Science to select papers for extra-secure review process.
In a notable move aimed at curbing fraud in scientific publications, the journal Science said last week that it will probably begin targeting certain “high-risk” papers for extra scrutiny.
The move comes in response to a report from an external committee convened by the journal to assess its handling of the papers behind the Woo Suk Hwang fraud scandal. And it turns on its head — for a handful of papers at least — the traditional presumption that manuscripts submitted to a journal are researched and written honestly.
“Until now, it has been assumed as a default that scientists are honest. The burden of proof is to show that they are not. Now, at least for a select number of papers where the risk factor is high, there is a new burden, to show that these papers are honest,” says Sheldon Krimsky, a bioethicist at Tufts University in Medford, Massachusetts.
Hwang, a South Korean researcher working at Seoul National University, published high-profile papers in Science1,2 in 2004 and 2005 that claimed to have generated embryonic stem cells by somatic-cell nuclear transfer. This is a key step to generating replacement tissues from a patient's own cells. Both papers turned out to have been fabricated, and Science retracted them in January3.
The review committee, headed by John Brauman, a chemist at Stanford University in California, released its report into Science's conduct last week. Although Science had followed high-standard editorial procedures “with exceptional care” in the Hwang case, Brauman says, “we suggested that the journal institute a policy we describe as risk assessment” in an effort to clamp down on fraud.
Writing in an editorial4, Science's editor-in-chief Donald Kennedy said that the journal is now developing criteria for deciding which papers deserve particularly careful scrutiny. “Papers that are of substantial public interest, present results that are unexpected and/or counterintuitive, or touch on areas of high political controversy may fall into this category,” he wrote.
Such papers, perhaps ten a year, would receive “special attention” that could include greater requirements for including primary data and more intensive evaluation of digital images. The journal would also demand explicit descriptions of each author's contribution to a paper.
Even as Kennedy announced the plans, Science reported doubts about the results of another high-profile paper it recently published in a controversial area of developmental biology. Pending the outcome of an investigation by the University of Missouri, Columbia, this will also probably be retracted.
In a paper5 that sparked debate from the moment it was published, R. Michael Roberts and colleagues, researchers at the university, claimed that mouse embryonic cells have distinct developmental fates from the first cell division onwards. This flies in the face of the broadly held view that in mammals such cells can still develop into any cell in the body. The paper was published on 17 February this year, and the university launched an investigation in April that is still continuing.
Kennedy cautioned in a press briefing that the social costs associated with loss of trust among scientists might be greater than those of the occasional retraction. But he said he would collaborate with Nature and other journals to draw up a common set of standards aimed at deterring fraud. Philip Campbell, Nature's editor-in-chief, declined to comment in detail on the committee's findings, but said: “We at Nature welcome the external review conducted by Science and are considering its recommendations.”
But many commentators seem to feel that the recommendations don't go far enough. “There should be very strict criteria not only for high-impact or controversial papers but for everybody,” says Thereza Imanishi-Kari, an immunogeneticist at Tufts. Imanishi-Kari's career stalled when her work became the subject of a high-profile US government investigation in the 1990s. She was exonerated.
Krimsky also favours increased scrutiny of papers: ”What could the negative impacts be? A less trusting community of scientists? A new McCarthyism? I don't think that's likely.” He adds that a further criterion should trigger extra scrutiny of papers: where an author or authors have an unusually high degree of commercial interest in the results.
Meanwhile, Arthur Caplan, director of the Center for Bioethics at the University of Pennsylvania in Philadelphia, says he would like to see the criteria expanded to include papers in areas with a history of fraud and those in highly specialized areas where relatively few other scientists would be equipped to detect such problems. He adds: “I think you have to be blunt and say that papers coming from certain countries also raise red flags. China has not yet been convincing and South Korea has problems.”
But Kennedy said in the press briefing that he would guard against any targeting of foreign scientists. “We don't want to engage in profiling. It would really be unfair if we started looking extra hard at some of the papers from emerging scientific powers in countries such as South Korea.”
Hwang, W. S. et al. Science 303, 1669–1674 (2004).
Hwang, W. S. et al. Science 308, 1777–1783 (2005).
Cyranoski, D. Nature 439, 122–123 (2006).
Kennedy, D. Science 314, 1353 (2006).
Deb, K., Sivaguru, M., Yong, H. Y. & Roberts, R. M. Science 311, 992–996 (2006).