Ganzfeld experiment

From WikiSynergy

This page has two parts. The first part is a description of the ganzfeld procedure (sometimes misspelled "ganzfield"), which (as used in parapsychology) is a way of detecting small telepathic effects. The second part describes the analysis of the ganzfeld results, and the debate surrounding them. It is not comprehensive, but gives a general overview. Most of this account was taken from Dr. Dean Radin's book The Conscious Universe. Some of the material is derived from previous editions of the Wikipedia article on the Ganzfeld.

"...when you have eliminated the impossible, whatever remains, however improbable, must be the truth." - Sherlock Holmes, in Sign of Four by Sir Arthur Conan Doyle

Contents 1 Experimental procedure 2 Phase one, preparation 3 Phase two, sending 4 Phase three, judging 5 Analysis of results 6 References

[edit] Experimental procedure

The ganzfeld procedure has three phases which have been developed over decades of critical analysis. It usually requires three participants: an experimenter, a 'sender' and a receiver. The sender tries to transmit information about a target to the receiver by telepathy (Palmer, 2003). The ganzfeld is designed to create a clear separation of the receiver, sender, and experimenter so that there can be no sensory leakage which might influence the outcome. The ganzfeld is meant to create results which are unambiguously telepathic. Each ganzfeld session requires about 90 minutes, and constitutes a single data point. The three phases of the experiment consists of preparation, sending, and judging the outcome:

[edit] Phase one, preparation

Although there are variations of the ganzfeld experiment, the basic procedure is as follows:

The receiver, who will try to obtain the telepathic "transmission" is put in a soundproof room, and seated in a comfortable reclining chair. The receiver listens via headphones to white noise which sounds something like radio static. Translucent hemispheres -usually ping-pong ball halves- are fixed over the receiver's eyes, and a red light is shined onto the receiver's face. A progressive relaxation tape lasting about 10 minutes is also played over the headphones.^[1] (Radin 1997: 74-77)

This procedure is designed to present a changeless sensory experience, because the nervous system responds primarily to changes. When exposed to this unchanging sensory field for a few minutes, the receiver's nervous system achieves a state similar to that of sensory deprivation. Because the brain is deprived of new input, mild to vivid imagery is ordinarily experienced.

At the end of the relaxation tape, the experimenter asks the receiver to report aloud any feelings or images which occur. The receiver is told to continue speaking aloud until the experimenter gives the instruction to stop, about 20 minutes later. After giving the instruction to speak, the experimenter shuts the door to the ganzfeld room and takes the sender to a distant, secure, and isolated room.

Earlier, the experimenter asked an assistant to randomly choose a "target pack" out of a large selection of packs. Each pack contains four pictures, and both the pictures and the packs which contain them are in opaque envelopes, which do not have any indication of their contents. Thus, the assistant cannot know the identity of the target pack before selecting it. Then the assistant selects a picture at random from the pack, which will be the target of the telepathic experiment. In this way, the experimenter remains ignorant of the target. Whether the experiment uses video or picture targets, the targets are selected to be as different from each other as possible (experiments which do not use visual targets often do not seem to pick up telepathy).

However, if the ganzfeld experiment is fully automated, a computer would automatically select a pack of video clips, and then select a video clip from the pack to be the target of the experiment.

There are two types of ganzfeld experiment, the manual ganzfeld and the autoganzfeld. In the manual ganzfeld, the target selection is by manual access to a computer or to random number tables, and other important aspects of the experiment are done by hand. As a consequence, the technique has limited safeguards against fraud or data selection, at least when compared to the autoganzfeld. In the autoganzfeld, most of the key parts of the experiment are fully automated and computerized, in order to prevent fraud, errors, or sensory leakage. The selection and presentation of the target, and the recording of the evaluation of the similarity between the target and the responses given by the recipient, are all automatic.

[edit] Phase two, sending

When the experimenter and the sender are outside the isolated room, the experimenter gives the sender the target, and then seals the sender into the room. In an automated experiment, the target would be presented by the computer through a closed-circuit video system in the isolated room. In the most sophisticated ganzfeld experiments, all the steps in the procedure are completely automated, to ensure that they are followed exactly the same way each time (Radin, 1997 p. 74-89).

