Duke Experimenters

This article was collectively written by J. B. Rhine (Professor of Psychology), J. G. Pratt (Instructor in Psychology), Charles E. Stuart (Prince Memorial Fellow), Burke M. Smith (Graduate Research Assistant) and Joseph A. Greenwood (Assistant Professor of Mathematics) of the Parapsychology Laboratory Department of Psychology at Duke University. It appeared in "Extra-Sensory Perception After Sixty Years: A Critical Appraisal of the Research in Extra-Sensory Perception" (1940, Henry Holt and Company, New York).

Test Conditions that Affect Performance

The Cognitive Situation | Modes of Response | Social Environment | Conditions Favourably Affecting Motivation | Conditions Adversely Affecting Motivation | The Effect of Drugs | Summary

- J. B. Rhine, J. G. Pratt, C. E. Stuart, B. M. Smith and J. A. Greenwood -

          THE NEXT step is to find, if possible, any special conditions or test procedures which limit ESP performance. To do so requires a survey of the general conditions, as far as they have been discovered, under which ESP occurs and of all special conditions found to affect score averages either favorably or unfavorably.

The ESP results available represent many researches, done largely without co-ordinated attack, upon the problems here considered. There is naturally some difficulty in finding a plan of presentation that co-ordinates these various conditions. There are, however, certain more cognitive aspects to the various test situations; there are different modes of response in the tests, varieties of social situation, and miscellaneous ways of affecting motivation, ways that induce positive and those that evoke negative deviations.

The Cognitive Situation [top]

Two sets of special conditions to be considered in this chapter belong under this general heading; the effect of the so-called "trance" state is one, and the other is the effect of number of possibilities for a choice, or the p-value* on each trial.

* = probability of success in each trial.

Trance(1). The discussion under this heading includes both self-induced and hypnotically-induced trance. The former is of particular interest because of the general association in the public mind between self-induced trance and ESP, owing to the fact that many professional mediums go into trance (or appear to do so, at least) before attempting to divulge information supposedly arrived at extra-sensorially. Also, as indicated in the preceding chapter, the hypnotic state and ESP performance have been associated historically since the days of Mesmer and his followers.

(1) A state of withdrawal of attention to (or responsiveness to) environment, with (usually) the exception that special rapport is maintained with a given individual (hypnotist, "sitter").

Hypnotic Trance. The amount of work done with a subject in the state of so-called hypnotic trance, conducted with the assistance of Lundholm and reported by Rhine(2), indicated that this state as induced by them was not successful for the subjects tested in bringing out any significant evidence of GESP ability. The results obtained were slightly positive in trend, but they are reported as being no more so than those obtained with the same and with other subjects in tests in the normal state. In a number of the earlier investigations of ESP, particularly in France and England(3), the hypnotic state was used and success was assumed to be associated with that state. There were, however, no adequate controls made as to what might have been obtained from the same subjects in the normal condition. This consideration, taken together with the fact that equally remarkable results have since been obtained on other subjects by other experimenters with the subjects in the normal state, suggests that hypnotic trance is not necessarily advantageous to experimental success(4). The word "trance" is not a very precise term and no acceptable criteria are known as yet for the determination of measure of the state whether self-induced or induced by suggestion. The above discussion is necessarily subject to all the reservations that attach to the term used.

(2) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.
(3) See; Myers, F. W .H. Human Personality and Its Survival of Bodily Death. London: Longmans, Green, and Co., 1903, 2 vols.
(4) In later work at the Duke Laboratory, Zirkle and Ownbey hypnotized three subjects, giving them, in all, 105 runs in the normal state before hypnotic state was induced, and 105 runs during the trance. The control average was 5.11 and the average during the trance was 5.55. The average during the trance is significant in critical ratio, but the difference between the two blocks is not significant. The results are inconclusive, in any case, in view of the fact that the subjects were not naive. They knew that the Purpose of the experiment was to obtain higher scores during the trance than before. The 5.55 average obtained during the trance was not above their normal average. The normal control of 5.11 was, in fact, below it.

Self-Induced Trance. Performance of trance mediums, as mentioned, have given the general impression that the ESP capacities involved in trance are very different from those possessed by the subject in the normal waking state. While adhering strictly to the ESP procedures, Rhine(5) studied the performance of a professional medium whom Pratt(6) found capable of demonstrating, under very well controlled conditions, unaccountable knowledge approximating (in type) that shown by the medium in her routine professional practice. In the Rhine investigation, the medium was subjected to routine telepathy and clairvoyance tests both in the waking state and in her customary self-induced trance with a purported spirit personality assuming control of her body. The results reported were approximately the same for the trance as for the waking state, being only slightly lower in the former, both in telepathy and clairvoyance tests. A wide discrepancy between score averages in telepathy and clairvoyance was shown in both the waking and the trance state, and the same trend in scoring was shown in both methods and for both personality states (see Graph 1, below).

