|
Science is the search for truth. Its tools are rationality, objectivity, experimentation, and the free exchange of reliable information. But what happens when a scientist reports unreliable or fraudulent information? How common is fraud in science? Can science distinguish the fraudulent from the genuine? What steps are scientists taking to police themselves? Science journalist Christopher King reports on these questions and other controversies surrounding fraud in science.
By Christopher King
It was supposed to be the greatest archaeological discovery in history. Approximately 90 years ago, an amateur geologist uncovered fragments of skull and jaw bones in a gravel pit on Piltdown Common in Sussex, England. At a scientific meeting in 1912, the assembled pieces, thought to be nearly a million years old, were presented to the world: 'Piltdown Man,' a humanlike skull with a large, apelike jaw. Here was proof of the prehistoric 'missing link,' the crucial step on the evolutionary progression from primitive apes to modern humans. And, best of all, he was British!
 |
|
Also on Encarta |
|
 |
|
|
|
|
Despite the excitement, however, many scientists doubted the authenticity of the skull. Their suspicions were finally justified in the 1950s, when modern chemical analysis showed that the bones were actually from two different sources: The jaw belonged to an orangutan and the skull to a human—neither of them much more than 600 years old. Some modern trickster, after carefully filing down the fragments and staining them to make them appear ancient, had planted them in the gravel pit to be 'discovered.' In the decades since the Piltdown scandal, historians have debated whether any of the scientists involved were simply tricked or if they actually took part in the forgery. But the perpetrator of the hoax has never been indisputably identified. Piltdown Man may have lost his place in the annals of paleoanthropology, but he lives on as perhaps the most notorious example of scientific fraud in history. Unfortunately, this was the not the first instance of scientific misconduct ever recorded—and certainly not the last.
According to a traditional view that goes back at least as far as the time of the Italian scientist Galileo Galilei some 500 years ago, the process of science is governed by rationality, logic, and truth. The scientist carefully and objectively observes, collects, and classifies information, then formulates a hypothesis in order to explain the data and to predict what might happen under various conditions. The scientist also performs experiments to test the hypothesis. Depending on the outcome, the hypothesis may be expanded, revised, or completely rejected. If the hypothesis proves sturdy enough to withstand a series of experiments, a scientist might develop a broader set of explanations and predictions known as a theory. In turn, even theories are subject to modification or replacement as new knowledge accumulates.
In science, an essential form of communication is the scientific paper—a detailed summary of an experiment, published in a specialized journal for fellow scientists around the world to read. Several steps ensure the integrity of the scientific paper. Before publication, the journal's editors typically send the paper to referees—experts in the field who evaluate the quality of the data and the soundness of the paper's conclusions. And each scientific paper includes specific information on how the work was done, in sections discussing materials, methods, and so forth. This enables any other scientist to perform an identical experiment—a process known as replication—in order to verify the results. Refereeing and replication are two elements that allow science to correct itself—to ensure that not only are honest errors corrected, but also that instances of deliberate cheating are promptly exposed. That, at any rate, is how the process is supposed to work.
|
|
Also on MSN |
|
|
|
|
|
|
Scientists, of course, are human beings. An important scientific achievement—a discovery, a cure, or some other breakthrough—can bring prestigious awards, worldwide recognition, and lasting fame (not to mention financial gain). The prospects of these kinds of rewards can be powerful motivators. Just as every teenager with a guitar dreams about writing a hit song, and every aspiring actor imagines accepting an Academy Award, it is likely that many scientists daydream about traveling to Stockholm, Sweden, to accept a Nobel Prize for a scientific breakthrough.
Daydreams aside, there is intense competitive pressure in science to be first to achieve some significant result—a distinction referred to as priority. 'Credit in science goes only for originality, for being the first to discover something,' write science journalists William Broad and Nicholas Wade in their book Betrayers of the Truth. 'With rare exceptions, there are no rewards for being second.'
