Extrasensory Perception

Brian J. Matkins


This chapter examines the myth of extrasensory perception (ESP). First definitions are given related to ESP. A glance at some early studies that are questionable are then explored. Next an innovative experiment involving virtual reality is examined in which significant results were not found. A study using classical conditioning follows again yielding poor results. Next an experiment carried out through the mail was cited which, again, does not yield significant results. A psychic named Uri Geller, who was popular in the 1970's is then shown to be fraudulant in his abilities to bend metal objects with his mind. Gellar is also shown to be a fraud on his proclaimed clairvoyance. Next it is shown that a large number of people want to believe in ESP phenomena and explanations are given as to why people want to believe. Neuroimaging techniques are then examined and show that there is no difference in brain activity during "psi hitting' verses 'psi missing. Ray Hyman's critique of Ganzfield experiments is then explored. James Randi is then looked upon in regard to his ESP testing website. Finally, Randi's $1,000,000 prize for demonstrating psychic abilities in the laboratory is called upon to leave the reader with no doubt that if someone had psychic abilities, this prize would have been collected long ago.

I. Introduction

Any discussion of extrasensory perception would warrant several definitions. First of all, extra sensory perception is defined by the American Psychological Association (APA) Dictionary of Psychology as “Alleged awareness of external events by other means then the known sensory cannels. It includes telepathy, clairvoyance, precognition, and, more loosely psychokinesis. Despite considerable research, the existence of any of these modalities remains highly controversial” (2007).
Telepathy is defined by the APA dictionary as “the alleged direct communication of information from one mind to another, in the absence of any known means of transmission (2007). Moulton and Kosslyn (2008) define clairvoyance as the direct sensing of remote events. Precognition is define by the APA as “… the purported ability to see or experience future events through some form of extrasensory perception” (2007). The APA defines psychokinesis as “… the alleged ability to control events or more or change the shape of objects through the power of thought…” (2007). Psi phenomenon is defined by APA as “unspecified mental functions held to be involved in telepathy and other parapsychological processes that currently defy scientific explanation (2007)”.

II. Pearce and Pratt

With such a starting point, a look at some early studies would be fruitful. The Pearce-Pratt experiment is a landmark in the study of clairvoyance. In this study Pearce and Pratt synchronized their watches and Pearce went across the quadrangle to the library at Duke University (Hansel, 1980). Pratt could see Pearce enter the library about a hundred yards away.
Pratt sat at a table with a pack of ESP cards and shuffled them. Pratt placed each card on top of a book for a total of one minute, and then placed it to the left. At the end of the trails, Pratt turned the cards to reveal their order (Hansel, 1980). The order was recorded in duplicate, one of which was sealed in a envelope and later returned to Rhine.
Pearce, the recipient, recorded his guesses about the identity of each card lying on the book. After recording 50 guesses he made a duplicate copy and sealed in an envelope. The envelopes were then usually delivered to Rhine before Pratt and Pearce compared their number of hits (Hansel, 1980). This procedure was carried out during thirty-seven sittings between August 1933 and March 1934.
Hansel (1980) states that the odds against the overall the result occurring by chance are in excess of ten to the 22nd power to one, and the result of each subset is statistically significant. Rhine stated in the Journal of Parapsychology in 1954 that the only alternative explanation besides ESP would involve collusion among the participants. Pearce was not supervised during the experiment which suggests he could have cheated in a number of ways. Pratt saw Pearce disappear into the library, then when the testing was over they met and the scores were checked (Hansel, 1980). It is possible that Pearce could have left the library and observed Pratt turning over the cards to record their order. There were glass windows through which Pratt could have been observed. There were also glass windows above the doors. It would have been possible to stand on a chair and look through the windows above the doors. There was also a trap door that would have made it possible for Pratt to be observed (Hansel, 1980). Pearce, after failing in earlier experiments, became suddenly unsuccessful after being moved one-hundred yards away or more (Hansel, 1980). However, as was reported in the Journal of Parapsychology, the subject was given an opportunity to enter the room to observe the order of the cards. There was, however, no time for this to occur because the subject was to show up with his record in his own hand writing ready to hand Rhine upon entering the room (Hansel, 1980). Rhine, on the other hand, was watching Pratt record the order of the cards and therefore could not have been watching the windows for Pearce. There were discrepancies between the results given in the various journals. But one thing is clear, the next experiments involved a distant of two miles and there was no success. This would suggest no psi phenomenon.

III. Computer Game Interface

In a study designed to ascertain whether ESP and psychokinesis (PK) are unitary phenomenon, a computer game interface was used in which a random number generator was employed to test four ESP and PK. Six animated grey hound dogs were run across the computer screen from left to right in a mock dog race (Roe, Davey, & Stevens, 2005).
In the ESP condition the computer knew before hand the outcome of the race. The participants were asked to state which dog they thought was the winning dog, and then watched a replay of the race. In the PK condition the race was run in real time by the random number generator. The subjects were to ‘will’ their selected grey hound to win (Roe et al., 2005).
Overall performance was at chance levels for both ESP and PK trails, and for true and disguised trails.
Despite the fact that both ESP and PK were not demonstrated the authors attempt to explain the results in terms of the fact that the participants may have been aware of the deceptive nature that was involved in some of the conditions. The authors, however, clearly state that “there is no psi in this study” (Roe et al., 2005).

