Phantasms of the Living - Volume I.. Frank Podmore
Owing to the rapid convergency of the series which we have to sum, it will be found sufficient to evaluate two or three terms.
3 Tables of Logarithms, and of the values of log (x + 1).
4 Here, for instance, is Professor Barrett’s record of a casual trial made on August 4th, 1882—only he and Mrs. Myers knowing the card selected. Eight cards were successively drawn from a pack; of these, three were guessed completely right—two of them at the first attempt and the third at the second attempt; in this last case the first guess was the nine of clubs, and the second the nine of spades, that being the card chosen. In addition to these the suit was given rightly three out of the remaining five times, the pips or court card twice out of the five. Immediately after this experiment the two younger sisters of the guesser were called in and allowed to know the card chosen by Mrs. Myers and Professor Barrett. The results, compared with the preceding, were as follows:—
In the absence of the sisters. Eight experiments. Two complete successes on the first attempt and one on the second.
With the assistance of the sisters as agents. Seven experiments. Two complete successes on the first attempt and one on the second.
And to make the coincidence more curious, the partial successes were identical in number in the two series.
1 Even the successes obtained when Mr. Creery was helping us were less remarkable than those which, according to his records, had been obtained in the earlier trials, when the whole affair was regarded as an evening’s amusement, and the children were without any sort of gêne or anxiety. Still, with his assistance, we have had such successes as the following. Out of 31 trials with cards (the chances against success by accident being in each case 51 to 1) 17 rightly guessed at the first attempt, 9 at the second, 4 at the third; 8 consecutive successes in naming cards drawn at random from a full pack; and the following series, where the names on the left hand, written down at random by one of ourselves, are what the agents silently concentrated their minds on, and the names on the right hand are what the percipient said, usually in two or three seconds after the experiment began:—
William Stubbs.—William Stubbs.
Eliza Holmes.—Eliza H—
Isaac Harding.—Isaac Harding.
Sophia Shaw.—Sophia Shaw.
Hester Willis.—Cassandra, then Hester Wilson.
John Jones.—John Jones.
Timothy Taylor.—Tom, then Timothy Taylor.
Esther Ogle.—Esther Ogle.
Arthur Higgins.—Arthur Higgins.
Alfred Henderson.—Alfred Henderson.
Amy Frogmore.—Amv Ereemore. Amy Frogmore.
Albert Snelgrove.—Albert Singrore. Albert Grover.
1 To illustrate these various points, I will give one series where the success was below the average.
Cambridge, August 3rd, 1882.
Miss Mary Creery was outside the closed and locked door,—a thick arid well-fitting one—and a yard or two from it, under the close observation of a member of the Committee, who observed her attentively. A card was chosen by one of the Committee cutting a pack; the fact that the card had been selected was indicated to the guesser by a single tap on the door. The selected card was placed in view of all the agents, who regarded it intently. After the guesser had named a card loudly enough to be heard through the door, the word “No” or “Right,” sas the case might be, was said by one of the Committee; otherwise complete silence preserved.
The cards chosen are printed on the left, the guesses on the right. Two guesses only were allowed.
1. Three of hearts.—Ten of spades (No). King of clubs (No).
2. Seven of clubs.—Nine of diamonds (No). Seven of hearts (No).
3. Ten of diamonds.—Queen of spades (No). Ten of diamonds (Right).
4. Eight of spades.—King of clubs (No). Ten of spades (No).
5. Nine of hearts.—Nine of clubs (No). Ace of hearts (No).
6. Three of diamonds.—Six of diamonds (No). Ten of diamonds (No).
7. Knave of spades.—King of spades (No). Queen of clubs (No).
8. Six of spades.—Six of spades (Right).
9. Queen of elubs.—Queen of diamonds (No). Ten of clubs (No).
10. Two of eiubs.—Ten of diamonds (No). Ace of diamonds (No).
Here there were only two complete successes; and in tabulating results and computing averages we should of course count all the trials except the third and eighth as complete failures. But the result numbered 7 was on the verge of complete success; in 5 and 9 the correct description was given piecemeal; and in 2 the number of pips was correctly given.
1 In an account of some experiments with words, which we have received from a correspondent, it is stated that success was decidedly more marked in cases where there was a broad vowel sound.
1 It should be remarked, however, that the introduction of any principle of selection, after one experiment, is always objectionable. For some more or less plausible reason could probably always be found for setting aside the less favourable results.
1 The rules to observe are these: (1) The number of trials contemplated (1,000, 2,000, or whatever it may be) should be specified beforehand. (2) Not more than 50 trials should be made on any one occasion. (3) The agent should draw the card at random, and cut the pack between each draw. (4) The success or failure of each guess should be silently recorded, and the percipient should be kept in ignorance of the results until the whole series is completed. The results should be sent to me at 14, Dean’s Yard, S.W.
2 For these calculations we have again to thank Mr. F. Y. Edgeworth. For an explanation of the methods employed, see his article in Vol. iii. of the Proceedings of the S.P.R., already referred to, and also his paper on “Methods of Statistics” (sub. fin.) in the Journal of the Statistical Society for 1885.
1 Report by Professors J. M. Peirce and E. C. Pickering, in the Proceedings of the American Society for Psychical Research, Vol. i., p. 19. This Society has also carried out 12,130 trials with the 10 digits—which similarly gave a result only slightly in excess af theoretic probability. But here the digits to be thought of by the agent were not taken throughout in a purely accidental order, but in regularly recurring decads, in each of which each digit occurred once; and consequently the later guesses (both within the same decad and in successive decade) might easily be biassed by the earlier ones. This system may lead to interesting statistics in other ways; but to give thought-transference fair play in experiments with a limited number of objects, it seems essential that the order of selection shall be entirely haphazard, and that the guesser’s mind shall be quite unembarrassed by the notion of a scheme.