612 A. Summary Of The Report On The Census Of Hallucinations

The "Census of Hallucinations" was undertaken in 1889 by a Committee of the S.P.R., under the direction of Professor Sidgwick, and consisting of himself and Mrs. Sidgwick, Dr. A. T. Myers, Mr. F. Podmore, Miss A. Johnson, and the present writer, and the full report of the Committee was published in 1894 (Proceedings S.P.R., vol. x. pp. 25-422). The object of the inquiry was -

(1) To ascertain approximately what proportion of persons have experienced sensory hallucinations:

(2) To obtain details as to such experiences with a view to examining into their cause and meaning.

It was especially intended to test whether the number of "veridical" hallucinations {i.e. hallucinations representing some external fact) was, or was not, sufficiently numerous in proportion to the whole to preclude us from regarding as merely accidental the coincidence of fact and phantasm.

The inquiry was conducted by the assistance of "collectors" (410 in number). Each of these was instructed to address the following question to twenty-five adults, to be chosen without reference to the probability of an affirmative answer: -

"Have you ever, when believing yourself to be completely awake, had a vivid impression of seeing, or being touched by a living being or inanimate object, or of hearing a voice; which impression, so far as you could discover, was not due to any external physical cause?"

The answer "No" and the answer "Yes" were to be recorded with equal scrupulousness; but in the case of an affirmative answer, a first-handwritten account of the details was, if possible, to be obtained. The scope of the inquiry was designed to exclude the hallucinations of sleep, delirium, and insanity.

A similar method of inquiry had been employed by Gurney (v. Phantasms of the Living, vol. ii. chap, xiii.), but on a scale perhaps hardly large enough to enable valid inferences as to the proportion of veridical and non-veridical cases to be drawn from the number of persons questioned (5705).

1 These experiments were afterwards described in full in a paper by Mrs. Verrall in Proceedings S.P.R., vol. xi. pp. 174-197.

In the present inquiry, conducted during the three years 1889-92, 17,000 persons were questioned. After deducting such affirmative answers as were found to refer to experiences outside the scope of the inquiry (e.g., dreams, impressions of inarticulate sounds, hallucinations occurring during any illness of a kind that is known to produce delirium) there remained 1684 affirmative answers {Report, p. 39), that is to say, 99 per cent, of the whole.

The question arises, How far does this result represent the true state of things? Is it to any considerable extent vitiated by possibilities of error and deception?

Intentional deception is not likely to have had any appreciable influence. The method of inquiry - the questioning by the collector of his acquaintances and the absence of any motive for deception, practically entitle us to leave it out of account.

A certain number of persons refused to answer the Census question. But the proportion of refusals reported was in general small; and a consideration of the reasons reported for refusing led to the conclusion that refusals probably only reduced to a small extent the proportion of affirmative answers.

An obvious danger was that of selection by the collectors of persons to be questioned according to what they were likely to say. The instructions to collectors (Report, Appendix A, p. 403), and the number of answers (25) to be obtained by each, were designed to obviate this error. In order, moreover, to check the results on this point, collectors were asked to mark in their lists the names of those persons whose answers were previously known. This check, unfortunately, was not fully operative, through the neglect of some collectors to carry it out. A further more efficient check was a comparison between the whole collection and -

(a) Answers obtained from complete unselected groups of persons, e. g. households or chance gatherings.

(b) Answers collected by the Committee and vouched for by them as unselected.

(c) Answers collected by persons trained in medicine or psychology. These three sets of answers are given in a table on p. 49 of the Report, and in this table the percentage of affirmative answers is above the average. It may then be inferred that the percentage of affirmative answers in the whole collection has not been increased by selection.