IV. Virtual Environment

In a study by Murray, Howard, Wilde, Fox, and Simmonds-Moore (2007), An Immersive Virtual Environment was employed to test the validity of telepathy. Here the subjects were involved in a high tech experiment in which they were physically separated to alleviate the possibility of fraud and sensory leakage.
One way that realism can be achieved is if the sender of the message and the receiver can experience the same environment where the target is located. If both sender and receiver could interact with the same objects it could facilitate the acts of sending and receiving. Actual touching of real objects could introduce extraneous variables in the transmission of telepathic information.
A sense of environmental immersion can be provided by using a head-mounted display (HMD). These can employ stereo images and sounds to create a realistic virtual environment. This type of technology allows researchers to increase mundane realism while minimizing damage to experimental control.
Participants included two-hundred subjects (88 males, Mage = 30.6, and 112 females, Mage = 27.99). The subjects were able to look around the virtual room by making movements of his or her head. The senders were able to hear the sound made by the object through an ear phone (left ear). They were also able to hear the spoken words of the receiver through speakers placed close by.
The interaction with the virtual environment was facilitated through the use of an instrumented glove which allowed the subjects “handle” the virtual objects. As well as being able to see the room at 360 degrees, the subjects were able to see their own virtual bodies. Targets were selected from the shelves through a gesture of the hand. When an object was selected, it affixed itself to the participant’s virtual hand. The participants could not see each other’s Virtual body at any time, although their environments were identical. Once the subjects felt comfortable in their virtual environment, were comfortable with the equipment, task, and selection/deselection gestures, the microphone from the sender’s room was unplugged.
The sender’s computer selected one of the objects sets and the sender saw one randomly selected object and three fixed opaque panels. In the receiver’s environment, four initial opaque panels were seen, hiding the objects from view. The sender could not only see the target object but could point a gesture with the instrumented glove. The object would come off of the shelf and affix it’s self to the senders virtual hand. At the same time, associated sounds were played through the headphones. For example, a hair dryer would make the associated blowing sounds.
At the end of the first trial, experimenter two switched off the speakers in the senders room and the subjects were allowed to remove their HMD. The subjects then signed a sheet to confirm what the target object is. Receivers, while wearing the HMD, were shown the four virtual objects, and asked to indicate whether any items they felt were not the target. The receivers were then asked to rate each object according to how much they felt each object was the target. This was indicated was a percentage from 0-100 for every object. These were then used to derive confidence ratings for each object. For trial two, sender and receiver reversed roles. Each object set would be expected to appear 50 times over 200 trials. This would be 25 times in each of the two sender receiver trials.
The ESP magnitude effects were scored by regrouping the ranks into extreme vs. middle. A binomial test was conducted to see if subjects were more likely to rate the object as one (indicative as psi hitting) or four (indicative of psi missing) rather than the middle scores of two or three. There were 200 trails of which 48% were extreme ranks and 52% were middle ranks. The mean chance expectation would be 50%. Murray et al. ( 2009) states that the study did not find support for the psi hypothesis.

