Parapsychology Research Paper

1. B. Rhine adapted the term parapsychology from the German word Parapsychologie in the 1930s to replace the earlier term psychical research. More than a change in terminology was involved. Psychical research covered a broad range of systematic investigations into spiritualistic phenomena, haunted houses, premonitory dreams, visions, and the like. Rhine’s program avoided much of this material, especially that which dealt with questions of survival after death. Instead, he emphasized controlled laboratory experiments using normal individuals as subjects. Thus, the change in terminology signaled that parapsychology aspired to become an accepted scientific discipline.

1. Historical Background

1.1 Scientists And Psychic Phenomena

In 1848, the two young sisters Margaret and Katherine Fox initiated modern spiritualism with their demonstrations of rappings and other material phenomena which they claimed were caused by the spirits of dead people. Soon many other individuals, calling themselves mediums, were regularly producing phenomena which they attributed to spirits of the dead. Because science deals with material phenomena, and because material phenomena were what the spiritualists were offering as support for their paranormal claims, some major scientists became interested in investigating these claims. In 1853, Robert Hare, an important chemist from the University of Pennsylvania, began an investigation of table-tilting. He attended seances where he became convinced that the spirits of the dead were moving the tables.

In 1865, Alfred Russel Wallace, the cofounder with Darwin of the theory of evolution through natural selection, began his investigations of spiritualistic phenomena. He quickly became convinced of the reality of the paranormal and for the remaining 48 years of his life he continued his investigations and outspokenly defended the reality of spiritualistic phenomena. Sir William Crookes, the discoverer of thallium and the inventor of the cathode ray tube, began his investigations of the paranormal in 1869. He concluded, to the dismay of his scientific colleagues, that a psychic force was operating in these seances.

1.2 The Society For Psychical Research

Inspired by the investigations of such eminent scientists as Hare, Wallace, and Crookes, a group of scholars and spiritualists formed The Society for Psychical Research in London in 1882. The philosopher Henry Sidgwick, its first president, stated that the goal was to collect more evidence of the kind that had been reported by those scientists who had investigated mediums. Although Sidgwick believed that the previous investigations of mediums had already provided evidence sufficient to prove the existence of the paranormal, he was aware that the majority of the scientific community were not convinced. Therefore, the aim of the Society was to accumulate enough additional evidence of the same sort so as to force the scientific community to admit that the case for the paranormal had been proved or to accuse the proponents of insanity or gross incompetence.

In his inaugural presidential address, Sidgwick claimed that, in fact, some members of the new society had already collected just the sort of evidence that proved beyond a doubt the existence of telepathy. Sidgwick was referring to the investigations involving the Creery sisters. Unfortunately, the Creery sisters were later discovered to be signaling to each other using a code. Although an attempt was made to claim that such a code could not have accounted for some of the results of their investigations, this body of evidence was subsequently removed from the republished proceedings of the Society for Psychical Research. A similar fate befell the very next set of investigations which the Society originally offered as solid proof for the existence of telepathy. These investigations involved two young men, Smith and Blackburn, who apparently could communicate with each other telepathically. Many years later, Blackburn published a confession of how he and Smith had tricked the investigators. This case was also deleted from later printings of the Society’s proceedings (Hyman 1985a).

2. From Psychical Research To Parapsychology

2.1 J. B. Rhine

Joseph Banks Rhine and his wife Louisa came to Duke University in 1927 and developed a program of research based on card guessing experiments. They termed their program ‘parapsychology’ and focused their efforts on making this discipline into an experimental science. They replaced the regular deck of playing cards with a deck consisting of five distinct symbols—a circle, a square, three wavy lines, a cross, and a star. Each symbol occurred five times, making a deck consisting of 25 cards. The same set of cards could be used as targets for experiments in telepathy, clairvoyance, and precognition. In addition, the symbols could be presented in a variety of ways.

In a typical clairvoyance condition, for example, the deck would be shuffled and the subject would be asked to guess the symbol of each card at each location in the deck. Since there were five different symbols, the subject had one chance in five of being correct on each guess just by chance. In other words, the subject would be expected to average five correct hits on each run through a deck of 25 cards. In a precognition test, the subject would make his or her guesses before the deck was shuffled. In a telepathy test, a sender would look at the symbol on each card at the time the subject was making his or her guess.