(5) Rhine, J. B. "Telepathy and Clairvoyance in the Normal and Trance States of a 'Medium'," Character and Personality, III (1934), 91-111.
(6) Pratt, J. G. "Toward a Method of Evaluating Mediumistic Material," Bulletin 23, Boston SPR, 1936.

Comparison of Scoring Rates in Trance and Normal States

Chronological comparison of scoring rates in telepathy (T) and in clairvoyance (C) by the trance personality U and the normal personality of Mrs. G. The points represent the score average of 3-day periods from A to F.

(From Character and Personality, 1934, 111, p. 107)

This is the only study of this nature in which extra-chance results were obtained. A series of tests which were given the same subject a year later by Rhine(7), using pre-shuffle card-calling tests (which had been successful with other subjects), yielded only chance results. Still later a series by Goldney and Soal with this medium, using GESP and PC procedures, is reported to have given chance. Yet the medium was still able to go into trance. It can at least be stated that it is suggested that the self-induced trance, like the hypnotic, is neither essential nor apparently advantageous in the ESP test performance. Some experimenters find that their subjects approach a light trance in their effort at concentration. Whether this is essential or is only important because the subject believes it to be so has not been determined and remains an important question.

(7) Rhine, J. B. "Experiments Bearing on the Precognition Hypothesis," Journal of Parapsychology, II (1938), 38-54.

Number of Choices, or p-Value. The second topic under cognitive relations has to do with the possible effect upon performance of the various numbers of possibilities which the subject has to keep in mind in the tests. These numbers represent the p-value. If p = 1/2, there are, of course, only two symbols or objects to which response is to be made. If p = 1/25, the task of judgment is presumably burdened with the necessity of keeping cognitively available as many as possible of the 25 objects relevant to the judgment. The question is: Have these differences any effect upon scoring?

It might logically be supposed that the number of available possibilities (hence the p-value) in the ESP test should be the number most conveniently kept in the foreground of awareness. In fact, the primary reason why a range of five symbols (or p = 115) was made the basis of the experiments reported by Rhine(8) was that most people can readily retain five symbols and that fewer than five symbols would tend to make fixed habits of patterning more likely than would the larger number. Although this was decided on theoretical grounds, its merits may now be tested quantitatively. With the introduction of a formula for the ESP quotient (per cent of trials most probably attributable to ESP alone, chance hits excluded) by Foster, there is made available a means of comparison of rates of scoring on trials with different p-values. Such comparisons are shown in the ESP quotients of Table 18 (below). These are obtained from the totals for the different p-values given in Table 5, the items being reproduced with appropriate quotients.

(8) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.

I. Results Obtained with all p-Values having over 10,000 Trials

A. Total Results B. Results Reported by Academic Men
Item p-Value Trials Dev. C.R. ESP Q Trials Dev. C.R. ESP Q
1 1/2 172,140 +925 4.45 1.1 165,818 +566 2.78 0.7
2 1/4 126,338 +1,789 11.61 1.9 110,381 +942 6.54 1.1
3 1/5 2,758,354 +52,732 79.37 2.4 2,493,094 +41,740 66.08 2.1
4 1/6 119,929 +1,444 11.17 1.4 52,992 +205 2.38 0.5
5 1/10 34,078 +577 10.41 1.7 30,965 +285 5.39 1.0
6 1/13 52,873 +525 8.57 1.1 42,578 +159 2.89 0.4
7 1/40 11,240 +52 1.22 0.5 11,149 +49 2.96 0.5
8 1/52 72,890 +884 23.94 1.2 46,437 +327 11.04 0.7

TABLE 18 - Continued
II. Results Obtained with all p-Values having less than 10,000 Trials

A. Total Results B. Results Reported by Academic Men
Item p-Value Trials Dev. C.R. ESP Q Trials Dev. C.R. ESP Q
1 1/3 324 +49 5.77 22.6 324 +49 5.77 22.6
2 1/9 41 +3 1.48 8.2        
3 1/16 143 +44 15.22 32.8 143 +44 15.22 32.8
4 1/20 41 +9 6.47 23.1        
5 1/25 35 +9 7.75 26.8 35 +9 7.75 26.8
6 1/26 40 +12 9.91 31.2 40 +12 9.91 31.2
7 1/29 134 +21 9.95 16.2 134 +21 9.95 16.2
8 1/32 59 +2 1.50 3.5        
9 1/36 612 +8 1.97 1.3 612 +8 1.97 1.3
10 1/48 187 +56.1 28.61 30.6 187 +56.1 28.61 30.6
11 1/81 2,402 +199 36.71 8.4 2,402 +199 36.71 8.4
12 1/90 3,259 +364 60.86 11.2 117 +13 11.50 11.2
13 1/100 56 0 0.00 0.0 56 0 0.00 0.0

