CONCEPTUAL AND PROCEDURAL FLAWS IN CERTAIN SO-CALLED ESP EXPERIMENTS

We simply list here, and discuss, the sorts of flaws that occur over and over in almost all supposed “ESP experiments,” or “studies in parapsychology,” which are represented as yielding “above chance” results.

INVESTIGATORS SHOW NO INTEREST IN THE UNDERLYING PHYSICAL PROCESSES, OR THE PHYSICAL AND PHYSIOLOGICAL BASIS OF SUPPOSED RESULTS, IF ANY. In no genuine field of scientific research do we find the indifference to any question of what the actual process supposedly under investigation is, that characterizes the field of parapsychology. Self-proclaimed ESP researchers never seem to worry about what physical stimulus is involved, or what part of the human body or brain is involved, or stimulated, and how. In fact, the fundamental problem of demonstrating directly that such a stimulus exists, independent of whether humans can detect it or not, is completely ignored!

THE EXISTENCE OF SUCH A STIMULUS WOULD NOT CONFORM TO ANYTHING KNOWN AT PRESENT ABOUT NATURE. No known interaction between physical systems in nature could account in any way for the huge variety of claims made in the parapsychology literature concerning ESP. In fact, the claims are inconsistent with the existence of any real, physical underlying basis. To see the problems clearly, try to imagine that there is an “ESP machine,” something like a Walkman or iPod, for sale at every electronics store, which when worn gives you ESP powers. How could a single device make it possible for you to read minds, move objects without touching them, foretell the future, learn the contents of sealed boxes without opening them, send your consciousness to distant locations, even other planets, find lost objects and hidden bodies of crime victims, identify killers, etc., etc., etc.?

MOST ESP EXPERIMENTS ARE IRRELEVANT TO THE WHOLE CONCEPT OF PERCEPTION. INSTEAD THEY ARE “TARGET GUESSING” EXPERIMENTS. A target guessing experiment cannot demonstrate the existence of a suspected channel of communication. Instead, it could only be used to study the signal-to-noise ratio of a channel already known to exist. [For example, over a noisy radio channel, one man could transmit a random list of letters and another man with a receiver could write down the letters as he interprets them through the noise. The noisier the channel the more errors the receiver will make.] In the absence of any concrete physical phenomenon to study, a list comparison, with one list arbitrarily called source and the other list receiver, provides no information. Suppose I have a computer generate two separate lists of random numbers from 1 to 5 (which would not be as easy as you might think!). If I then place the lists side by side and compare them line by line, this comparison has precisely the same structure as any target-guessing ESP experiment. By construction of the experiment, agreement between “source” and “receiver” lists, or “target” and “guess” lists, is due to chance.

THE ACTIONS OF THE INDIVIDUAL “RECEIVER” IN AN ESP EXPERIMENT, IN PROVIDING GUESSES, ARE IN NO WAY CONSISTENT WITH THE USE OF ANY CHANNEL OF INFORMATION. For instance suppose you were asked to guess the order of the cards in a deck of 40 playing cards (court cards removed) face down in a box in a drawer in some unknown room in some unknown building on the UT campus. You would of course have no trouble in making 40 guesses, as fast as you can call off the 40 different cards. But consider the problem someone would face, using any conceivable real channel of communication. [To be concrete, imagine you have a cellphone and you have a friend who can roam the campus at will, talking with you continuously on his or her own cellphone. How does this help you or your friend in any way?!?] How do you find the building? How do you find the right room in the building? How do you find the piece of furniture in the room that holds the cards? How do you find the cards in the drawer? Having located the cards, how would you “see” their faces? You have to “see” through walls and floors, through the metal or wood of the desk, and through the cardboard of the box and through the cards themselves above the particular card you are supposedly “seeing.” Does it even make conceptual sense to imagine a form of perception that can see through rock and concrete and metal and cardboard, yet can see the ink on the face of the card? After all, card ink cannot be felt or smelled or tasted. Printed symbols are accessible to only one sense, sight. The card is just a bunch of molecules, cardboard and ink and plastic coating. You are just a bunch of molecules. What is the connection between one bunch of molecules, the card, and the other, you, not involving vision, that could possibly allow you to “see” the ink on a given card, near the bottom of a stack inside a desk? And how is it that you are able to tell that you are “seeing,” for instance, the 12th card and not the 11th or 13th? Or the 15th or the 9th? The experiments do not appear in any way consistent with the existence or study of any actual process, established or hypothetical. Instead, one is simply asking the tested individual to guess. But asking someone to guess is not the same as asking someone to perceive in some way! If the participant were asked to write down what he actually perceives, he would have to write down nothing. To get any data to work with, the experimenter must always call for mere guessing, not perception.