When Rhine published his first monograph, Extra Sensory Perception, in 1934 he had collected data on a total of 91,174 guesses in three years. The average number of hits per run of 25 guesses was 7.1 as compared with the chance expectation of 5. Although the results were barely 2 hits per deck higher than the expected 5, the probability of such a departure from chance, given the huge number of trials, is very low.

The scientific community, at first, took great interest in this outcome. Never before had so much scientific data been amassed in favor of a paranormal claim. After some years, however, skepticism and criticism from the scientific community prevailed. In part, this was because several attempts by independent investigators to replicate these findings failed. In addition, criticisms were aimed at both the statistical procedures as well as the experimental controls. For the most part, the statistical criticisms turned out to be unfounded or insufficient to explain away the results. The methodological critiques were more serious, and gradually parapsychologists improved their procedures for pre- venting sensory leakage, randomizing the targets, specifying the criteria in advance, and the like.

The paradigm that Rhine initiated under the rubric of parapsychology transformed psychical research into a serious contender for a place among the accepted sciences. It also established the basic procedures for parapsychology for a period covering the years from 1930 to around 1970. During this time the ESP cards were the predominant stimulus for experiments in telepathy, clairvoyance, and precognition. Rhine also initiated the study of psychokinesis (PK) by having subjects attempt to influence the outcomes of dice tosses. The full story of the first couple of decades of this formative period is recounted in Mauskopf and McVaugh (1980).

2.2 S. G. Soal

As already indicated, the problem of replicability haunted the early results of Rhine’s card-guessing experiments. Rhine and other US parapsychologists claimed that their results, when they were successful, were equally good for telepathy and clairvoyance trials. The British parapsychologists, however, claimed that their successful experiments only involved telepathy and never clairvoyance. Perhaps the strongest critic of Rhine’s early work was the British mathematician, vs. G. Soal. During the years 1934–9, Soal managed to amass 128,350 guesses from 160 percipients. He reported that the number of hits was just what would be expected by chance.

A fellow parapsychologist, Whately Carington, eventually convinced Soal to review his data to see if some of his percipients might reveal a ‘displacement effect.’ Perhaps, he suggested, some of the subjects’ guesses were systematically off by one or two targets. Soal, indeed, discovered that two of his 160 subjects showed significant hitting above chance when their guesses were matched not against the intended target but, instead, against the target for the next trial.

Soal realized that finding such displacement patterns in two out of 160 subjects after the fact could be a statistical quirk. He was able to find and persuade one of these subjects, Basil Shackleton, to participate in 11,378 more guesses during the years 1941–3. As was the case in the original experiment, Shackleton’s guesses, when matched against the actual target, were consistent with chance. However, when compared with the symbol that came after the target (precognitive hitting), his guesses were well above chance. Soal reported the odds against chance for such a pattern as 10³⁵ to 1. Beginning in 1945, Soal was able to gather 37,100 guess from the second subject, Gloria Stewart. Unlike her previous results or those of Shackleton, Mrs. Stewart showed no displacement effect. Instead, her guesses of the intended target were significantly above chance. Soal calculated the odds against chance of this outcome as 10⁷⁹ to 1.

Such strikingly successful results over a period of years, especially from a strong critic of Rhine, caught the attention of both parapsychologists and skeptics. Parapsychologists hailed the experiments as definitive proof of psi. Critics looked for possible flaws or weaknesses. The early critics managed to find some suspicious patterns, but nothing conclusive. Some proposed fanciful scenarios, involving the collusion among several investigators, to account for the results.

It wasn’t until 1978 that Betty Markwick published her findings which convinced most parapsychologists and all the skeptics that Soal had indeed cheated (Hyman 1985a). Ironically, Markwick made her discovery while she was carrying out an elaborate investigation to vindicate Soal’s reputation. As part of her plan, she searched tables of logarithms for matches to sequences of targets in Soal’s experiments. Because Soal had claimed that he had used Chambers’ tables for his random sequences, Markwick reasoned that if she could find matches between Soal’s target sequences and sequences in Chambers’ tables, this would show that, at least, Soal had not tampered with the target sequences. What she found were partial matches. Many of Soal’s sequences had extra insertions within the sequence and these extra insertions usually corresponded with hits. The strong implication was that Soal had deliberately constructed his target sequences with blank spaces inserted systematically. In recording the percipient’s guesses, he could surreptitiously insert target numbers in the blank spaces that would match the guess for that trial. Such insertions would suffice to bias the results towards a highly significant, but spurious, outcome.