Since this comparison is of very general character and is made on groups of results obtained by different experimenters with different subjects, and with a diversity of other conditions, it is recognized to be hazardous to base conclusions on it. In the interest of a more uniform comparison, however, those p-values which have over 10,000 trials are selected to make Section I-A of the table. This will insure that a greater number of series, subjects, and investigators be included in each item, and should make any comparison more representative and correspondingly safer. Also as a further step toward getting a more uniform basis for comparison, the results obtained by investigators of the psychological and academic groups segregated in Table 4 have been treated separately, again subject to the minimum of 10,000 trials, and the ESP quotient is given for each item of this selection (in Section I-B). (In Section II of the table are given the p-values with smaller series, ranging from 41 trials to 3,259 and in p-value from 1/3 to 1/100. These are mostly represented by a single report each, all by earlier work, and frequently by a single subject in each series. Any conclusion as to relative merit of p-values from these data would be hazardous; they are presented to complete the record.) It will be seen that there is no important difference between the ranking of p-values and quotients of this more selected group of experiments and the general total, and that fact is itself of interest. From this section of the table, it will be seen that in both sub-sections, A and B, the highest ESP quotient is found for p = 1/5 and that in both directions from this value the quotient falls away.

There is an interesting exception to the trend, however, in the case of p = 1/10. Here we have to do with perhaps the most familiar of all symbols, the cardinal numbers, and here presumably we have the factor of familiarity reducing the cognitive burden of the subject in keeping all the elements of possible choice before his attention. It might, indeed, be wished that a more deliberate experimental study of this point were available for the comparison under discussion, but the analysis just made nevertheless leaves it clearly suggested that 1/5 is the most advantageous p-value. It is more speculatively suggested that ease of retention combined with maximum of variety together make 1/5 the best range of choice.

Modes of Response [top]

Matching and Calling. With reference to the overt expression of ESP judgments, it is again profitable to refer to nonexperimental origins. There have been developed in the various practices of the occult (in which ESP might be involved) two general types of response: One of these consisted of oral communication of information supposed to be revealed through hallucinations, more commonly visual (exemplified in the use of the crystal ball) but sometimes auditory (represented in the experiences reported of Socrates and Joan of Arc). The other type consisted of motor, usually manual, movements such as those involved in the Ouija board, the exploratory pendulum, the dowsing rod, and the planchette and automatic writing.

The experimental invasion into this field of questions has followed these two general lines. There has been, on the one hand, considerable work done with test procedures which depend upon manual response by the subject in registering his judgment. The various matching methods and ESP shuffle, as well as Tyrrell's machines, represent this type. In the calling methods, on the other hand, the subject registers his mental impression (hallucination, image) orally or in writing. It has been established that ESP performance is obtainable with either of these two lines of response: either with matching methods such as those used in the Pratt and Woodruff series, as well as a number of others; or by the originally more common method of mentally choosing the card and either writing it down or calling it orally to a recorder. The greater number of series of tests under the best conditions were done by the latter method, more commonly with the subject writing down his own responses.

Comparison of the test performance for the matching and calling procedures has reached the point at which it can be said to be indicated that one is as successful as the other. Tyrrell(9) obtained with his pointer apparatus an average of 6.26 per 25; with the electrical machine, 5.90; and in 1,000 card trials, 5.70. Card trials were of the calling procedure, whereas the responses made on the two types of apparatus were manual. Tyrrell points out that his subject does not usually do as well during the initial stage of a new procedure and there may be some allowance made on that account for the slightly lower average made on the card trials. Tyrrell emphasizes the greater cognitive activity involved in calling the cards which he thinks may be unfavorable for some subjects. Woodruff and George made a study of the two types of response in a series offering a closer comparison. They compared the same subject on BT* tests and two kinds of matching, OM** and BM***. They found that the choosing and the (oral) calling response of the BT method showed very little difference from the averages obtained from OM. With BM, however, which presumably involves less of conscious experience as a basis for judgment, the average was lower than it was for either BT or OM. The Woodruff and George experiment consisted of two sections: that done with the use of screens, and that which was unscreened. The same relationship held in both of these divisions. The bar graph shown below, reproduced from the Woodruff and George report, gives a convenient summary of the results of interest here.

* Before Touching. The technique in which the top card of the face-down deck is called and, after being called, is laid aside for checking at the end of the run. Each card in the deck is treated in the same way.
** Open Matching. The technique in which the subject matches a deck of ESP cards to five key cards which are face up before him.
*** Blind Matching. The technique in which the subject matches a deck of ESP cards to five key cards which are laid face down before him in an unknown order. Unless otherwise stated, the order is also unknown to the experimenter.
(9) 322 Tyrrell, G. N. M. "Further Research in Extra-Sensory Perception," Proceedings of the Society for Psychical Research, XLIV (1936), 99-168.

Comparison of Test Procedures

Woodruff and George's comparative results from OM, BT, and BM tests with screened cards (black bar) and unscreened cards (white bar).

(From Journal of Parapsychology, 1937, 1, p. 23.)

