The Science of Plagiarism : STEALING INTO PRINT: Fraud, Plagiarism and Misconduct in Scientific Publishing, <i> By Marcel C. LaFollette (University of California Press: $30.00; 298 pp.)</i>
The old academic saw “publish or perish” became publish and perish last month at Montreal’s Concordia University when a professor of engineering was arrested for allegedly shooting three colleagues dead.
Before the shootings, Valery Fabrikant had accused certain colleagues of exploiting him and others by insisting that they be included as co-authors on scientific papers on which they had done no work. Fabrikant filed suit, demanding that two colleagues in particular withdraw their names from 35 published papers. The two countersued for contempt after Fabrikant voiced his complaints on an electronic bulletin board. Exactly what happened next is not clear: Only one of the three victims was named in the accusations, and the two chief culprits, as Fabrikant saw things, were not even wounded.
Only time, and further investigation, may tell whether Fabrikant’s complaints were justified, even though his solution obviously was not. But the case illuminates many of the issues Marcel LaFollette raises in her fascinating book. Who is wronged in scientific publishing? How should accusations be investigated? Is punishment appropriate and if so, what kind?
Part of the problem is the sheer size of the scientific enterprise: Every 30 seconds, 24 hours a day, 365 days a year, a manuscript arrives at one of the 40,000 or more research journals published every year. Each claims to represent an advance, and each carries a massive burden. Scientists are supposed to be original, honest seekers-after-truth, and yet some of those manuscripts are not what they appear. They are, one way or another, fakes.
Raw data may have been invented wholesale, or gently manipulated to present a false picture. People named as authors may have had nothing to do with the research described. Genuinely involved researchers may have been left off the list of authors. Authors and collaborators might not exist at all. The author may have plagiarized, stolen another’s ideas (and violated their copyright) without giving credit. Sometimes, compounding the felony, information is misappropriated by someone in a position of special trust, a referee, perhaps, who is deciding whether a grant proposal should be funded or a manuscript published.
These thefts are not new, as LaFollette ably documents. Some are as old as scientific publishing. Nor are they restricted to science. But they do seem to be particularly problematical in the realm of science, probably because science sets such store by truth. If a financier is a fraud, well, that’s tough, but not entirely unexpected, given the motivating power of greed. If a novelist borrows more than he or she gives credit for, that too is reprehensible but understandable. Scientific fraud, in any of its many guises, is painful precisely because science is supposed to be a way of uncovering truth.
That must be why people have become so exercised about fraud in science, but is it a problem peculiar to science? No simple answer emerges from “Stealing Into Print,” although not for want of trying. LaFollette surveys the history of fraud in scientific publishing, especially the recent headline-grabbing cases, and as the editor of a journal, she can offer insights from the other side of the desk, juggling authors, referees and publishers and trying to protect each from the possible malpractices of the others.
It is here, really, that the book scores. Few people, scientists included, understand the totality of scientific publishing. The power of referees, protected by anonymity and given privileged access, is made plain. “Authors” who have contributed nothing to a piece of research are happy to share credit, but seldom share responsibility when a fraud is exposed. Publishers are reluctant to pursue miscreants, perhaps for fear of litigation. Editors, too, have been known to invent reasons for rejection rather than voice fears about the probity of a piece of work. Some have even gone so far as to publish work about which they have doubts, only to denounce it later; one can only guess at the reasons for such odd behavior.
By and large, LaFollette’s catalogue of crimes and misdemeanors makes for entertaining reading, although someone not familiar with the cases might want a bit more detail. There are also a few instances in which LaFollette sets out the simple rules by which scientific publishing ought to be governed, and those could be helpful in focusing discussion. In the end, though, the central issue is: Does it matter?
To many people--not least Professor Fabrikant and his victims--it certainly does, although the question seldom seems to have exercised apologists for science, usually eminent scientists guarding their patch. One (admittedly not very eminent, and probably not representative either) confided to me that Fabrikant was “probably loopy anyway,” and that fraud in science had nothing to do with his admittedly astonishing behavior. I suspect that such sentiments are typical of these defenders of scientific probity, who have long trumpeted a series of lame-brained excuses.
First, that there is no fraud, only honest error. That one has been thoroughly scotched by the many original works on which LaFollette draws. Then, that science is as subject to fraud as any other endeavor. I don’t know how they know, but I do raise, again, the point that because science is about uncovering truth it is widely perceived as having some kind of duty to be truthful. Finally, say the apologists, it doesn’t matter anyway, because in the long run science is self-correcting. Vital experiments are repeated, and if the results cannot be replicated, then the reputation of the original work and its authors will wither.
LaFollette, to her credit, goes beyond these simplistic views to examine the other impacts of fraud: careers blighted, funds swindled, employers and colleagues bamboozled, the public disillusioned. To be sure, nobody ends up dead--that will have to await a revised edition --but nobody comes out of her analysis entirely intact either.
Analysis alone, however, is not enough. What is to be done? In scientific publishing, as in the rest of human life, much is taken on trust. And because all scientific publishing originates with authors, it is authors, LaFollette says, who must take the ultimate responsibility.
That, really, is the heart of the matter. Of course it will not eliminate fraud, nature being what it is, but it would make a good start to reemphasize individual responsibilities. When fraud is feared, it ought to be possible to voice suspicions, and investigate them, without offending anybody. Unfortunately, these ideals, like all ideals, are hard to live up to.
As a case in point: How should I have reported the basic stats of scientific publishing mentioned in the beginning of this piece? With a direct quotation, complete with quotation marks? Too wordy. By mentioning the original source, a 1987 article by Gerald Bracey in the Chronicle of Higher Education? But I hadn’t read it. In any case, a moment’s thought made me question the numbers. If Bracey was right, each journal would be receiving only 26 papers a year. That has to be a gross underestimate, even if no paper is ever truly rejected as unfit for publication anywhere but simply trickles down the pyramid of prestige to end up at a level that matches its quality precisely.
Naturally, as a responsible author, I decided to check the original publication. Nothing wrong with the numbers, as cited by LaFollette. Nothing wrong with them, therefore, as I have recited them. And yet, I continue to have my doubts.
I haven’t quietly asked Bracey directly, as LaFollette would have me do, because, frankly, it just wasn’t worth it. In any case, as he is berating authors for sloppy work, I felt sure I could rely on him: all that stuff about motes and beams. I just have to hope that I am not guilty of handling stolen goods.