2.3 The Post-Rhine Era

Around 1970, Rhine’s specific targets and procedures began to be discarded in favor of new targets and procedures. The striking successes of the early 1930s were no longer in evidence. Results were inconsistent, replicability was still elusive, and the hit rate—even in the few successful outcomes—had become very small. Cynics attributed the declining hit rate to better controls and other methodological and statistical improvements.

Most parapsychologists, however, blamed the weak or nonexistent effects to the austereness of Rhine’s procedures. The targets on the ESP cards were meaningless and abstract. The constraint on the subjects’ responses to just one of a few choices and the rigidity of the experimental plans all conspired, in the view of many parapsychologists, to dampen the delicate ESP ability.

So parapsychologists began using targets that were emotionally and otherwise meaningful—actual photographs, motion picture clips, actual physical sites, etc. They also eschewed the forced-choice methods of Rhine in favor of free-responding, i.e., the percipients were encouraged to say whatever came to mind and to describe their ‘mentation’ in whatever words they chose. In addition, they began to take advantage of new technological breakthroughs, especially video and computer developments. These more natural targets and fewer constraints upon the subjects’ responding made problems of control and scoring more complicated, but parapsychologists felt that this was a small price to pay in order to achieve what they call psi-conducive conditions.

3. The Contemporary Scene

Contemporary parapsychology has generated a variety of novel approaches. Computers are central to much of the research. A line of research that goes back to the late 1960s involves having subjects attempt to influence psychically the output of random number generators (RNG). Computers are also used to control many of the aspects of conducting experiments such as selecting targets, recording the guesses, reporting the outcomes, etc. Advances in neuroscience are also exploited and many experiments use physiological indices instead of, or in addition to, verbal responses. Among these new approaches, three lines of research predominate. The research on remote viewing began in 1972 at the Stanford Research Institute. Until 1995, much of that research, conducted under the auspices at first of the CIA and later the Defense Intelligence Agency was classified. The work was declassified in 1995 and evaluated by a panel consisting of the statistician, Jessica Utts, and the psychologist, Ray Hyman. Hyman and Utts disagreed about the scientific status of remote viewing. Utts concluded that the evidence clearly supported the existence of psi and Hyman that the scientific claim had not been proven. Both agreed, however, that the results of remote viewing were too unreliable to justify using them for information gathering.

The second line of research, initiated by Helmut Schmidt in 1969, involves having subjects attempt to influence or bias the outputs of random number generators (RNGs). Although many different investigators and laboratories have published research on apparent psychokinesis or PK effects on RNGs, the overwhelming number of trials in this line of research has come from the Princeton Engineering Anomalies Research (PEAR) program at Princeton University under Robert Jahn and his associates. In 1987, Radin and Nelson (Radin 1997) conducted a meta-analysis of the RNG experiments conducted from 1959 through 1987. A meta-analysis is a method for combining or pooling the results from several previously published reports involving similar procedures. They reported that the combined outcome produced an effect with odds against chance of more than a trillion to one. Radin claims that the results of such studies completed since 1987 are still consistent with that meta-analysis.

Despite these claims for the RNG research, the studies have been criticized on a number of grounds. Although the meta-analysis pools the data from all the laboratories, the pooling does not deal with the various and different patterns of outcomes from separate laboratories. For example, the PEAR researchers claim to have found a distinct pattern: The RNG outputs slightly more positive units when the subject is trying to influence it in that direction and slightly more negative units when the subject is trying to bias the output in the negative addition. In the neutral condition, the output is evenly divided but the variability is less than would be expected by a truly normal distribution. When the results from these three conditions are pooled the distribution becomes perfectly normal with a mean of zero. No other laboratory has replicated this interesting pattern. The replicability that Radin and Nelson claim is a rather weak one. In effect, any departure from the chance baseline, no matter the specific pattern, is considered equal to any other departure. The one physicist who has deliberately tried to replicate Jahn’s RNG findings, Stanley Jeffers, has failed to do so.

