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

Princeton U. Press
Princeton, NJ
, 2010. $22.95 (168 pp.). ISBN 978-0-691-13966-1

David Goodstein discusses the science enterprise and its strengths and weaknesses in his latest offering, On Fact and Fraud: Cautionary Tales from the Front Lines of Science. In this short book, he draws from a series of anecdotes, many of which are based on his extensive experiences as an experimental physicist and as a vice provost at Caltech. As vice provost, Goodstein handled the cases of faculty members accused of scientific fraud.

On Fact and Fraud, the result of a course the author taught on scientific ethics, is easily accessible to physicists of all disciplines. Goodstein starts by pointing out the flaws in some of the reasonable ethical principles suggested for the practice of science. He addresses the federal government’s attempt to define scientific fraud, then presents the accepted definition of scientific fraud and the three psychological characteristics that most cases contain. Later, he discusses at length the histories of high-temperature superconductivity and cold fusion as examples of the range of episodes in the science enterprise that reveal surprising, courageous, and sometimes misguided actions.

The book’s primary strength is the telling of those individual stories, which the author does with keen insight and understanding. Of particular value is the history of Robert Millikan’s measurement of the charge on the electron—Millikan has been accused of fraudulently manipulating his data. The author analyzes that accusation in depth and comes to a persuasive conclusion that fraud was not committed. However, I strongly disagree with some of the general principles the author draws from other histories. In particular, I take issue with two of his stances.

The first involves reproducibility in experimental measurements, a principle that the author seems to consider of secondary importance. He takes a sympathetic view toward scientists working in cold fusion—a prime example of a field characterized by unverified results. By contrast, I strongly believe that reproducibility of experimental measurements should be at the heart of science. Simply put, if an experimental group cannot reliably reproduce a result in its own laboratory, it should never submit the result for publication in a scientific journal, and no scientific journal should ever publish a paper based on such unverified experiments. That is more than a philosophical difference I have with the author. There are serious consequences to publishing unverified results; for example, other scientists will read those papers and try to reproduce the results, almost always unsuccessfully. Those unsuccessful attempts are not only a waste of a scientist’s precious time and resources but, in some cases, can even end a scientific career.

The second point on which I part ways with the author is whether full disclosure is required in a publication. If, for example, collaborating researchers conduct five different experiments and four support their thesis, are they required to mention the fifth experiment that doesn’t? The author says that is too much to expect, but I think it is absolutely required. The reason is more than just honesty. That missing information could be invaluable to other scientists working in the field in helping them choose which experiments to run or how to interpret past ones.

Despite our differences, I strongly recommend Goodstein’s On Fact and Fraud to all physicists. The issues raised by this book are extremely important and need to be discussed. Also, physicists who read this review should also read Goodstein’s positions on the specific issues I’ve highlighted before they come to their own conclusions. If his book generates a healthy debate about those issues, it will have served a valuable service to the physics community.