V. Classical Conditioning

In a study conducted by Vassy (2004), telepathy was studied through the use of classical conditioning. This employed the used of measurement of skin conductance for a measure of presentiments (Precognition) and telepathy. The aim of Vassy’s experiment was to use skin conductance for detecting telepathy and asking specific questions about its nature. Vassy assumed that telepathy involves an unidentified physiological process in the brain of the receiver. Using a behavioral output, such as skin conductance, was used as a way to operationalize this process. Electric shock causes a sudden rise in skin conductance and is here established as the unconditioned stimulus. This skin conductance response to electric shock does not habituate. It causes a response even after many trials. The conditioned stimulus in this experiment was reception of a message from the sender (Vassy, 2004). A signal was sent to the sender via computer that he was to telepathically warn the receiver of an impending shock. After a constant delay, the shock was given to the receiver.
The main premise of this study was based upon detection of the conditioned response (the telepathic message), through the measurement of skin conductance (as was the unconditioned shock). The interval between the telepathic message and the electric shock was set to be longer than the latency time of the skin response, so the conditioned stimulus (telepathic warning) could perceive the electric shock.
In experiment 1 (of three), the sessions were approximately 10 trials each and lasted about 15 to 20 minutes. This length time was considered optimal to avoid boredom or nervousness in the subjects. Fifty sessions were planned with various sender-receiver pairs. This would hopefully determine if the occurrence of the conditioned response could be elicited. In the 2nd experiment, the sender only sent the warning stimulus in about half the trials. These were simply omitted to test if the receiver could detect the shock using precognition. The third experiment was a attempt to replicate the first. During the first experiment, the sender and receiver were in different rooms, with a room in between them behind closed doors. When the experimenter was also the sender, there were only two people involved. When the sender was another subject, the experiment sat in the middle room. The electric shocks were administered by a Hungarian-made device called “PC Simulator” which was originally designed for medical muscle rehabilitation. This was controlled by a computer, which also sent the “trigger” word to the sender: the word “send” on the screen on the PC. Skin conductance of the receiver was measured with two different devices. The first device was called “relax 96” and was obtained through the same company as the shock stimulator. The relax 96 was originally used for bio feedback relaxation. After the 24th run, a British made device was used. This was called the SC5 SA made by contact precision instruments. Both of the conductance instruments were connected to an IBM PC compatible computer. This recorded the skin conductance data and the shock stimulus markers.
Each subject was encouraged to come with a friend and could decide whether they wanted to be senders or receivers. If an individual came alone, that individual would be the receiver and the experimenter would be the sender. This is bad science in the sense that the experimenter is introducing himself into the experimental milieu.
In the administration of the electric shock, the skin conductance shows clear bumps on a graph. The latency time was typically 1.5 to 2 seconds. There are also two smaller bumps that occur without apparent reason. Vassy (2004) calls these casual bumps. The most relevant bumps are those that precede the shocks. These are called prestimulus regions and are the ones where the condition responses (telepathy) are expected. The visual triggers on the sender’s computer screen preceded the shocks by 6 seconds. Time was expected to elapse between the onset of the trigger and it’s precognition by the sender. Consequently, no skin responses from the receiver were likely during the two seconds following the onset of the trigger signals (Vassy, 2004). Therefore the length of the prestimulus period had to be four seconds at most (6 minus 2).
Because humans have an unconscious clock they can react to periodic stimuli. Thus the shocks had to be delivered with random intervals of time between them. Otherwise the responses to the periodic stimuli (shock) could be confused with the conditioned responses (telepathic warning message).
During the first 25 runs of the experiment, the pseudo-random number generator built into the Visual Basic Program was used. During the second 25 runs the random number generator of Marsaglia (Maclaren & Marsaglia, 1965; Vassy, 2004) was used. These generators created a time interval after each unconditioned stimulus (shock), after which the next prestimulus region began. The shortest possible interval was 20 seconds, where the longest was 60 seconds. The distribution of the timing of the shocks was uniform between the longest and shortest interval. In the second 25 runs, the timing of the shock was defined the same way, however another pseudo-random generator decided if there would be a shock or a non shock control stimulus maker. Thus, the probability of a shock being delivered was 0.5. Therefore, the interval in between shocks could be rather long. The skin conductance of and individual often varies without apparent reason, especially during an uncomfortable situation. Knowing that one is about to receive electric shocks is one such situation. How then does the experiment distinguish between these noise bumps and a conditioned response? This supposedly depended on the amplitude over a various threshold level and the actual absolute level of skin conductance. There are also bumps that are caused by the shocks themselves. These start at about 1.5 seconds to 2 seconds after the shock and last about 5 to 10 seconds with a fast decay period. Thus all 10 second long post shock periods are automatically omitted from the analysis.
If a bump starts in a prestimulus region it may not reach it maximum value in that region. It may, instead merge into the larger unconditioned response (shock) without its own maximum at all. In this case, the strategy would be to use the end of the bump at the moment of the shock. The moment after the shock it’s self can be the cause of mechanical or electrical artifacts this was used in experiments one and three.
The amplitude of the prestimulus bumps can be underestimated in this way, resulting in a false omission when the amplitude of the bump falls below threshold. A more realistic approach would be to take into account that the latency time of the bump is always greater than 1.5 seconds. Consequently the end of the bumps can be set at 1.5 seconds later than the shocks onset, if the bump is still rising. Sounds complicated? Try sifting through the some of the statistical analysis of the experiment! Noise and bumps and amplitudes of shock. It is all kind of nebulous. The relative amplitude threshold was set for its optimum value, .005, which was suggested by an overall analysis. All relevant points were counted in non prestimulus and non post stimulus regions where a bump could start (all points where the skin conductance was either unchanging or decreasing with time). The point wise response rate takes into account the subtraction of the estimate frequencies of casual responses.
The number of receivers in experiment 1 was 199. Each worked with their own sender. Multiple analyses were done on the results which are confusing in themselves. You would have to be a statistician to understand them. Vassy (2004) does state, however, that all but one threshold gave significant results. Vassy (2004) also states that the outcome of the experiment was significantly positive.
The main thesis of experiment one was ‘the finding that there were more skin reactions in the prestimulus regions than elsewhere’ (Vassy, 2004). Vassy goes on to state that the method applied seem appropriate for the investigation of telepathy (or precognition if interpreted as such). Vassy (2004) states that the best results were found at an amplitude threshold of 0.005. This amplitude was chosen for experiment two. The four second wide prestimulus time window was deemed to be adequate. The average time of recognized shocks was about 2.6 seconds.
In order to determine whether the prestimulus responses were due to telepathy or precognition, the trigger stimulus “Send” was omitted in approximately half of the trails in experiment 2. If the number of prestimulus responses turns out to be the same in both trials, it means that the sender has no role in the experimental process, and all of the noncasual responses are precognitive. If the number of prestimulus responses in greater in the telepathic trials, it is assumed that the receiver responds to telepathic messages.
The set up for experiment 2 was similar to that of experiment 1 except that the Contact Precision Instrument SC5 SA designed to measure skin conductance was used throughout. The computer program randomized the decision to either show the trigger signal ‘Send” to the sender, or not. Two sender receiver pairs were selected from the subjects of experiment 1. Both pairs were from among the best performers in experiment 1. These subjects were paid for their participation. The dependant variable was the number of prestimulus responses of at least 0.005 amplitude, both in the telepathy trials and the precognitive trials.
Vassy (2004) states that of the 101 telepathic trials, 12 ‘telepathic’ responses were elicited. In the precognitive trails, four “precognitive” responses were elicited. It doesn’t take fancy mathematics to show that this is not astounding! After analysis, Vassy (2004) states that while more trials are needed the telepathy hypothesis seems to be more likely than the precognitive one.
Experiment 3 was an attempt and replication of the findings of experiment 1. The only difference in the methods between experiment 1 and experiment 3 was that the SC5 SA conductance instrument and the Marsaglia random number generator were used throughout. The receivers used partly overlapped with the subject pool of experiment. They’re where a total of 19 receivers, 12 of which worker with their own senders.
The results of the attempt to replicate the former experiment were unfruitful. There was no significant difference between prestimulus responses and chance expectation. The author states that the “prestimulus region” does contain significantly more bumps than the others. (Vassy, 2004) I cease to be impressed. Vassy (2004; p. 348) justifies all this by stating that “the positive outcome of experiment 1 was not caused by classical conditioning in the first place, especially because there in another convenient interpretation of precognition of the future shock.” This sounds like self justification to me. Vassy (2004; p. 349) goes on the say that “… it may be that after some trials that the conditioned stimulus does not arrive, or arrives more and more rarely, to the brain of the receiver. Then no reinforcement can take place, and even the association that developed earlier becomes inhibited.” He even adds that there may be a suppression mechanism that comes into play. Anyway, it looks as if the overall study is a bunch of speculative assumptions based on insufficient data.
Postal and World Wide Web Study

