Scientific Fraud

Peter Medawar

  • Betrayers of the Truth: Fraud and Deceit in the Halls of Science by William Broad and Nicholas Wade
    Century, 256 pp, £8.95, July 1983, ISBN 0 7126 0243 7

Some policemen are venal; some judges take bribes and deliver verdicts accordingly; there are secret diabolists among men in holy orders and among vice-chancellors are many who believe that most students enjoying higher education would be better-off as gardeners or in the mines; moreover, some scientists fiddle their results or distort the truth for their own benefit.

None of these, though, is representative of his profession – and only people young enough to be cynical believe them to be so. The number of dishonest scientists cannot, of course, be known, but even if they were common enough to justify scary talk of ‘tips of icebergs’ they have not been so numerous as to prevent science’s having become the most successful enterprise (in terms of the fulfilment of declared ambitions) that human beings have ever engaged upon. The profession, sticking together (which is not such a bad thing to do), believes that cheating in science is a curious minor neurosis like cheating at patience – something done to bolster up one’s selfesteem. Rather than marvel at, and pull long faces about, the frauds in science that have been uncovered, we should perhaps marvel at the prevalence of, and the importance nowadays attached to, telling the truth – which is something of an innovation in cultural history, if by the truth we mean correspondence with empirical reality. The authors of the more lurid travellers’ tales would have been taken aback if someone had described them in modern vernacular as ‘bloody liars’, but so they were, many of them. They were telling stories, and wanted to tell good stories. Aristotle’s conception of poetic truth was one in which correspondence with reality played little part, and his biology gave an account of what he thought ought to be true in the light of his deep conception of the true purposes of nature. Thus it ought to be true according to the hebdomadal rule that male semen is infertile between the ages of seven and 21 – a pathetic absurdity of which Aristotle would not have been guilty if he had had any real sense of empirical truth. Aristotle was a pioneer, perhaps, in what I believe to be the commonest form of self-deception in science: the kind of attachment to a dearly loved hypothesis that predisposes us (yes, all of us) to attach a special weight to observations that square with and thus uphold our pet hypotheses, while finding reasons for disregarding or attaching little weight to observations and experiments that cast doubt upon them. There is no one who does not roll out the welcome mat with a flourish for those who bring evidence that upholds our favourite preconceptions.

The most puzzling fraud of all – for such in effect it was – was that of the segregation ratios (3:1; 9:3:3:1) as reported by Gregor Mendel in his plant-breeding experiments. As R.A. Fisher was the first to point out, these ratios conformed far too closely to theoretical expectations to be plausible, having regard to the numbers of plants and seeds involved. The explanation could be as simple as that Mendel was a nice chap whom his gardeners and other assistants wanted very much to please, by telling him the answers which they suspected he would dearly like to hear: moreover, as Mendel was an abbé, his assistants may have felt that there was an element of heresy in securing results other than those the Reverend Father was convinced were true. This is a subject on which the authors of the present book write amusingly.

I do not suppose that personal advancement is a principal motive for cheating in science: rather it is the hunger for scientific reputation and the esteem of colleagues. And I believe that the most important incentive to scientific fraud is a passionate belief in the truth and significance of a theory or hypothesis which is disregarded or frankly not believed by the majority of scientists – colleagues who must accordingly be shocked into recognition of what the offending scientist believes to be a self-evident truth.

Two scientific theories or viewpoints are notorious for arousing this passion: the doctrine of the inheritance of acquired characters associated with the name of J.-B. P.A. de Monet, le Chevalier de Lamarck, on the one hand, and the farrago of sillinesses that may be compendiously called ‘the IQ nonsense’, on the other. Consider Lamarckism first. One kind of Lamarckian inheritance is so commonplace and obvious as to be recognised for what it is by anyone who gives the matter a thought: it is that in which parents or members of a parental generation impart to their children or in general to a filial generation the knowledge and skills they had themselves acquired during their lifetimes. This is heredity all right, but it is ‘exogenetic’ in character, in the sense that it is not mediated through the genetic plant of chromosomes and genes, but through precept, example and deliberate indoctrination. Unlike ordinary or endogenetic heredity, this other kind is reversible and is Lamarckian in style, for that which is acquired in one generation may be transmitted to the next and so on, cumulatively. The existence of this mode of heredity has given people a powerful incentive to believe that ordinary or genetic heredity works in this way too, as it seems only natural justice that it should, and even professional biologists have been taken in by the parallel between exo- and endogenetic heredity and by what looks like a constitutional inability to realise this is not how nature works. The mechanism of heredity is selective, not instructive: what happens in an organism’s lifetime, even if it is a profound bodily modification brought about by an adaptive response, cannot be imprinted upon the genome. There is no known or even conceivable genetic process by which DNA can be taught anything. It seems most unjust that this should be so, but so it is, for in heredity a person’s exertions to improve his body or mind to adapt himself to new environments all go for nothing.

