Flim-Flam! Psychics, ESP, Unicorns, and Other Delusions - James Randi [75]
SRI colleagues questioned the two scientists about a judging procedure in the Price tests. When three judges who had been chosen failed to come up with results good enough for Targ and Puthoff, they chose two others, who obliged with favorable findings. They chose no more. Were the announced results based upon all five judges' results? Targ and Puthoff do not tell us. A proper procedure in the remote-viewing tests with Price would have been to have him choose from among a number of photographs to identify the correct target, rather than just ramble on about his general impressions of the required answer. This would have been definitive and easily judged without ambiguity. When a psychologist at SRI suggested this, the idea was ignored. Theirs was a good test, according to Targ and Puthoff, and there was no need for such a process. (A 1979 remote-viewing test at Metropolitan State College in Denver, Colorado, used this more rigorous method. The results were negative.)
Of all the experiments—if I may use that term loosely—that Targ and Puthoff performed with Geller, the thirteen ESP tests are best known to students of these matters. They were the highlight of the article in Nature. The experimenters chose their targets from a dictionary, using a fairly acceptable random method. Geller, sealed in a test area, was supposed to guess the targets, which were drawings of the chosen words. He had the option of "passing," which was the usual arrangement in these tests. Fair enough—if that was the actual procedure.Geller was supplied with paper and pen and asked to make a drawing that corresponded to the target. We are told that he was required to stay in the room until each test was declared terminated, at which time he would emerge and submit his effort to the experimenters before being shown the target. Again, a good procedure, if it was the method used.
Over a period of several days, Geller made thirteen tries at ESP guessing. Contrary to usual scientific procedure, the tests were conducted under widely varying conditions that changed every moment. Geller was able, we are told, to identify seven of the thirteen targets. That's 54 percent success, with odds of millions to one since the target pool consisted of a very great number of possibilities. Sounds impressive, until you remember the usual reporting standards of the two in charge. Actually, Geller correctly identified only three of the thirteen targets, and there's very little mystery about how he got two of them.
Targ and Puthoff conducted these tests and drew their conclusions in a way previously unknown to scientists but often used by bunglers. The scorecard read:
The report Targ and Puthoff issued listed only three passes in the thirteen tries. Actually, in the cases of the "camel" (8), "bridge" (9), "kite" (11), and "church" (12), Geller passed, though this was not reported. According to the rules of Targ and Puthoff, a pass is allowed only if you miss! There were several responses to "camel," for example, and Puthoff chose the one closest to it, a horse, as the winner. They reported that "all drawings were published," but there were many that they chose to omit apparently in order to bolster the results. But, we are told, even with all this skulduggery, they did submit these results for "double-blind" decisions by people who did not know the expected findings. Yep. And in doing so, they left out numbers 5, 6, and 7, since they were passes, but included the selected responses to numbers 8,9,11, and 12—and these were passes, too! But—when carefully trimmed and weeded—they were very good evidence in favor of Geller's ESP powers! The double-blind safeguards do not take into account such careful "weeding."
At this late date, with the detailed and careful obfuscation in the reports that has been applied in the interim to conceal the needed information, it is impossible to say just how Geller fooled the