Post hoc, Steincamp (2005), explains there may have been good reason why the different experimental settings may have produced different results. In the laboratory condition subjects listen to a relaxation tape suggesting “that they were in a timeless zone in which they could see past, present, and future at once” (p. 343). These subjects were also subjected to white noise for 20 minutes while they verbalized any thoughts that arose concerning they’re precognitive target. Afterward they were shown four pictures to the rated on their similarity to their thoughts. This connotes a strong suggestion to the subjects in favor of precognition.
In the postal experiment, subjects were send a sealed envelope with four possible targets and were asked to take time at home to gain an insight into the picture that they would receive in the mail. They later opened the seal envelopes and rated the four pictures in terms of similar to their mentation. Steincamp (2005) suggests that because the subjects had their four targets already in hand, and because they had not been subject to the relaxation tape, the subjects may have been more focused on the envelope at hand rather the image they were to receive in the mail. This may be why the laboratory experiments showed results in the precognition experiment whereas the postal experiment showed my significance in the clairvoyance experiment.
The fourth study in the group was a replication of the postal experiment in which Steincamp (2005) made this condition more closely resemble the laboratory experiment. In this study participants were sent the same relaxation tape as in the laboratory experiment. Also, subjects were not sent the target materials until after they had sent in their responses. The experimental power had been increase so that if an effect was present, it could be more readily detected. Each subject in this condition was given only one trial.
Eight hundred pictures were created (200 target sets of 4), and around 60 sets consisted of pictures that had been used with success in other studies. The other pictures were printed from the WWW or were post cards. Some of the pictures were cartoons, some were abstract art, other were photographs or realistic art and pencil drawings. These pictures could be either black and white or colored. Each picture was mounted on a 9 by 6 in. grey cardboard and had its own random number on its back for identification purposes. The first information pack send to the subjects consisted of a welcome letter explaining the experiment, and empty envelope, a relaxation tape similar to the one used in the previous laboratory experiment, a sheet of paper where they could record their impression, and optional questionnaire, an addressed envelope with postage paid for their reply.
The second information packet consisted of: a letter explaining what the subject was to do next, a sealed envelope containing their four possible targets, a ratings form, and a prepaid envelope for returning the materials. On receiving their packets, subjects were asked to write their name on the empty envelope, draw a picture, or otherwise personalize it. They were informed that this was a way of bonding with their envelope, because it would later contain a picture that the experimenter would mail them. Subjects, at their own convenience would listen to the relaxation tape, attempt to get impressions about the pictures they were to be sent, and then record their impressions on the provided sheet. The subjects then returned the tape, their impression, their envelope that they had personalized, and optionally the questionnaire in the prepaid envelope. When the experimenter received the envelope, it was given to a colleague unopened. The colleague then opened up the received materials, checked that everything was completed, and then photocopied the mentation (impression) and put the impression in the second information pack so that the subject would be reminded of his impressions. The experimenter was allowed to see none of this.
Upon receiving the second packet, subjects opened the sealed envelope containing the target set of four pictures and rated each picture as to how well it fitted the impressions they had gained and gave a percentage rating. The picture to receive the highest rating was the picture the subject thought they would be later receiving via the mail. The subjects were instructed not to give and two pictures the same rating and to return the ratings and the pictures in the mail.
The subject’s envelope was opened and the rating that the subject had given each of the four pictures was converted into a rank. If, for instance, the subject gave the highest percentage rating to the target picture it was converted to a rank of 1. If it had gotten the second highest percentage rating, it was given a rank of 2, and so on.
The subjects where then sent the target picture in the envelope they had individual personalized along with feedback about whether the picture was their first, second, third, or fourth choice. Each subject completed only one trial for this experiment and it was deemed to be a test of true precognition. Three reminders were sent to have the subjects send their rating. When all three had been sent, the experiment was deemed to be completed. The response rate in this study was similar to the aforementioned previous postal study. This was 57% and 55%, respectively.
Steincamp (2005) states: “The main aim of this study was to see if evidence could be obtained for true precognition. The study yielded no evidence” (p.347). Steincamp (2005) explores the possibility that because a third of the subjects had to be sent at least one reminder before the experiment was completed; perhaps they had lost the motivation to complete the experiment. However, there was no correlation found between the number of reminders the subject had to be sent and the rank to the target. Consequently, neither this postal study, nor its previous attempt yielded any evidence for precognition.

