On Fact and Fraud: Cautionary Tales from the Front Lines of Science

David Goodstein, professor of physics and applied physics and vice-provost at the California Institute of Technology, was responsible for developing Caltech’s research-governance policy, and has promoted research integrity through his course on scientific ethics.

The power of this book derives from his personal testament. Although I do not agree with all his views, the opinions expressed are delivered in a careful and thoughtful way.

Most books on research fraud start with a historical review and this text follows that model. Those familiar with the field will perhaps find it a little tedious to be reminded yet again of the misdemeanours of Sir Cyril Burt, William Summerlin and John Darsee, and the case of Piltdown Man. Perhaps irritatingly for those who do not know the field, these old chestnuts are not referenced, so such readers will be driven to a search engine to learn more.

Goodstein then takes us on a philosophical adventure through 15 diktats relating to research integrity. Much of this thinking can be found in published research codes, but not all. For instance, he dismisses the first diktat, namely that “a scientist should never be motivated to do science for personal gain, advancement or other rewards”, as simply impracticable.

He explores the dilemma of presenting “all of the evidence”, even that which is contrary to the investigator’s own hypothesis, at the same time acknowledging that, inevitably, there will be selectivity in the presentation of results. I particularly enjoyed his thoughts on the underlying motives that encourage scientists to commit research misconduct.

Goodstein relates the case of Robert A. Millikan in great depth. Millikan, a Nobel laureate in 1923, was a distinguished physicist and a Caltech founder, but until the age of 40 had “failed to make his mark” in the world of physics. However, between 1911 and 1917, he took centre stage in his field when he succeeded in measuring the charge of the electron.

In the paper announcing the discovery, he states clearly that he had reported all observations taken during the 60 consecutive days of the experimental period, but, on examination of his laboratory notebooks years later, it became clear that this was not the case. He had carefully selected the experiments that he finally reported.

Goodstein performs a forensic defence of Millikan, indicating that he selected what he judged to be the best and most representative experiments, and contends that even if he had included all of the data, the fundamental observation would not have changed materially. Millikan’s failing, therefore, was not that he was wrong or corrupt, but that he failed to describe precisely what he had done.

In the early 1990s, there was rising concern in the US about research misconduct and the need to establish a uniform approach to the definition of fraud and the investigation of allegations within the research community.

Goodstein touches on some of the difficulties during those years and uses examples of two high-profile cases in the laboratories of Robert Gallo and David Baltimore. Again, these are well-trodden paths in the world of research misconduct, but they serve as important reminders of how our approach to the investigation of the matter has matured during the past two decades.

Towards the end of the book, we are entertained royally by Goodstein’s retelling, peppered with personal anecdotes, of the cold fusion story. His account is generous to those concerned and suggests that rather than corruption or fraud, some scientists just “believed” that cold fusion had occurred, and that their experiments were not sufficiently robust to confirm or deny their assertions.

His final physics story is that of the moral maze of high- and low-temperature superconductivity. It is an example of the breakthrough that was not a breakthrough, but I shall leave you to discover the delights of the story and the concluding postscript, which involves an amusing joke about a pope’s response to the Second Coming that is not totally irrelevant to the field.

For Goodstein, what constitutes fraud is clear “fabrication, falsification and plagiarism”: the rest is just sloppy science. Personally, I think that selectivity in data reporting - which includes “trimming” information in individual experiments and the failure to report those that “haven’t worked” or do not support the prevailing hypothesis - may be, on a quantitative level, more important.

Serious research fraud is still relatively uncommon and usually does not cause serious harm, whereas selective reporting may have a much more insidious effect by undermining the scientific endeavour and potentially causing harm, particularly in the fields of health and biomedicine.

The strength of this slim volume is that the author knows and understands his subject well and can talk from experience and from the heart. Its publication comes at an important time as the UK Research Integrity Office prepares to move from being primarily focused on health and biomedical research towards a broader remit that will include the physical sciences and engineering, the arts and humanities and the social sciences.