But not everyone was convinced. In 1955, dr. George Price of the department of Medicine at the University of Minnesota published in Science an article highly critical of the card guessing experiments. Price relied on david Hume’s famous statement that it was more reasonable to believe that witnesses of miracles were deceived or lying than to accept violations of the well established laws of physics. On this basis, Price (1955) argued that Rhine and Soal’s results, because they violated the laws of physics, must be the result of undetected fraudulent behavior. But some years later Price (1972) wrote a letter to Science about his 1955 article, saying, “during the past year I have had some correspondence with J. B. Rhine which has convinced me that I was highly unfair to him in what I said.” He regretted his accusations of fraud. It is possible, of course, that cheating or inadvertent cueing of the subjects may have been involved. But Rhine and Soal had gone to great lengths to prevent such things.
Typical card tests made use of a deck of 25 cards. The cards were each marked with one of five symbols (star, wavy line, square, circle, or cross) so that each symbol was represented by five cards. In the earliest tests, experimenters gave a subject a shuffled deck of cards and asked the subject to guess the top card. After guessing, the subject turned over the top card, checking to see if the identification was correct, and then guessed the next card. critics suggested that printing presses may have left impressions on the backs of the cards. The subjects could have detected these impressions by touch, and used this information to correctly guess the cards. To rule out this possibility, the cards were put into opaque envelopes. critics suggested that the subjects could mark the cards with their fingernails, and feel the marks through the envelopes. Experimenters arranged things so that the subject no longer handled the envelopes. critics suggested the experimenters might be giving subtle cues to the subjects. To prevent this, the subjects were separated from the experimenters and cards by opaque screens. Experimenters were later placed in remote rooms or buildings. critics suggested that in recording the experimental results the experimenters often made errors, errors in favor of paranormal explanations. To solve this problem, experimental designs incorporated duplicate recording of results and doubleblind checking. Monitors were employed to insure that experimenters followed procedures and did not engage in fraud. critics suggested that experimenters sometimes stopped recording data when the results looked good. To solve this problem, experiments were run with a fixed number of trials (Radin 1997, pp. 94–95).
In 1997, dean Radin published the results of his study of 34 card guessing experiments carried out with high levels of security. The experiments were conducted by two dozen researchers during the years 1934–1939, and involved 907,000 separate trials. The chance expectation would be correct guesses in one out of five trials, for a hit rate of twenty percent. Radin (1997, p. 96) arranged the studies in four groups, according to the kinds of security measures employed, and found that the hit rates were significantly above chance for all four groups. critics propose that these hit rates might be the result of selective reporting. In other words, for every published report with favorable results, there might have been other studies with unfavorable results that the experimenters did not publish but kept in their file drawers. This is called “the file drawer problem.” But in order to eliminate the positive results from the 34 published reports, there would have had to have been at least 29,000 unpublished studies, a ratio of 861 to 1 (Radin 1997, p. 97). Such a massive number of unpublished studies is exceedingly unlikely. Radin further noted: “If we consider all the ESP card tests conducted from 1882 to 1939, reported in 186 publications by dozens of investigators around the world, the combined results of this four-million trial database translate into tremendous odds against chance—more than a billion trillion to one.” To eliminate this positive result, the number of unpublished studies in the “file drawer” would have to have been 626,000, for a ratio of more than 3,300 unpublished reports for every published report (Radin 1997, p. 97).
In 1974, nature, the world’s foremost scientific publication, printed a paper by physicists Harold Puthoff and Dr. Russell Targ, about paranormal experiments carried out at the Electronics and Bioengineering Laboratory of the Stanford Research Institute, associated with Stanford University. Targ and Puthoff sought to test the ability of a subject to give information about drawings of objects or scenes shielded from ordinary sense perception. The subjects included Uri Geller, whose psychic achievements were surrounded by accusations of fraud. Whatever one may think about those accusations, one should still be prepared to independently judge particular experiments as to whether or not adequate precautions were taken to prevent deception. Targ and Puthoff stated (1974, p. 602): “We conducted our experiments with sufficient control, utilising visual, acoustic and electrical shielding, to ensure that all conventional paths of sensory input were blocked. At all times we took measures to prevent sensory leakage and to prevent deception, whether intentional or unintentional.” Thirteen remote perception experiments were carried out with Uri Geller. In the first ten, either Geller or the researchers were placed in a shielded room. In the majority of cases, Geller was in the acoustically and visually isolated room, which had double steel walls with double locking doors. Only after this isolation procedure was carried out were target drawings made by the researchers and selected for Geller to identify. Geller did not know the identity of the researcher selecting the target or the method by which targets were selected. In most cases, the target drawings were made by SRI scientists who were not part of the experimental group. The target drawings were kept in a variety of locations, ranging from 4 meters to 7 miles away from the viewing site. Experimenters provided Geller with a pen and paper, and asked him to reproduce the target drawing, giving him the option to pass if he felt he could not detect the target. If he did produce a drawing, the researchers collected it before Geller was allowed to see the target drawing. In an additional three cases, drawings were made by computer. In one case the drawing was visually displayed on a computer screen. In another case it was kept in computer memory but not displayed on the screen. In the final case the target drawing was displayed on the screen but the contrast was adjusted so that the image was not actually visible to the human eye. during these three computer-screen tests, Geller was kept isolated in a faraday cage, designed to weaken electrical signals. Geller gave responses to ten of the thirteen tests. To evaluate how well Geller’s drawings matched the targets, they were submitted to two SRI scientists not part of the research team. The judges were asked to match subject drawings with target drawings. Targ and Puthoff (1974, p. 604) said, “The two judges each matched the target data to the response data with no error. for either judge such a correspondence has an a priori probability, under the null hypothesis of no information channel, of P = (10!)-1 = 3 × 10-7.” In other words, in each case he submitted a drawing, or set of drawings, Geller was able to match the target.