Frequently, when someone claims a positive psi result, Dr. Richard Wiseman appears in the media giving reasons why it probably isn’t psi, often quoting a similar experiment that he has done which has failed. To his credit, he has been exploring his consistent failure with Marilyn Schlitz (1998) and they have discovered that whether in his lab or hers just by having Richard involved as experimenter the experiment will fail. However, I have never yet heard Richard mention this when telling the media of yet another of his failed experiments. I also know him as an accomplished stage magician and member of the Magic Circle who performed at the SPR Christmas meeting in 1998. The use of magicians in psychical research is important. They are aware of all the tricks that can be used to manipulate people and can thus help separate true results from false ones. However, it seems dubious that they should be performing experiments themselves. With their powers of manipulation they could easily, even subconsciously, be getting the results they want. Perhaps this is the origin of the Wiseman experimenter effect.
The first I heard of “The World’s Largest Esp Experiment Ever” was on the evening of the sixth of December 2000 when I heard Richard on BBC Radio 4’s Midweek describing how thousands of people were expected to take part in a public experiment to test telepathy the next day. Richard’s press release (13/11/2000) showed he was replicating the well-known Ganzfeld telepathy experiment but using many ‘senders’ based on a 1971 experiment with 2000 senders at 6 ‘Grateful Dead’ concerts, described by Ullman, Krippner and Vaughan (1973).
Richard was to attempt 10 trials and he accurately pointed out, 6 direct hits were required for significance (p = 0.0197, binomial p:0.25). The Ganzfeld with single senders has an expected hit rate of 1 in 4 or 0.25 and an observed hit rate of about 0.30 but Ullman et al. concluded of many senders that “Certainly no particularly striking improvement in accuracy was noted when compared with what ordinarily occurs when single agents are used”. So, based on the two past experiments that he was trying to replicate, psi can be expected to manifest with a hit rate of only 0.30, not the 0.60 required for 6 hits. Indeed, the probability of getting 6 hits is still significantly unlikely using an expected of 0.30 (p = 0.047, binomial p:0.30). In other words, here is an experiment which should fail whether psi exists or not.
Despite Richard being a Council member he apparently hadn’t informed the SPR of the experiment, thereby foregoing the potential involvement of interested volunteers and independent expert invigilators. Suspicious that someone didn’t want psychic investigators to attend and intrigued by the Wiseman experimenter effect, I organised a tape recorder and camera and managed to arrive just as the first trial was starting barely 12 hours later.
The first point to make was that there was no-one there. Well, Richard and his technicians and venue staff were and perhaps 5 or 10 members of the press but absolutely no-one else. This was more like a press conference than an experiment to test many senders.
The experiment was apparently well-financed and technologically sophisticated. It involved the co-experimenter Matthew Smith putting a ‘receiver’ through the standard Ganzfeld procedure: relaxation while white noise was played through headphones and a red light was shone at halved ping pong balls placed over their open eyes. The receivers had been chosen, not for any psychic abilities but for their artistic and extravert temperaments and were in an acoustically isolated room on the 19th floor of a tower block. Meanwhile any senders started looking at a slide randomly selected from 4 in the nearby “Museum Of The Unknown”. A one way audio link transmitted whatever the receiver said they visualised back to the senders and was noted by Matthew. At the end of about 10 minutes the session finished and Matthew read back to them what they had said. They were then shown the correct image mixed with 3 decoys and had to sort them into order of correspondence with what they had visualised. During the course of the day 10 trials, each with a different receiver, were attempted; 8 inside the museum and 2 with many more senders in an adjacent park. The first trial was a ‘direct miss’: the correct image being placed in 4th place. The next, 11 o’clock, trial may have had a couple more members of ‘the public’ and the receiver’s visualisations seemed much closer to the image, placing it 2nd. I then heard that an extra trial had been done live on GMTV at 8:30 a.m. that morning and been a ‘direct hit’: first place. For me, this was exciting because I knew this was right in the middle of the local sidereal time window which James Spottiswoode (1997) has found enhances the Ganzfeld effect size dramatically. I am finding a similar enhancement in my current Lottery project.
During the day Richard was relaxed, entertaining and seemed fair. For example, he made a point of not mentioning the results of the earlier failed trials to the senders before each trial.
Of the eight small scale indoor trials there was 1 direct hit, 2 in second place and 5 direct misses: noticeably below chance. However, the next two outside trials attracted more senders: 50 to 100. The first outside target was an image of the famous sculpture of George Washington cut into Mount Rushmore. Although the early feedback from the receiver was not particularly appropriate, towards the end he came up with “America, like a 3D virtual image/map….. old man..”. The receiver then selected the correct image easily. The crowd was very encouraged and were cheering enthusiastically.