VI. Uri Geller

Anyway, of the above mentioned studies, one can only find a glimmer of hope that such a thing as ESP exists. It appears people want to believe such a thing. While being subjective, I have a strange memory of my mother being able to guess a number that I had randomly picked. Since there were no other people involved, it seems strange that she could guess a number between 1 and 100, 8 out of 10 times. Perhaps this involves fabricated memory on my part, but I went through my childhood believing that my mother was capable of telepathy. This caused me to believe strongly in the existence of such a phenomenon. In turn, I was impressed with such people as Uri Geller who “demonstrated” such feats as bending a spoon, spike, key, or other metal objects “with his mind.”
Uri Geller was an Israeli born man who moved to the United States when he was “discovered” by Dr. Andrija Puharich. Uri usually had a man named Shipi Shtrang with him when he preformed his “feats” of mental powers.
James Randi, a noted magician and skeptic of anyone who claims to have psychic powers, managed to meet with Geller on several occasions (Randi, 1982). During a visit with Geller on a metro liner, Geller manages to take a nail file bend, given to him by Randi. Seemingly, Geller only had to run his finger over the file to get it to bend then break.
Uri appeared on the Mike Douglas Show and bent keys for Hugh Downs and some of Douglas’s production staff. Uri has women swooning in the audience and when in public he always had women slipping him their phone number. What a name he had made for himself through his purported psychokinesis. The most amazing thing about Geller is that he is a fraud. The “Amazing Randi” (James Randi) is able to duplicate all of Geller’s feats and can duplicate all of Geller’s feats using only normal magician’s techniques (Randi, 1982).
In 1972 the Stanford Research Institute was conducting research into parapsychological phenomena. Both of the institute researchers, Harold E. Puthoff and Russell Targ, were laser specialists. These two men were both into the pseudoscientific beliefs of Scientology (Randi, 1982). Their famous research report “Information Transmission Under Conditions of Sensory Shielding,” included Geller in their 6 week study. It began circulating in scientific journals “like an academic gone with the wind” (Randi, 1982, p 30).
Nature magazine was skeptical about printing the report and wrote and introductory editorial that “made it quite clear that the; SRI report was published as an example of the kinds of research currently being done in the field of parapsychology and in no way endorsed the report or the research methods used to perform the test therein” (Randi, 1982, p. 31).
In this introduction it was made clear that “there was agreement that be paper was weak in design and presentation, to the extent that details given as to the precise way in which the experiment was carried out was disconcertingly vague” (Randi, 1982, p. 31). According to Randi (1982), even an avowed believer in Geller said that the researchers “gave in to whim of Geller and that he was allowed to jump up and run about at will refusing of postponing attempts at any test, returning to other uncompleted ones and, in general, thoroughly misdirecting their attention” (p. 31).
Geller was to perform tests to determine whether he could reproduce pictures drawn by the experimenters in a electrically shielded room, reproduce drawings stored in the memory of a computer, and reproduce drawings enclosed in seal envelopes (the contents of which were unknown to the experimenters).
After Geller was in the shielded room, a ‘target’ would be chosen. This was done to prevent the possibility of advance cueing. He was not told who would be doing the choosing, or how the choosing would be done. One method of choosing was to open a dictionary (randomly) and selecting the word seen first that could be represented by a picture (this was posted on the wall, by the way!). Sealed pictures were also given to the experimenters, as mentioned before. A target pool was also used on which the experiment gave little information.
According to Randi (1982), Geller either attempted or “passed” on the drawings and he was allowed to pass on any test he was not sure of. His drawing was taken away from him before he was allowed to see what the target drawing had been. Through examination of the room no sensory leakage was found, though nothing is said in the report about what methods where used to determine this insulation. A second set a tests Geller had done were designed to test his clairvoyance. In these tests no one knew what the test object was so telepathy was ruled out and thus any results would be considered clairvoyant.
One hundred drawings were sealed up in envelopes. On all of these trials, Geller “passed.” However, everyday he made 12 drawings that he claimed were associated with the entire set of drawings. On each of three days two of his drawings would be associated with two targets in the day’s test drawings. But this did not produce any effect greater than chance.
As to Geller’s possessive psychokinetic “powers” Randi’s youngest student could solve the trick immediately. Geller would wave his hands over a compass, after careful examination with a probe to be sure he had no concealed magnets. The compass needle deflected! However it moves not with his hand movements, but in response to the movement of this head. Apparently the magnet was in his mouth.
The experimenters watched Geller stroke a spoon in order to make it bend. The spoon did not bulge at all. However you can see the spoon bending trick in film divided into five segments. In each segment Geller can be seen stroking the spoon and turning it every direction while concentrating steadily. Every time a segment would end “to wind the camera or put a new film in” the spoon returns to the picture with a new bend in it!
Gellers’ assistant, Shippi, was often there, according to Hamlon, during the experiments passing as Geller’s brother or cousin. The steel room referred to as the “Faraday Cage” is not anywhere near sound proof. Taps on the outside to the room can be easily heard inside. This is supposed to be a test done under controlled conditions. One of the telepathic trials included 30 responses to a simple target, of which only two seemed to correspond to the transmission. These were tridents: “three pronged forks.” Curiously enough of all the drawing responses were done carefully, comma except for the tridents. These could have been added very quickly as Geller emerged and saw the target. On another test Geller had made a number of drawings, then he was shown a television set. On the back of his response to another drawing are a few faint lines that look like a part of a square. Geller grabs the pen from the experimenter and draws in a television over his lines, adding two little knobs, and his name. Geller said “I mean look, you saw it,” as he puts down the pen. “The pen is down here… I got your television set!”
On the tridents they are the only sloppy drawings in a set of 30, one of them overlaps another drawing, carelessly. Randi (1982) concluded that Geller came from the room saw the drawing on the wall and captured the one theme he could quickly sketch, “the trident.”
Puthoff and Targ (See Randi, 1982) allowed loose conditions to exist during the test. Geller could demand to be let out of the sealed room and Geller’s “cousin” Shipi was readily accessible. The two experimenters were so eager to keep Geller around, that they met his every whim. If he threatened to walk off they would do what he wanted. Bad science!
The room had refrigerator type doors. However, in order to get cables in for the electroencephalograph (EEG) test, a square hole had been cut into the steel wall and the space around it had been stuffed with gauze. It was reported that gauze was found on the floor after Geller had correctly made a drawing of some grapes. Randi (1982) believes that Shipi made a drawing of the target and pushed it through the hole to Geller. Geller had chosen not to try many of the other transmissions, but made a very suspicious direct “hit” on this one.
A rabbit, a tree, and an envelope were all “passed” by Geller, though they were simple basic objects that one would think a psychic could “pick up.” However in the case of the 100 envelopes under real experimental conditions, Geller was unable perform. No one was able to signal the answer to him. Only in these tests where there was no possibility of transmission of information by a confederate did Geller refuse to try the test or plainly failed it.
In the “SRI” report the 13 drawings were reported but it fails to say anything about the 100 envelopes Geller would not try. Randi assumes, because the account was not to Geller’s credit, it was not mentioned. The “SRI” did not report exactly what happened. It could not have because there were seemingly unimportant occurrences they did not attend to, and lapses in their accuracy. In the room they assume perfect shielding of Geller so he was not searched for the presence of electronic equipment.
On an occasion when Randi and Geller were in a hotel lobby Geller suggested that he try bending metal. Randi gave Geller his house key which he worked with unsuccessfully. Geller suggested that they move to the next room. There were a few soft chairs near the door. Geller suggested he try bending it against a metal ashtray (by rubbing the key against it) in which he was unsuccessful. Finally, Geller suggested that they move into the corner and sit down on a sofa. A gentleman was sent to get Geller’s jacket. Geller sat down first and Randi walked around the table and was just sitting down as the gentleman was walking across the floor with Geller’s jacket. Therefore, neither of them was watching Geller closely. Suddenly, Geller launched forward and spread his legs so rapidly that be split his pants. After making a joke about his ripped pants, he held the key from the point end, enclosing most of it in his hand. Geller’s hand was slightly arched and Randi could see that the key had already been slightly bent. Geller suddenly said “it was bending,” and slowly moved his hand down the key to expose the bend. It was not a large bend but he demonstrated it putting the key on the coffee table.
Randi states that he cannot actually say that Geller did not bend his key by non paranormal means. But he can offer an explanation that he find more plausible then previously unidentified mental forces. First it should be noted that keys are surprisingly easy to bend, especially with someone who has strong hands like Geller.
Sitting in a chair with your legs slightly spread, one is able to reach down to the bottom of the chair set and feel part of the chair frame. If you hold the head of the key in both hands and put the point on top on the frame, pressing down, the key bends quite easily. With practice this can be done with a quick casual movement. The Johnny Carson show in the United States, proceeded to broad cast Geller’s failure and his discomfort, from a tape that Geller had made previously. The original show had to be cancelled and the tape scrapped, because Geller looked bad. Time magazine reported that a psychologist named Merarari agreed with Geller to examined his “powers,” with the agreement that if his results were not positive no report would be issued.
Geller uses the ploy of failing a certain amount of the time, often much more than he succeeds. This is to give the impression that miracles don’t just always happen when we want them to the most. Also audiences figure that if Geller’s abilities were just a trick, they would work every time. Since this is not the case the demonstration must be the real thing.
Randi believes that whether Geller had the power of mind over matter could not really be ascertained without a doubt. He had not seen Geller fix watches with his own eyes, but he believed the testimony of those who had. The experiments in telepathy did not move Randi. He had seen magicians give similar performances using trickery.
Randi likes to believe that telepathy exists as well as psychokinesis but he has never experienced psychokinesis. The proposition that human consciousness is a form of real energy seems to be self-evident. So why shouldn’t consciousness be able to affect the physical properties of things?
Geller made convincing appearances with Merv Griffin and Jack Paar. On the Jack Paar show he bent a heavy metal spike. In the mean time, favorable articles about him were being published all over the place. “Time” magazine was the only magazine that accused him of being a fraud.
Someone asked Geller to try to bend a fork. Geller replied he did not like to work with silverware, it bended too easily. As he spoke he picked up the fork by the middle in a casual way and the fork drooped over his hand. Geller said, “My god! Look at that! I was not even trying to do it!”
Evidently there is a lot of trickery and even anecdotal accounts that kept Geller in the mainstream. It has already been mentioned that women would swoon over Geller, he was a charming personality and a good looking man. This only adds to his mystique.
The New Zealand Rationalist and Humanist magazine published an article about Geller in 1978. Geller was exposed as a Fraud by his friends and coworkers. This was after he had already convinced Apollo-14 astronaut Ed Mitchell to help fund the research at SRI. Nature magazine had published an article that consisted of an “SRI report” that documented Geller’s remarkable “powers.” The New York Times, Newsweek and even Barbara Walters treated Geller with awe.
In Israel, however, his friends, relatives, his girlfriends and his managers swear that Geller is a cheat and a liar; and they are able to perform Geller’s tricks with their own hands. Geller’s father was an officer in the Israeli army, but Geller told that he was from a circus family. Itzhaak Saban, a close friend of Geller’s, revealed that Geller used him as a confederate in the audience. He used to sit in the front row during Geller’s performance, and gave Geller all kinds of signal motions to enable Geller to succeed.
Shipi’s sister, whom Geller used to date, told how Geller and Shipi met at summer camp where Geller was a counselor. Uri told Shipi about a magic book and the two kids decided to make a business at it. Uri and Shipi used a method whereby Shipi use to sit in the audience and pass signals to Geller that had been practiced beforehand. After Geller became known for his “abilities,” he got a manager named Baruch Cotni. Geller presented Shipi as his younger brother.
Geller became a success in Israel because of his “supernatural powers.” Baruch Cotni later shared that Geller use to have the mandate that Shipi sit in the first row, or at least in the middle of the second row, Shipi’s sister became part of the act in that she would sit next to shipi where she passed signals to Geller.
Geller’s chauffer admitted to many tricks that Geller had revealed to the chauffer. He said Geller had confessed in a conversation that everything in the act was a bluff. A magician named Ayalon began replicating all of Geller’s tricks and claimed that he was also the “real thing.” A professor of physics from the University of Tel Aviv whom was also a magician was able to identify most of Geller’s tricks. The professor invited the producer of Geller’s show to his own house where he performed most of Geller’s feats and the manager was shaken. The manager refused to expose Geller because of obvious business reasons. Need I say more? People obviously wanted to believe in Uri Geller, but he has been exposed as a fraud.