A scientist must also be concerned with a career—with job placement, promotion, and obtaining funds for research. 'Publish or perish,' goes the old saying, and for many a scientist trying to build a career in a competitive world, it is barely an exaggeration. In recent years, competition in science has created pressure to stand out from the crowd by having a long list of published papers to one's credit—preferably, papers in prestigious journals reporting significant findings.
Science itself has changed in recent decades. Before World War II (1939-1945), the federal government provided minimal support for science. Now, through such agencies as the National Institutes of Health (NIH), the National Science Foundation (NSF), and the Department of Defense (DOD), the United States government funnels upwards of $70 billion annually into scientific research at universities and other institutions. Competition for federal dollars provides yet another source of pressure on scientists to produce noteworthy results. And still another financial lure lies in patents for drugs, chemical compounds, new materials, and other fruits of basic research. It is not uncommon for scientists to be employees (or even founders) of companies in such areas as biotechnology and pharmaceuticals. This raises the possibility of a conflict of interest in which an objective report, for example on the results of a drug trial, might jeopardize the financial success of the company.
In legal terms, fraud usually refers to a deliberate, knowing attempt to misrepresent facts or otherwise deceive someone in order to win advantage or gain at the other person's expense. In the world of science, however, defining fraud or misconduct is not so clear-cut. Certainly, fabricating data would count as fraud—describing results of experiments that were never performed or listing clinical observations of test subjects who do not exist. An equally serious activity would be changing or falsifying existing data in order to achieve some predetermined goal. And plagiarism, knowingly stealing someone else's published work and presenting it without attribution as one's own, is another unmistakably fraudulent act.
A whole range of behavior, however, falls somewhere between the scientific ideal and outright misconduct. Scientists, being human, simply make mistakes sometimes. A radio astronomer, for example, detects evidence of some dense object at the outer reaches of the galaxy; the published paper describing the observation creates a fanfare. Later, however, it is discovered that the signals came not from outer space but from a malfunction in the astronomer's own equipment. An unfortunate mistake, certainly, but it is not misconduct.
Scientists can also deceive themselves. The Piltdown hoax provides an example, with a cast of British scientists all too eager to accept evidence that modern humans arose on British soil, instead of in Africa. Sometimes the desire to achieve a certain result may blind scientists to weaknesses in procedure. Behavioral scientists studying communication between humans and animals, for example, may fail to realize that their own head movements and body postures are providing subtle cues to the animal subjects.
Other mistakes might reflect sloppiness, carelessness, or poorly designed experiments—a mislabeled graph in a paper, for example, or statistics incorrectly copied from a lab notebook. Poor science, however, does not necessarily constitute fraudulent science.
Perhaps the most difficult aspect of scientific fraud to pin down is the manipulation, or massaging, of data. As far back as 1830, the British mathematician Charles Babbage described techniques of data manipulation that were already well known to scientists. These included data trimming (deleting data that do not fit in with the desired result) and cooking (selectively reporting observations so as to make the data appear more convincing). The idealized view of science holds that scientists are objective and will report all results completely and honestly. Unfortunately, this is not always the case. There is a difference between a scientist who simply cleans up data for presentation, removing numbers that may have resulted from statistical quirks or other random factors, and a scientist who knowingly deletes essential data in order to get a graph to go in the 'right' direction.
Suspicion of misconduct, or at least of highly questionable behavior, reaches back across several centuries, touching some of the greatest names in science. For example, as Broad and Wade recount in Betrayers of the Truth, the 2nd-century astronomer Claudius Ptolemy appears to have engaged in unethical practices. Ptolemy's theories about the positions of the planets and other aspects of astronomy were influential for more than a thousand years. Yet many historians, upon examining Ptolemy's work and comparing it with that of his contemporaries, now believe that Ptolemy did not make the celestial observations he claimed. Instead, the evidence indicates that he appropriated the observations of an earlier astronomer, Hipparchus of Rhodes.