At this stage, experimental protocols seemed to be forgotten. Chronologically my criticisms are:
1. For each trial the image to project had been decided by volunteers selecting balls from gold and silver magician’s bags. When this was completed for this trial Richard said “..which translates on mine to the animal..”. In other words, Richard already knew which image would have been used as a result of which number. The protocol of Ganzfeld experiments has been laid down by Hyman and Honorton (1986) in a rare spirit of agreement between sceptics and psi proponents. This protocol makes it clear that two people must be used to prevent the experimenter having any idea which image will result from the randomisation. When it is being done by a magician using magician’s paraphernalia, the requirement seems much more crucial. One reason for this part of the protocol is that otherwise it would be possible for an unconsciously biased experimenter to make sure that the chosen target image was one which had more (or potentially less) chance of being selected.
2. On selection of an Elephant image. Richard immediately said “Earlier today we had a moose and it didn’t go at all well”. Not only did this break the admirable protocol of not giving the results of previous trials to the senders but it clearly undermined the task at hand.
At this stage Richard announced that he had just been told that the next trial was to be shown live on Channel 4 news.
3. Richard then informed the senders about the 1971 experiment in which the receivers were professional psychics, saying “one psychic was not bad and the other was absolutely dreadful”. In this experiment, one psychic was the well-known Malcolm Bessent(1944-1997). The senders had been shown slides saying “Try using your ESP to ‘send’ this picture to Malcolm Bessent. He will try to dream about the picture. Try to ‘send’ it to him. Malcolm Bessent is now at the Maimonides Dream Laboratory in Brooklyn.” (45 miles away). The second psychic: Felicia Parise was a control psychic and wasn’t mentioned to the senders.
Of the 6 trials Felicia got 1 direct hit: exactly chance, as you might expect from a control psychic whom the audience knew nothing about. Malcolm got 4 direct hits.
Crucially, Richard omitted to mention that Felicia was a control. As a scientist who has spent much of the last 20 years developing control groups it upsets me to see a colleague diminishing an experiment by presenting a control group as if it were the main experimental group. I cannot think of a worse misrepresentation.
4. Even ignoring the unforgivable error of failing to mention the control, no scientist could fairly summarise a psi experiment where one subject was four times chance and the other exactly chance, using the terms “not bad” and “absolutely dreadful” respectively, unless they come from a world where psi is the accepted norm.
However, only three hours earlier at the 4 p.m. trial, Richard had said of exactly the same psychics one “was fairly successful, one wasn’t”
Playing down the significance of the only previous mass experiment in such a blatantly unfair and totally inconsistent way was bound to affect the confidence of the senders. Many in the crowd must presumably have arrived for the outdoor trials initially sceptical that it would work. However, having just witnessed the dramatic success with the George Washington image, they could have started to believe it might just be possible. Since then, Richard had improperly or inconsistently informed them of 3 previous results as “didn’t go at all well”, “not bad” and “absolutely dreadful”. This was certain to influence many of the senders and Richard had to be aware of this, because in the short write up of the 1971 experiment Ullman et al. wrote “For those who find it an exciting and novel challenge, ESP may actually be enhanced; for those who feel they are attempting something impossible and do not expect success, there may be no results.”
5. The main conclusion of the 1971 study was that the use of Malcolm Bessent’s name had established a rapport between him and the senders. Ullman et al. wrote “… it is easier for one subject to make telepathic contact if there is rapport.”. During the day Richard had told us the names of the receivers and they appear on all of the 5 previous trials I taped. However in the crucial final trial the senders were never made aware of the name of the receiver.
The result of the final trial was a miss with the elephant being placed third in the rank ordering.
It is my opinion that the most logical reason for so many irregularities, misrepresentations, and inconsistencies occurring at this one time is that after the first large scale trial had succeeded dramatically, Richard, perhaps unconsciously, feared that the final trial might succeed, perhaps even live on national television. Clearly this would have compromised his consistent message that psi probably doesn’t exist and therefore presumably tarnished his growing media profile as a debunker of psi.
I trust this analysis helps show the type of distinct but subtle influences which produce the Wiseman experimenter effect. In the interest of balanced scientific endeavour it seems important that Richard continue his ‘experimenter effect’ work in order to decide whether it is indeed true that any experiment he is involved in will fail. In the interim, it seems unfair to present others’ experiments as failures, sometimes in a blaze of publicity, when the most likely explanation is not that the experiment does not manifest psi, rather that Richard as the experimenter has subconsciously manipulated the experiment to fail.
Acknowledgements: Thanks are due to Pat Harris and several anonymous reviewers whose comments helped meet a tight deadline and to Richard Wiseman himself who clarified some other concerns.
References:
Hyman, R. & Honorton, C. (1986) A joint communiqué: the Psi Ganzfeld Controversy. Journal Of Parapsychology, 50, 356.
Spottiswoode, S. J. P., (1997) Apparent Association Between Effect Size in Free Response Anomalous Cognition Experiments and Local Sidereal Time. Journal of Scientific Exploration, 2, 109-122.
Ullman, M., Krippner S. & Vaughan A. (1973). Dream Telepathy. Turnstone, 174-177. (1983 2nd edition 134-137)
Wiseman, R. & Schlitz, M. (1998) Experimenter effects and the Remote Detection of Staring Journal Of Parapsychology, 61, 197-208.
Mick O’Neill: m.on@virgin.net
Go to “A Reply to O’Neill” by Richard Wiseman below: