Hyman the Terrible, A Flawed Conclusion
by Guy Lyon Playfair

To summarise for new readers: Natasha was asked to match seven diagnoses with seven patients and correctly matched four out of seven. Not bad going, you might be forgiven for thinking. Yet it was not good enough for the CSICOP hit squad. Although Hyman correctly pointed out that “any scientific hypothesis… cannot be confirmed or disconfirmed by one test or one experiment” and that “scientific investigation requires a series of experiments.” (May/June, p. 30), readers – and viewers of the Discovery Channel film – were clearly led to believe that Natasha’s ability had been disconfirmed good and proper. As Wiseman gloated with sickening smugness, “She had the claim, we tested it, she didn’t pass the test.”

That sounds pretty clear. But hang on – Hyman now reveals that the original test was “designed and intended to be the first step in a potential sequence of tests” (Sept/Oct, p. 58), and there would only be further steps if Natasha passed the first one which, according to Wiseman, she didn’t.

Hyman, whose experience as a consultant to the CIA has no doubt familiarised him with the world of dis- and misinformation, goalpost-shifting, data-fudging and general deceit, explains that although a probability rating of .05 (that is, one in twenty, or as we say informally “odds of 20 to one against chance”) is “the conventional level of significance”, he had decided to raise the goalpost bar to a level of 0.01, or 100 to 1 against. It was, he explains without giving us a reference, J. B. Rhine, no less, who suggested this level for psi experiments. “He advocated using the .01 level of significance for parapsychological hypotheses”, he alleges.

No, Dr. Rhine didn’t. As you can see from Rhine and Pratt’s book Parapsychology – Frontier Science of the Mind (1957, p. 186) what he actually wrote was that “most workers in parapsychology accept a probability of .01 as the criterion of significance” (that was fifty years ago; .05 is a level generally accepted today). He admitted that this level was ‘arbitrary’ and added this important qualifier:

“The acceptance of a criterion of significance does not mean that proof of the occurrence of any given phenomenon can be claimed merely because a result at the .01 probability level is obtained. Establishment of a scientific principle must always wait upon repetition and confirmation” (of which there was none in this case).

Rhine did not need to add that neither can a negative conclusion be claimed merely because a result just over 0.01 is obtained, which what Hyman and Co. did. You can be sure that any of Rhine’s subjects who performed as well as Natasha did on her first attempt would have been thoroughly tested further.

Natasha only performed at the 0.0139 level so she was presented as having failed despite the statistical significance of her score – failed by a mere 0.0039 points, surely an extreme case of academic nit-picking even by CSICOP standards? Hyman notes that Natasha named four out of seven diagnoses correctly but spin-doctors this away by alleging “a combination of external clues and luck”.

He may have a point regarding external clues – Natasha, it seems, turned up early and saw some of the subjects arrive, so could easily have guessed, for example, that somebody who limped slowly upstairs had a foot problem. This is inexcusable carelessness on the part of the investigators who should have seen to it that Natasha had no chance to see anybody before the test. Here, Hyman is quite rightly criticising himself and his colleagues for their sloppy research methods.

Now over to the Skeptical Inquirer readers. There was some typical loud-mouth abuse from Californian Joseph A. Soyo, according to whom Natasha is a charlatan who has “finally been exposed”. This unsupported accusation of charlatanism is repeated by Canadian reader B. Abbott, in whose opinion “no sane or honest person would contend that anyone has X-ray eyes”. I hope Natasha has a good lawyer.

Fortunately, some readers sound both sane and honest. Psychology professor David Sheskin of West Connecticut State University reckons that “it would seem to me that a reasonable argument could be made for further investigating the girl’s abilities.”

As for Hyman’s ‘lucky guess’ hypothesis, this was resoundingly knocked on the head by biology professor John D. Nagy of Scottsdale Community College, Tempe, Arizona. He described a computer simulation test he had run a million times in which the random guesser got 2 or more right less than 9 percent of the time. “Apparently, Natasha was not just guessing,” he wrote, “and the effect was relatively large. So why conclude that no further testing is warranted?” He added that “her claim to paranormal powers if of course incredible, but the post hoc arguments explaining away her score are equally unconvincing.” Ouch.

A final voice of reason from Dennis Hall of Berkeley, California. “A few more simple tests with improved experimental controls and a control group would have greatly reduced the possibilities for continued debate.”

If CSICOP, Discovery or anybody else in the U.S. wants to do some more testing of claims of X-ray vision, there may be no need to go all the way to Saransk, Natasha’s home town. As viewers of “The Paranormal World of Paul McKenna” (1997) saw, there was and perhaps still is another Russian, Gregori Antyukhin, living in Milwaukee. He did a couple of diagnoses live on camera in the presence of the patients’ G.P., who reckoned one to be 75% correct and the other 85% to 90%.

And some news just in. I am told by a fairly reliable source that Monica Garnsey, producer-director of the Discovery film, decided to see if she and one of the inquisitors could match the diagnoses by pure guesswork, external clues or whatever.

Both, in a splendid display of psi-missing, got them all wrong.

Go to the next page below: