ODDSEAMINDEPENDENT FIELD ARCHIVE

The Soal–Goldney experiments

Open on the shelf ↗

FRAUD DETECTED

Psychic research / England · run by S. G. Soal of the University of London

The Altered Target Sequences

In forty card-guessing sittings between 1941 and 1943, directed by S. G. Soal, Basil Shackleton scored above chance on the card one ahead of the one being looked at, which was taken to suggest precognition. A 1973 Nature abstract called the series "long quoted as a model of meticulous technique". By the Society for Psychical Research's account, the statistician Betty Markwick later re-examined the score sheets and found evidence that Soal "had surreptitiously introduced figures into the target sequences to match the corresponding guesses".

The explanation on file is experimenter fraud. The Society for Psychical Research, which published the original work, says that after Soal's death it "was discredited by a devastating critique" in its Proceedings by Markwick (1978). Markwick's paper itself was not opened for this entry; the finding is given in the Society's words.

AI illustration · not an evidence photograph — The Soal–Goldney experiments
AI illustration · not an evidence photograph
File
BR-0018
Date
Forty sittings, 1941–43 · exposed in 1978
Place
England · run by S. G. Soal of the University of London

AssessmentD · a specific mundane cause, established by re-examination of the original records

The record on file

Witnesses
S. G. Soal, then a lecturer in mathematics at the University of London, directing, with K. M. Goldney; the subject Basil Shackleton; many independent witnesses called in to the sittings
Evidence
The score sheets, copied at each sitting and later independently checked, as described in Scott and Haskell's 1973 Nature abstract; Betty Markwick's re-examination of them in 1978, described here from the Society for Psychical Research's own account
Record made at the time
The score sheets of the forty sittings, 1941–43. Scott and Haskell's 1973 Nature abstract says copies of them "were made at each sitting and later independently checked". Soal and K. M. Goldney published the results in the SPR's Proceedings, vol. 47, in 1943 (cited in Nature's reference list, not opened).
Agency
Society for Psychical Research

Source on file
Society for Psychical Research, on West and Markwick, Dr Soal: A Psychic Enigma (Proc. SPR vol. 60, part 224) (archived)

The full file

Between 1941 and 1943 S. G. Soal, then a lecturer in mathematics at the University of London, ran forty card-guessing sittings with Basil Shackleton, a London portrait photographer. His co-experimenter was K. M. (Mollie) Goldney of the Society for Psychical Research. For three decades the results were counted among the strongest experimental evidence for extrasensory perception. They are now known mainly as a case of experimenter fraud, established from the experiment's own records thirty-five years after the last sitting.

What was reported

The method was meant to rule out cheating by the subject. Soal prepared the target sequences in advance, as the digits 1 to 5, written in columns on the score sheets. An experimenter sat on one side of a screen with the score sheet. At each trial he or she showed a number through an aperture in the screen to the agent on the other side. The agent had five animal picture cards in a row, shuffled before each run, and looked at the card in the position shown. Shackleton was in the next room, and wrote down his guess when a call told him one was due. Goldney served as one of the two experimenters throughout. Most of the sittings used the same agent, Rita Elliott. As Scott and Haskell summarised it in Nature in 1973, Shackleton scored above chance "on the card one ahead of that being looked at", which suggested precognition, and when the calls were speeded up the success moved to the card two ahead. Soal and Goldney published the series in the Society's Proceedings under the title "Experiments in precognitive telepathy". The volume is 47, but Nature's own reference lists date it to 1943 in one letter and to 1942 in another.

The record, and where it is held

The case is on this shelf because of the score sheets, made at each sitting in front of witnesses. At the end of each sitting a handwritten copy of the completed sheets was posted to C. D. Broad, the Cambridge philosopher. The originals did not survive. In 1956 an American psychologist asked to see a score sheet, and Soal replied that he had lost the records on a train. Every later analysis, including the decisive one, depended on Broad's copies. The Society's archive has been held at Cambridge University Library since 1989, and it includes Goldney's and Donald West's papers on the Soal–Shackleton work. No catalogue entry for the Broad copies themselves could be confirmed for this write-up.

