A Collection of Strange Beliefs, Amusing Deceptions, and Dangerous Delusions

From Abracadabra to Zombies

Edgar Mitchell's ESP experiment

The followiing is from my review of Entangled Minds (EM), a sequel to Dean Radin's 1997 defense of psychic phenomena The Conscious Universe (CU).

the Edgar Mitchell ESP fiasco

Another example of Radin's distorted history of psi research is his claim that astronaut and psi enthusiast Edgar Mitchell conducted a "successful ESP card experiment from the Apollo 14 space capsule" (p. 76). Mitchell is a founder of the Institute of Noetic Sciences, where Radin is employed as a senior scientist. The experiment was unauthorized, so NASA did not set aside a nice block of time for Mitchell to conduct his experiment. News reports based on interviews with Mitchell indicate that he went through a hand-made Zener ESP deck several times while four friends on Earth tried to get psychic messages (either from Mitchell's mind [telepathy] or from the cards [clairvoyance]. Mitchell didn't use a Zener deck, however. He said:

The well-known experiment in the laboratory was to use cards with the five Zener symbols, but the actual cards aren't important. It was easier for me to use random number tables than carry the physical cards. Instead, all I did was to generate four tables of 25 random numbers just using the numbers 1 to 5. Then I randomly assigned a Zener symbol to each number. For each transmission, I would then check the particular table of random numbers and think about the corresponding symbol for 15 seconds. Each transmission took about 6 minutes.*

We know that there was a prearranged time when Mitchell and his friends would do their trials, but that problems prevented things from going as planned. As a result, the recorded guesses on Earth were made before Mitchell went through the trials. No problem. In his published paper on the experiment, Mitchell just changed the goal to a study of precognition! (Mitchell's paper, "An ESP Test from Apollo 14," was published in the Journal of Parapsychology in 1971.) He also stated in an interview that the fact that the timing was off didn't matter:

That didn't seem to make any difference. We took off forty minutes late but I didn't try for an exact time anyway, just in the evening. We now understand why that should work, because the sequence is important but having the precise time is not.*

This is fantastic news for psichologists. It means that in an ESP study one can use data collected anywhere and any time to use as proof of psi. How could you not like such a cooperative thing to study?

Original news reports made it sound like there were 200 trials, leaving one to conclude that Mitchell went through his tables eight times (there are 25 cards in a standard Zener deck). Since there are five symbols (or numbers) in each table, chance guessing would be one of five, or 20%. Mitchell reported that two subjects performed better than chance and two performed worse than chance. What a shock! This is what you'd expect from chance. News reports indicated that the group got 51 correct, which would be a hit rate of 25.5%, not bad for an individual guessing 200 times, but not statistically significant for a group of four. No problem. Mitchell provided the New York Times with the number 3,000 to 1 as being the odds against getting the results he got. Maybe this is where Radin learned to use statistical odds as a substitute for evidence.

A closer look reveals this story is typical of the history of psi: much ado about nothing except manipulation and deception by the defenders of psi. Here is an abstract of Mitchell's experiment (S=receivers or subjects, E=the experimenter and sender):

Preflight arrangements were made with 4 Ss noted for ESP ability for a short "unofficial" experiment, to be carried out while the author, who served as the E, was on the Apollo 14 mission. On 6 different days the Ss were to guess the symbol order in 1 target run of 25 symbols being concentrated on by the E. Actually, E was able to carry out only 4 target runs, and the timing arrangements could not be met under flight conditions. The Ss, a, b, c, and d, made 6, 6, 1, and 2 runs of guesses, respectively, (a total of 15) to be checked against the 4 target runs brought back by E. The major problem was to determine, before checking, how the Ss probably oriented their efforts to guess the series of 4 target columns. Three independent analyses were conducted. In the 1st, the view was that as in a precognition test the sequence of 4 target runs on the record sheet would be the natural aim of the S. There were 8 runs (4 each by a and b) that could be checked by this plan; they gave a deviation of +11. The odds are 20:1 against chance. The 2nd analysis, initiated by E, aimed at determining the relationship between the guesses and the time proximity of the targets. There were 12 runs in this group (2 runs that overlapped were omitted) and these gave a very negative deviation having odds of 3,000:1. A 3rd analysis, also with below-chance results having odds of 25:1, followed an independent plan that cut across the other 2 analyses. The 1st 2 analyses, committed in advance, are independent of each other. The deviation in the 1st was positive, that in the 2nd, negative, indicating a psi-differential response by the Ss. This effect is further revealed in the reversed direction of distribution curves of hits within the runs for the 2 analyses.*

Cutting through the gobbledygook, we find that the 3000 to 1 odds was derived by an afterthought analysis of the data that found the subjects did not guess correctly. Parapsychologists call this psi-missing. If you guess better than chance, that supports the psi hypothesis. If you don't guess better than chance, that also supports the psi hypothesis. Psi works both way, positive and negative. Rather than canceling each other out, they reinforce each other in what is called "a psi-differential response." What other scientific field would tolerate such nonsense?

Why do I spend so much time chewing on this old rag? To emphasize to the reader that the claim by Radin that an experiment was "successful" or has been "replicated" in several labs should be taken with a quantum of entangled sodium chloride. (I know; there is no such animal. Allow me some slack.) This tactic of proclaiming success and replication, even though the claim never withstands scrutiny, works. Mitchell used the same tactic in an interview where he claimed that his experiment "did show that what had worked in the laboratory also worked in space with the same very positive results."* I've had several people e-mail me about Radin's claim regarding presentiment experiments, which he claims have been replicated in several labs, a sentiment apparently parroted by Alex Tsakiris on his podcast, Skeptiko.

further reading

Why do people believe in the palpably untrue?

What if Dean Radin is Right?

A Short History of Psi Research

my review of The Conscious Universe

Last updated December 27, 2010


Pseudoscience and the Paranormal
Why psychics don't win the lottery.

The Conscious Universe Exposed
What does the science really show?

The Skeptic's Shop
No shirts, no mugs, no tinfoil hats.


Print versions available in Dutch, Russian, Japanese, and Korean.

This page was designed by Cristian Popa.