The sender is then allowed to view the target, and attempts to telepathically send the target to the receiver. If the target is a video, the sender takes a break while the video rewinds. The sender attempts to send the target for about 20 minutes. The sender is instructed to try to become "immersed" in the target, and to send the receiver the full experience (Radin, 1997 p. 74-89).

In some experiments, the sender can listen to what the receiver says via headphones on a one-way audio link. The purpose is to reassure the receiver that they are not totally alone, and to provide feedback to the sender. The session may also be recorded for the purposes of future research (Radin, 1997 p. 74-89).

[edit] Phase three, judging

After 15 or 20 minutes, the experimenter terminates the sending phase. It is time for the receiverto view the material which was sent. The sensory-field apparatus is removed or turned off, and the experimenter (or computer) then gives the receiver the four pictures from the pack (or allows the receiver to view the videos). In a fully automated system, the computer presents the targets to the receiver in random order via a video monitor.

In the more recent autoganzfeld video experiments, separate copies of the videos are used, so that the receiver cannot identify the target by noticing which tape is most worn (Radin, 1997 p. 74-89).

The experimenter, who does not know the identity of the target, asks the receiver to rank the targets 1 to 4, according to how well they match the impressions received during the sensory-field period. After the targets are ranked, the experimental session is over. At this time, the experimenter, sender, and receiver convene, and the actual target is revealed (Radin, 1997 p. 74-89).

A "hit" is assigned if the receiver ranks the real target as number 1. Otherwise, the session is a "miss". Statistically, this experiment should by chance result in a "hit" once in every 4 sessions, for a 25% "hit' rate. Any hit rate which is greater than this means that information about the target was able to overcome the double-blind controls which were put in place in order to prevent any sensory leakage. That is to say, the experiment dected telepathy (or at least an anomaly which looks very much like it) (Radin, 1997 p. 74-89). When many such experiments are combined, the results may become statisitically significant, meaning that it is very highly unlikely that the results are due to chance. If the results are not due to chance or to flaws in the experimental procedure, then it is likely that telepathy was detected.

[edit] Analysis of results

Over many ganzfeld experiments conducted in many labs around the world, the average hit rate has been closer to 32% rather than the 25% expected by null hypothesis. Such a result is highly significant statistically.

In 1982, Charles Honorton presented a paper at the annual convention of the Parapsychological Association which summarized the results of the ganzfeld experiments up to that date, and concluded that they represented sufficient evidence to demonstrate the existence of psi. Dr. Ray Hyman, Professor Emeritus of Psychology at the University of Oregon, disagreed. The two men later independently analyzed the same studies, and both presented meta-analyses of them in 1985. Honorton thought that the data at that time indicated the existence of psi, and Hyman did not (Hyman, 1985; Honorton, 1985; Radin, 1997 p. 77-89).

In 1985, experiments using the ganzfeld had been published in 34 papers by 10 different researchers, describing 42 experiments. 28 of these experiments were suitable for meta-analysis, and 23 of those resulted in hit rates greater than chance. When Hyman and Honorton analyzed the results of the 28 studies using meta-analysis, they returned odds against chance of ten billion to one. Hyman and Honorton agreed that these results were not due to chance. It was also determined that no single laboratory was responsible for the significance of the results. This was the case even when Honorton's own results were excluded from the analysis. The positive ganzfeld were determined to have been replicated independently by 8 other laboratories (Radin, 1997 p. 77-89).

One possible criticism of these results was that there might be a file-drawer problem, which means that many studies which had null (insignificant) results had not been reported. In 1980 however, the skeptic Susan Blackmore found that the file-drawer problem was not a significant issue. Hyman and Honorton also agreed that it did not represent a significant problem, perhaps due to the fact that it would have taken 423 unreported experiments with null results to make the results of the 28 studies insignificant.

