If you have ever studied evolutionary biology, you probably know who Robert Trivers is. His groundbreaking articles on reciprocal altruism, the evolution of sex ratios, and parent-offspring conflicts are staples in every Evolutionary Biology course. They make great reading, and his theories pop up all the time in all sorts of disciplines. I’m sure his work has inspired many budding scientists.
Currently, Trivers is a professor at Rutgers. In 2005, he co-authored an article on body symmetry and dance that made the cover of Nature. A couple years later, he and his colleagues started to suspect that the first author on the paper may have fabricated data. In 2008, they contacted Nature about their suspicions and tried to retract the article. According to Trivers, because at least one co-author anonymously disagreed with the decision to publish a retraction, Nature didn’t publish one, although this month they did publish a great piece by Eugenie Samuel Reich about the saga in which Trivers has become embroiled. Frustrated that he couldn’t get the word out about the faked data, in 2009 he and two other scientists self-published a short book called The Anatomy of a Fraud, detailing evidence that the data in the paper had been forged. After the book came out, Rutgers was forced to investigate the matter, and in 2012 they issued their conclusions. In the Research Advisory Board’s report, which Trivers posted on his website, they agreed that it was pretty clear that the first author of the paper had manipulated the data in order to make a good story. The fall out from this episode has been major. A glance at his website shows how aggrieved Trivers is about the failure of so many institutions (Nature, Rutgers) to take the accusations of fraud seriously. And in 2012, Trivers was banned from campus for months after an unpleasant exchange with a collaborator who is also on faculty at Rutgers.
While there has been a lot of chatter about the aftermath of these accusations (he said, she said type-stuff about the way the co-authors have interacted with one another), it seems like the real issue at stake is how hard it is to blow the whistle on scientific fraud. No one looks forward to admitting that they co-authored a doctored study–especially when that study appeared in a high profile journal like Nature. It takes integrity to investigate whether fraud was committed in your own lab and, after finding it was, to promptly try to retract the affected work. Since scientific progress is cumulative, with current studies building on past findings, this kind of self-policing is exactly the type of behavior all of us want to encourage. Trivers and colleagues published their book outlining the problems with the symmetry and dance study in 2009, but that didn’t stop the citations. According to Google Scholar, the article has been cited 128 times, many of the citations occurring as late as 2013. It’s well-known that articles continue to be cited even after they are retracted, but certainly a retraction in Nature would have helped in this example–the odds that someone doing a literature search would link this article to a little known, self-published book are extremely low. While the problem of retractions (their growing frequency, their inability to effectively remove false findings from the literature once the original articles have been published) is attracting increasing attention, especially with help from blogs like Retraction Watch, retracting a fatally flawed article is certainly better than not retracting it. Unfortunately, retractions are an unpleasant business for all of the important players involved: the journal, the academic institution, and the co-authors. If a very prominent scientist has this much trouble trying to reveal a case of academic fraud in his own research group, what hope is there for everyone else? Hopefully all of the attention this case has received will put the pressure on Nature to rectify this situation, but I’m afraid events like this may dissuade junior scientists from coming forward with unpleasant but important information.