningful research undergoes tends to start from the default assumption that data in a reviewed paper is real, and instead focuses on whether the paper represents a meaningful advance and is correctly positioned with respect to other research. Almost no reviewer is going back to check to see if what is described in a paper actually happened.
Fraud, therefore, is often caught only when other researchers actively try to replicate a result or take a close look at the data. Science watchdogs who find these fraud cases tell me that we need a strong expectation that data be made public — which makes it much harder to fake — as well as a scientific culture that embraces replications. (Given the premiums journals put on novelty in research and the supreme importance of publishing for academic careers, there's been little motivation for scientists to pursue replication.).
It is these watchdogs, not anyone at Harvard or in the peer-review process, who caught the discrepancies that ultimately sunk Gino.
Crime and no punishment
Even when fraud is caught, academia too often fails to properly punish it.
When third-party investigators bring a concern to the attention of a university, it's been unusual for the responsible party to actually face consequences. One of Gino's co-authors on one of the retracted papers was Dan Ariely, a star professor of psychology and behavioral economics at Duke University. He, too, has been credibly accused of falsifying data: For example, he published one study that he claimed took place at UCLA with the assistance of researcher Aimee Drolet Rossi. But UCLA says the study didn't happen there, and Rossi says she did not participate in it.
In a past case, he claimed on a podcast to have gotten data from the insurance company Delta Dental, which the company says it did not collect. In another case, an investigation by Duke reportedly found that data from a paper he co-authored with Gino had been falsified, but that there was no evidence Ariely had used fake data knowingly.
Frankly, I don't buy this. Maybe an unlucky professor might once end up using data that was faked without their knowledge. But if it happens again, I'm not willing to credit bad luck, and at some point, a professor who keeps "accidentally" using falsified or nonexistent data should be out of a job even if we can't prove it was no accident. But Ariely, who has maintained his innocence, is still at Duke.
Or take Olivier Voinnet, a plant biologist who had multiple papers conclusively demonstrated to contain image manipulation. He was found guilty of misconduct and suspended for two years. It's hard to imagine a higher scientific sin than faking and manipulating data. If you can't lose your job for that, the message to young scientists is inevitably that fraud isn't really that serious.
What it means to take fraud seriously
Gino's loss of tenure, which is one of a few recent cases where misconduct has had major career consequences, might be a sign that the tides are changing.
In 2023, around when the Gino scandal broke, Stanford's then-president Marc Tessier-Lavigne stepped down after 12 papers he authored were found to contain manipulated data. A few weeks ago, MIT announced a data falsification scandal with a terse announcement that the university no longer had confidence in a widely distributed paper "by a former second-year PhD student." It's reasonable to assume the student was expelled from the program.
I hope that these high-profile cases are a sign we are moving in the right direction on scientific fraud because its persistence is enormously damaging to science. Other researchers waste time and energy following false lines of research substantiated by fake data; in medicine, falsification can outright kill people. But even more than that, research fraud damages the reputation of science at exactly the moment when it is most under attack.
We should tighten standards to make fraud much harder to commit in the first place, and when it is identified, the consequences should be immediate and serious. Let's hope Harvard sets a trend.
—Kelsey Piper, senior writer