A skeptic might also contend that in a perfectly designed experiment, the results would, on average, be null. Also, the better the experimental design, the less significant the results should be. It is possible to test this hypothesis using meta-analysis, by assigning judges to rate each study according to how well it is designed. If the results of the better designed studies are closer to chance than the studies with worse designs, it is an indication that the positive results of the studies are due to poor design.

The main possible flaw in a ganzfeld experiment is sensory leakage, that is, any way in which the receiver could know what the target is by normal means. It was found that one study among the 28 did contain such a flaw (though the flaw may or may not have influenced the results). However, in that study the participants performed slightly below chance expectation, rather than above. In the end, Hyman and Honorton agreed that there was no systematic relationship between any such flaws in the studies and the outcomes of the studies.

Hyman said that the ganzfeld papers did not describe optimal protocols, nor were they always accompanied by the appropriate statistical analysis. He presented in his paper a factor analysis which he said demonstrated a link between success and three flaws, namely: Flaws in randomization for choice of target; flaws in randomization in judging procedure; and insufficient documentation.

Honorton asked a statistician, David Saunders, to look at Hyman's factor analysis. Saunders concluded that the number of experiments was too small to complete a factor analysis and also that Hyman had chosen his three flaws from a list of nine, and since there are 84 ways to select three elements from nine, Hyman had not corrected for multiple analysis (Saunders, 1985).

In the end, ten statisticians and psychologists gave commentaries on a 1986 published debate between Hyman and Honorton. None agreed with Hyman that randomization flaws could account for the results of the ganzfeld experiments, and two psychologists and two statisticians agreed explicitly with Honorton. Later, in a separate analyses by behavioral scientists Robert Rosenthal and Monica Harris of Harvard University, Hyman's own flaw ratings were used, but the no significant relationships were found between the alleged flaws and the study outcomes. Thus the general agreement among those who analyzed the data was that the ganzfeld study results were not due to chance, selective reporting, or to sensory leakage. Only Hyman thought that the results might have been due to randomization procedures (Radin, 1997 p. 77-89).

In 1986, Hyman and Honorton published A Joint Communiqué, in which they agreed that though the results of the ganzfeld experiments were not due to chance or selective reporting, replication of the studies was necessary before final conclusions could be drawn. They also agreed that more stringent standards were necessary for ganzfeld experiments, and they jointly specified exactly what those standards should be.

In 1983 Honorton had started a series of autoganzfeld experiments at his Psychophysical Research Laboratories. These studies were specifically designed to avoid the same potential problems as those identified in the 1986 joint communiqué issued by Hyman and Honorton. Ford Kross and Daryl Bem, both professional mentalist magicians (magicians whose specialty is simulating psi effects) examined Honorton's experimental arrangements, and said they provided excellent security against deception by subjects (Bem & Honorton, 1994).

In addition to randomization consistent with the specifications of the communiqué, and computer control of the main elements of each test, these autoganzfeld experiments isolated the receiver in a sound-proof steel-walled and electromagnetically shielded room.

The PRL trials continued till September 1989. Of the 354 trials, 122 produced direct hits. This is a 34% hit rate, and is similar statistically to the 37% hit rate of the 1985 meta-analysis (25% is expected by chance). The 34% hit rate is statistically significant with a z score of 3.89., meaning that there was 1 chance in 45,000 that the results were due to chance (Honorton et al, 1990).

Concerning these results, Hyman wrote that the final verdict of whether psi can be demonstrated in the ganzfeld awaited the results of future experiments conducted by other independent investigators.

Other studies were also run, such as those of Professor Dick Bierman of the Department of Psychology at the University of Amsterdam, professor Daryl Bem of Cornell University's Psychology department, Dr. Richard Broughton at the Rhine Research Center, and Professor Adrian parker at the University of Gothenburg, Sweden, and Rens Wezelman at the Institute for parapsychology in Utrecht, Netherlands, and Kathy Dalton of the Departmet of Psychology at the University of Edinburgh Scotland. Only the studies of Kathy Dalton at Edinburgh replicated Honorton's results in a statistically significant manner as of 1997 (above a 95% confidence interval). However, all the studies were over chance expectation. The overall hit rate of all ganzfeld experiments, including the autoganzfeld as of 1997 was 33.2%, which is statistically significant, indicating that such results would occur by chance one time in more than one million billion.