Gibson's results fairly closely approximate those of Woodruff and George, using GESP results as a substitute for BT, since Gibson did not use the BT method. Table 19 (below), reproduced from Gibson's paper, offers a comparison of the two sets of results. Price(10) likewise found matching methods and calling methods to be approximately similar in score averages.

(10) 220 Price, Margaret M. "A Comparison of Blind and Seeing Subjects in ESP Tests," Journal of Parapsychology, II (1938), 273-286.

Comparison of Independent Results

Woodruff and George Present Experiment
Technique Trials Av. per 25 Technique Trials Av. per 25
BT 6,825 6,92 GESP 4,535 5.97
OM 10,975 6.97 OM 6,875 6.67
BM 8,700 5.43 BM 20,690 5.51

So far, then, as to whether there is any great difference between calling (BT or GESP) and matching methods (OM), the evidence is thus far negative. The fact that the more remarkable series, such as that of Riess, was done by the calling procedure, while matching methods used by Pratt and Woodruff yielded only a 5.2 average, superficially suggests a difference that closer and more comparative study does not bear out.

DT and BM. Before leaving the question of the relative merits of ESP methods, the question of the lower level of scoring in DT than in BT (or GESP), and in BM than in OM, deserves consideration. With but one exception(11), and that not involving a great number of trials, BM averages below OM, DT below BT. This may be rated as indicated on the basis of the work shown above in Table 19 (above), so far as matching methods are concerned; and in DT comparisons the work of Rhine(12), MacFarland, and Gibson confirms the lower trend of that method compared to BT or GESP. In this connection should be mentioned the ESP shuffle method which, though less familiar, has been tried in comparative studies sufficiently to warrant its consideration here. In a study of children by Pegram(13), in which OM, BM, DT, GESP, and the ESP shuffle techniques were compared, BM, DT, and the ESP shuffle gave averages respectively of 5.29, 5.21, and 5.21, whereas OM yielded 5.48. There were about 400 runs made with each method. The GESP averaged 4.99, a fact which the experimenter attributed to the circumstance that it came first in the experimental period and suffered from the initial need of adjustment of the subjects to the experimenter. This comparative study would appear sufficient to indicate on experimental grounds a rating of the ESP shuffle along with DT and BM, especially in view of the fact that the scoring that has resulted in the rather extensive work reported by Rhine, Smith, and Woodruff, and in the ESP shuffle work of Pratt and Price more closely resembles the DT and BM averages in general than the BT and OM.

(11) Price, Maragret M. "A Comparison of Blind and Seeing Subjects in ESP Tests," Journal of Parapsychology, II (1938), 273-286.
(12) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.
(13) Pegram, Margaret H. "A Comparison of ESP Test Procedures." [Unpublished MS.]

The following hypothesis is suggested to explain the lower score rating of the DT, BM, and ESP shuffle methods. The three methods under consideration probably do not permit as sharp a focus of attention as do the other methods with which they have been compared. By this is not meant a poor focus of attention to physical detail, but instead a definite mental limitation or lack of orientation. In the case of DT, the U-curves reported by Rhine(14), Pegram, Gibson, and Sharp and Clark - though not found by MacFarland and George or by Martin and Stribic(15) - suggest such a difficulty in focus of attention in the middle of the deck. This is indicated to be a mental difficulty rather than a physical obstruction by the fact that the cards in the lower part of the deck, which should offer (to the subject located above the level of the deck) even more physical difficulty than those in the center, average about as many hits as do the cards located toward the top of the deck and more than the cards in the middle. The cards towards both ends of the runs naturally stand out cognitively above those of the interior and are more easily located and identified with the same acuity of attention.

(14) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.
(15) Martin, D. R., and Stribic, F. P. "Studies in Extra-Sensory Perception: I. An Analysis of 25,00 Trials," Journal of Parapsychology, II (1938), 23-30.

In BM, as contrasted with OM, there is again less upon which to fix attention since the five key cards are no longer visibly presenting themselves to render the choice vivid for the subject. Their standing out in OM presumably facilitates the focusing of attention. Likewise in the ESP shuffle, the effort at concentration of attention upon a given card objective is greatly frustrated, more so perhaps than in any of the tests. It is only a broad, dull, and very vague focus that can be achieved at best. On the other hand, in GESP, BT, and PT, even though there is not sensorially present any key card for the assistance of attention, there is the temporally and physically isolated event as a target upon which the subject can concentrate in the act of making his response.

Social Environment [top]

Group Tests with Adult Subjects. In searching for relations between the text situation and ESP performance, the presence of other persons in the subject's field of experience appears to be important. In fact, somewhat more positive statements can be made in this connection than has been true of the questions raised earlier. One of the most important of these problems is that of the relation of isolation of the subject during the test. The evidence shows that it is established that the testing of the adult individual in isolation is, with present test procedures, superior to group performance. It might even be inferred that successful extra-chance results in group tests of adults are not to be expected with present techniques although new conditions and methods may succeed where the old ones have thus far failed.

