Most people would not have been interested in the sins of Ariel Fernández.
In 2013 someone suggested that Mr. Fernández, an Argentine scientist, had contributed bad data to a genomics paper. Two of the institutions affiliated with Mr. Fernández had investigated; one had found his data credible, the other had not. “Interpret the data with due caution,” wrote the editors of BMC Genomics, the journal that had published the paper two years earlier, in a note to readers.
The implications of the note were hard to parse. What exactly had gone wrong? Could the paper be trusted, or not? What did “due caution” mean?
Retraction Watch was set up to answer questions like those. By that time the thorny little blog had already planted itself in the side of journal editors and researchers who preferred that errors in the scientific record be dealt with discreetly. Its founders, a pair of veteran science writers, were not just interested in big-ticket fraud cases; they were determined to apply scrutiny to scientific screwups of all kinds, including the obscure ones that tended to slip through the cracks.
So when BMC Genomics posted its note, Retraction Watch wanted answers. “One of Fernández’s three institutions, we don’t know which, found cause for concern with his results,” wrote Adam Marcus, one of the blog’s founders, in a post about the journal’s note. “Another did not (why only two are referenced here is a mystery). What, we wonder, did Fernández have to say about all this?”
He soon got a message: Take down the post, or face a lawsuit.
Retraction Watch later quoted several emails that its editors said Mr. Fernández had sent to Mr. Marcus and to editors at BMC Genomics. The messages threatened legal action against the blog and asked the journal to help stop Retraction Watch from damaging Mr. Fernández’s reputation. (In an email to The Chronicle, Mr. Fernández denied writing the messages. “Someone is using my email address,” he said, adding, “I don’t read blogs.”)
The messages asserted that Mr. Fernández’s paper should not have been written about on a blog called Retraction Watch because technically the journal had issued an “expression of concern,” not a retraction. When Mr. Marcus explained that he had made the distinction clearly in his post, he received a reply, in all caps, insisting that his post amounted to libel.
It was not the first time a scientist had threatened to sue Retraction Watch, and it wouldn’t be the last. Over the last five years, Mr. Marcus and his partner, Ivan Oransky, have gotten under the skin of plenty of researchers and journal editors by turning retraction-spotting into a spectator sport. In the process they have earned a few enemies — along with many fans, including a few powerful grantmakers.
Unexpected Influence
Armed now with a bona fide reputation and $700,000 in foundation funding, Retraction Watch finds itself in a position of unexpected influence at a time when scientific researchers are struggling to maintain their credibility in the public eye. The past decade has seen an boom in research-fraud cases, some of which have made national headlines. A recent meta-study of 100 psychology papers found that less than half of the published findings could be replicated. People looking for excuses to distrust scientists no longer need to look very hard.
Dr. Oransky and Mr. Marcus are not antiscience. But they say the scientific literature does nobody any favors by hiding its dirty laundry. In order to reaffirm science’s ability to fix its own mistakes, say the Retraction Watch founders, researchers and journals need to be more forthcoming when mistakes happen.
Science is made by human beings. And yet when it comes to the scientific literature, all traces of humanity tend to disappear. The passive voice reigns: Bunsen burners are lit. Variables are manipulated. Measurements are taken. The researchers themselves — men and women with hopes, fears, ambitions, and personalities — are obscured in method and process, the supposed bulwarks against human error.
Journals have sometimes reinforced that sense of separation between science and the flawed people who make it. They haven’t always felt compelled to tell readers when scientists mess up. And even when they did so, they didn’t provide much of a backstory.
Recently, when a trio of researchers scoured PubMed, a federal archive maintained by the National Institutes of Health, the earliest retraction they found was a 1973 paper that was retracted in 1977. “Retraction is a relatively recent development in the biomedical-scientific literature,” they wrote, “although retractable offenses are not necessarily new.”