VII. Why do we want to believe?

People want to believe in all of this. Jones et al. (1977; as cited in Blackmore & Troscianko, 1985), found that 58 % of a sample of students believed in about half of paranormal phenomena and 27 % claimed to have experienced one. Blackmore (1984; as cited in Blackmore & Troscianko, 1985), found that 25 % of those who responded claimed to have experienced telepathy, 36 % believed in ESP, and of these 44 % said they believed in ESP because of their own experiences.
Errors in judgment about probabilities can be a main cause of belief in the paranormal. People will show a lot of confidence in their erroneous assumptions even in the face of contradictory evidence. People can underestimate the probability of a coincidence. Believers in the paranormal could make greater underestimations than non-believers. Selective forgetting is another mechanism by which errors of probability may occur. An individual may forget all the dreams that don’t come true but remember those that do (Blackmore & Troscianko, 1985).
Another illusion involves the ‘illusion of control’ (Langer, 1975; as cited in Blackmore & Troscianko, 1985). This is relevant to psychokinesis. People may feel they are under control of chance events. If no normal attribution is made, the illusory control may be attributed to psi. Langer (1975; as cited in Blackmore & Troscianko, 1985), believed that the illusion of control was highest when subjects though the task involved a degree of skill. Lander and Roth (1975; as cited in Blackmore & Troscianko, 1985), showed the conditions which produced this effect (such as early success), and showed a tendency for remembering a higher number of successes than had actually occurred. In an ESP task, Ayeroff and Ableson (1976; as cited in Blackmore & Troscianko, 1985) found that perceived success was primed by the subject’s involvement in picking out the testing materials and being allowed to “warm up.” These things create an illusion of control. In that test, only subject confidence in success was affected. The actual ESP scores were at chance level.
Jones et al. (1977; as cited in Blackmore & Troscianko, 1985), found that the relationship in the belief in psi and illusion of control was found to be such that subject with a high paranormal belief score gave higher ratings of their performance on a psychokinesis type task than those who scored low on the paranormal belief scale.