To see if other, post-Joint Communiqué experiments had been as successful as the PRL trials, Julie Milton and Richard Wiseman did a meta-analysis of ganzfeld experiments carried out in other laboratories. They found no psi effect, with a database of 30 experiments and a non-significant Stouffer Z of 0.70 (Milton & Wiseman, 1999).

This meta-analysis was criticized for including all ganzfeld experiments, regardless of the methods being used. Some parapsychologists believed that certain researchers had used protocols that were not part of the standard ganzfeld set up, such as targets consisting of music (traditional ganzfeld experiments use visual targets) (Bem, Palmer, & Broughton, 2001). It was these experiments which did not return significant results. A second meta-analysis was conducted by Daryl Bem, John Palmer, and Richard Broughton in which the experiments were sorted according to how closely they adhered to a pre-existing description of the ganzfeld procedure. Additionally, ten experiments that had been published in the time since Milton and Wiseman's deadline were introduced. Now the results were significant again with Stouffer Z of 2.59 (Edge & Schmeidler, 1999). In science, replicability is one of the main requirements before the data can be accepted. However, the replicability of one type of experiment is not put in question by other types of experiments. The results of the ganzfeld experiments can only be brought into question by failed replications of the original experiments.^[2]

In a 1995 paper discussing some of the challenges, deficiencies and achievements of modern laboratory parapsychology Ray Hyman said,

Obviously, I do not believe that the contemporary findings of parapsychology, [...] justify concluding that anomalous mental phenomena have been proven. [...] [A]cceptable evidence for the presence of anomalous cognition must be based on a positive theory that tells us when psi should and should not be present. Until we have such a theory, the claim that anomalous cognition has been demonstrated is empty.[...] I want to state that I believe that the SAIC experiments as well as the contemporary ganzfeld experiments display methodological and statistical sophistication well above previous parapsychological research. Despite better controls and careful use of statistical inference, the investigators seem to be getting significant results that do not appear to derive from the more obvious flaws of previous research. -The Journal of Parapsychology, December, 1995, Evaluation of Program on Anomalous Mental Phenomena. By Ray Hyman

Since that time, the debate has continued, without conclusion. Parapsychologists generally consider psi to be a proven fact, though they cannot explain it theoretically. Skeptics continue to challenge.

[edit] References

Bem, D.J., and Honorton, C. (1994). Does psi exist? Replicable evidence for an anomalous process of information transfer Online version. Psychological Bulletin, 115 (1), 4-18.

Bem, D. J., Palmer, J., & Broughton, R. S. (2001). Updating the Ganzfeld database: A victim of its own success? Journal of Parapsychology, 65(3), 207-218.

Edge, H., & Schmeidler, G. R. (1999). SHOULD GANZFELD RESEARCH CONTINUE TO BE CRUCIAL IN THE SEARCH FOR A REPLICABLE PSI EFFECT? PART II. EDITED GANZFELD DEBATE. Journal of Parapsychology, 63(4).

Honorton, C. (1985). Meta-Analysis of Psi Ganzfeld Research: A Response to Hyman Journal of Parapsychology 49

Honorton, C., Berger, E. R., Mario, P. V., Quant, M., Derr, P., and Schechter, I. E. (1990). Psi Communication in the Ganzfeld. Journal of Parapsychology 54

Hyman, R. (1985). The Ganzfeld Psi Experiments: A Critical Appraisal Journal of Parapsychology 49

Milton, J. and Wiseman, R. (1999). Does Psi Exist? Lack of Replication of an Anomalous Process of Information Transfer Online version. Psychological Bulletin, vol 125, no 4, p 387-391

Saunders, D (1985). On Hyman's Factor Analysis Journal of Parapsychology 49