Lamarckian inheritance is a topic upon which literary people have for some reason felt themselves entitled to express an opinion. It is entirely understandable that George Bernard Shaw should have done so, but less obvious why Samuel Butler should have been among their number, especially as he expressed better than anyone else the essence of the teaching of that August Weismann who overthrew Lamarckism. Butler said that according to Weismann ‘a hen is simply an egg’s way of making another egg.’ Lamarckism has on at least one occasion been the subject of a scientific fraud, as recounted in Arthur Koestler’s Midwife Toad. Koestler disclaimed being a Lamarckist, but created an atmosphere favourable to Lamarckism by representing Darwinian geneticists as the spokesmen for a dull and unperceptive establishment of conventional belief. Biologists and psychologists who have been won over to Lamarckian thinking have published a whole number of experiments which purport to demonstrate Lamarckian inheritance, but all such demonstrations have been faulty. Either they have not excluded an orthodox Darwinian interpretation, or they have been open to explanations of other kinds, or they have been technically faulty. I myself was involved as a spectator in one such attempt. There was no dishonesty, indeed nothing more culpable than self-deception – in this case, the enthusiastic selection, from results that were all over the place, of only those that fitted the hypothesis the experimenter was seeking to corroborate.

Why should Lamarckism arouse such passionate conviction of its truth? I believe that the well-known association of Lamarckism with the sinister and indeed evil opinions of Trofim Denisovich Lysenko points to a political explanation. Lamarckism seems only fair: is it not right that mankind should benefit from their exertions and utterly wrong that man’s genetic provenance – his breeding, in fact – should determine absolutely his character, capabilities and deserts? It was his well-founded suspicion that his teachings tended to question the pre-eminence of a man’s breeding that caused Napoleon’s contemptuously dismissive attitude towards Lamarck. To a man convinced that Lamarckian inheritance is true because it is fair and socially just, it seems that selectionist theory presents an attempt, in Condorcet’s words, to ‘render nature herself an accomplice in the crime of political inequality’.

The second of the two major causes of that passionate belief that can conduce to fabrication was that which I referred to as ‘the IQ nonsense’, by which I mean the complex of beliefs arising out of a contention of H.J. Eysenck’s which I lose no opportunity to hold up to public ridicule. ‘Clearly the whole course of development of a child’s intellectual capabilities is largely laid down genetically ...’ (The Inequality of Man, p. 111, 1973).

The most shocking deception arising out of a passionate acceptance of the idea that intelligence is susceptible to a scalar measurement and is 90 per cent heritable was the lengthy and studied scientific frauds of Sir Cyril Burt in his measurement of the IQs of twins reared together or separated from birth. This fraud was uncovered by Dr Leon Kamin and a skilful geneticist turned investigative journalist, Dr Oliver J. Gillie. In the present book, which gives a lot of attention to this case, Broad and Wade illustrate the inefficacy of scientific monitoring within the profession itself – of the procedures which those of us who maintain the integrity of scientists believe prevent or rectify scientific fraud. But the reason Burt’s findings were not subjected to intent and independent critical scrutiny is simple and understandable. There was no effective check of Burt’s findings because he told the IQ boys exactly what they wanted to hear. The fault lay not with the scientific monitoring system but with the bigotry and deep-seated misconceptions of the champions of the IQ concept.

The present authors greatly enlarge our understanding of the Burt frauds by recounting how a graduate student of Iowa State University, Leroy Wolins, ‘wrote to 37 authors of papers published in psychology journals asking for the raw data on which the papers were based’. No fewer than 21 reported that their data had been misplaced, lost or inadvertently destroyed.

The difficulty of laying hands on the 28 sets of data that were ‘lost’ or withheld was made somewhat more comprehensible by the horrors that emerged from the nine sets made available. Of the seven that arrived in time to be analysed, three contained ‘gross errors’ in their statistics. The implications of the Wolins study are almost too awesome to digest. Fewer than one in four scientists were willing to provide their raw data on request, without self-serving conditions, and nearly half of the studies analysed had gross errors in their statistics alone. This is not the behaviour of a rational, self-correcting, self-policing community of scholars.