The class tests reported by Rhine(16) from the Duke Laboratory yielded only chance results. Sharp and Clark, and later Rothera(17), likewise obtained chance results when testing groups. Sharp and Clark made comparisons of individuals who had participated in group tests. However, the individual tests gave significant deviations, while the group test did not. These results, taken together with the fact that the case for ESP has rested entirely upon individual tests with the subject working alone and with not more than two experimenters present, appear to warrant the rating given above of the experimental establishment of this difference with adult subjects between individual and group tests.

(16) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.
(17) Rhine, J. B. "Rothera Experiments," Journal of Parapsychology, II (1938), 325-326.

Group Tests of Children. A distinction is indicated, however, between the effect of the group upon children and adults. L. E. Rhine and Price and Pegram have described successful ESP performance with child subjects taking the tests in the presence of their playmates. In such instances the whole procedure is much like a game. Children already accustomed to playing with one another appear to be relatively uninhibited by each other's presence. Also, Bond gave GESP tests to 22 fourth grade children in the schoolroom, with significant results, and Richet(18) refers to an anonymous schoolmaster who gave similar tests to his pupils with successful results.

(18) Richet, C. Thirty Years of Psychical Research. New York: The Macmillan Co., 1923, p. 91.

Visitors. It is indicated that the presence of a visitor unoccupied in the conduct of the test will adversely affect the scoring of the subject, presumably through distraction of attention and motivation. This observation was first made by Lodge(19), it was confirmed by Bender(20), and experimentally demonstrated by Rhine(21). Brugmans and Rhine both observed certain improvement in scoring with the isolation of the subject even from the experimenter's presence. It should be mentioned that in each case the subject was an adult, thoroughly experienced with and deeply interested in the research program. The case might well have been different with a child subject or with one whose interest was less intense. In such case, isolation might lead to a decline of interest.

(19) Lodge, Oliver J. "An Account of Some Experiments in Thought-Transference," Proceedings of the Society for Psychical Research II (1884), 189-216.
(20) Bender, Hans. Zum Problem der Ausserinnlichen Wahrnehmung. Leipzig: Johann Barth, 1936.
(21) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.

Subject's Relation to Experimenter. That the subject-experimenter relationship is important in ESP tests has been suggested in the work of a number of observers in the history of ESP. Sharp and Clark pointed out that different experimenters obtained different results with the same subjects. One experimenter consistently obtained only chance scores, while another fairly consistently got extra-chance scores with the same subjects. The most successful experimenter, too, obtained considerably lower scores than usual during a period of intense strain and worry occasioned by the severe illness of a member of his family. Pratt and Price brought the question of the subject-experimenter relationship to an experimental focus, and their results indicated that this relationship may well determine success or failure for the subject. Following upon a long series of chance results by one experimenter and a long series of extra-chance results by the other, the two experimenters collaborated first under a procedure laid down by the experimenter who had previously been unsuccessful. The results during that regime were still unsuccessful so far as statistical significance of results was concerned. The successful experimenter then took complete charge of the relations with the subjects although the other experimenter was continually present and was equally responsible for all the safeguards. The results during this period rose to the level customary for the more successful experimenter and to a significant critical ratio as well.

In the Pratt and Price study, the experimenters were present in the room in which the tests were given. In MacFarland's investigation, the subject was isolated and a signalling system was used. Two experimenters were at the other end of the communication system, which was several rooms away on a different floor. In GESP and in DT tests, both experimenters had packs of cards and both packs were set up as targets for the subject. One experimenter consistently obtained higher scores for his pack than did the other, even in the DT series. In fact, the second experimenter obtained results but little above expectation. The experimenter who succeeded was nominally in charge of the experiment and presumably had more prestige. He also was more experienced in the research. That these considerations, however, are not sufficient as explanation for all cases is shown by the fact that the opposite was true in the Pratt and Price study, the more successful experimenter being younger, less experienced, and less dominant in the experimental situation.

It is obvious that much remains to be discovered as to the proper subject-experimenter relationship and a generalization will be difficult. But it has been indicated that the experimenter's relation to the subject is a determining condition in ESP performance. This much can be said without attempting to define more closely just what in a given situation the accompanying relationship actually was.

Conditions Favourably Affecting Motivation [top]

Some of the conditions already considered may well have affected motivation, especially those under the heading of social environment. Those now taken up are perhaps better called miscellaneous conditions presumably affecting performance through the subject's motivation.