Other immortal names have been tainted. The 17th-century scientist Galileo, for instance, has generally been regarded as an outstanding example of the rational, thorough experimenter. After attempting to recreate some of his investigations, however, some historians doubt whether Galileo actually performed all the experiments for which he is famous. His real powers may have been in persuading people of the truth of his theories, rather than in demonstrating it through experimentation. Still another giant—British physicist Sir Isaac Newton—has come under suspicion for manufacturing data to fit his theories, as well as for actively trying to discredit his scientific competitors in order to secure his own place in history. Newton was certainly not alone in this preoccupation. As the American science sociologist Robert Merton has written: 'The fact is that almost all of those placed firmly in the pantheon of science—Newton, Descartes, Leibnitz, Pascal, or Huygens, Lister, Faraday, Laplace, or Davey—were caught up in passionate efforts to achieve priority and to have it publicly registered.'
Even the work of Gregor Mendel, the 19th-century monk whose experiments with peas gave rise to the modern field of genetics, has been called into question. Modern researchers examining Mendel's work have noted that the statistical precision reported in the results of his pea-breeding experiments appears too neat to be true. Historians lack sufficient evidence, however, to determine the extent to which Mendel may have doctored his results.
The thread of scientifically questionable work extends into the modern era as well. Some of this work, before being reexamined, had a significant influence on social policy. The case of British psychologist Cyril Burt during the middle part of this century provides one example. Burt supposedly carried out 'nature vs. nurture' experiments—in some instances, on twins who had been raised apart—in support of the theory that intelligence is primarily inherited at birth rather than a trait that can be affected by upbringing or other factors. Burt's findings on intelligence quotients (IQ) were very influential in British educational policy; an IQ test administered to 11-year olds, for example, determined whether a child would be placed in an academic or vocational school. During his life, Burt was regarded as one of the United Kingdom’s greatest psychologists. It was not until after his death in 1971 that scholars reviewed his data and found major flaws. In many instances Burt appeared to have simply fabricated data, describing subjects who never existed. He also apparently borrowed older data from obscure sources, liberally modifying figures to suit his aims. Today, Burt's work has been almost entirely discredited. But troubling questions persist. Why were the obvious fabrications in his work not caught prior to publication? How did the process of referee review fail?
Another factor that has complicated the detection of fraud in recent years is the sheer size of the scientific literature. With more than 40,000 scientific and scholarly journals around the world publishing hundreds of thousands of papers per year, even if every paper is peer reviewed, not all of them can be thoroughly inspected for scientific integrity. This problem contributed to the infamous case of Elias A. K. Alsabti in the late 1970s and early 1980s. Alsabti, an Iraqi national, received medical training in his native country and came to the United States in 1977 in search of a research job in his field of immunology. First at Temple University in Philadelphia, and at several institutions subsequently, Alsabti was caught in acts of misconduct. In at least one case, coworkers observed him making up data for a paper. In several other instances, Alsabti committed acts of plagiarism, retyping existing papers from journals, and resubmitting them in the form of manuscripts to more obscure journals. In some cases, alert readers happened to catch the plagiarism, while at other times Alsabti's own carelessness tripped him up when he neglected to remove telltale signs of a paper's original author. When these acts were discovered, Alsabti was confronted and dismissed by his employers. However, over the course of several years he managed to stay one step ahead of his reputation, acquiring positions at six different institutions and achieving licenses to practice medicine in two states. As Marcel C. Lafollette notes in the book Stealing Into Print, none of the affected institutions ever launched a formal investigation into Alsabti's conduct. Instead, the journals Nature and the British Medical Journal publicized his misdeeds. Alsabti ultimately lost his medical license in Massachusetts, and his reported death in an automobile accident in 1991 ended the case.
Another case from the early 1980s, that of John Darsee, also highlighted serious problems in the process by which scientists publish research. Darsee was a young cardiologist performing research at one of the world's most prestigious institutions, the Harvard Medical School in Boston, Massachusetts. By all accounts he was bright, capable, and motivated enough to have published nearly 100 papers and abstracts in his two years at Harvard. In 1981 three of Darsee's coworkers reported to the head of the lab that they had witnessed him fabricating data for an experiment. They also voiced their belief that he had faked data on several other papers. When confronted, Darsee admitted to a single act of fabricating information. Subsequent investigation ultimately discovered that Darsee had faked data not only at Harvard, but in his previous position at Emory University in Atlanta, Georgia, and even as an undergraduate at Notre Dame University in South Bend, Indiana. Darsee was dismissed from Harvard and barred from receiving government funds for research. Journals that had published the falsified papers printed retractions.
