Scientists have a macabre fascination with fraud. Some of the most famous recent cases – including those of the physicists Victor Ninov and Jan Hendrick Schön, who falsified results in nuclear physics and nanotechnology, respectively – have remained hot topics for years, continually generating investigations, articles and invited talks at scientific conferences. Part of the reason for this fascination is the damage created by fraud, which not only pollutes the scientific sea, but also causes other scientists, including graduate students, to pursue research in erroneous directions. That cost in time, funding and hampered or even destroyed careers is not even calculable.

Another part of the fascination is that the typical scientist, when confronted with clear fraud, often remains in denial. As scientists, we train ourselves to detect what Irving Langmuir called “pathological science”, in which practitioners park their scientific method outside their laboratories and replace it with wishful thinking. We have learned to review articles and listen to presentations while carefully considering any possible over-interpretations, including those that stem from ignoring data points that do not fit the theory the author or speaker “believes” to be correct. Being human, we have all been guilty of pathological science to some degree; being scientists, we rely on colleagues to guide our way to new knowledge and understanding through discussions, reviews and reproducibility. Actual premeditated lying – such as pulling data points out of thin air – is, however, in a different realm and is more consistent with sociopathic behaviour. We generally have no defence against this, and as history has shown, we are slow to detect it.

David Goodstein’s book On Fact and Fraud: Cautionary Tales from the Front Lines of Science offers a short and engaging education for those who want to know more about understanding and detecting true fraud. The author is well known in the field of condensed-matter physics, being recognized both for his research and for his outstanding, indeed seminal, 1975 textbook States of Matter. In 1988 he became the vice-provost of the California Institute of Technology (Caltech), where he pursued issues of science and society, with a focus on scientific misconduct. Not only does he have vast experience, but he also co-developed and taught a course on scientific ethics for more than 10 years. Perhaps most importantly, while at Caltech he drafted one of the first formal policies on scientific misconduct. His thoughtful and groundbreaking procedures have since guided many of our great universities in dealing with such difficult matters.

After discussing how to detect pathological science, Goodstein presents a set of 15 precepts on how science works, including the charge that scientists must “bend over backwards” to report any way in which they might be wrong. He then succinctly describes our scientific “reward system” – a mechanism for promoting star performers that is closely linked to the scientific “authority structure”. Goodstein defines this structure as an 11-step ladder that starts with being admitted to a prestigious college and culminates in prizes such as named professorships, membership in the National Academy and a Nobel prize.

The recipe for success becomes less well defined at each level, as the gatekeepers, who are generally a few rungs higher, gain ever more power over the rewards. As Goodstein explains, this system has its roots in the 17th century with the first experimental physicist (Galileo) and the first scientific research laboratory (created by Sir Robert Boyle). Although this reward system has many virtues, one of its faults is that it tends to benefit those who are already successful and necessarily lends itself to self-promotion, including the over-interpretation of results. This is clearly a less serious violation than fraud, but it nevertheless gets in the way of good science.

A particularly ingenious way that Goodstein educates us about detecting misconduct is by describing a case where accusations of fraud have turned out to be false. Robert Millikan was first to determine the charge of a single electron with his ingenious oil-drop experiments, which won him the 1923 Nobel Prize for Physics. The accuracy of Millikan’s results was not doubted until 1984, when the scientific research society Sigma Xi published a booklet entitled Honor in Science. This booklet (compiled by Sigma Xi executive director C Ian Jackson) calls Millikan’s work “one of the best-known cases of cooking” and claims that Millikan selected the droplets that gave the “right answer”. Goodstein takes us through Millikan’s published paper and notebooks, and based on their evidence, he exonerates him of scientific misconduct and other distasteful behaviour. In doing so, Goodstein reminds us that any inquiry into scientific misconduct must investigate all sides of the story, and must itself apply the scientific method.

Goodstein also describes unambiguous fraud cases in some detail. Of these, the Schön case is the obvious poster child. As eloquently described in Eugenie Reich’s book Plastic Fantastic (see Physics World May 2009 pp24–29, print edition only), Schön did not extrapolate results such as single-molecule transistors as if he were a victim of pathological science, but instead clearly fabricated data. Goodstein asks “Did he believe he knew the right answer?”, but in fact, it is so obvious that Schön committed premeditated fraud that the answer to that question is irrelevant.

Cold fusion presents a less clear-cut case. Although Stanley Pons and Martin Fleischmann’s 1989 breakthrough-that-wasn’t is often believed to be fraudulent, Goodstein puts forth arguments that both men were victims of extreme pathological science. He cites the Mössbauer effect as another implausible nuclear phenomenon; if that can exist, he argues, why should it be “obvious” that cold fusion cannot? In fact, although the Mössbauer effect (in which atoms bound in a solid absorb and emit gamma-ray photons without any recoil) was considered surprising at the time of its discovery in 1957, it was reproduced quickly, explained theoretically and remains an important tool in condensed-matter physics today. In contrast, cold fusion has been vigorously investigated for more than 20 years but has not been reproduced, nor explained theoretically. As a 1989 report from the US Department of Energy stated, “Nuclear fusion of the type postulated would be inconsistent with current understanding and if verified, would require theory to be extended in an unexpected way.” Still, Goodstein provides a very interesting historical account, noting that there were many mistakes on all sides. Evidently, this case remains under debate.

In contrast to cold fusion, the phenomenon of high-temperature superconductivity (HTS) is described as something that seemed too good to be true but was not. From the discovery of superconductivity in 1911 until 1986, the critical temperature for its onset, T_c, increased only incrementally, from about 4 K to just less than 30 K. During that period there were many claims of HTS, all of which were easily shown to be erroneous; in cases where the “proof” was not as simple as demonstrating a measurement error, other scientists’ inability to reproduce the supposed effect was certainly enough.

Then, in 1986, Georg Bednorz and Alex Müller reported that they had measured a T_c of about 40 K in a LaBaCuO compound. Their discovery was first made known to most of the community at a Materials Research Society meeting in Boston, where two other independent and eminent scientists, Ching-Wu (Paul) Chu from Houston and Koichi Kitazawa from Tokyo, reported similar findings. Many of those present were convinced enough to repeat the experiments. Within weeks, the results were being reproduced in dozens of laboratories worldwide. The following January, a new compound with T_c ~ 90 K was announced; it too was widely reproduced within a month. Today, any new claim of HTS is met with well-equipped and capable laboratory scientists all over the world, so if the claim is not reproduced broadly and quickly, then it is not taken seriously. We have learned from this field that the most important diagnostic for determining scientific fact is reproducibility.

Goodstein concludes his book with a clear summary of what we have learned. Perhaps most importantly, he includes, as an appendix, Caltech’s pioneering policy on misconduct. This is the policy foundation of how to deal with scientific misconduct, already adopted by many universities, and it should be required reading for scientific researchers and administrators.

Several excellent books and articles have been written on the broad subject of fraud in science. In addition to Langmuir’s seminal articles and Reich’s book mentioned earlier, the books Flim-Flam by James Randi, Voodoo Science by Bob Park and the short article “How to judge flawed science” (2005 MRS Bulletin 30 75) by Ivan Schuller and Yvan Bruynseraede have been particularly helpful in educating our community. Since scientific fraud is not going away, we need greater understanding and education to help us detect and deal with it. David Goodstein’s book fulfils an important need. This is a valuable book and one not to be missed.