EXPERIMENTS GIVE NON-REPRODUCIBLE RESULTS; EXPERIMENTS INVOLVE DATA SELECTION, SUPPRESSION OR MANIPULATION; THE FACT THAT HIGH SCORES CAN BE ACHIEVED BY CHANCE IS DELIBERATELY DISGUISED IN PRESENTATION OF RESULTS. For instance, by sheer accident an individual might do very well in target guessing for a brief period, even though his overall score over a long run corresponds very closely to chance results. Instead of admitting that this is what happens, parapsychologists tend to invent a scenario that allows them to ignore the actual chance outcome. For example, that scores will decline after a brief period of high scores is called “the decline effect,” and is a supposed characteristic of ESP! If scores fall well below chance expectations for a brief period, which is just as likely as a brief period of high scores, this is “negative ESP,” and still evidence of ESP! Parapsychologists often claim that ESP is the only human ability which does not improve with effort, concentration, experience, exercise or training! If a person being tested achieves scores that remain close to chance expectations throughout, he is said to have “negative attitudes” that are affecting his performance, or to be “deliberately hiding his talent, for fear of ridicule.” One favorite tactic of experimenters is to find an excuse to fail to mention average or low scores for an individual, giving the impression that the person tested had uniformly high scores. For instance, the experimenter might stop the tests when an individual's score falls to chance levels, from an accidental above-chance start, claiming that he is “tired” and has “lost the ability,” then count and record only the short above-chance run. Another tactic is, at the end of the experiment, to label the periods of average or low score arbitrarily as “test runs,” and count only the high scores as “the actual experiment.” Further, individuals who get high scores are generally never tested again, apparently for fear the scores would then “average out.” We are often told that an individual's scores “scared him,” and that he refuses to be tested any further! When one makes guesses of a target, the results in the short run almost invariably fall above or below the result expected in a very long run, and fluctuate back and forth as the run continues, averaging to the expected long-run result. ESP mythology is clearly simply providing a pseudoscientific (and mathematically illiterate) framework in which to give supernatural significance to these chance fluctuations.

EXPERIMENTAL PROCEDURES AND PROTOCOLS ARE NOT CLEARLY DESCRIBED; OR STATED PROCEDURES ARE IN FACT NOT FOLLOWED; OR LAX CONTROLS THAT PERMIT HIGH SCORES BY CHEATING OR SENSORY CLUES ARE MAINTAINED THROUGHOUT, WITH THE SUBJECT ALLOWED TO CONTROL COMPLETELY THE CIRCUMSTANCES OF HIS “PERFORMANCE.” At times one encounters ESP experiments where the high scores are not due to chance, but rather to cheating, fraud or collusion. The usual situation is that an “informal” study is done, in which the tested subject is allowed many ways to obtain the desired information. His scores are then very high. When the “formal” experiment is begun, the loose controls are tightened up one by one, and the subject's scores decline over the long run to chance expectations. The experiment is then abandoned, and the “informal” and “formal” results are averaged together, but in the published results the impression is created that the above-chance results were obtained entirely during the “tightly controlled” part of the experiment. It also happens that tight controls are never actually put into effect, and the experimenters simply lie about the situation. It also happens that the tight controls do not apply to one person in the experiment, for instance a data recorder or “observer,” who can then cheat to produce high results that mystify both subject and other experimenters. More than one ESP experimenter has been detected simply inventing all his results, without taking the trouble to go through with the pretense of an “experiment.”