Novelty of Material. Among the conditions indicated as affecting ESP performance, there is the previously mentioned work of Pratt and Woodruff in which it was found that novelty of the materials used, consisting of the introduction of different sizes of the ESP symbols, was associated with better performance. This was true regardless of size of symbols and is an instance of apparent dominance of the psychological over the physical differences. An investigation by Zirkle, conducted with Rhine's subject, H. P., after the latter had suddenly lost his capacity for high scoring, confirms the advantage of novelty in test performance. The subject was given a series of 272 runs with 21 different experimental conditions, the aim being to change from one to another as soon as the subject appeared to be bored with the method. The variations in technique were often small, such as staring at a fixed point, looking into a crystal ball, following a rhythmic tapping, or walking about the room during the tests. The results of each experimental condition were divided as nearly as possible in the middle, and the average determined for the first half and the second half of the 21 series taken together. The average for the first half was 5.44 and for the second half, 4.90. The difference is significant.

Rewards. The effect of rewards, too, has been investigated in experimental series, and it is indicated to be a positive one. Woodruff and George introduced small rewards (movie tickets) during one month in the midst of a four-month series. They found that the scores in general were improved to a significant extent during the month when the rewards were offered. For example, the critical ratio of the month preceding was 9.54; that of the month during which the rewards were given, 16.25; and that of the month following, 6.3. Approximately the same number of trials were made each month, thus making these critical ratios comparable. During the month in which the rewards were used, they were offered for only two weeks out of the four. There was, however, no significant difference between the reward weeks and the non-reward weeks. Accordingly, if the rewards are to be associated with the greater success obtained during the month, it can only be on the ground that the introduction of rewards brought in a different experimental atmosphere, perhaps making the experiment as a whole more interesting even to the degree that the control week was itself subject to the general effect. Most important, perhaps, on the speculative aspect of this experiment, is the possible competitive spirit that was introduced by offering a reward only to the winner(22).

(22) In an experiment conducted by Rhine in "The Effect of Rewards upon Performance in ESP Tests." [Unpublished MS], a similar general enhancing effect was obtained, but with the same lack of specific relation to the actual period of application of reward. That is, if favorable results are to be attributed to the use of rewards, it is as if the effect is not well localized, but spreads to the entire experiment during the series in which rewards are used.

Price and Pegram and L. E. Rhine used candy or some form of refreshment in their experiments with children. There was, however, no control as to the effect except an incidental one when Pegram(23) mentions that once she forgot to provide the candy, and on that occasion the results were the lowest of the series, being somewhat below chance.

(23) Pegram, Margaret H. "An Investigation of ESP Ability with Mentally Deficient Children." [Unpublished MS].

Competition. The competitive condition was mentioned above as a possible factor in the Woodruff and George results. Price(24), using carefully screened open matching, made a definite study of the relative value of competition with 25 children as subject. Three conditions were used: first, the subject working alone; second, two subjects working together, but in such a way that each had no concern with the results of the other (the checking was done in a manner that did not permit easy comparison); and the third condition, competitive - the two subjects working opposite each other and expecting to compare scores at the end of each run. In the solitary situation, 220 runs were given, resulting in an average of 5.2; 250 runs in the social situation averaged 4.5; 220 runs in the competitive situation averaged 6.5. The competitive tests gave not only the highest result, but the only significant one. This indicates that competition affects favorably ESP scoring by children. It might be suggested that the competitive condition was the one which would appear to the children as most like a game, and hence the one in which they would feel most spontaneous interest. It is of interest to note that the apparent effect of competition did not "spread" as did that of reward, though it is not clear why this difference should have occurred. It is plain that these effects raise further problems.

(24) Price, Margaret M. "An Investigation of the Effect of Competition in ESP Tests with Children." [Unpublished MS.]

Conditions Adversely Affecting Motivation [top]

Under this title is introduced work indicating that unusually long runs, delay in checking and scoring, mechanical timing of rate of response at an abnormal pace, and highly routinizing experimental procedures will inhibit ESP.

Long Runs. Richet(25) pointed out that long runs were a hindrance to good scoring and submitted experimental work to show that when the run was excessively long (over 100 trials) the hits were relatively few. Estabrooks reported better success in the first 10 of 20 trials. Rhine(26) found that runs of 50 and 100 trials gave definite decline curves. Sharp and Clark compared DT* 25 with DT 75 and DT 125. While the averages are not given, inspection of their Graphs 3 and 4 shows that the longer runs averaged below the shorter one of 25. This relation is indicated, although there is much to be learned yet about the relative length, rate, and other circumstances of the test before the relation is at all clearly understood.

* Down Through. The technique in which the cards are called down through the deck before any are removed or checked.
(25) Richet, C. Thirty Years of Psychical Research. New York: The Macmillan Co., 1923.
(26) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.

Delay in Information Regarding Scores. The first suggestion that delay in checking might inhibit scoring came from an investigation by Pratt(27). Rhine(28) found in the analysis of work conducted with sealed opaque envelopes that when the results were separated into three conditions: (a) immediately checked, (b) brief delay in checking (that is, less than 24 hours), and (c) long delay in checking (24 hours or longer), that the score average declined from the first to the third of these conditions. They were significantly positive in the first, slightly positive in the second, and significantly negative in the third. A negative effect with delay in checking is thus indicated.

