The case for and against self-regulation amongst the scientific community
This article was published in Living Ethics: issue 84 winter 2011
All over the world researchers open copies of the latest journals, containing the latest scientific papers, confident that what is described there is what was actually carried out. Their confidence does not depend on their knowledge of the people involved. it does not depend on the eminence or fame or academic position of the authors. it is a matter of trust.
However implausible that mutual trust might sound it is in fact the basic premise of the whole enterprise: that one can accept at face value, published scientific research. Thus the existence of scientific fraud or, to use a broader term, scientific misconduct should be a matter of grave concern to all researchers and one might have expected that there would be major efforts applied to its prevention and detection. Instead [in 1993], until very recently, there has been almost none. Why not? Because of two beliefs about science itself.
The first of these is that science is self-correcting; that errors will either be detected automatically during the processes of peer review, the process by which papers are selected for publication or that they will come to light as soon as someone else tries to repeat the research. The second belief is that since science is essentially a search for truth, fraud and misconduct can almost be ruled out as inherently unlikely behaviour.
The two beliefs are linked, of course. If science is self-correcting then there is little mileage to be gained by scientific misconduct. Even if a scientist did, for example, invent experiments on beagles and pretend to have carried out descriptive studies on women taking a new oral contraceptive, as Professor Michael Briggs is believed to have done at Deakin University (N Swan, Baron Munchausen at the Lab Bench?, Fraud and Misconduct in Medical Research, 1993), the system would detect it and the researcher would gain nothing.
As the President of the National Academy of Sciences declared in United States Senate hearings in 1981: fraud happens rarely and when it does “it occurs in a system that operates in an effective democratic and self-correcting mode: that makes detection inevitable” (W Broad & N Wade, Betrayers of the Truth: Fraud and Deceit in the Halls of Science, 1982).
In this scheme of things there is absolutely no role for either conscience or regulation. Both are redundant.
That view is still extant and still powerful but it has been weakened in the last decade by a relatively public discussion of specific cases of scientific misconduct: Robert Slutsky, who published one scientific paper every 10 days while at the same time being a hospital radiology resident and a clinical associate professor (RL Engler et al, Misrepresentation and Responsibility in Medical Research, New England Journal of Medicine 317, 1987); John Darsee, caught in the act of falsifying an experimental print out, who turned out to have been engaged in scientific misconduct not just as a postdoctoral researcher at Harvard Medical School and previously at Emory University but during his undergraduate life as well (WJ Broad, Harvard Delays in Reporting Fraud, Science 215, 1982; BC Culliton, Emory Reports on Darsee’s Fraud, Science 220, 1983); Stephen Breuning whose publications from 1980 into 1983 constituted a third of all papers related to mental retardation and psychopharmacology, but from who ‘only a few’ of the many thousands of subjects reported on had ever been studied (AJ Hostetler, Investigation Fraud Inquiry Revives Doubt: Can Science Police Itself?, The APS Monitor 18, 1987); Charles Glueck – prominent cholesterol researcher – whose paper on the safety and efficacy of cholesterol lowering drugs in children included data on their physical development and school performance later described by the man himself as findings which would have been better thought of as ‘anecdotal clinical impressions’ (C Holden, NIH Moves to Debar Cholesterol Researcher, Science 237, 1987).
Australia has not been left out: as well as Michael Briggs we have had Dr William McBride, experimenting on a very widely used drug for morning sickness, whose paper about the effect of the drug on pregnant rabbits included six serious differences from the facts of the matter. These were: adding on two imaginary experimental rabbits and eight controls, doubling the number of rabbits with abnormal embryos, misreporting the amount of the drug the rabbits had been given, describing a second experiment which had, in fact been unsuccessful (two rabbits died, three did not become pregnant – out of six) and describing the abnormalities of the foetal rabbits seen at dissection when all of them were still intact, several years later. The phrase ‘inventing rabbits’ has even become shorthand for forms of extreme scientific misconduct.