ESP EXPERIMENTS HAVE TYPICALLY USED GAMBLING DEVICES. Cards and dice have been used by gamblers for many centuries, and a huge body of lore concerning subtle techniques for beating the odds and cheating is in existence. Similarly, there is an enormous amount of literature on magic tricks and stunts that can be done with cards and dice, together with a huge repertoire of slight-of-hand methods of working miracles with cards and dice. ESP experimenters are not trained in such techniques and are easily fooled by “streetwise” subjects. Consultation with professional gamblers, gambling experts, and magicians in the course of the design of such ESP experiments would be almost a necessity--- but it is never done. Further, why don't the experimenters simply examine the records of large gambling casinos, where huge numbers of runs are made with very tight controls and surveillance? [The answer is that casino records give no evidence whatsoever of beyond-chance results! The income of a given casino depends on the customers' hunches and guesses conforming very, very precisely to chance in the long run, so that the “house odds” pay off a precisely predictable amount toward casino operation.]

ESP EXPERIMENTS APPEAL TO THE TRUTH CRITERIA OF SCIENCE WHILE SIMULTANEOUSLY DENYING THEIR VALIDITY. The overwhelming majority of ESP experiments done by psychologists since the 1920s have had reasonably adequate controls and reasonable statistics, and have thus found NO effects (i.e., pure chance effects). All the hundreds of such studies are ignored and only the occasional crude early, or deeply flawed, or otherwise invalid studies, which gave “spectacular results,” are publicized or cited. The impression is created that all ESP experiments give inexplicable or exciting results, when in fact the exact opposite is true.

PUBLICIZED RESULTS, IF FOLLOWED UP, REDUCE TO ABSURDITY. In a famous study done by pseudoscientist Helmut Schmidt, “evidence” of ESP continued to exist when human subjects were replaced by cats, and later by cockroaches! The obvious conclusion, that the high scores were created by data selection and were independent of the subject, was not drawn by the experimenters. Similarly, a researcher named Walter J. Levy continued to get high ESP scores when he replaced human subjects with egg yolks! Why not cigarette ashes? Or rocks? Levy was later fired when he was caught by suspicious co-workers making up his data, but if he hadn't been caught, he'd probably still be going strong. The experimenters themselves tend to demonstrate unmistakably sooner or later that their procedures and results are invalid. The “research” is then totally abandoned, without further comment, and the original meaningless results are allowed to stand unquestioned.

Two classic instances of such reductions to absurdity are provided by the “retroactive psychokinesis” experiments of Helmut Schmidt and of the Princeton Engineering Anomalies Research Center founded by Robert Jahn. Both Schmidt and Jahn claimed that one could record a series of random numbers, a statistical baseline, onto magnetic tape, and that at any time in the future someone could concentrate on biasing the hitherto-uninspected (according to Schmidt) or already inspected (according to Jahn) results of the experiment away from the baseline, so that when the tape was finally inspected (or inspected again) the recorded results would now actually show a bias! But if this were possible, the results obtained in any experiment of any kind would always be meaningless because there could be no statistical baselines for comparison! Surely, with the entire future before us, someone somewhere could have concentrated and changed all the “control runs” that were used to establish statistical baselines! In fact, Schmidt and Jahn should never have obtained positive results, because I personally have mentally focussed for years on setting their non-chance results back to pure chance! Also, if it is possible to mentally affect information recorded on magnetic tape or disk, it should be equally possible to affect the printed texts of articles, to cause even them to change from positive (above-chance) results to pure chance results! Indeed, parapsychologists frequently do blame the large number of skeptical scientists for the very small “above chance” results seen in such experiments. A powerful psychic is clearly needed to offset even slightly the negative thoughts of thousands of “white-coated, atheistical materialist” scientists.

Such considerations are so childishly ignorant, hopelessly absurd and internally contradictory that one feels like a fool even mentioning them, but parapsychologists, far from being shamed, find such considerations to be “the most exciting aspect of parapsychology.”


The typical bungling ESP “researcher,” J. B. Rhine



Scientologist and ESP “researcher,” Hal Puthoff



The typical bungling ESP “researcher,” Rupert Sheldrake

(Click on each image for more information)