3.1 The Ganzfeld-Psi Experiments

The line of research that has had the most influence since the early 1970s and is at the center of parapsychology’s claim of a demonstrable psychic anomaly is the ganzfeld-psi program. The March 1985 issue of The Journal of Parapsychology was devoted to ‘The Ganzfeld Debate.’ This debate consisted of a 47-page critique by Ray Hyman of the 42 ganzfeld-psi experiments known to exist at that time. This was followed by a 47-page response by Charles Honorton, one of the pioneers of this type of experiment.

Hyman (1985b) agreed that, taken as group, the overall hit rate in the ganzfeld experiments exceeded what would be expected by chance. However, he identified a number of statistical and methodological flaws characterizing the experiments. These were flaws which parapsychologists recognize as ones that should be avoided. Yet, not one of the experiments in this data base was free of all the flaws. Hyman concluded that, given the various flaws, the alleged significant levels had been vastly overinflated and the set of experiments could not be used to draw any conclusions about the existence of psi.

In his response, Honorton (1985) devised a different method for assigning flaws and concluded, contrary to Hyman, that the flaws, while there, had no impact on the outcome of the experiments. The controversy continued with parapsychologists supporting Honorton’s analysis and the skeptics supporting Hyman’s. Rather than continue with further responses and rebuttals, Hyman and Honorton collaborated on a joint paper to focus on areas of agreement and to suggest guidelines for future ganzfeld-psi research (Hyman and Honorton 1986).

In 1990 Honorton and his colleagues published an article in the Journal of Parapsychology (Honorton et al. 1990) which summarized six years of new ganzfeldpsi experiments supposedly meeting the guidelines devised by Hyman and Honorton. The new set of experiments, mainly controlled by computers, became known as the autoganzfeld experiments. The hit rate was consistent with the previous set of ganzfeld experiments and was highly significant. In 1994, a major psychological periodical, the Psychological Bulletin, published the article ‘Does psi exist? Replicable evidence for an anomalous process of information transfer’ jointly authored by the psychologist Daryl Bem and Charles Honorton (Bem and Honorton 1994) (unfortunately, Honorton died before the article was published). The same issue contained a commentary by Hyman (Hyman 1994) as well as a ‘Response to Hyman’ by Daryl Bem (Bem 1994).

Hyman questioned the claim by Bem and Honorton that the autoganzfeld findings were consistent with the original ganzfeld data base. He discovered two peculiar and striking patterns in the autoganzfeld data. While Hyman found these two patterns suggestive of an artifact, Bem responded that if it turned out to be a property of psi, parapsychologists should call it the Hyman Effect.

The ganzfeld-psi research continues in many laboratories around the world. Radin (1997) describes a meta-analysis of all the ganzfeld-psi studies through early 1997. He concludes that ‘The overall hit rate of 33.2 percent is unlikely with odds against chance beyond a million billion to one’ (p. 88). This contrasts sharply with the conclusions from a meta-analysis of 30 ganzfeld experiments by Milton and Wiseman (1999). These experiments were conducted after the autoganzfeld experiments and yielded an effect size consistent with zero. Milton and Wiseman ‘conclude that the ganzfeld technique does not at present offer a replicable method for producing ESP in the laboratory.’

4. The Case For Psi

The strongest arguments for the reality of psi have been presented by Jessica Utts (1991) and Dean Radin (1997). Both base their case upon meta-analyses of previously published articles. Utts relies upon metaanalyses of the ganzfeld-psi experiments, early precognition card-guess experiments, RNG studies, early dice-throwing experiments, and studies relating extroversion to ESP performance. Radin includes these as well as several other meta-analyses.

The meta-analyses are used to draw two conclusions. The first is that the pooled effect size emerging from each meta-analysis, although small, is sufficiently large to reject the possibility of being due to chance. The second is that no difference in effect size can be traced to differences attributed to flaws in the experimental procedures because the effect size is essentially the same for methodologically strong and weak experiments.