VIII. Neurobiological studies

Turning now to neurobiology, a study by Moulton and Kosslyn (2008) uses functional magnetic resonance imaging (fMRI) to show that psi does not exist. They state that many people subscribe to the view that psi is a very real phenomena based upon peoples personal accounts. This study tends to debunk these accounts through using these imaging techniques. In this experiment the authors operationalize the psi hypothesis by asking: “Does the brain respond selectively to psi stimuli?” (p.183). Under the null hypothesis psi and non-psi phenomena should evoke indistinguishable neuronal responses.
Research in cognitive neuroscience suggests that there are two general effects that psi might have on the brain. On the one hand, psi might provide subjects with specific, implicit knowledge of stimuli. If this were the case, one would expect a “suppressed brain response to psi stimuli compared to non psi stimuli” (p.183). On the other hand, “psi might increase participants’ attention to stimuli without providing them with stimulus-specific content. In this case, we would expect an enhanced brain response to psi stimuli, given the evidence that attention enhances brain activity” (Corbetta, Shulman, Miezen, & Peterson, 1995; see Moulton & Kosslyn, 2008).
Moulton and Kosslyn’s approach may circumvent the problems that arise in psi research. This would involve going directly to the source (the effects of psi on the brain). Moulton and Kosslyn (1985) state that “because this paradigm uniquely minimizes assumptions about the source of knowledge, the kind of processing, or the nature of mental content responsible for psi, any ensuing null results will be qualitatively more informative than those from behavioral methods” (p. 183).
In Moulton and Kosslyn’s (2008) study, each trial consisted of sender participants viewing a randomly selected target stimulus from outside the scanner. The participant senders tried to send this information to the receiver by mental means alone. Participant receivers were completing a simple binary guessing task, and fMRI was used to monitor brain activity. On each trial, the receivers sequentially viewed two stimuli, guessing which one was the stimulus being “sent.” The receivers then saw the psi stimulus a second time. This allowed Moulton and Kosslyn (2008) to cover all of the proposed mechanisms of psi: telepathy, clairvoyance, and precognition. The sender was the potential telepathic source, the computer monitor provided the potential source of clairvoyance, and the second presentation of the psi stimulus was the potential source of precognition.
This experiment was taken from earlier experiments using event related potentials (ERP). This is based upon the proposition that psi stimuli differs psychophysiologically from non-psi stimuli. However, the dependant measure of blood oxygen dependant (BOLD) hemodynamic response is much more informative than low density ERPs. fMRI provides far more information about the location of brain activation than does low density ERP’s and affords the opportunity to detect more possible “sources” for an effect. To further enhance the possibility of a psi effect, biologically or emotionally related participants (e.g., identical twins) were recruited. This was due to anecdotal and experimental evidence that suggests they are more prone to psi (Sheldrake & Smart, 2003; Taylor, 2003; Playfair, 1999; see Moulton & Kosslyn, 2008). The experimenters also used emotional stimuli which have been implicated in psi as well (Sherwood, Dalton, Steinkamp, & Watt, 2000; Cornell, 1999; Moss & Gengerelli, 1967, 1970; Myers, 1903; see Moulton & Kosslyn, 2008).
The behavioral task served two purposes: it enabled participants to detect the psi stimuli, and it provided structure for the BOLD contrasts. There are no behavioral tasks that can be considered as “standard metrics” of psi, and it is precisely this failure that has motivated this neuroimaging study. All behavior requires neural activity, but not all neural activity activates behavior. “Thus in the current experiment, although the unexpected presence of a behavioral effect mandates a neural correlate, the expected absence of a behavioral effect does not necessarily imply null neuroimaging results” (Moulton & Kosslyn, 2008, p. 184).
Nineteen pairs of individuals including six couples, five emotionally close pairs, three identical twin pairs, one mother-son pair and one pair of sisters were recruited for the experiment each session was about two point five hours, with one point five hours devoted to the experimental task and the rest was used for paperwork, instructions, training, and debriefing. Not only were the participants compensated for traveling expenses, they were also awarded $77-$134 depending upon task performance and a flat rate of $50 for the senders.
The International Affective Picture System (IAPS) stimulus set (Lang, Bradley, & Cuthbert, 1995; as cited in Moulton & Kosslyn, 2008) was used. Two hundred and forty pairs of photographs was used as test stimuli. Negative high-arousal pictures (e.g., snake) were paired with either positive low-arousal or neutral pictures; positive high-arousal (e.g., erotic couple) were paired with either neutral of negative low-arousal pictures. Pictures were randomly assigned half of the psi stimuli to each position, either first or second, to ensure that psi assignment was not confounded with stimulus presentation.
All stimuli that were presented during fMRI scanning were generated using a Macintosh G4 Computer and Psy Scope software (Macwhinney, Cohen & Provost, 1977; as cited in Moulton & Kosslyn, 2008). An LCD projector and a mirror attached to the head coil were used to present visual stimuli. Behavioral responses were recorded by using a mr-compatible button box.
Participants completed a practice version of the experimental task for about ten minutes in order to become familiarized with the stimuli, procedure, and minimum response window (2 seconds). An initial scan of the receiver participants was implemented for approximately 25 minutes, while they viewed a screen saver. After the initial scan five series of functional scans were administered while the participants completed five sets of the experimental task. Each set consisted of 48 trials, and each trial contained three sequential exposures two unique AIPS stimuli. First and second exposures, lasting one second each, presented the assigned psi and non-psi stimuli to the participants from each stimulus pair (with the presentation order varying across trials), and were separated by a random period of fixation (2,4, or 6 seconds). From the perspective of the participant each of the first two stimuli was a potential psi stimulus. After the second exposure participants saw (press key now) on the screen and guessed as to which item was the psi stimulus by pressing one of two buttons. After this forced-choice response, participants were presented with the correct psi stimulus (for one second) and feedback (“Correct! + $1,” or “Wrong!,” 2-6 seconds). Receivers were told that they would receive an extra dollar for each correct response.
The senders were in another room where they could not see or hear the scanner. At the start of each series, the investigator indicated for the sender to begin, and the sender started sequentially viewing the psi images. The receiver and sender scripts were synchronized in time such that throughout the each entire trial of the receiver’s task, the sender viewed that trial’s psi stimulus. The senders spend ten-twenty two seconds viewing each stimulus depending upon timing of the receiver’s stimulus presentation. The investigator asks senders to “influence the receiver” with the psi stimulus. The investigator did not provide senders with real-time feedback. Both reviewers and senders were informed of the overall hit rate at the end of the experiment.
Brain and non-brain image data were segmented using FSL’s Brain Extraction Tool (Smith 2002 a; as cited in Mouton & Kosslyn, 2008). Each extraction was inspected manually and conducted again with new parameters (Center of Gravity, Fractional Intensity Threshold) to ensure maximal segmentation data were convolved in 3-D with a 5-mm (full width half maximum) Gaussian kernal to increase signal-to-noise ratio and allow later GRF-based statistical tresholding. The accuracy of each registration was visually inspected and each brain volume was manually checked for evidence of scanners spiking or other image anomalies. For each series neural responses to the psi stimulus, non-psi stimulus, and feedback stimulus, as well as their temporal derivatives were modeled with a double-gamma canonical hemodynamic response function (Glover 1999; as cited in Moulton & Kosslyn, 2008).
The behavioral data yielded exactly at chance on the guessing task. None of the results from any individual deviated from expectation on the basis of chance variation alone. For the fMRI results analysis of group data revealed no evidence of psi. Psi and non-psi stimuli evoked wide spread but in distinguishable neuronal responses. Moulton and Kosslyn (2008) state that “null findings are always difficult to defend, in only because they could simply indicate that the technique lacks adequate sensitivity.” Moulton and Kosslyn contrasted activation elicited by the psi stimuli with that elicited by the feedback stimuli. Psi stimuli were novel and likely to provoke response preparation. On the other hand, the feedback stimuli were recognizable and probably provoked reward anticipation. “A combination of repetition suppression, fMRI adaptation, and attentional enhancement effects might also underlie any differential activation. This analysis revealed numerous regions of the brain that discriminated between psi and feedback stimuli” (Moulton & Kosslyn, 2008, p. 189). However note that the number and strength of the results indicate that the technique in fact sensitive to subtle changes in information processing.
On half the trials, low arousal pictures preceded high arousal pictures, and on the 2nd trial high arousal pictures were presented before low arousal pictures. Three areas of significant brain activation were found to contrast in high arousal versus low arousal pictures. These were the left occipital temporal regions, right occipital temporal regions, and the right precuneu region..
The brains of the participants reacted to psi and non-psi stimuli in a statistically indistinguishable manner. Even if the psi effect were transient, “it should have left a footprint that could be detected by fMRI- as did the other subtle effects that [were] detected” (Moulton & Kosslyn, 2008, p. 189). The large activation of the arousal contrast shows that if a psi effect exists, it must be substantially smaller than the effect of arousal on brain activity.
This study incorporated the technology of cognitive neuroscience in conjunction with behavioral data. Methodological variables widely considered to facilitate psi were used. This means that the null hypothesis results not only fail to support the psi hypothesis, but the results offer strong evidence against it.