THE TRADITIONAL CATEGORIES OF ESP PHENOMENA ARE NOT DISTINGUISHABLE AND DO NOT MAKE SENSE. ESP PHENOMENA ARE DEFINED NEGATIVELY. The phrase “ESP” can mean almost anything, for instance: telepathy, or reading the mind of another; clairvoyance, or seeing without the use of the eyes; precognition, or awareness of future events before they take place; astral projection, or the ability to send one's consciousness outside the body to distant locations; and psychokinesis, the ability to move or otherwise somehow affect external objects and phenomena by mental concentration. The problem with this labelling can be seen by asking, if someone draws a card, looks at it, and asks me to guess it, and if I get it right, did I use telepathy, or clairvoyance, or precognition, or astral projection? Or did I get it right by chance? If someone draws a card, does not look at it, and asks me to guess it, and if I get it right, did I use clairvoyance, precognition, astral projection, or what? Or did I get it right by chance? If I correctly guess the contents of a sealed metal box, did I use telepathy, clairvoyance, precognition or astral projection? Or did I get it right by chance, or by some sneaky magician's tactic? If I decide to try to use psychokinesis to “bias” a random number generator toward the number 4, and later when the results are examined 4 does come up more often, did I use psychokinesis or did I use precognition? Or was there something wrong with the “randomness” of the generator? And so on! The fact that the ability to be tested is not defined in any meaningful way makes the test inherently meaningless, no matter how carefully it is done. If I appear to possess any information that the experimenter doesn't think I could have obtained, no matter how I actually obtained it, this is ESP! The phenomena are defined negatively. But absence of knowledge about how information was obtained cannot be made the basis of an argument for the existence of new ways of obtaining information! This is the familiar pseudoscientific technique of argument from ignorance. If you don't know, and make no effort to find out, you are stuck with the conclusion that you don't know, and there is no way without further effort to go beyond that lack of information.

Three types of ESP experiments that have been done very often in the past two decades are (1) “ganzfeld” tests, (2) “random number PK” tests, and (3) “remote perception” tests. These are worth discussing in more detail since most current bogus ESP studies fall into one or another of these categories.

In the “ganzfeld” tests [the German phrase translates roughly as “single object” (of concentration)] subjects are isolated from “ordinary” sensory inputs. Eyes are covered, ears are fed white noise through headphones, and the subjects recline on soft beds, in isolated, perfumed rooms. While “senders” elsewhere look at pictures or patterns, supposedly chosen randomly from a set, the subjects describe into a tape recorder whatever pops into their minds. After 30 minutes or so, the subjects are shown all the pictures in the set and asked which pictures match their recorded imagery. If more matches occur than would be naively expected by chance, this is proclaimed “evidence” of ESP. The role of subjective validation in matching actual specific pictures to unrelated, vaguely-reported imagery should be obvious, and is rarely absent at any moment from such studies.

In the “remote perception” tests, experimenters visit a variety of locations, while subjects in the lab attempt to “see” the remote location as if through the eyes of the experimenters. In reality, this is just a disguised version of the usual target-guessing experiment. And again subjective validation runs rampant in deciding which actual location is a match for which vague and ambiguous description. No one who has carried out such studies with claimed above-chance results seems even vaguely aware of double-blind protocols.

In the “random number PK” experiments, random numbers are supposedly generated by a computer, or some mysterious electronic device, while subjects attempt to “bias” the outcome in some way, for instance by concentrating on a specific number. If the target number comes up more often than “chance” would indicate, this is proclaimed evidence of “psychokinesis” (PK). Yet, by construction of the experiment, the only thing being studied is the bias inherent in the random-number generator itself. However, the situation is sometimes more subtle than this, and is worth an extended discussion.

RANDOMNESS IS THE KEY!

Note the basic problem remains in all three of these studies. One is just looking for correspondences between two supposedly “random” lists of items. If more correspondences consistently show up than chance would indicate, and there is reasonable survelliance against cheating, the most likely hypothesis is that the lists are not actually random.

In recent years, psychologists have realized that the issue of randomness is a key issue in almost all reported ESP experiments. In the “ganzfeld” experiments, note that people asked to choose pictures or patterns from a set will not choose randomly, but will be influenced by various psychological biases. Such biases will affect both “senders” and “subjects.” Again, lists of sites to visit will not likely be chosen truly randomly from the set of “distinctive” locations within easy driving distance of the laboratory. Again, suppose a computer is used to prepare a list of random numbers, and the pictures are matched against the numbers in order to “shuffle” them randomly before they are given to the “senders” or “subjects.” How do we know the list of numbers generated by the computer is genuinely random? In fact, the generation of truly random numbers via computer remains an unsolved problem. All widely used methods of generating strings of “random” numbers, such as the FORTRAN function RAN(X), available during the 1960s through the 1980s, yield only pseudo-random numbers. If the set of numbers generated is large enough, subtle patterns emerge. When such patterns happen to favor psychologically appealing pictures, the ganzfeld experiments can give seemingly inexplicable results. Similarly, if the remote viewing sites are in any way stereotypical, the rambling descriptions of the subjects can be matched in some way in most cases. Only very recently has some progress been made in the actual subsequent “randomization” of pseudorandom numbers initially generated by computers or other similar electronic means.

