June 21, 1988. /p
a

To the Editor: ee

Nicholas Wade's editorial (June 13), chiding scientists for their
unhealthy claim of infallibility, attacks a straw man. Scientific research
does eventually yield objective, reliable knowledge about the external world;
but at the growing points it is an inefficient, often fumbling process. That
is why experienced scientists do not get excited about uncorrected errors in
the literature. Errors are inevitable, both in observation and in
interpretation. Only the most important ones deserve formal retraction or
refutation, while the others are bypassed. It is Mr. Wade who seeks
infallibility, in his demand for a pure, consistent literature.

Mr. Wade further suggests that if researchers had to make their data
freely available to others, without challenging motives or credentials, "at
might encourage them to report their data objectively". Hut apart from the
question whether any self-appointed inspector has a lien on a scientist’s (or
an editorial writer‘s) records and time, the very concept of reporting data
objectively is simplistic. Most experiments do not work, and the data are
likely to have a good deal af inconsistency until the experiments are
improved, The scientist must therefore exert judgment in deciding which data
‘to publish.

The editorial also discusses a recent.Congressional hearing chaired by
John Dingell. This hearing deserves a more critical analysis, for along with
some cases of well authenticated fraud it has set a dangerous precedent by
considering a new accusation of improper publication, based on a dispute
between a young investigator and her mentors (including a Nobelist). Several
senior investigators, whom she consulted, all concluded that the disputed

paper fell well within the range of acceptability; any error that might be
present would later be settled by the results of further experimentation.
Most researchers in training would have accepted this resolution, but this
one, with a strong conviction about purity in the literature, did not; hence
the issue finally landed in the hearing.

Mr. Wade recognized that Congressional hearings are not the ideal forum
for adjudicating scientific claims. But this is quite an understatement. For
one thing, we should not devalue whistle-blowing pver fraud by confusing it
with disputes over differences in judgment. (This issue is even more
important for scientists than for congressmen, because fraud involves not anly
taxpayers’ money but the very foundations of the search for objective
knowledge.) Second, by publicizing only one side of this dispute the hearing
damaged reputations without due process -- and due process is a precious
tradition even if it is not legally required of Congressional hearings.
Finally, by discussing only fraud and incompetence, and ignoring the ambiquity
that is inevitable in even the best research, both the hearings and the
editorial have given the public a distorted picture of the scientific process.

To laymen the apparently casual attitude of scientists toward honest
error may seem sloppy. But it has evolved as a pragmatic, effective solution
-- with passionate battles, sometimes lasting years, over contradictory
findings or interpretations. find despite the resulting imperfect literature,
biomedical research is now flourishing as never before. It has done so
through generous external support and through the tradition of autonomous
regulation within the scientific community. To be sure, some institutions
have treated real cases of fraud too casually, and so legislators are
justified in pressing for a more vigorous response. But we must also preserve
the tradition of trust and openness in science. We will not advance the field

by treating the rare cases of fraud as a major problem, by using the heavy
hand of legislation or of governmental regulation to try to contain fraud, or

by confusing it with normal ambiguity.

Bernard D. Davis