Several cases
of malpractice by scientists, the consequence of either intentional hoax or
sloppiness, have been discovered over the last forty years, and more have been
indicated by historians of science. Since the scientific enterprise is usually
regarded as the heaven of intellectual honesty, fraud in science strikes us as
a sort of blasphemy and it is an important task to discover its occurrences.
Were it only because of this reason, The Great Betrayal by H.F. Judson
would be worth reading; but indeed this book has greater merits: it is well
written, makes pleasant reading, and has extensive documentation.

The reader will be surprised to learn that such
famous scientists as Newton, Darwin and Pasteur (to name but a few) did
occasionally misreport their data in order to make their case more compelling.
Here and there Judson inclines to be sensational rather than critical. The case
against Darwin, by the admission of Judson himself, is a minor one (p.61-64),
is a good example: in "The expression of the emotions in man and
animals" Darwin resorted to the then pioneering art of photography to
illustrate the facial expressions associated to inner emotions. Some, or
perhaps most, of these photographs have been demonstrated to be artifice (as it
was common for the portraits of the time). Judson fails to acknowledge that the
photographs were not Darwin’s "data". Darwin’s data, right or wrong,
were long hours spent observing animals (and humans); photographs could have
been substituted for by paintings without touching in the least the scientific
value of his book.

The attempt
to indict great scientists of the past is a wrong move for Judson’s cause, for
it gives the impression that modern hoaxers stand on a par with Newton and
Mendel, which is quite the opposite of common sense. This impression is
strengthened by the cursory (and unargued) dismissal of the argument that the
standards of the XIX century were not those of today. Mendel was biased in his
analysis of peas by his theories (p.52-59), and today we would demand that he
asks a fellow monk, unaware of the crossings, to judge whether a pea is green
or yellow (i.e. that "blind" methods are resorted to). But in
Mendel’s times unconscious psychological biases were not fully appreciated and
blind methods had not been developed; yet his biases did not prevent him from
doing a beautiful job and from discovering the fundamental laws of genetics.

If the cases
of great scientists of the XIX century may prove misleading for the uncritical
reader, that of Freud is probably wrong. Freud cheated with his clinical
observations, declared cured patients who were almost as ill as before entering
psychoanalysis, accepted spurious or irrelevant data as proofs of his theory,
published his hypotheses on books destined to the lay public rather than in
scientific journals and so on. Such gross violations of scientific rules did
not escape notice of Freud’s contemporaries, and psychoanalysis was denied the
status of a science from its very beginning. Psychoanalysis may qualify as an
intellectual fraud, but surely never as a scientific one, unless we accept
Freud’s and nobody else’s judgment on its scientific nature. Indeed, several
authors maintain that psychoanalysis is a hermeneutical, rather than
scientific, enterprise, and as such is not at all fraudulent, but subjective
and true to its believers (the reader may refer to the works by J. Habermas and
P. Ricoeur on this point).

Only when one
comes to the recent cases of scientific fraud can the merits of Judson’s book
be fully appreciated, for his review is succinct and yet accurate, and poses
the right questions in an unescapable way. Judson demonstrates that scientific
fraud is not exceptional, on the contrary he proves it to be quite widespread,
and to be seldom handed properly by the competent institutions.

The frequency
of scientific fraud is difficult to determine, and Judson quotes estimates
between 3 and 30% (p.171), that seem both exceedingly high and exceedingly
variable. An obvious problem to solve, in my opinion, is the rigidity of the
definition of scientific fraud: "… fabrication, falsification,
plagiarism … does not include honest error…", according to the US
Public Health Service (p.172). This definition does not leave room for
involuntary but culpable errors: overstatement, poor statistics, incomplete
representation of the results and so on. Two recent and important cases of
culpable errors in science, committed by honest researchers who failed to run
the opportune control experiments and heavily trusted poorly reliable
measurements, are Benveniste’s water memory and Fleishmann and Pons’ cold
fusion; neither of these appears in The Great Betrayal, and the reason
for their exclusion is probably that they do not fit the definition of fraud.

