The phony homeopathy study
`WHEN to believe the unbelievable.'' With that provocative headline on an editorial disclaimer, the June 30 issue of the premier scientific journal Nature published a research paper whose validity the journal questioned. A solution of a biologically active chemical diluted to the point where none of the chemical remained seemed to retain its potency. This ``unbelievable'' conclusion would support the claims of homeopathic medicine. Nature promised to make its own investigation. As its current (July 28) issue reports, it found the research to be flawed. The ``unbelievable'' conclusions appear to be the result of wishful thinking.
Thus skeptics, including Nature's editor, Dr. John Maddox, can feel vindicated. Physical law, which the research seemed to challenge, stands triumphant. But now we have a new issue - when to publish the unbelievable.
Maddox has explained that Nature had already held up the paper for two years while it was rewritten after critical review; that it contained no obvious flaws; and that another laboratory had seemed to confirm its findings. Furthermore, the results had leaked to the French press. Not wishing to suppress the paper or have its findings reported piecemeal, Nature published it so ``vigilant members of the scientific community ... may be able to suggest further tests.''
That seems reasonable. But the inevitable result was that Nature's decision to publish a paper it disbelieved without waiting for the results of its own investigation became news in its own right. News media around the world picked it up. Despite caveats citing Nature's reservations, the main message was that some intriguing research challenged physical law. It fed an appetite for news of paranormal and oddball science.
And the research is oddball. Prof. Jacques Benveniste of the University of Paris-Sud and colleagues in several countries have investigated the well-known interaction of human white blood cells with certain antibodies (proteins) at increasing dilutions of the antibody solution. Even at dilutions where probably no antibody molecule remained, they claimed to have an interaction. The re-searchers speculated about a reorganization of water molecules to give such extreme dilutions a ``memory'' of the antibody.
Besides Maddox, Nature's investigators included two Americans - Dr. Walter Stewart of the National Institutes of Health, a biochemist who investigates scientific fraud, and magician James Randi, a MacArthur fellow skilled in analyzing scientific fakery and self-delusion. This team found that the claimed results were not repeated under carefully controlled conditions. Instead, it concludes that ill-controlled research and selective rejection of experiments that did not fit the researchers' expectations led them to believe in a phenomenon that does not exist. The team found no evidence of fraud, only of wishful thinking.
Benveniste protests that Nature's investigation was itself flawed and biased. Further research can settle that point. But for the moment, there is little rea-son to believe in his ``unbelievable'' claim.
Nevertheless, those who want to believe will continue to do so.
The early news reports have probably created an impression of a far-out discovery fighting for recognition against hidebound skepticism. It would have been better for the news media to wait - as this newspaper did - for Nature's study. But that is asking too much when Nature's own action became the news.
If the paper in question was half-baked science, then to publish it with only the promise to provide a balancing report of a follow-up investigation was half-baked journalism. A scientific journal of Nature's stature has an obligation to the general public, which cannot evaluate technical work, as well as to the scien-tific community. Nature should have waited for its own investigation. It should not put the media in a position where they have to report an obviously titillating but unbalanced scientific presentation.
A Tuesday column. Robert C. Cowen is the Monitor's natural science editor.