The most common reason for a retraction is misconduct. Sometimes that means a researcher has gotten two different journals to publish the same article, but more often the sin is graver, with implications for the integrity of the science itself. A 2012 analysis of all retracted biomedical and life-sciences articles in PubMed — 2,047 of them — found that 43 percent had been withdrawn because scientists falsified data.
Spiking Retractions
Sometime around the turn of the millennium, retractions began to spike. Fraud, in particular, increased sevenfold from 2004 to 2009, according to R. Grant Steen, a researcher in North Carolina who later participated in the PubMed study.
Were scientists becoming more dishonest as they faced mounting competition for grants and tenure appointments? It’s possible. But another theory, at once more hopeful and disturbing, was that the scientific record had always been more fraught than its keepers had cared to acknowledge.
Retraction Watch emerged in this era of doubt with little fanfare.
The blog was the brainchild of Dr. Oransky, who at the time was executive editor of Reuters Health, a medical news service. A medical-school graduate with a newsman’s penchant for cutting through sanitized narratives, Dr. Oransky was not intimidated by the complexities of reporting on scientific and medical research, and occasionally going toe to toe with the scientists behind it.
He enlisted Mr. Marcus, a fellow science writer who had played a key role in covering the downfall of Scott S. Reuben, a Boston anesthesiologist who faked data in a number of influential pain-management studies and later went to prison. Together they launched Retraction Watch in August 2010.
It was conceived as a lark. Neither founder quit his day job. There was no budget, no revenue, and no sign that anyone else would be interested.
“We literally didn’t have any plans,” says Dr. Oransky. “We started this thing really just to tell stories, uncover some things, and shine a light where not a lot of people wanted light shone.”
‘None of Your Damn Business’
Some journal editors were less than thrilled to hear from Retraction Watch.
“It’s none of your damn business,” said L. Henry Edmunds Jr., editor of The Annals of Thoracic Surgery, when Retraction Watch asked him to spill the details behind a discredited study of hypertension. “If you get divorced from your wife,” the editor added, “the public doesn’t need to know the details.” (Dr. Edmunds did not respond to an email from The Chronicle.)
John Maunsell, then the editor of The Journal of Neuroscience, also declined to cooperate with Retraction Watch. He did not think that capitulating to a blog’s thirst for human-interest stories was part of his job.
By issuing a retraction, “the journals are fulfilling their only real responsibility, which is to maintain the hygiene of the scientific literature,” said Mr. Maunsell in an interview with The Chronicle. “Ultimately a retraction expunges the article from the literature. It’s gone, it’s no long citable. There’s actually no point in worrying what’s in it.”
But the blog proved difficult to ignore. The same irreverent tone that annoyed some readers drew in others. Mr. Marcus and Dr. Oransky mocked journals for their opaque retraction notices, but they also cheered authors and editors who were forthcoming about their mistakes. And the reporting was solid.
Above all, it was fun to read. Retraction Watch’s writers named names. They showed faces. They made puns. They translated journal jargon for lay readers: In their very first post, they referred to an “expression of concern” as “a Britishism that might be better expressed as ‘Holy shit!’”
Making Retractions ‘Sexy’
Mr. Steen and other researchers had done the yeoman’s work of quantifying how big retractions had become as an issue, but Retraction Watch’s founders figured out how to make people pay attention.
“Retraction Watch has made retractions sexy,” says Elizabeth Wager, an editor at Research Integrity and Peer Review, an ethics journal. “I think if you kept it dry and dusty, the journals would still have gotten away with doing it badly, and nobody would be talking about it. The way to get people to talk is to bring that human-interest story.”
People talked. Thomas Reller, the head of global corporate relations for Elsevier, a multinational company that publishes many journals, remembers getting constant emails from editors who wanted to know how to handle inquiries from the blog.
He recommended that the journals work with Retraction Watch. He had dealt with Dr. Oransky before, and thought he was a fair reporter. Beyond that, Mr. Reller also suspected that scientific publishing had reached a critical point in its relationship with the modern news media. Blogs had started to matter, whether journal editors liked it or not.