Selection of
experimental data is an especially delicate issue. On the one hand selecting
for representation those experiments that conform to the scientist’s
expectations and discarding contrary ones is clear-cut falsification. On the
other hand discarding those experiments that the scientist believes are
contaminated or in other ways unreliable is part of the honest judgment of
every scientist. Often, doubtful or unexpected results are repeated several
times and accepted only if reproducible, whereas expected or confirmatory
results are more easily accepted. All scientists every now and then are
confronted with data of uncertain quality and may decide that these should not
be published; their decision in such a matter spans the whole range from
prudent and honest judgment to outright fraud (p.189), and culpable errors may
occur, perhaps frequently. If culpable errors of this type are counted as
frauds, the incidence of scientific fraud may be significantly overestimated.

An important
point raised in The Great Betrayal is the inadequacy of the peer
reviewing process in stopping frauds; and although on the pars destruens
of Judson’s critics is flawless, the pars construens leaves much to be
desired. Of peer reviewing, like of democracy, it can be said that it has
obvious defects, but no acceptable alternative is in sight. Judson suggests
that open refereeing (i.e. not anonymous) is superior to anonymous refereeing,
and that the availability of scientific journals on line for free is going to
improve the process.

Another
relevant point Judson addresses is the rules and roles of the funding
institutions, but here he goes astray and praises beginning-of-the-century
practices and munificent millionaires choosing the right men and the promising projects
out of their quasi-divine foresight, against state funding agencies and peer
reviewing of grant applications. Because of his stubborn faith in the methods
of the Rockefeller foundation, Judson gets caught in a futile paradox: the
worst case of scientific fraud reviewed in his book, that of T. Imanishi Kari
and D. Baltimore, although financed by federal fundings, appears to him as
based on Rockefeller methods gone astray: "That was an anomaly."
(p.253). The least we can say is that there is no obvious relationship between
funding and scientific fraud, and, pace Judson, no obvious corrective of the
latter by the former. Another defect of the present grant evaluation
procedures, and here we cannot but concur with Judson, is elitism: only top
scientists from top institutions pass scrutiny and emerging is very hard for
young scientists from provincial institutions. This selection against
potentially promising people is already imposing a heavy toll on the
recruitment of young scientists.

An important
evolution of the concept of authorship in scientific literature, that Judson
thoroughly reports and praises, is championed by those journals which require
that the published paper reports the specific expertise and contribution of
each author. This practice may prevent or diminish "ghost authorship"
(omission of a person who contributed to the paper but for some reason does not
want to appear) and "guest authorship" (presence among the authors of
people who did not contribute but is included for convenience or political
reasons).

The second
last chapter of The Great Betrayal bears the title "The rise of
open publication on the internet" and expresses the confidence that free
on-line journals will solve the problems of peer reviewing and reduce the
incidence of fraud. Here I must have missed something, because, granted that
having on-line journals is very convenient, the reasons why fraud should
diminish and reviewing become less relevant thanks to the web is by no means
clear. Electronic publishing is convenient and cheap, but why the shift from
paper to magnetic support should change or make unnecessary the refereeing
procedures? Why should it prevent fraud? Fraudsters will publish online as well
as they publish on paper and the larger audience guaranteed by e-journals will
not be at any advantage to detect frauds than the referees. After all, in
several cases listed in The Great Betrayal fraudsters were detected by
those refereeing procedures that Judson criticizes and hopes to dismiss.

Indeed, most
e-journals employ the same peer-reviewing procedures of their ink-and-paper
counterparts, and are as liable as those to all types of fraud except dull
plagiarism, that is more easily detected by computers. An unrefereed, globally
open system, as some advocate, would rapidly degenerate into a massive
collection of junk, where important contributions would be dispersed to the
point of being lost.  Free open
e-archives would probably grow faster and be of lesser quality: after all the
refereeing process sets a standard and many papers that in the present system
would be rejected, would appear in the unrefereed e-archives; more errors and
more frauds would be published, and would escape notice for longer times.
Judson does indeed present pro and cons of the open e-archives, but in the end
his judgment is that online e-archives would be more exposed to the control of
the community and this fact would discourage fraud.

A very interesting chapter of the book is devoted to
the comparison between the logics of laws and that of science, applied both to
the cases of forensic science and of trials of scientific fraud. This subject
is beyond my expertise, and I found Judson’s analysis interesting and
educative.

Overall, The
Great Betrayal is a good book on an important subject, but requires
critical reading, and it is to be praised because it is stimulating and
thought-provoking despite its flaws.

©
2005 Andrea Bellelli

Andrea Bellelli
has an MD and a degree in psychology, and teaches biochemistry in theMedical School
of the University of Rome, Italy.