What investigators concluded

Doubt came early. Critics were calling the results too good from about 1949, and in 1960 C. E. M. Hansel published a critique suggesting ways the evidence might have been falsified. That same year it became public that Gretl Albert, the agent at two of the sittings (at one, according to the Society's encyclopedia), had said she saw Soal altering figures on a score sheet. The philosopher C. W. K. Mundle summarised her claim in Nature in 1973. During sitting 16, he wrote, she had seen Soal changing figures "1" into "4" or "5".

In 1971 R. G. Medhurst, in the Society's Journal, pointed out inaccuracies in Soal's account of how the random target series had been constructed. In 1973 and 1974 C. Scott and P. Haskell showed that the data for sitting 16, and for sittings 8 and 17, fitted closely the pattern the alteration Albert described would leave. They argued that the pattern could not plausibly be put down to ESP.

The decisive analysis was Betty Markwick's, published in the Proceedings in 1978. Working from the Broad copies with a computer, she found that runs of target digits, some up to 25 long, recurred from one sitting to another. Within these repeats, single extra digits had been inserted, and those extra digits almost always produced a hit. Writing later, she said the evidence of fraud was "virtually conclusive" for two of the forty sittings and "suggestive only" for six more. Soal had died in 1975.

The leading explanation: Soal altered the targets

The conventional explanation is experimenter fraud: Soal changed or added target digits after the guesses were known. It has been accepted by the Society that first published the work. The Society now says the series was "discredited by a devastating critique" and that Soal had "surreptitiously introduced figures into the target sequences". Its strength is that three methods, all independent of one another, point to the same thing: an eyewitness, the statistical pattern in her sittings, and the repeated sequences in the target lists. Its weakness is scope: the direct evidence touches only some sittings. Goldney was not implicated.

What remains disputed

What is disputed is how far the fraud reached, not whether there was any. Markwick's strongest evidence covers eight of the forty sittings at most, and no one has shown, sitting by sitting, how the rest of the high scores were produced. With the originals gone, the Society's own encyclopedia says it is impossible now to determine whether any of the scores are genuine. In 2018 West and Markwick published Dr Soal: A Psychic Enigma. The two authors also found that the records of Soal's tests with Gloria Stewart had been altered and recopied, and the book left open the question of what Soal himself believed.

Sources

Revisions

Last revised . Every correction to the archive

  • Corrected The full file: An edit earlier on 10 October had reduced Goldney's role to 'assisted'; the Society's encyclopedia says she assisted throughout as one of the two experimenters, so the original wording has been restored.
  • Corrected The full file: The text said one sceptical source dated Soal's loss of the original records to 1945; no cited source gives that date, so it has been removed.
  • Corrected The full file: Hansel's 1960 work was described as a critical review in the Proceedings; the cited source describes a 1960 critique suggesting ways the evidence might have been falsified and does not name the journal, and the text now says that.
  • Corrected The full file: Gretl Albert was said to have told Goldney what she saw; the sources say only that she said she saw Soal alter figures. The text also now notes that Nature says she was the agent at two sittings and the Society's encyclopedia says one.
  • Corrected The full file: Medhurst was said to have failed to find the targets in Soal's logarithm tables; the cited source says that in the Society's Journal in 1971 he pointed out inaccuracies in Soal's account of how the target series was built, and the text now says that.
  • Corrected The full file: Markwick's repeated target sequences were said to recur 'sometimes reversed'; no cited source mentions reversal, so the words have been removed.
  • Corrected The full file: The 2018 West and Markwick study was said to have found a 'similar pattern' in Soal's tests with Gloria Stewart; the Society's encyclopedia describes Stewart records that had been altered and recopied, and the text now says that.
  • Added The full file, a long-form write-up with its sources, added.

Also filed under Psychic research