Terence Hines - Pseudoscience and the Paranormal

Other nonparanormal sources of information existed about the design on the cards in the Rhines’ experiments. The small shield did not eliminate clues when the agent wrote down which card was presented on each trial. The sound of a pencil writing provides information about what is being written. Try this experiment: Have someone write the numeral 2 and then the numeral 4 behind you while you listen carefully. Now have the person write “lines” and “square,” names of two of the Zener cards. In both cases, you can hear the difference. Observing the top of a moving pencil or pen also provides information about what is being written. Magicians use this technique in several tricks; with a bit of practice, one can become quite good at it. Obviously, since there are only five alternatives in the Rhine experiments, even a little information based on such sources would boost a subject’s score above chance, with no paranormal powers needed to explain the results.

Other procedural flaws in the Rhine experiments have been detailed by Zusne and Jones (1982) and Hansel (1966, 1980). Zusne and Jones point out that “the keeping of records in Rhine’s experiments was inadequate. Sometimes, the subject would help with the checking of his or her calls against the order of cards. In some long-distance telepathy experiments, the order of the cards passed through the hands of the percipient [subject] before it got from Rhine to the agent” (p. 375). In other words, the subject was given a list of the cards that he or she was later to attempt to guess. The opportunity for nonparanormal transfer of information in such situations is obvious. As early as 1939, Kennedy concluded that the vast majority of the seemingly positive experiments by the Rhines were due to poor experimental control. Three studies that Kennedy did not fault on procedural grounds have since been questioned on the basis of opportunity for cheating or use of clues that Kennedy was not aware of (Hansel 1966; Zusne and Jones 1982).

The procedural errors in the Rhine experiments have been extremely damaging to his claims to have demonstrated the existence of ESP Equally damaging has been the fact that the results have not replicated when the experiments have been conducted in other laboratories. In a very real sense, the Rhines and ESP have been the Blondlot and N rays of the middle part of the twentieth century. Crumbaugh’s (1966) comments on his failure to repeat the Rhines’ findings, even after years of effort, have been noted previously; other researchers fared little better when trying to repeat the Rhines’ work. By 1940 “six different researchers, using some 500 subjects in experiments totaling about half a million trials demonstrated nothing but chance scores” (Zusne and Jones 1982, p. 375). See also Hansel (1966, 1980) for detailed accounts of the failures to replicate Rhine’s findings.

Occasionally, of course, apparent replication would occur. It should be remembered, however, that replications were also reported of Blondlot’s N ray findings. As in the N-ray case, reports that seemed to support the Rhines’ work were few and could be attributed to some of the same procedural problems as found in the original work. In addition, it should be remembered that what is taken as evidence for ESP is above-chance performance on some sort of card-guessing task. If one tests enough subjects long enough, sooner or later one of them will score above chance at a statistically significant level, for at least some set of trials. Does this mean that, for this brief time, the subject possessed ESP? No, it simply means that chance is operating as expected. Let us assume that out of two hundred subjects, one scores significantly above chance at the .01 level. That is, the deviation from chance would be expected to occur once in one hundred times. This will impress no one who has any knowledge of statistics, since the fact that an event that is expected to occur by chance once in one hundred times does so when you give it two hundred opportunities to occur is not evidence of anything extraordinary.

Real-world analogues to the card-guessing experiments occur frequently. One can consider every spin of the roulette wheel, every throw of the dice, every draw of the card in gambling casinos the world over as a single trial in a worldwide ongoing study in parapsychology. At gambling casinos, the odds are in favor of the house by only a tiny margin (if the margin were greater, people would lose much more frequently and be much less willing to play). Nonetheless, this tiny margin is enough to produce huge amounts of money for the house. Over the billions of “trials” in this real-world “experiment,” there has been no hint of any deviation from the strict laws of chance. State-run lotteries offer another opportunity to look for ESP in real-world situations. Billions of state lottery tickets have been sold since New Hampshire introduced the modem state lottery in 1964. Skolnick (1985, personal communication) examined the data from the New York State Lottery for several years. He found that New Yorkers were not winning the lottery at a rate higher than chance. If ESP had been operating, even in only a minority of players, the rate of winning would have been higher than chance.

The argument is often made by proponents of paranormal claims that these powers cannot be used for profit, so when one tries to use ESP to foresee the outcome of a spin of a roulette wheel, a football game, or the movement of a company’s stock on Wall Street, psi powers promptly vanish. However, not all parapsychologists agree with this position. Rhine himself contended that highly motivated subjects did better in ESP experiments (Rhine and Pratt, 1962). For most people, money is a strong motivating factor.

In England, S. C. Soal attempted to replicate the Rhines’ findings. At first Soal believed he had failed, stating, “I have delivered a stunning blow to Dr. Rhine’s work by my repetition of his experiments in England… there is no evidence that individuals guessing cards can beat the laws of chance” (Thouless 1974, quoted in Markwick 1985, p. 287). Soal’s studies included 160 subjects and a total of 68,350 trials. Later, however, Soal reanalyzed his data for what has been termed a displacement effect . That is, he compared the card guessed, not with the card actually present on that trial, but with either the card that had been present on the previous trial or with the card that had appeared on the succeeding trial. When this was done, 2 subjects out of 160 showed hit rates well above chance. The two subjects were Mrs. Gloria Stewart and Basil Shackleton.

Soal’s further work with Shackleton lasted for forty sessions and totaled twelve thousand trials. Cards with a picture of one of five animals were used instead of Zener cards with symbols. Each design was coded with a digit from 1 through 5. Before a session, Soal reported, he would obtain a random list of the digits 1 through 5 from a book of logarithms. He described in detail the procedure he used to obtain these lists of digits. This was important, as it permitted other investigators, years later, to find the exact source of the random digits used and to attempt to duplicate the tests. The random digits were written in a “target” column on the score sheets for the session and were used to determine the order in which the five different designs would be presented. Thus, if the digit 1 was the code for the elephant design, when a 1 appeared in the target column on a trial, the elephant design would be used. Shackleton would make his guess by naming the animal, not naming a digit. His guess was also coded into the appropriate digit and that digit recorded on the score sheet. Thus, if he guessed “elephant,” a 1 would be recorded in the “guess” column.

Soal’s work with Shackleton (reported in Soal and Goldney 1943) was extremely successful. It was long considered the best evidence for the existence of ESP by both proponents and skeptics. As might be expected, critics spent considerable time and effort attempting to refute the results. Frequently these attempts were unfair, distorted, or simply wrong (see Markwick 1985, for a brief review). Then in 1960 it was revealed (Soal and Goldney 1960) that in 1941, Soal had been charged with changing some of the l’s in the target column to 4’s or 5’s after a session. The accuser was a Mrs. Gretl Albert, who had taken part in the experiment as an “agent,” the individual who attempted to send the identity of the card to Shackleton. Soal originally denied the charge and did not permit it to be published until 1960.