Sunday, March 26, 2006

Fraud in science

Book of the Day: Reading Journal Entry: The Great Betrayal: Fraud in Science by Horace Freeland Judson "Book of the Day" features a thoughtful review of Horace Freeland Judson's latest (?) book, on fraud in science. Mapletree 7 is (rightly, I hope) sceptical about Judson's claims that there is much unidentified fraud commited by scientists. (One of those "how can he know?" questions.) Mapletree 7 comments on Judson's identification of peer-review of grants and of publications as the two key areas in need of reform. She is absolutely right to say that scientists are just like everyone else, and hence no more or no less likely to cheat as everyone else; the difference lies, she says, in that scientists deal with experiments and data --- the interpretation of evidence. She concludes that Judson's book would have been better had it focused more on science and less on becoming a "general polemic" about fraud in other areas of society. The peer-review system is not designed to detect fraud; it was designed to judge results that were obtained honestly. I think it is very hard to design a workable system that is going to detect fraud in the first place. It is less hard to design a system that reacts to fraud once it is suspected -- the Hwang (Korean stem-cell) affair will add impetus to current efforts, I am sure. Journals can act promptly when informed of suspicion of fraud, and can use the power of the internet to label retrospectively the online version of suspicious papers. Abstracting and indexing services such as Medline can do the same, so that "suspicious" science is clearly identified to any future reader. Funding bodies, whether of grants or employers of the scientists concerned, can also act promptly. Plenty of minds and committees are currently focused on this question, and plenty of electronic and real ink is being expended by journalists, who have a range of opinions about what should be done. At the end of the day, catching anyone cheating is difficult in any profession. Proving a suspicion without a witchhunt mentality is also difficult, although as Mapletree7 implies, scientists should always keep accurate records of their research and be prepared to show them if asked. It is interesting in this context that scientific journals are often accused of failing to appreciate the importance of innovative research. Nobel prizewinners nowadays often tell of the difficulties they had in getting their groundbreaking research published. (My own journal, Nature, famously rejected the Krebs cycle, and reduced the report of the discovery of monoclonal antibodies from an Article to a Letter.) Everyone agrees that scientific research is about innovation and discovery. But there is no simple answer to the question of how the entire system of professional science can be designed to welcome unexpected new discoveries, yet be responsible about ensuring such claims are justified.