“With the Internet, the politicalization of science, and the rise of ethics journalism, the conduct of science has more visibility than ever,” Mr. Reller wrote in a 2011 note to journal editors.
“The inquiries and coverage from science skeptics is not always comfortable,” he wrote, “but in this new era of science media, it can be a positive development for retaining public trust.”
Mr. Steen worries a lot about the public trust these days. It’s one reason he doesn’t study retractions anymore.
“I decided that, to the extent I was getting in the way of science getting funded and plunging ahead, I didn’t want to participate,” he says.
Mr. Steen worries about the “anti-intellectual” bias of modern American life, citing the antivaccination movement and persistent skepticism about climate change and evolution. He wonders if Retraction Watch might be inadvertently supplying ammunition to the enemies of science.
It’s true enough that researchers are only human, says Mr. Steen, but as human endeavors go, science has more integrity than most. When he and his colleagues looked in the PubMed database of biomedical research, they found 25 million papers and only 2,047 retractions.
“That’s roughly equivalent to a weatherman who is right 30 years in a row and then gets it wrong once,” he says. “The interest in retractions far outstrips the seriousness of the problem.”
‘We Wanted to Be Comprehensive’
Ferric C. Fang, a professor of laboratory medicine and microbiology at the University of Washington, worked with Mr. Steen on his retraction research. He too worries about harming the public perception of science, and he does not care for the “voyeuristic aspect” of some of Retraction Watch’s work.
But Dr. Fang ultimately reached a different conclusion about the blog. Retractions are embarrassing, he says, but journals should not be afraid to explain how they happened. When Dr. Oransky asked him to join Retraction Watch’s Board of Directors, Dr. Fang accepted.
“It probably makes people a little more nervous to think that there are these extra eyes on what they’re doing,” he says, “and I don’t think that’s necessarily a bad thing.”
Many people are now watching Retraction Watch. The blog runs two or three articles each weekday and regularly gets more than 150,000 unique visitors every month.
The blog was widely cited for its role in spreading the news that Michael LaCour, a political-science researcher at the University of California at Los Angeles, had probably fabricated data in one of the most-talked-about studies of the year. Dr. Oransky and Mr. Marcus have been asked to contribute to The New York Times, The Wall Street Journal, and other mainstream publications. Their army of tipsters has grown.
But it is not nearly enough.
When they started the blog, the goal had been to cover every single retraction notice. “We wanted to be comprehensive,” says Dr. Oransky. “That’s actually part of why this will be useful to people.”
They found out early that there are simply too many retractions. They can’t keep up.
Still, Retraction Watch is growing. The blog recently won a pair of hefty grants, from the John D. and Catherine T. MacArthur Foundation and the Laura and John Arnold Foundation, which has allowed it to expand its staff. The blog now has an editor, a staff writer, and a researcher in addition to its two founders.
The ultimate goal is more ambitious: to build a comprehensive database of retractions that researchers can check before they cite an article.
Jeff Ubois, a program officer with the MacArthur Foundation, says such a tool will take a long time to build. “This is a long-term project,” he says. “This grant was intended to get it off the ground.”
In the meantime, the blog continues apace. On Thursday it flagged yet another “expression of concern” on a paper by Mr. Fernández, the Argentine scientist. When a staff writer for Retraction Watch emailed him for comment, he answered her questions patiently.
Mr. Fernández never sued Retraction Watch. But he has not retracted his disdain for the blog.
“I thought about suing RW,” he told The Chronicle in an email this month, “then I quickly realized that nobody with scientific credentials takes RW seriously.”
Steve Kolowich writes about how colleges are changing, and staying the same, in the digital age. Follow him on Twitter @stevekolowich, or write to him at steve.kolowich@chronicle.com.
Update (9/30/2015, 6 p.m.): This article has been updated to further emphasize that Mr. Fernández denied sending emails that threatened Retraction Watch with legal action.