Burt’s is only one of the many notorious cases of fraud the authors deal with. All the old favourites are to be found in the index: Piltdown, Paul Kammerer of The Midwife Toad and the infamous William T. Summerlin. This last is the only case of a fraud in which I have been involved as a witness for the prosecution: Summerlin worked at the largest cancer research institute in the world, the Sloan-Kettering in New York, under the patronage of its Director, Dr Robert A. Good. Summerlin claimed that by a short period of cultivation in vitro a skin graft would be so altered as to make it transplantable from one mouse to another mouse of a different strain, and that even a graft into a member of a different species would become acceptable – such as a human cornea into a rabbit’s eye. These claims were all highly implausible. I discussed them with my London colleague, Professor Leslie Brent in the Department of Immunology of St Mary’s Hospital Medical School, and we decided there was nothing for it but to go over the same ground again. We did so using very familiar techniques on which we were, I suppose, the world’s principal authorities. We found no reason whatsoever to believe Summerlin’s results to be valid. When Dr Lewis Thomas, at that time President of the entire Memorial-Sloan Kettering Cancer Centre, passed through one day, we told him of our findings, upon learning which he looked very grave and clearly resolved in his mind to take some action. At the next meeting of the Scientific Advisory Board of the Institute, of which I was a member, Summerlin accordingly presented his results – an occasion I remember vividly. Summerlin demonstrated a rabbit which, he said, had received a limbus-to-limbus corneal graft – in effect, an entire corneal graft – from a human donor. In a report on the matter I said that the rabbit contemplated the Board through an absolutely clear and transparent eye and with the candid and steady gaze which only a rabbit with a perfectly clear conscience would be capable of. I saw no evidence that the rabbit had been the subject of an operation. But I did not have the moral courage to say what I was thinking: namely, that the man was a liar and a fraud. It is a little difficult to say something that will ruin a man’s career unless one is very certain of one’s ground. After this demonstration the Board repaired to Summerlin’s laboratory, where we were shown mice that were alleged to bear surviving grafts from mice belonging to other strains. The quizzing of Summerlin caused some amusement among the technical staff present – I know now they were aware that Summerlin was a fraud. I believe that the technical staff had a twofold source of amusement: it is always fun to see the discomfiture of a visiting bigwig, though they were pleased also with the likelihood that Summerlin would be exposed (and it was indeed a technician who in the end gave him away). The crime for which Summerlin became notorious was that of trying to convince his boss of the success of one of his skin grafts, a graft from a black animal onto a white mouse, by touching up the graft with a black felt pen. One can imagine the surprise of a technician upon finding out that this particular skin graft was soluble in alcohol. Summerlin was suspended instantly from work on full pay ($40,000), ostensibly to give him an opportunity to receive psychological treatment. The exposure of Summerlin showed everyone in a bad light. The enemies which a man as successful as Robert Good is bound to accumulate now at last set aside their whetstones and plunged their daggers into him. Summerlin abetted them by representing himself as a victim of the Director’s pressure upon him to publish his results.

A shocking story. Yet it is not the authors’ intention to shock, though in fact they do so: no, the purpose is rather to show that research is not a wholly rational and explicitly logical procedure but subject to the confinements and constraints that afflict other professional men trying to make their way in the world. Moreover it questions our comfortable assumption that scientific cheating is very rare – an exceptional event that does not become a serious threat because science is protected by a whole number of built-in professional safeguards which bring it about that fraud is soon uncovered and the culprits punished.

The authors are very experienced professional science writers, and have made a highly responsible and well-argued contribution to the sociology of science. In spite of these sterling virtues, their book contrives also to be interesting and readable, and suitable for a lay readership. Even science writers are sometimes frauds, and the present authors, though they must have been aware of it, make no mention of a book by a science writer giving an account of a professedly authentic example of human cloning, ingeniously tricked out with quasi-scientific references put in ‘to add corroborative detail to an otherwise bald and unconvincing narrative’.

What lesson should the scientific profession learn? Should we henceforth go around on our guard, doubting and questioning, looking for fraud and misrepresentation with the air of men expecting to find evidence of it? No, indeed not. Listening for a second time to Sir Kenneth Clark’s splendid series of television broadcasts on ‘Civilisation’, I was again struck by the importance that Clark attached to confidence as a bonding agent in the advance of civilisation, as it is indeed throughout professional life. Do not lawyers, bankers, clergymen, librarians and editors tend to believe their fellow professionals unless they have a very good reason to do otherwise? Scientists are the same. The critical scrutiny of all scientific findings – perhaps especially one’s own – is an unqualified desideratum of scientific progress. Without it science would surely founder – though not more rapidly, perhaps, than it would if the great collaborative expertise of science were to be subjected to an atmosphere of wary and suspicious disbelief.