George Walford: Scientifitricks
Science has been showing up badly lately, “dozens” of its practitioners in the US having been exposed as willing to stretch a point – or two, or three – if there was profit or kudos to be had. Things are no better in Britain, although stricter libel laws have restricted publicity, and the early scientists seem to have been no more scrupulous than their successors. Some of the greatest names carry the taint; Ptolemy, Galileo, Newton, Dalton and Mendel all doctored their data.
Undiscovered scientific cheating is like unreported crime in having to be estimated from indirect evidence, but some indication of its prevalence is given by the refusal of three out of four scientists to provide their original results for verification when requested. One common practise is, having learnt unofficially that another scientist has solved your problem, to publish his result as your own, faking data to support it; another, to select from your experiments only those giving the data desired. So long as they support the right answer, data are seldom queried since the famous process of replication, often supposed to guarantee the validity of all scientific work, is applied only to the most important findings; no credit is to be won by checking somebody else’s work.
The significance of all this for systematic ideology seems to be twofold. First, it confirms that earlier ideologies do not disappear as the more sophisticated ones emerge; expedience is only suppressed, not eliminated, in the development through domination to precision, and sometimes it bursts its bonds. Second, it confirms that science belongs to the eidostatic rather than the eidodynamic, its practitioners tending to hold together as a group rather than to display the ferociously competitive intellectual individualism found out towards the revolutionary and repudiative end of the range. [Data taken, without replication, from an article by Stuart Sutherland in the Observer 20 August.]
from Ideological Commentary 42, November 1989.