The recent appearance of a National Academy of Sciences report concerned with misconduct in the conduct of scientific research , and of a book on fraud, plagiarism, and misconduct in scientific publishing  have drawn needed attention to these seamy sides of the scientific endeavor. Need more be said? I believe so.
We all progress in science by standing on the shoulders of others: our teachers and other scientific progenitors. This progression need not, however, and should not, lead to our stepping on the faces of our scientific ancestors or contemporaries as we continue the enterprise of science and try to communicate something we think is new and valuable.
Here I classify and discuss some pathologies of scientific authorship, ones which, inevitably, arise from some aspects of human nature. Science is universal in scope, but only people do science. Therefore, one must expect that their productions may sometimes reflect the frailties of human nature. In particular, the “publish or perish” syndrome, which has developed over the past fifty years, has led to great pressures to publish, especially for young academics near the beginning of their careers in science. Such pressure, in turn, can sometimes lead them to accept ethically improper shortcuts.
Thus, although it should not be too surprising that occasionally corners are cut and unjustifiable claims made, these and other kinds of misconduct reflect badly upon the perpetrator when they are recognized and, more importantly, can result in appreciable costs to society even when they aren’t recognized. As Edmund Burke cautioned in Reflections on the Revolution in France, those who make themselves the most conspicuous are not necessarily the most important. This is a lesson worth heeding in many other areas as well as science.
Such serious scientific frauds as fabrication or alteration of data and plagiarism have been much in the news in recent years. Here, I discuss three other less obvious types of problems that often appear in scientific publications. Although the pathologies discussed here do not usually involve overt fraud, they nevertheless cause damage to individuals, to science, and to society as a whole. Even though the costs arising from any single publication which exemplifies one or more of these three problems are likely to be small, they occur so widely across all areas of science that their total hidden costs may even exceed all those arising from overt fraud.
The three types, in order of increasing implicit costs, are: (a) duplicate or multiple publication of the same or nearly the same work, (b) serial publication of many papers on the same subject which contribute little or nothing new not contained in the first paper in the series or which divide a complete paper into several smaller parts, and (c) publication of work implicitly or explicitly claimed to be new without adequate reference to the existing relevant literature in the field.
The ethics guidelines of the American Physical Society are relevant to these three generic problems. In part, the guidelines say, “It should be recognized that honest error is an integral part of the scientific enterprise. It is not unethical to be wrong, provided that errors are promptly acknowledged and corrected when they are detected. Professional integrity in the formulation, conduct, and reporting of physics activities reflects not only on the reputation of individual physicists and their organizations, but also on the image and credibility of the physics profession as perceived by scientific colleagues, government, and the public.”
Multiple publication is worst when two or more virtually identical papers are published in different, readily available journals covering the same field. Journal space is thus needlessly preempted without the likelihood of a much increased audience. Thus, costs are increased, and science is diminished. Although it is difficult for editors and reviewers to recognize multiple publication when it is nearly contemporaneous, this author’s artifice to increase number of publications is explicitly prohibited by most journals and so is fairly rare — but it still happens. I believe that multiple publication should not be considered particularly reprehensible, however, when the first paper appears in a low-circulation conference proceedings or when several versions of a paper are published in different journals with small readership overlap.
Because a premium is set by some editors and referees on short papers, the probability of acceptance of a short paper may be higher than that for a long one. Although many authors would probably prefer to write a complete paper on their subject and findings, they can increase their publication count by dividing the work into an ongoing series of short parts, LPU’s (least-publishable units). Furthermore, they can publish the first part sooner, rather than have to wait to complete the unity of the work. Finally, such early but incomplete publication also helps one claim priority for new ideas and results.
Serial publication of this kind is thus advantageous for the author but often occurs at the expense of readers and society. First, the reader must initially judge the value of the work only on its first stages and must wait, sometimes for years, for its completion. Since such a series of parts must have connecting and background material in each part, its totality will nearly always be appreciably greater than that of a single longer paper covering the same material. Sometimes seriality can be carried to absurd lengths. I know of an example in which at least eleven papers, published over a seven-year period, were devoted to the analysis and discussion of the same solid-state data set. One or two would certainly have been sufficient. Unfortunately, because of the advantages mentioned above of serial publication, seriality is not likely to diminish significantly in future. Of the present three pathologies, the most damaging and costly one is inadequate or missing attribution and unjustifiable claims arising from such inadequacies. Sometimes a paper which claims new results actually re-does, by the same or not significantly different methods, work already published by others. If citation of the earlier work is included, this is perfectly acceptable. If the overlap is great, however, a responsible referee would likely recommend rejection of the manuscript. If the lack of attribution is based on ignorance of the earlier work by the author and by the referees of the paper, no serious scientific misconduct is present, provided the author moves quickly to acknowledge omissions and to withdraw unjustified claims of novelty once they have been discovered.
