Critics Say Sting on Open-Access Journals Misses Larger Point

Perhaps months from now, when the dust settles and academics really look back at it, they’ll find some hard lessons in the elaborate Science magazine exposé this week by the journalist John Bohannon.

After more than a year of work, in which Mr. Bohannon, who has a Ph.D. in biology, crafted a fraudulent cancer-research article and painstakingly tracked the responses to it from more than 300 journals, he gave his industry the embarrassing news that 157 of them had agreed to publish it.

“The data from this sting operation reveal the contours of an emerging Wild West in academic publishing,” Mr. Bohannon wrote in Friday’s issue of Science.

For now, however, allegations of flaws—at least in the way the magazine promoted the piece, if not how the study was constructed from the start—are commanding the bulk of the attention.

Mr. Bohannon offered his fake science submission only to open-access journals, a growing model in which published articles are made freely available rather than restricted to readers with a paid subscription.

More than a dozen critiques have been posted to various news sites and blogs, some of them suggesting a bias by Science, which charges for subscriptions, against the open-access model.

The pique is less about Mr. Bohannon’s 4,200-word article, which suggests he confirmed a problem throughout academic publishing, than his magazine’s 200-word press release (read it here; scroll down to see it), which repeatedly emphasized his findings as an indictment of the open-access model. The sting operation, Science said in its promotion, “exposes the dark side of open-access publishing.”

The controversy has made for “a painful 24 hours,” Mr. Bohannon said on Friday. “I worked in secrecy for an entire year on this project, it generated important results, and so far the loudest voices have been from people who don’t seem to have even read the article.”

Certainly there are those who are outraged over what they read. They include Lars Bjørnshauge, founder and managing editor of the Directory of Open Access Journals, which contains articles from more than 9,900 freely available journals. Mr. Bjørnshauge suggested Mr. Bohannon selected a small subset of open-access journals to discredit the entire genre. Any conclusions from the study, his organization said in a statement, “cannot be applied to the wider open-access community,” nor was there evidence to suggest the situation was any better at subscription-based journals.

Mr. Bohannon, who signed off on Science’s press release, was troubled by some of the complaints, but said he was coming to understand some of the reasons for them. The wording of the release, he acknowledged, was “a little bit critical of open access.”

He said his original plan for the sting encompassed both subscription journals and open-access alternatives, to allow for a comparable sample of both. In total, he said he planned five separate papers covering a range of scientific areas, including physics, computer science, chemistry, and the social sciences. “I came up with this grand plan,” he said, “and it didn’t take me long to realize that it was just incredibly overambitious.”

After sending out his bogus papers to an initial selection of open-access journals—each slightly different, but obviously flawed, versions of a piece purporting to describe the anticancer properties of a chemical extracted from a lichen—Mr. Bohannon said he realized the time and complexity involved in pursuing them all, and abandoned plans for a more extensive set of submissions.

His findings, while therefore limited, still represent the most extensive such attempt to test the responses of scientific journals, acknowledged one critic, Michael B. Eisen, an associate professor of biology at the University of California at Berkeley. Prior stings by others had been confined to one or two journals, Mr. Eisen said.

Editors at Science involved in the process defended their handling of it. The magazine’s news editor, Tim Appenzeller, said the press release “accurately reflects the story.” Mr. Appenzeller said that he recognized the study was limited in scope, but that it found real problems. “We don’t know whether peer review is as bad at traditional journals,” he said. “Then again, OA is the growth area in scientific publishing.”

And, lost in all the controversy over open access, Mr. Bohannon lamented, are at least two major findings that the academic community should consider seriously, regardless of journal financial model, given how much value universities place on peer-review publication.

One finding is that while it was clear that many journals automatically accepted the fraudulent paper without analyzing it, some accepted it even after they had actually put it through their peer-review process.

The second alarming finding, he said, was that some journals published the study even after he immediately wrote back in the name of the faux author, as part of his study protocol for handling any acceptances, asking for it to be withdrawn.

Mr. Eisen, who dissected Mr. Bohannon’s work in a blog posting, said the problems he had identified were mostly at journals widely understood to be of low quality. If measured by their volume of publications, Mr. Eisen said, the most prominent open-access journals fared well in Mr. Bohannon’s test. Those at PLoS, Hindawi, BioMed Central, and other publishers, which together represent about 80 percent of open-access journals, all rejected it, he pointed out.

And at least one of those journals caught in the sting was published by Elsevier, a leading provider of subscription publications. The company, in response, is reviewing its quality criteria, said an Elsevier spokesman, Tom Reller.

Another publication that accepted Mr. Bohannon’s fake paper, the Journal of Natural Pharmaceuticals, is being shut down altogether. “We are deeply concerned about this incident,” said Connie Hughes, a spokeswoman for its publisher, Wolters Kluwer Health.

In a statement, Ms. Hughes said Wolters Kluwer was “taking quick action to review” its processes with its publishing-services company, Medknow Publications. Wolters Kluwer and Medknow had “no involvement in the creation or writing of content or in the peer-review process,” she said, and such matters are “entirely for the editor and editorial boards of the applicable client journal who act independently of Medknow and Wolters Kluwer.”

Mr. Bohannon said he had heard the opposite from the journal editors he had asked to explain their decision to publish his sham science. Most of them explained that the process was outside their control, and some asserted that the journal owners were simply using their names on the publications, he said.

“It’s a shell game,” Mr. Bohannon said. “Wolters Kluwer Health can blame Medknow. Medknow can blame the journal editor. The editor can blame Wolters Kluwer Health. Who knows who’s truly to blame?”