The 1985 meta-analysis cited by Bem was conducted by Honorton (1985). A second study of the same cases was conducted by the skeptic Ray Hyman (1985a, 1985b). Even Hyman was forced to conclude that the results were not the result of improper use of statistics, sensory leakage, or cheating (Radin 1997, p. 82). He suggested that improper randomization techniques might have been responsible, although Honorton had given arguments against this. Radin (1997, p. 83) noted, “In this case, ten psychologists and statisticians supplied commentaries alongside the Honorton-Hyman published debate . . . everyone [including Honorton and Hyman] agreed that the ganzfeld results were not due to chance, nor to selective reporting, nor to sensory leakage. And everyone, except one confirmed skeptic [Hyman], also agreed that the results were not plausibly due to flaws in randomization procedures.” nevertheless, Honorton, in a joint communiqué with Hyman, did agree to modify the ganzfeld protocols to take into account Hyman’s concerns (Hyman and Honorton 1986; in Radin 1997, p. 84). This had the good effect of getting Hyman to give his conditions in writing, so that he could not in the future simply raise vague possible objections to the ganzfeld results.
It turned out that Honorton had been conducting ganzfeld studies that complied with Hyman’s stringent conditions since 1983, when the process of image selection had been taken out of the hands of human experimenters and given over to computers. The entire procedure of data recording was automated. Measures were taken to isolate the receiver more effectively, and the whole physical set-up and experimental protocol were reviewed by two professional magicians, who attested the experiments were not vulnerable to cheating (Radin 1997, pp. 85–86). Between 1983 and 1989, 240 persons participated in 354 automated ganzfeld experiments (Honorton and Schechter, 1987; Honorton et al. 1990). In these sessions, the hit rate was 37 percent. The odds against chance were 45,000 to one (Radin 1997, p. 86).
Hyman (1991), ever the skeptic, asked for independent replication. The Honorton experiments at the Psychological Research Laboratories were in fact later replicated by several researchers: Kathy dalton and her coworkers at the Koestler chair of Parapsychology, department of Psychology, University of Edinburgh; Professor dick Bierman, department of Psychology, University of Amsterdam; Professor daryl Bem, cornell University department of Psychology; dr. Richard Broughton and coworkers at the Rhine Research center, durham, north carolina; Professor Adrian Parker and coworkers at the University of Gothenburg, Sweden; and doctoral candidate Rens Wezelman, Institute of Parapsychology, Utrecht, netherlands (Radin 1997, pp. 87–88). When combined with the experiments documented in the 1985 meta-analysis and the 1983–1989 series of Honorton, the total number of sessions was 2,549. “The overall hit rate of 33.2 percent is unlikely with odds against chance beyond a million billion to one,” said Radin (1997, p. 88).
More recently, Bem was coauthor, with Honorton, of an important report on ganzfeld studies of telepathy. Science news said of the report, “new evidence supporting the existence of what most folks refer to as telepathy . . . boasts a rare distinction: It passed muster among skeptical peer reviewers and gained publication in a major, mainstream psychology journal” (Bower 1994). In the Science news article, Bem said, “I used to be a skeptic, but we met strict research guidelines and the results are statistically significant. We hope the findings prompt others to try replicating this effect.” The actual report was published in the January 1994 issue of Psychological Bulletin. Bem and Honorton, using statistical meta-analysis, combined the results of 11 studies involving 240 subjects. The hit rate was one in three, compared to the chance expectation of one in four. In one study, 29 dance, drama, and music students got a hit rate of one in two. One of the reviewers of the article for Psychological Bulletin was Robert Rosenthal, a psychologist at Harvard. He said, “Bem and Honorton’s article is very sophisticated statistically and you can’t dismiss their findings” (Bower 1994).
not everyone was convinced. At a lecture on parapsychology at the Royal Institution, which I attended on february 5, 2000, psychologist Richard Wiseman of the University of Hertfordshire claimed that his own meta-analysis of the post 1987 ganzfeld studies gave a combined hit rate of 27 percent, which, according to him, does not exceed chance expectation. But he also admitted a subsequent study by one of his coworkers of the most recent studies, 1997–1999, yielded a hit rate of 37 percent. Altogether, it seems that there is a genuine paranormal effect in the ganzfeld experiments.
Modern Research into Psychokinetics
Having considered some of the recent scientific work in remote viewing and telepathy, let us now look at psychokinetics (pk), which according to my proposal involves the action of a subtle mind element on ordinary matter. In the nineteenth century and the early twentieth century, prominent scientists reported singular macro-psychokinetic effects such as floating tables. Later researchers have concentrated on reproducible micro-psychokinetic effects.
dice tossing experiments, in which subjects attempt to mentally influence the results, are one example. In 1989, psychologist diane ferrari and dean Radin, then at Princeton, did a meta-analysis of all such experiments in English-language journals up to that time (Radin and ferrari 1991). They found 73 reports, from 52 different investigators, published between the years 1935 and 1987. These reports recorded the results of 2.6 million dice throws by 2,569 subjects in 148 experiments. The reports also recorded control studies, in which the subjects did not try to mentally influence the outcome of the dice throws. The hit rate for the control studies was 50.02 percent, about what would be expected by chance, while the hit rate for the experiments was 51.2 percent. Radin (1997, p. 134) noted: “This does not look like much, but statistically it results in odds against chance of more than a billion to one.”
Radin and ferrari tested their statistical results against various criticisms. Were most of the positive results concentrated in only a few of the many studies? After removing from the database the studies with the most positive results, the remaining studies still indicated a positive result with odds against chance of more than three million to one. Were positive results concentrated in a large number of studies done by a few researchers? Radin and ferrari found that when they removed from their database the researchers who had done the most studies, the remaining results were still positive, with odds against chance of a billion to one. Were the positive results caused by selective reporting? It would have taken 17,974 unpublished studies with negative results to eliminate the positive results. This would amount to 121 unpublished studies for every published study (Radin 1997, 134–135).
One problem recognized by early researchers was the tendency of high numbers, like six, to turn up more often than lower numbers. The face of a die with six on it is made by scooping out six small depressions. This face is therefore lighter than the opposite faces with lower numbers and is more likely to turn up when the dice are thrown. So if the throwers were trying to get sixes, they would be likely to get a result higher than chance, not because of any paranormal mental influence but because of the natural tendency of sixes to come up more than lower numbers. Researchers established experimental protocols to control for this by varying the target numbers in a way that was carefully balanced. Of the 148 studies in the total sample analyzed by Radin and ferrari, 69 were performed with this balanced protocol. They reported that for these studies “there was still highly significant evidence for mind-matter interactions, with odds against chance of greater than a trillion to one” (Radin