An example of the last situation was recently discussed by A. K. Finkbeiner . A distinguished member of the Princeton Physics Department, in speaking about why his 1965 prediction of about 10 K for the cosmic background radiation temperature did not cite the 5 K 1948 prediction of Alpher, Gamow, and Herman, said “I like to decide first what I think and then find out what other people think. How do you know what you think if you consult other people?” Nevertheless, he added that he should have tracked down and cited the earlier work. Many of us can sympathize with the desire to create knowledge independently of the work of others. But, once completed, the obligation remains to fit it within the existing framework of the field.
Even in the honest-error case, there is possibly a large cost borne by both the authors of earlier pertinent work and by society as a whole. First, published corrections do not reach all readers of the new paper, either near the time of the corrections or into the indefinite future. There is thus a good chance that the new work may become the standard, and only it will be referenced in future – causing obvious damage to and erosion of the intellectual property of the authors of the earlier work.
The cost to society can likewise be considerable. Suppose that the new work is the product of a research group supported by government grants or contracts, the usual case in academic science today. The referees of the original proposal to carry out an investigation in a particular field clearly did not cite the earlier work, which, in extreme cases, may have already achieved all the goals proposed for support. Thus the support is approved, including summer salary for the principal investigator and salaries for at least several postdoctoral and graduate students. After two or three years of such support, the paper describing the work appears and someone points out that it has all, or nearly all, been done before.
In this extreme, but nevertheless realistic example, nearly all the support costs were wasted, and, given adequate initial information, they could have been devoted instead to truly new investigations. Since research support funds are extremely limited, I believe that the present scenario depicts a true example of scientific misconduct, primarily by the principal investigator. By not carrying out an adequate initial literature search, for example by means of an inexpensive search through one or more of the many currently available computer-based scientific data bases, the principal investigator has wasted a great many thousands of dollars of public money and has failed in discharging his obligations to his research group and students.
In the worst case, that of outright fraud by plagiarism, the authors of the paper knew about the earlier work, but they probably considered it sufficiently deeply buried in the literature that they could plagiarize it with impunity and, anyway, they needed another publication or two! LaFollette , in a book highly relevant to the present subject, has quoted characterizations of scientific plagiarism as possibly worse than outright fabrication of data and as “malignant.”
Unfortunately, one cannot usually distinguish this criminal approach to authorship from the honest but lazy one described above. But a few guidelines may be helpful. The situation is most likely to involve fraud if one or more of the following characteristics apply. First, the authors may refuse to acknowledge the priority of the earlier work and may even refuse to state whether or not they knew of it before writing their paper. Second, if one finds that the same authors have published several other papers on different subjects, again with inadequate or entirely missing attribution to earlier relevant work, one can probably reach correct conclusions about their general mode of operation: superlatively creative or just criminal.
Finally, very tardy acknowledgment of citation errors and incorrect claims may suggest fraud, but it probably is only characteristic of the type of person who doesn’t do his homework in the area of investigation before publishing its results. I am aware of a case where the authors of a paper based on work supported by the federal government claimed 12 “firsts,” in a single paper, many of them actually “seconds!” These authors did not publish corrigenda for well over two years after being made aware of the earlier relevant work. By such delay, they knowingly or unknowingly prolonged the period during which credit for their work could be used to aid in personal advancement.
There are, unfortunately, some authors who continue to publish papers involving ideas or methods that they or others developed earlier, without such authors acknowledging that the earlier ideas or methods have been superceded by new work in the field, even when they are aware of it. By not referring to such work, or not identifying and crediting corrections when they do make them, these authors do not alert a referee, who may be unaware of the earlier work or of the fact that corrections are being made, thereby increasing their chances of publication but, concomitantly, damaging the integrity of the scientific record and hampering the advancement of science.
The practice of claiming “firsts” and novelty is a dangerous one and should be totally eschewed. For one thing, it is unnecessary since acceptance for publication is itself evidence of novelty. Goethe said, “No one can take from us the joy of the first becoming aware of something, the so-called discovery. But if we also demand the honor, it can be utterly spoiled for us, for we are usually not the first.” Let the marketplace make the judgment and thus avoid possibly embarrassing self-aggrandizing claims.
Following my own precepts, I wish to state that it is likely that everything said here has been said before in other ways, and I certainly do not claim priority for any of it. There is thus no need for anyone to cite this work. But if it helps reduce some of the pathologies discussed above, these words will prove worthwhile. In conclusion, the following words of Francis Bacon in the Preface to his Maxims of the Law, should be a lifelong guide to all professionals:
“I hold every man a debtor to his profession; from the which, as men of course do seek to receive countenance and profit, so ought they of duty to endeavor themselves by way of amends to be a help and ornament thereunto.”
- Responsible Science – Ensuring the Integrity of the Research Process, Volume 1. Panel on Scientific Responsibility and the Conduct of Research, E. E. David, Chairman. National Academy of Sciences, Washington, 1992.
- M. C. LaFollette, Stealing into Print – Fraud, Plagiarism, and Misconduct in Scientific Publishing, University of California Press, Berkeley, 1992.
- A. K. Finkbeiner, The Sciences (New York Academy of Sciences), September/October 1992, p. 9.