Fraud Leads to Procedural Reviews at Scientific Journals
RENEE MONTAGNE, host:
Two different investigations have concluded that an Afghan hound named Snuppy really is the world's first cloned dog. The investigation into Snuppy's authenticity came after his South Korean creator admitted fabricating data in two other landmark papers on cloning. The papers were later retracted by the scientific journal that published them which led to the question how did the fraud slip past editors in the first place?
And that has prompted journals to take a close look at their editorial practices as NPR's Joe Palca reports.
JOE PALCA reporting:
To uncover the South Korean fraud, a team of investigators from Seoul National University required researchers to turn over all their raw data and to open their storage freezers so investigators could see what was really there. Natalie Dewitt says journals aren't in a position to carry out such a thorough investigation. Dewitt is an editor at Nature, the journal where news of the cloned dog Snuppy originally appeared. She says Nature publishes about 1,000 scientific articles a year.
Ms. NATALIE DEWITT (Editor, Nature): I don't think that we're prepared to investigate all of our authors and try to verify that there is no potential that there could be misrepresentation of such a blatant nature.
PALCA: But there are some things journals, including Nature, have begun to do to catch what's called data beautification: that's where a scientist uses digital tricks to make pictures look better than they really do.
Mr. MIKE ROSSNER (Managing Editor, Journal of Cell Biology): We can, using very simple techniques, get some idea about whether they've been manipulated or not.
PALCA: Mike Rossner is Managing Editor of the Journal of Cell Biology. He says there is software that can detect digital manipulation. Rossner says his journal routinely uses it.
Mr. ROSSNER: We do it for all figures of all accepted manuscripts and yes we do find examples of manipulation.
PALCA: When a manipulation is detected, Rossner says the authors are asked to submit their original data. About one percent of accepted papers are subsequently unaccepted because of what the Journal feels is fraudulent data manipulation. Rossner says more journals should adopt this policy.
Mr. ROSSNER: Editors have an obligation to do what they can do to protect the published record and this is something they can.
PALCA: Although journal editors have the final say about whether to publish a paper, they typically rely on a process known as peer review to assist them. When a paper arrives at a journal, the editors will send the paper to another scientist working in the same field. The theory is that a researcher's scientific peers are in the best position to judge whether experiments described in the paper were done appropriately and that the results justify the conclusions.
Ms. BARBARA COHEN (Editor, PLoS Biology): The peer review system as a whole is based on trust that authors faithfully report what they've done.
PALCA: Barbara Cohen is the editor of the journal PLoS Biology--she says peer review is only a protection against the most amateurish attempts at fraud. A carefully crafted fabrication would slip through as it did in the Korean cloning scandal. But Cohen says even when that happens; the scientific process has a way of detecting it.
Ms. COHEN: Peer review is a first step of trying to assess the originality to some extent the credibility of the results and then the more important second step actually happens once the results are published and shared within the community.
PALCA: Cohen says when other scientists try to repeat the experiments or build on the results, they'll fail and the fraud will be exposed. That's the good news. The bad news is that scientists will have wasted they're time pursuing a bogus result.
Joe Palca, NPR News Washington.
NPR transcripts are created on a rush deadline by a contractor for NPR, and accuracy and availability may vary. This text may not be in its final form and may be updated or revised in the future. Please be aware that the authoritative record of NPR's programming is the audio.