The fact that “random number generators” do not in fact generate random numbers is the basis of the operation of the tiny remote transmitters that everyone uses to lock and unlock his or her car. Starting with the same initial number, both the receiver's generator in the car and the transmitter's generator in the remote control generate precisely the same “random number” from it, a unique code for that particular transmitter-receiver combination!


Prof. S. G. Soal, a mid-century ESP researcher who notoriously fabricated all his “above chance” results.




Robert Jahn

Note how the same problems emerge with the “random number PK” experiments. The numbers generated by the computer are not actually random. It is equally as important to realize that people asked to generate a string of random numbers generally cannot do so. For example, if asked to generate a string of random numbers from 1 to 5, most people will not follow a 4, say, by another 4, because this “is not random.” But in fact a 4 should be followed by another 4 one time out of five. Such ignorance affects experimenters as well as subjects. For example, one of the immortal “Three Stooges” of ESP research, Charles Tart, once had a random number generator that lit one of four colored lights, red, green, blue and yellow. When he found the random number generator often (hopefully one time out of four!) lit a red light again after a red light had just been lit, for instance, he rewired the circuits so each successive light was guaranteed to be a different color from the one before! Where a pattern in the pseudorandom “target list” happens to match a pattern in the lists generated by the experimental subjects (e.g., the chosen “targets” they select “randomly” to attempt to bias the generator toward), above-chance results will occur.

There's a good discussion of these questions by British psychologist Susan Blackmore [see New Scientist, Sept. 22, 1990]. She gives many examples of psychological tests conducted by herself and others that reveal the poor understanding of the concepts of probability that most people have, and how such misunderstandings are relevant to ESP experiments. ESP experimenters have long divided subjects into two groups, “sheep,” who do well on ESP tests, at least initially, and “goats,” who score poorly (i.e., at chance level) from the beginning. Blackmore points out that in experiments done by psychologist Peter Brugger at the University of Zurich, he divided subjects into “sheep” (who had a poor comprehension of probability and randomess) and “goats” (who had at least a fair understanding). He found that the sheep created patterns, “response bias,” in their guessing which were generally avoided by the goats. In the opinion of Brugger and Blackmore, this single fact explains most of the major findings of the pseudoscience of parapsychology. For instance, children are claimed to do better at all ESP tests than adults--- and Brugger found children always show more response bias than adults. Similarly extroverts are claimed to score consistently better than introverts--- and extroverts consistently display the most response bias. Similarly, mentally disturbed, retarded or brain-damaged individuals are alleged to score consistently better--- and show the most response bias. Finally, even the “decline effect” is explained on this basis. It is found that high ESP scores tend to decline even when the experiment is automated to rule out arbitrary data selection--- and Brugger found that response bias always tends to decline from an initially high level as the subjects grow bored or fatigued and get careless about “making it really random.”

In her own studies, Susan Blackmore also found that the well known inability to guess probabilities correctly plays a major role in convincing individuals they have ESP. Probabilities are almost always underestimated. For example, she asked subjects to play a game in which they should get 10 correct responses by chance. Her “goats” correctly estimated that they should get 10 hits, but her “sheep” estimated that “chance” would give an average 7.9 correct responses. Thus, a sheep who got his expected 10 hits would, in Blackmore's words, “think he had done very well and look for an explanation. Obviously no reasonable explanation will be found since only chance was operating. So the obvious leap is to the paranormal,” and to ESP. Of course, many other flaws of thinking that we have discussed in the class play a role as well. A woman who thinks she is using psychokinesis to find parking places in a crowded lot will count the times she only has to circle a few times before a place opens up as successes, and simply ignore the times she had to circle for many minutes.