IX. Ray Hyman

Turning once again to behavioral data, Hyman (2010) looks at the overall data produced by the Ganzfield experiment, which is considered by many parapsychologists to be the greatest indicator of psi phenonmenon. Randi (2010) defines the Ganzfield experiment as “ [a] technique used in the field of parapsychology to test individuals for extrasensory perception (ESP). It uses homogenous and unpatterned sensory stimulation to produce an effect similar to sensory deprivation” (p. 2).
The primary way that researchers are providing statistical significance is through meta-analysis. Some parapsychologists have admitted that evidence in their field is “inconsistent, unreliable, contradictory, and elusive” (Atmanspacher & John, 2003; Bierman, 2001; Kennedy, 2001, 2003; Lucadou, 2001; as cited in Hyman, 2010, p. 487). Hyman (2010) goes on to say that “parapsychology’s Achille’s heel is its persistent inability to come up with even one example of prospective replicability” (p. 487).
In attempting to replicate former Ganzfield experiments, Broughton and Alexander (1997) found a hit rate of 25.8% (chance = 25%). The authors admit that the replication failed. At about the same time, Bierman (2001; see Hyman, 2010) conducted experiments that also failed to replicate the original Ganzfield. Storm et al. (2010; see Hyman, 2010), manufactured his own significance by relying on meta-analysis which masked rather than revealed the true outcome of the Ganzfield studies. They “create ‘homogenous’ databases by removing outliers” (Hyman, 2010, p. 487). This makes the effect size less variable but does not change the original experiments heterogeneity.
A scientist should remember that “an effect size is simply a standardized discrepancy between an observed outcome and an outcome expected by chance. The combining of effect sizes from different studies makes sense only if one can show that the separate effect sizes are conceptually coherent- that they all can be attributed to the same underlying cause” (Hyman, 2010, p. 487). After searching through the original Ganzfield experiments Hyman (2010) shows that the meta-analysis that was performed, consisted of a mixture of above- chance results with other experiments that were consistent with chance. The bottom line is that this is an arbitrary mixture of results and studies have failed to provide independent, replicable evidence. Hyman concludes that “despite over a century of failing to come up with even one replicable experiment, parapsychological claims are still with us, whereas the others occupy science’s discard pile” (p.489).

X. Conclusions

James Randi (2010) has a website where an individual can print Zener cards which are the standard cards used for testing for ESP. Randi offers a test for students in grades 9 through 12. Randi also offers some important guidelines for testing. First, avoid accepting only favorable results or stopping the trials prematurely. Second, we are looking for replicable results. This requires sufficient sample size and tightly controlled conditions. We are looking for statistical significance. Third, the test should be double blind whereas a third party (unaware of expected results) evaluates the results. This avoids testing bias.
Randi (2010) also states 5 rules: First, declare in advance if a test is a dry runor an actual test. An actual test should be counted in the final evaluation. Second, always decide beforehand the number of trials that will be run. Third, the number of trials should be as large as practical. Fourth, keep a careful record of test conditions. If there is a confounding factor or condition that is affecting the tests, either positively or negatively, perhaps this can be pinpointed. Fifth, use an outside judge to tabulate results.
Randi has established a foundation to fund research into paranormal claims. He provides scholarships to students who exhibit critical thinking and research grants for educators. Randi also provides seminars and workshops through the public media to examine the consequences of blindly accepting paranormal claims.
Randi’s foundation has set up a $1,000,000 prize for any individual who can demonstrate psychic, paranormal, or supernatural ability under mutually agreed upon scientific conditions! If all of the preceding information doesn’t cause you to doubt the credibility of psychic claims, it would seem that the $1,000,000 offer would cast doubt. If any individual is able to engage in psychic ability, why would he not come forth for a prize of that magnitude? I hope this chapter has aroused your internal skeptic and managed to entertain as well. Aw, I knew you were going to say that.



American Psychological Association (2007). APA Dictionary of Psychology. Washington, D.C: American Psychological Association.
Blackmore, S., & Troscianko, T. (1985). Belief in the paranormal: Probability judgments, illusory control, and the ‘chance baseline shift.’ British Journal of Psychology, 76, 459-468.
Hansel, C. E. M. (1980). ESP and Parapsychology: A Critical reevaluation. Buffalo, New York: Prometheus Books.
Hyman, R. (2010). Meta-analysis that conceals more than it reveals: Comment on Storm et al. Psychological Bulletin, 136(4), 486-490.
Moulton, S. T., & Kosslyn, S.M. (2008). Using neuroimaging to resolve the psi debate. Journal of Cognitive Neuroscience, 20(1), 182-192.
Murray, C. D., Howard, T., Wilde, D. J., Foxx, J., Simmonds-Moore, C. (2007) Testing for telepathy using an immersive virtual environment. Journal of Parapsychology, 71(1), 105-121.
No Author Cited. (1978). Uri Geller- a psychic fraud. New Zealand Rationalist and Humanist, 39(12), 1-4
Randi, J. (1982). The truth about Uri Geller. Buffalo, New York: Prometheus Books.
Randi, J. (2010). Do you have esp? Retrieved from http://dl.dropbox.com/u4648762/Do%20you%20have%20esp%20final.pdf
Roe, C. A., Davey, R., & Stevens, P. (2005). Are ESP and PK aspects of a unitary phenomenon? The effects of deception when testing the relationship between ESP and PK. Journal of the Society for Psychical Research, 69(1), 18-32.
Steinkamp, F. (2005). Does precognition foresee the future? Series 4: A postal replication. Journal of Parapsychology, 69(2), 341-351.
Vassy, Z. (2004). A study of telepathy by classical conditioning. The Journal of Parapsychology, 68, 323-350.