(27) Unreported, except as pooled total, because conditions lacked something in completeness. It is mentioned here with the reservation expressed in Chapter X (Introduction: Relations and Evidence).
(28) Rhine, J. B. "ESP Tests with Enclosed Cards," Journal of Parapsychology, II (1938), 199-216.

In addition to the delay in checking and informing the subject of his success, there were other conditions that might have contributed to the negative deviation. The sealed opaque envelopes used were sent to the subject by mail, and were used over and over again, the subject sending the record of his calls together with the code number on the envelope back to the experimenter for checking. Along with the delayed checking, there should be considered other conditions which might have contributed to a general feeling of frustration on the part of the subject. While Riess's subject was never told of her scores, even after the experiment was finished. she was presumably confident of her ability to begin with, by reason of earlier experiences, and was told each time that her work was interesting - i.e., the encouragement was simply not numerically expressed. Rhine's subjects were told that much too, but the receipt of the information by mail and only after several days had elapsed, may make all the difference between frustration and encouragement.

Formality of Procedure. Rhine(29) reported in 1934 a total of 18 out of 27 subjects, given at least 100 trials in individual tests, as having given a critical ratio of 2.5 or more. On the other hand, in a later series by Ownbey, Pegram, and Rhine, mentioned above, with the same safeguards and using the method that had been most successful in the earlier work, 43 subjects were each given, on the average, 425 trials with routine schedule and formal procedure. Out of these 43, only one subject approached the critical ratio of 2.5. The three experimenters involved had each obtained significant results with other subjects in earlier work. The only known difference was the fixed routine and the atmosphere of an experimental task rather than an informal game. Pratt and Price found the more informal manner of handling their child subjects was the more successful of the two sets of conditions used. L. E. Rhine referred to the advantage of a spirit of fun. Barrett, nearly sixty years ago(30), said. "But of the favorable effects of freedom from constraint and of a spice of pleasurable excitement, we can speak with entire assurance"(31). Many warnings may be found in the literature on ESP against restraint, formality, over-seriousness, and interference with spontaneity (this need not be confused with relaxation of caution). Pratt and Woodruff report an average of 5.34 with Woodruff working alone, and an average of 5.20 with both experimenters present and with an elaborated routine that requires considerable space even to review. In view of the fact that both common sense and what is known in psychology would lead to the expectation that an easily inhibited process would function best under informal conditions, and in view of the general agreement of expression by experimenters concerning this point, the experimental work is sufficient to warrant the rating of indicated for the hypothesis that formal routinized tests are less successful than the more game-like procedures with the same degree of precaution involved.

(29) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.
(30) ISS note: This article was written in 1940.
(31) Committee on Thought-Reading, SPR. "First Report," Proceedings of the SPR, I (1882), p. 29.

Frustration. Rhine(32) pointed to the negative effect of frustration of the subject by prevailing upon him to continue the tests when he was strongly bent on leaving. His subject A. J. L. had been making, for the preceding 600 trials, an average of 9.9 (ESP Q = 25); but upon reaching the time arranged for his departure, persuasion to participate further against his will was accompanied by a drop to an average of 4 for the next succeeding 400 trials (ESP Q = 20)(33). Estabrooks obtained positive deviations that were significant in his first three series. In the fourth, he asked subjects who had taken part in one series to participate in another in which they were taken to a still greater distance with four closed doors between them and the agent. The deviation was significantly negative (C.R. = 2.7 for suit; 640 trials). Estabrooks himself appears to have expected decline with distance. If the subjects felt as he did, the experiment provided excellent frustration conditions. Negativistic motivation may reasonably be supposed to result from undertaking a seemingly impossible task under mere constraint of courtesy. This work points the way not to a conclusion but to an experimentally suggested relation and to the need of further experiments in its verification.

(32) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.
(33) Assuming that frustration induces negative direction of effort, the subject may be expected to miss the card instead of hit it when frustrated, if ESP operates at all. Hence, if experimentally induced, the results may be treated as low-aim tests and the formula for low-aim ESP Q applied (Q = 20D, where D is the negative deviation).

Mechanical Restraints. Any mechanical restraint in the achievement of a delicate judgment might be expected to interfere with success.

Mechanically Fixed Tempo. That the restraint of a fixed rate of response may adversely affect the ESP judgment has been indicated by Stuart's(34) tempo experiments. In these the timing of the subject's response was fixed by a metronome, and the rate of response varied from one run to another, above and below the normal or favorite tempo of the subject, as ascertained at the beginning of the sitting. Blind matching and screened touch matching methods were used. The results showed that when the rate of response was fixed at that which was normal to the subject, successful ESP performance was possible, but when the rate was abnormal, only chance results were obtained. A mechanical restraint upon movement, either in speeding up or slowing down beyond the subject's normal rate, was found to be inhibitory.

(34) Stuart, Charles E. "The Effect of Rate of Movement in Card Matching Tests of Extra-Sensory Perception," Journal of Parapsychology, II (1938), 171-183.