One of the many troubling aspects of the Darsee case was the fact that he often listed several coauthors on his reports, including the names of senior faculty. If these scientists were not sufficiently involved with the research to be aware of Darsee's misconduct, why would they allow themselves to be listed as authors? Some critics questioned whether the practice of 'honorary' or 'gift' authorship—awarding credit to fellow scientists who actually did little or no work on a paper—could be considered another possible form of misconduct. Similar questions were raised at about the same time in connection with Robert Slutsky, a researcher at the University of California, San Diego. Slutsky came under suspicion when faculty members on a review committee happened to compare two of his papers side by side and noticed irregularities in the data. In a subsequent review, a committee examined 147 of Slutsky's papers, determining that nearly half the papers were questionable at best. In an echo of the Darsee case, many observers criticized the practice of honorary authorship, wondering how Slutsky's coauthors could sign their names to a paper when they clearly had incomplete knowledge, and little apparent concern, as to the integrity of the data.
Other cases of misconduct in the 1980s involved financial conflicts of interest. As Robert Bell recounts in the book Impure Science, controversy surrounded an article published in the Journal of the American Medical Association describing the effectiveness of the skin treatment Retin-A, a cream for reversing skin wrinkles. Financial records, obtained after an investigative article appeared in Money magazine, showed that some of the scientists who took part in tests of the treatment had received research grants and other sums of money from Retin-A's manufacturer. Another case involved Scheffer C. G. Tseng, an eye researcher at the Harvard-affiliated Massachusetts Eye and Ear Infirmary. Tseng, who had formulated a drug to treat dry eyes, was investigated for improperly testing the drug on hospital patients, and for making claims about the drug that were not supported by data. During these tests, Tseng helped form a company to market the drug, making large sums of money in sales of company stock.
In March 1981 the House Committee on Science and Technology held hearings on scientific misconduct; the sessions were chaired by then-Congressman Al Gore, Jr., of Tennessee. As a result, the federal government took a more active role in policing scientific integrity. A law, passed in the wake of the Darsee case, required federally funded institutions to report incidence of fraud without delay. In 1989 the Department of Health and Human Services founded the Office of Scientific Integrity (OSI), an investigative body within the NIH assigned to deal with cases of misconduct.
In the mid-1980s government and science collided in a long, bitter, and extremely complex investigation of scientific misconduct that came to be known as the 'Baltimore affair.' The case began in 1986 at Tufts University in Boston, Massachusetts. Margot O'Toole, a post-doctoral researcher in the laboratory of immunologist Thereza Imanishi-Kari, was unable to replicate key data that the team had just reported in a paper published in the journal Cell. Along with Imanishi-Kari, one of the paper's authors was American molecular biologist David Baltimore, winner of the 1975 Nobel Prize in physiology or medicine. O'Toole expressed her concerns to Imanishi-Kari and Baltimore, but the senior scientists assured her that the data, and the paper's conclusions, were sound. Refusing to let the matter drop, O'Toole ultimately took the matter to the NIH, which launched its own investigation. As the years dragged on, the matter also came to the attention of the House Oversight and Investigations Subcommittee, chaired by the combative Michigan Congressman John Dingell. Even the U.S. Secret Service got involved; the agency's document experts analyzed Imanishi-Kari's laboratory notebooks to try and determine if pages had been falsely back-dated and inserted to cover up weaknesses in the data.
The case took a toll on everyone involved. Baltimore, who stood by Imanishi-Kari, was never charged with wrongdoing but saw his reputation tarnished by the case. In 1991, unable to escape the controversy, he reluctantly resigned his post as president of Rockefeller University in New York City. Earlier that year, when the Secret Service analysis of the lab notebooks seemed to indicate misconduct on Imanishi-Kari's part, he quietly requested that Cell retract the disputed paper.