What these cases had in common was that none were detected by peer review, the process by which journals review papers submitted for publication. None were detected by someone trying, and failing, to repeat the experiments. In every case, the misconduct came to light through ‘whistleblowing’, usually by junior colleagues as in the case of McBride, but occasionally by other outsiders.
The self-regulation which science believed would operate didn’t. There were no back up systems to support and constrain the conscience of the individual researcher. Heads of large laboratories were so busy, probably writing and reviewing grant proposals, that they did not oversee what was going on in their labs. They were unfamiliar with the primary data from experiments. Others became co-authors of papers, not only without making a contribution to the work, but without being able to verify the contents of the paper and certainly without being willing to accept responsibility for the contents when they were questioned. In the McBride case, the fraud first came to light when the junior colleagues who ended up as whistleblowers found their names on a published paper which was at odds with their recollection of the experiments.
The response of institutions to the initial accusation of misconduct has not been supportive either. After a short phase of denial, the reaction has been redefinition. The first way is to minimise what has been done by describing the researcher’s behaviour as ‘sloppy’ or ‘careless’ or ‘misdirected enthusiasm’ and to use phrases like ‘trivial’ or ‘minor inconsistencies’ or ‘faults that don’t alter the conclusions of the paper’. The second response has been to characterise the behaviour as ‘bizarre’, ‘an isolated instance’, ‘an aberration’ a ‘single foolish act’. The redefinition is either towards error (something acceptable if undesirable) or towards madness (something beyond rational control) and this makes it neither a matter for conscience nor a matter of self-regulation.
Publishers of scientific papers have been concerned about scientific misconduct but some have refused to retract papers unless all the authors (including the guilty party presumably) signed a letter asking them to do so.
The (former) US Vice President, Al Gore, then a young Congressman, convened Senate hearings on scientific fraud early in 1981. “I cannot avoid the conclusion” he said, “that one reason for the persistence of this type of problem is the reluctance of people high in the science field to take these matters very seriously” (W Broad & N Wade, Betrayers of the Truth). Implicit in these hearings was the threat of regulation imposed from outside science if more effective processes of self-regulation were not established.
The institutional response has been the development of guidelines and recommendations. The National Health and Medical Research Council published a statement on scientific practice in November 1990 after extensive consultation with the Australian Vice-Chancellors’ Committee. The statement draws attention to the potential conflict of interest which exists when the Chief Executive Officer of an institution has a central responsibility and role in relation to judgements about scientific misconduct and recommends instead that an independent person be nominated for this role. Unfortunately, the comments on ethical behaviour leave something to be desired: it is the research project which must conform to ethical standards; research workers “have a duty to ensure that their work enhances the good name of their institution and the profession to which they belong” (Australian National Health and Medical Research Council Statement on Scientific Practice, 1993).
It is too soon to know how effective the new procedures and guidelines will be but one United States case suggests that the penalty for inaction may be severe. David Baltimore, Nobel prize winner and head of a department at the Massachusetts Institute of Technology, supported for several years a scientist accused of scientific misconduct in relation to a paper of which he was a co-author. He did not believe the data were faked but the group of co-authors never met to consider the allegations seriously and never inspected the primary data.
He accused investigators from the National Institute of Health’s Office of Scientific Integrity of carrying out “inappropriate prying” and seeing themselves as “guardians of scientific purity”. He believed, he said, in the principle that the United States Congress and others outside science should not insert themselves in scientific disputes. Baltimore wrote a nine page letter to 400 fellow scientists around the world asking for their support against this small group of outsiders who he described as absolutely unqualified to investigate and as redressing an imagined wrong. Six years after the first complaint was ignored, the findings of a secret service style investigation (DP Hamilton, Verdict in Sight in the ‘Baltimore Case’, Science 251, 1991) led to the retraction of the paper and a statement of apology by David Baltimore (D Baltimore, Baltimore says ‘Sorry’, Nature 351, 1991).
Lack of effective regulation may mean for scientists only heavy handed intervention by government agencies. If conscience and regulation between them are seen to be ineffective it will certainly lead to a further growth in cynicism and mistrust of science and scientific evidence throughout the whole community.
latest articles
