Under the Microscope/Prof William Reville: The three main types of fraud in science are: fabrication, falsification, and plagiarism. Fabrication is the outright creation of observations or facts - in other words, telling lies.
Falsification means the adjustment of experimental data so that they point more clearly to the desired conclusion. Plagiarism means copying another's work without attribution.
Of the three types of fraud, falsification is probably by far the most common. Plagiarism is not unusual but, mercifully, outright fabrication remains rare. Indulgence in fraud is a capitulation to human frailty and it occurs in all fields of endeavour. The incidence of fraud in science has increased greatly in recent decades because of increased pressures in the system.
The process of science has an inbuilt mechanism for detecting error, and therefore any seriously incorrect descriptions of the natural world introduced by fraud will, sooner or later, be corrected. But although the potential for fraud to lead us into fundamental error is limited, serious fraud can delay scientific progress. More importantly, a significant incidence of scientific fraud could destroy the credibility of science in the eyes of non-scientists.
Proven examples of scientific fraud are relatively rare, but many undetected cases are probably published every year. For example, an American survey in 1993 found that about 8 per cent of scientists questioned were aware of results that had been plagiarised or falsified. In 1995 a Norwegian poll of 300 scientists found that 22 per cent were aware of "serious breaches of research ethical guidelines" in their communities. The consensus is that very few intentionally set out to falsify their research but intense pressures tempt many to cut research corners. The biomedical science industry is probably most at risk, with intense competition for breakthroughs in the lucrative drugs market. Promising results can cause a company's share price to soar.
Because science is a human activity it has always been attended by some degree of fraud. Even some of the greatest scientists who ever lived are suspected of indulging in limited fraud, including Ptolemy, Galileo, Newton, Mendel, Pasteur, Millikan, and others. Most of this fraud is suspected to have been falsification, where experimental data was either adjusted to better favour the hypothesis being tested or where awkward data in the form of observations that did not support the hypothesis were ignored. In some cases, plagiarism also plays a part. For example, the great Charles Darwin has been accused of failing to give credit to other workers who made valuable contributions to evolutionary theory.
Fraud in science is very difficult to detect because the great majority of the fraud will be in the form of falsification in work that supports the conventional scientific wisdom. The system is naturally trusting of such work and there is little incentive to question it. Yet if someone were silly enough to publish a revolutionary new theory supported by fabricated data, scientists from all over the world would eagerly examine the new theory hoping that it would open up new and fruitful avenues of research for themselves and, of course, would quickly expose the fraud and its perpetrator. Of course, fraud isn't the only route to hot water - as Voltaire said: "It is dangerous to be right on matters on which the established authorities are wrong."
In recent decades the developed world has frankly realised its dependence on science-based technology and the fact that future economic development depends on science. As a consequence, unprecedented money has been poured into science. Unfortunately, not much thought went into the creation and development of stable career paths for research scientists. Much of the money that goes into scientific research funds short-term contract research positions that last for a few years and then may or may not be renewed. Renewal of a contract position, or winning a new contract position when the current one expires, requires a record of high scientific productivity in the form of published papers. Scientific research is notoriously unpredictable. Sometimes you are lucky, it goes well, results come in quickly and papers are published quickly. At other times, despite your best efforts, the work goes slowly, results are difficult to get and papers are few and far between.
Now imagine you find yourself in the latter situation and you have just started the third year of a three-year research contract. You have only published one short paper on the work thus far and you know that to have any chance of having your contract renewed at the end of the year you will have to publish at least two more substantial papers. You and your spouse have just had your first baby and you hope to get your first house soon. The prospects of unemployment arriving in nine months is a nightmare. The temptations are obvious.
Restructuring research positions so that they offer better security of employment to researchers should greatly reduce the incidence of falsification fraud in science. Offering greater security will also serve to encourage more bright young people to study science with a view to pursuing a science career. I am convinced that a major reason for the decline in interest in science, particularly in physics and chemistry, among young people in recent decades is because neither they, nor their parents, are convinced that science competes well with other professional areas in offering stable well-paid careers. So, enhancing career prospects in science should improve both honesty and quality.
William Reville is associate professor of biochemistry and public awareness of science officer at UCC