And Margot O'Toole, the 'whistleblower' whose concerns had launched the case, found herself exiled by the scientific community. Her research fellowship at Tufts was not renewed, and she was unable to find a scientific position for several years. Were her difficulties a result of her whistleblowing? No one knows, but her difficulties seem similar to those faced by some others who have exposed cases of misconduct.
The Baltimore case wore on. In 1991 the OSI announced its finding that Imanishi-Kari had committed fraud. In 1996, however, ten years after the case began, the decision was reversed by a federal panel: Imanishi-Kari was cleared of any misconduct. In 1997 Baltimore became president of the California Institute of Technology. Criticism now turned against weaknesses in the Secret Service analysis of the lab notebooks, and against Representative Dingell and the strong-arm investigative tactics of the OSI. Critics, among them Bernadine Healy, who had been NIH director from 1991 to 1993, expressed dismay over what they viewed as abuses of power by Dingell and his fellow investigators. Before her appeal, for example, Imanishi-Kari had never been formally presented with the evidence against her, nor had she been allowed to cross-examine witnesses.
In 1992 a new office, the Office of Research Integrity (ORI), was created to replace the OSI. Although its image suffered as a result of the Imanishi-Kari acquittal and reversals in other cases, the office continues to investigate fraud. The ORI provides specific guidelines for government-funded institutions to investigate and deal with allegations of misconduct. Protecting whistleblowers is also a priority. The ORI received about 1500 allegations of misconduct in its first five years and pursued about 90 cases a year.
But the problem of fraud persists. In 1996, for example, Francis S. Collins, the highly regarded scientist in charge of the U.S. government project to map the entire human genome, announced that a graduate student in his own laboratory had been caught falsifying data in research involving the genetics of leukemia. As a result, five papers had to be retracted. Collins, who personally investigated the matter when the student first fell under suspicion, told reporters that it is virtually impossible to protect against an intelligent, resourceful person who is intent on committing fraud. And the problem is certainly not confined to the United States. In 1997 a German investigative committee uncovered evidence that two biomedical scientists had falsified data in as many as 37 publications between 1988 and 1996. And in 1995 Stephen Lock, editor of the British Medical Journal, wrote in an editorial, 'the time has come for Britain to abandon its lax approach to scientific fraud.'
Despite the considerable attention paid to scientific misconduct, no one knows for certain how extensive the problem is. A 1990 survey of 4000 doctoral students in graduate departments at 99 American universities indicated that 44 percent of students and 50 percent of faculty were aware of two or more types of misconduct and questionable research practices. Between 6 and 9 percent of the respondents reported direct knowledge of faculty plagiarizing. Twenty percent of chemistry doctoral students reported falsification of data by their peers. In Betrayers of the Truth, Broad and Wade speculate that for every major of case of fraud that comes to light, perhaps 100 go undetected, and that for each major case, another 1000 lesser violations are committed. Aside from such guesses, however, or indications in surveys, exact figures do not exist.
Whatever the precise dimensions of the problem, the scientific community is at least taking some action. Courses at such institutions as Indiana University in Bloomington, the University of Pittsburgh in Pennsylvania, and Stanford University in California now promote ethical conduct in research on the part of students and faculty. Different methods of evaluating researchers—placing less emphasis on the number of published papers—have also been discussed.
Undoubtedly, most scientists are honest, and most scientific research derives from sound, ethical methods. Many scientists and administrators believe that, despite the high-profile fraud cases and disastrous publicity of recent years, the ongoing process of science still corrects itself, ultimately exposing and negating fraud and poor research. The exact number of dishonest scientists, however, and precisely how much their actions affect the reliability of research as a whole, must be counted among the many unknowns that science faces every day.
About the author: Christopher King is the editor of ScienceWatch, a newsletter that tracks trends and performance in basic research, published by the Institute for Scientific Information in Philadelphia, Pennsylvania.
Appears in
Medical Ethics; Science; Fraud
|
Hotmail
Messenger
My MSN
MSN Directory
Appears in