Experimental science demands repeatability of results, in part because there are so many ways that experiments can go wrong. There are several famous experiments whose results were later retracted or discredited. The most common problem is simply overlooking an important source of noise or bias; such sources sometimes become apparent only with extensive experience with complex equipment or theories. Some errors are introduced when the experimenter's desire for a certain result unconsciously influences selection of data (a problem which today is avoided by double-blind protocols). There have also been cases of deliberate scientific misconduct.
A reported faint visual effect that experimenters could still "see" even when the supposed causative element in their apparatus had been secretly disconnected.
Kaufmann (1906) – claimed experimental disproof of special relativity
Published in Annalen der Physik and said to be the first journal paper to cite Einstein's 1905 electrodynamics paper. Kaufmann's paper stated that his results were not compatible with special relativity. According to Gerald Holton, it took a decade for the shortcomings of Kaufmann's test to be realised: during this time, critics of special relativity were able to claim that the theory was invalidated by the available experimental evidence.
A number of earlier experimenters claimed to have found the presence or lack of gravitational redshift, but Adams' result was supposed to have settled the issue ("definitively established", RWL "Relativity" )[clarification needed]. It is no longer considered credible, one of the more charitable interpretations being that his data may have been contaminated by stray light from Sirius A. The first "reliable" confirmations of the effect appeared in the 1960s.
Originally reported in Nature in 1955 and later. Diamond synthesis was later determined to be impossible with the apparatus. Subsequent analysis indicated that the first gemstone (used to secure further funding) was natural rather than synthetic. Artificial diamonds have since been produced.
In 1970 Joseph Weber, an electrical engineer turned physicist, and working with the University of Maryland, reported the detection of 311 excitations on his test equipment designed to measure gravitational waves. He utilized an apparatus consisting of two one ton aluminum bars, each a separate detector, in some configurations being hung within a vacuum chamber, or having one bar displaced to Argonne National Laboratory, near Chicago, about 1,000 kilometers away, all for further isolation. He took extreme measures to isolate the equipment from seismic and other interferences. But Weber’s criteria for data analysis turned out to be ill-defined and partly subjective. By the end of the 1970s Weber's work was considered spurious as it could not be replicated by others. Still Weber is considered one of the fathers of gravitational wave detection and inspiration for other projects such as LIGO.
Data from Fermilab in 1976 appeared to indicate a new particle at about 6 GeV which decayed into electron-positron pairs. Subsequent data and analysis indicated that the apparent peak resulted from random noise. The name is a pun on upsilon, the proposed name for the new particle and Leon M. Lederman, the principal investigator. The illusory particle is unrelated to the Upsilon meson, discovered in 1977 by the same group.
^Wood, R.W. (29 September 1904). "The N-Rays". Nature. 70 (1822): 530–531. Bibcode:1904Natur..70..530W. doi:10.1038/070530a0. After spending three hours or more in witnessing various experiments, I am not only unable to report a single observation which appeared to indicate the existence of the rays, but left with a very firm conviction that the few experimenters who have obtained positive results, have been in some way deluded. A somewhat detailed report of the experiments which were shown to me, together with my own observations, may be of interest to the many physicists who have spent days and weeks in fruitless efforts to repeat the remarkable experiments which have been described in the scientific journals of the past year.
^Labinger JA, Weininger SJ (2005). "Controversy in chemistry: how do you prove a negative?—the cases of phlogiston and cold fusion". Angew Chem Int Ed Engl. 44 (13): 1916–22. PMID15770617. doi:10.1002/anie.200462084. So there matters stand: no cold fusion researcher has been able to dispel the stigma of 'pathological science' by rigorously and reproducibly demonstrating effects sufficiently large to exclude the possibility of error (for example, by constructing a working power generator), nor does it seem possible to conclude unequivocally that all the apparently anomalous behavior can be attributed to error.