Over the past two decades, dozens of behavioral scientists have risen to prominence pointing out the power of small interventions to improve well-being.
The scientists said they had found that automatically enrolling people in organ donor programs would lead to higher rates of donation, and that moving healthy foods like fruit closer to the front of a buffet line would result in healthier eating.
Many of these findings have attracted skepticism as other scholars showed that their effects were smaller than initially claimed, or that they had little impact at all. But in recent days, the field may have sustained its most serious blow yet: accusations that a prominent behavioral scientist fabricated results in multiple studies, including at least one purporting to show how to elicit honest behavior.
The scholar, Francesca Gino of Harvard Business School, has been a co-author of dozens of papers in peer-reviewed journals on such topics as how rituals like silently counting to 10 before deciding what to eat can increase the likelihood of choosing healthier food, and how networking can make professionals feel dirty.
Maurice Schweitzer, a behavioral scientist at the Wharton School of the University of Pennsylvania, said the accusations were having large “reverberations in the academic community” because Dr. Gino is someone who has “so many collaborators, so many articles, who is really a leading scholar in the field.”
Dr. Schweitzer said that he was now going through the eight papers on which he collaborated with Dr. Gino for indications of fraud, and that many other scholars were doing so as well.
Behavioral work is common in psychology, management and economics, and scholars can straddle these disciplines. According to her résumé, Dr. Gino has a Ph.D. in economics and management from an Italian university.
Questions about her work surfaced in an article on June 16 in The Chronicle of Higher Education about a 2012 paper written by Dr. Gino and four colleagues. One of Dr. Gino’s co-authors — Max H. Bazerman, also of Harvard Business School — told The Chronicle that the university had informed him that a study overseen by Dr. Gino for the paper appeared to include fabricated results.
The 2012 paper reported that asking people who fill out tax or insurance documents to attest to the truth of their responses at the top of the document rather than at the bottom significantly increased the accuracy of the information they provided. The paper has been cited hundreds of times by other scholars, but more recent work had cast serious doubt on its findings.
Dr. Gino did not respond to a request for comment, and Harvard Business School declined to comment. Reached by phone, a man who identified himself as Dr. Gino’s husband said, “It’s obviously something that is very sensitive that we can’t speak to now.”
Dr. Bazerman did not respond to a request for comment for this article, but told The Chronicle of Higher Education that he had had nothing to do with any fabrication.
On June 17, a blog run by three behavioral scientists, called DataColada, posted a detailed discussion of evidence that the results of a study by Dr. Gino for the 2012 paper had been falsified. The post said that the bloggers contacted Harvard Business School in the fall of 2021 to raise concerns about Dr. Gino’s work, providing the university with a report that included evidence of fraud in the 2012 paper as well as in three other papers on which she collaborated.
The blog — by Uri Simonsohn of ESADE Business School in Barcelona, Leif Nelson of the University of California, Berkeley, and Joseph Simmons of the University of Pennsylvania — focuses on the integrity and reliability of social science research. The post on Dr. Gino noted that Harvard had placed her on administrative leave, a fact that was reflected on her business school web page, though no reason was given. The Internet Archive, which catalogs web pages, shows that Dr. Gino was not on leave as recently as mid-May.
The 2012 paper was based on three separate studies. One study overseen by Dr. Gino involved a lab experiment in which about 100 participants were asked to complete a worksheet featuring 20 puzzles and were promised $1 for every puzzle they solved.
The study’s participants later filled out a form reporting how much money they had earned from solving the puzzles. The participants were led to believe that cheating would be undetected, when in fact the researchers could verify how many puzzles they had solved.
The study found that participants were much more likely to report their puzzle income honestly if they attested to the accuracy of their responses at the top of the form rather than the bottom.
But in their blog post, Dr. Simonsohn, Dr. Nelson and Dr. Simmons, analyzing data that Dr. Gino and her co-authors had posted online, cited a digital record contained within an Excel file to demonstrate that some of the data points had been tampered with, and that the tampering helped drive the result.
Last week’s post was not the first time the DataColada watchdogs had found problems with the 2012 paper by Dr. Gino and her co-authors. In a blog post in August 2021, the same researchers found evidence that another study published in the same paper appeared to rely on manufactured data.
That study relied on data provided by an insurance company, to which customers reported the mileage of cars covered by their policy. The study purported to find that customers who were asked at the top of the form to attest to the truthfulness of the information they would provide were significantly more honest than customers who were asked to attest to their truthfulness at the bottom of the form.
But through analysis of the raw data, Dr. Simonsohn, Dr. Nelson and Dr. Simmons concluded that many of the data points were created by someone connected to the study, not based on customer information. The journal that published the 2012 paper, the Proceedings of the National Academy of Sciences, retracted it the month after the blog post appeared.
In that case, another of the paper’s co-authors, Dan Ariely of Duke University, was the scholar who procured the data from the insurance company. Dr. Ariely, one of the world’s best-known behavioral scientists, said in an email on Friday that he had been “stunned and surprised” to learn that some of the insurance data in the paper had been fabricated, “which led me to proactively retract it.”
DataColada has since published blog posts laying out evidence that results were fabricated in two other papers of which Dr. Gino was a co-author. The bloggers have written that they plan to publish one more post laying out issues in an additional paper on which she collaborated.
In interviews and comments on social media, several scholars said they had not suspected fraud in Dr. Gino’s work. But some noted that the findings in the genre of behavioral research that she specializes in, which is closer to psychology, often resemble findings generated by questionable research methods.
One category of questionable methods, said Colin Camerer, a behavioral economist at the California Institute of Technology, is p-hacking — for example, testing a series of arbitrary data combinations until the researcher arrives at an inflated statistical correlation.
In 2015, a team of scholars reported that they had tried to replicate the results of 100 studies published in prominent psychology journals and succeeded in fewer than half the cases. The behavioral studies proved especially hard to replicate.
Source: New York Times