The Effect of Drugs [top]

The narcotic drugs, alcohol and sodium amytal, and the stimulant, caffeine, have been used in ESP experiments. Several hypothetical relations between drugs and score effects are experimentally indicated.

Small Amount of Alcohol. It should be remembered that the effect of narcotic drugs is in general very dependent upon the amount administered and upon the individual taking them. Small amounts give results which in some respects are opposite to those of larger doses. Brugmans reports getting the best results of his series after the subject had been administered a small dosage of alcohol (30 grams). The 29 trials given following this dosage gave an ESP quotient of 75, which is almost of the order of Riess's remarkable series (ESP Q = 91). The 38 hits of the other 151 trials in Brugmans' series gives the much lower quotient of 18. The amount of alcohol given the subject presumably was too small to have adversely affected the acuity of discrimination, but it may have been sufficient to add not only to the agreeableness of the experiment but to the relaxation and abstraction from surroundings regarded by most experimenters as important in ESP tests. The importance of this relaxation and the passive state of mind was emphasized by Brugmans himself. His interpretation is that the alcohol counteracted the normal inhibitions.

Sodium Amytal. Rhine(34) reported in 1934 the results of experiments with two subjects using the BT procedure, and with one subject using the PT* procedure, in which large doses of sodium amytal (6 to 15 grains) were used. All three subjects were given the drug after lengthy preliminary series which permitted an estimate of scoring levels in the normal state. Two subjects were informed what drug they were given, but had no knowledge of its effect beforehand. The third was informed that the drug in capsule form might be either caffeine or amytal. In all three cases, the score average dropped abruptly and to a striking degree. The subject given the largest dose, 15 grains, dropped to chance average.

* Pure Telepathy.  The word "pure" emphasizes the exclusion of clairvoyance.
(34) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.

Confirmatory results were obtained by Clark and Sharp(36) using the screened touch matching procedure. Subjects were not informed as to which of three capsules they were given: sodium amytal, caffeine, or a control.

(36) Unpublished report.

Caffeine. Rhine(37) reported the use of caffeine with one subject on occasions following depressed scoring and the report by the subject that he was drowsy or fatigued. Comparison of scores before and after administration of the drug showed a significant difference. Another subject, working with the PT procedure, was administered a 5-grain dose of caffeine citrate in capsule form following three hours after a 6 grain dose of sodium amytal which had produced a highly significant drop in score average. The effect of the administration of caffeine on a 300-trial series made one hour after ward was to produce a highly significant rise in score average from that of the 300 trials preceding, made under the influence of the sodium amytal.

(37) Rhine, J. B. Extra-Sensory Perception. Boston: Bruce Humphries, 1934.

This marked drop with the administration of sodium amytal and recovery with the later administration of caffeine was likewise confirmed in the work of Clark and Sharp under the more safeguarded method of the screened touch matching procedure.

Summary [top]

1. Neither hypnotically-induced nor self-induced trance is found, in the limited exploration thus far made, to be advantageous to success in ESP tests. The investigations of these points are, however, limited to the extent that these findings cannot be regarded as more than suggested.

2. It is on the other hand suggested that there is an advantage to a p-value of 1/5. Success, however, is by no means confined to that p-value.

3. Both matching and calling procedures are established as successful. It is indicated that these are approximately equal in advantage.

4. The DT and BM methods are indicated as averaging below BT and OM. ESP shuffle is indicated to belong likewise in the lower-scoring group.

5. It is established that individual tests are superior to group tests for adult subjects.

6. It is indicated, however, that children may constitute an exception to this rule.

7. It is likewise indicated that visitors merely witnessing the performance of the subjects, and taking no part in the experiment, may adversely affect performance in the tests.

8. It is indicated that the subject-experimenter relation may be determinative of success.

9. The following conditions presumably affecting motivation are indicated to be of importance in ESP tests (the first three favorably, and the rest adversely): (a) novelty of material, (b) reward, (c) competition, (d) length of run. (e) delay in information regarding scores, (f) formalizing the procedure.

10. It is indicated that mechanical restraints limiting the subject's rate of performance affect the scores adversely.

11. Narcotic and stimulant drugs in different doses have been indicated to have an effect upon performance: (a) small dosage of alcohol, a favorable influence; (b) large doses of sodium amytal, a lowering of scores; and (c) caffeine, an effect of counteracting the effect of sodium amytal.


The article above appeared in "Extra-Sensory Perception After Sixty Years: A Critical Appraisal of the Research in Extra-Sensory Perception" (1940, Henry Holt and Company, New York) by J. B. Rhine, J. G. Pratt, C. E. Stuart, B. M. Smith and J. A. Greenwood.

Related Material

Home | About Us | Latest News | Biographies | Articles | Experiments | Photographs | Theory | Online Library | Links | Recommended Books | Contact Us | Glossary | Search


Some parts The International Survivalist Society 2004