It is on this basis that parapsychologists such as Utts and Radin conclude that the evidence for psi is conclusive. Radin argues that the case for psi has been made conclusive several times in the past and that the only mystery is why this is not better known. He implies that the public is unaware of the scientific case for psi because of the attempts by skeptical ‘gatekeepers’ to prevent the public from knowing about it.

5. Grounds For Caution

Despite such strong claims by Utts and Radin, the skeptics insist that there are good reasons for doubting that the evidence for psi is adequate. The meta-analysis of the most recent ganzfeld studies by Milton and Wiseman (1999) strongly suggested that the previous ganzfeld successes cannot be replicated. Worse, Milton and Wiseman found that most of the claims for correlates of psi performance also did not replicate. Although their meta-analysis is controversial, the controversy itself points to major problems with trying to base the proof of psi on meta-analysis of previously published studies. Critics have pointed out many problems with meta-analyses.

When used properly, meta-analysis can be a useful tool to help generate patterns and hypotheses about a body of data. Such a usage is consistent with what statisticians call exploratory analysis. The mistake the parapsychologists seem to make is to use the same meta-analysis both to find possible patterns and then prove the existence of such patterns. This is like having your cake and eating it too. A meta-analysis is inherently retrospective. The results are best used prospectively to make predictions about specific patterns to be found in new sets of data. So far, parapsychologists have not succeeded in doing this.

6. Prognosis

Although some key parapsychologists argue that their evidence strongly supports the reality of psychic phenomena, the main body of the scientific community remains unconvinced. At best the parapsychologists have demonstrated, by pooling results from previously published studies, that significant departures from the chance baseline occur more often than expected. They have yet to demonstrate, however, that these departures—if real—originate from a common cause. Nor have they demonstrated any lawfulness in these deviations. Worse, they have yet to come close to specifying a set of conditions and procedures that would enable believers and skeptics to reliably observe these so-called psi events. During the 15 years of conducting meta-analyses on parapsychological experiments, the parapsychologists claim that effect sizes have remained relatively stable. However, this cannot be taken as support for the existence of psi. In a progressive science, as the research develops and the sources of error are tamed, the effect sizes should appreciably grow larger with time. Until this happens in parapsychology, the case for psi will make little headway in the scientific community.

Bibliography:

1. Bem D J 1994 Response to Hyman. Psychological Bulletin 115: 25–7
2. Bem D J, Honorton C 1994 Does psi exist? Replicable evidence for an anomalous process of information transfer. Psychological Bulletin 115: 4–18
3. Honorton C 1985 Meta-analysis of psi ganzfeld research: A response to Hyman. The Journal of Parapsychology 49: 51–91
4. Honorton C, Berger R E, Vargolis M P, Quant M, Derr P, Schecter E I, Ferrari D C 1990 PSI communication in the ganzfeld experiments with an automated testing system and a comparison with meta-analysis of earlier studies. The Journal of Parapsychology 54: 99–139
5. Hyman R 1985a A critical historical overview of parapsychology. In: Kurtz P (ed.) A Skeptic’s Handbook of Parapsychology. Prometheus Books, Buffalo, NY, pp. 3–96
6. Hyman R 1985b The ganzfeld-psi experiment: A critical appraisal. The Journal of Parapsychology 49: 3–49
7. Hyman R 1994 Anomaly or artifact? Comments on Bem and Honorton. Psychological Bulletin 115: 19–24
8. Hyman R, Honorton C 1986 A joint communique: The psi ganzfeld controversy. The Journal of Parapsychology 50: 351–64
9. Mauskopf S H, McVaugh M R 1980 The Elusive Science: Origins of Experimental Psychical Research. Johns Hopkins University Press, Baltimore, MD
10. Milton J, Wiseman R 1999 Does psi exist? Lack of replication of an anomalous process of information transfer. Psychological Bulletin 125: 387–91
11. Radin D 1997 The Conscious Universe: The Scientific Truth of Psychic Phenomena. Harper Edge, San Francisco
12. Utts J 1991 Replication and meta-analysis in parapsychology. Statistical Science 6: 363–403