PSEUDOSCIENTISTS WORK WITH THE NOISE, NOT WITH THE SIGNAL

Note that the crucial question in most ESP experiments, where no other flaws happen to be present, is the question of whether a random list is truly random, i.e., whether noise is really noise. Imagine how far radio would have progressed since the turn of the century, if experimenters had ignored the design and building of powerful transmitters and receivers, and instead tried to show that some primitive, impossibly feeble transmitter was actually transmitting to some impossibly insensitive receiver by making a study of the static and noise output of the receiver, and trying to show that, at a level very slightly above chance, the static or noise in the receiver happened to correspond sometimes, maybe, to the feeble pulses of the transmitter! Radio would still be a laboratory curiosity, with great debate and doubt among scientists as to whether or not it even existed. Sound familiar?

Scientists can make real, rapid progress with real phenomena. Research into ESP began in the 1880s and 1890s, about the same time as research into electromagnetic radiation. In the next century, the world was utterly transformed by scientific and technical applications of electromagnetic waves. Where is ESP? Still in the lab, still looking at the noise and pretending to find signals. Based on what we have learned in this course, it seems unlikely that ESP experiments will ever reveal any actual, existing phenomenon of any sort, currently unknown to science.

Links:
  • Skeptic's Dictionary: ESP
  • Skeptic's Dictionary: Parapsychology
  • Skeptic's Dictionary: extraordinary human functions
  • 2000 Interview with Joe Nickell
  • ``Psychic Illusions," by Susan Blackmore.
  • A former "psychic" on psychic ability.
  • Another one bites the dust!
  • Want to learn how to move pieces of paper, or even huge pieces of lumber, by the mystical powers of the unknown mind? Or something? It's free!
  • Suggested Reading:

  • “Extrasensory Perception,” by J. E. Alcock, in Encyclopedia of the Paranormal, ed. by Stein, G., Promethus, NY, 1996.
  • Secrets of the Psychics, Massimo Polidoro (Prometheus, NY, 2003).
  • Pseudoscience and the Paranormal, Terrence Hines (Prometheus, NY, 2003).
  • Guidelines for Testing Psychics, Richard Wiseman and Robert L. Morris (Prometheus, NY, 1995).
  • Psychic Sleuths, ed. by Joe Nickell (Promethus, NY, 1994).
  • The Illusive Quarry, Ray Hyman (Prometheus, NY, 1989).
  • The Search for Psychic Power: ESP and Parapsychology Revisited, C.E.M. Hansel (Prometheus Books, NY 1989).
  • Parapsychology: Science or Magic?, James Alcock (Pergamon, NY, 1981).
  • The Psychology of the Psychic, David Marks and Richard Kammann (Prometheus Books, NY, 1980).
  • Enhancing Human Performance, Part III, "Parapsychological Techniques," National Research Council, ed. by Daniel Druckman and John A. Swets (National Academy Press, 1988).
  • The Adventures of a Parapsychologist, Susan Blackmore (Prometheus Books, NY 1986).
  • Flim-Flam! James Randi (Prometheus Books, NY 1982).
  • Leaps of Faith, Nicholas Humphrey (Copernicus/Springer-Verlag, NY, 1996, 1999).
  • "Science and the Supernatural", George R. Price, Science 122 (1955): 359-67.
  • "Investigating the Paranormal," by David F. Marks, Nature 320, 13 March 1986, 119 - 124. [One of the best articles of its kind!]
  • "Psychic Sleuth Without A Clue," Joe Nickell, The Skeptical Inquirer, 28, No. 3, May/June 2004, 19-21.
  • Astrology!
    Avoiding Facing Death!
    Blavatsky, Queen of Pseudoscience!
    Cities on the Moon and Mars!
    Creationism and Intelligent Design!
    Distinguishing Science from Pseudoscience!
    ESP!
    Flying Saucers (1947–1985)
    Fortean Phenomena
    The Fourth Dimension?!?
    Ghosts
    Gods from Outer Space!
    Hollow Earth!
    Interstellar Travellers!
    Kirlian Photography and the Aura!
    Martian Canals!
    Medical Quackery!
    Monsters! and Ape Suits!
    Mystery Spots?
    Mystical and Bogus Physics!
    The New Age!
    Pareidolia!
    Postmodernism vs. Science!
    Prophecy and Prophets!
    Psychic Detectives!
    Pyramid and Crystal Myths and Powers!
    Science Fiction and Pseudoscience!
    Space Brothers!
    Spiritualism!
    UFOs 1985-2005!