ESP Over Distance: A Survey of Experiments Published in English

Reprinted from the Journal of the American Society for Psychical Research Volume LIX — January 1965 — Number 1 ESP OVER DISTANCE: A SURVEY OF EXPERIMENTS PUBLISHED IN ENGLISH BY KARLIS OSIS and APPENDIX TO DR. OSIS’ ARTICLE: ASTATISTICAL MODEL FOR EXAMINING THE RELATION BETWEEN ESP AND DISTANCE BY M, E. TURNER Jr.

ESP Over Distance: A Survey of Experiments Published in English Karts Osis! classes and evaluated by new statistical methods devised by Dr, E. Turner, Jr., diene? of Statistics at vane ant ‘Turner's — (see Appendix) is a method for describing any tionship distance, found that the scores in su ine si with The decline does not follow the cus inverse square law of the known radiant energies, but rather an “inverse two-fifth’s law”; i.¢., the scores decline by somewhat less the inverse root, However, the inverse two- fifth’s law cannot be extended outside the surveyed sample to ESP in general because of severe limitations in the data le. ‘THe PROBLEM AND ITS PERSPECTIVE One of the most fascinating characteristics of ESP is that it can overcome long distances in space, and seemingly in time also, penetrating the future. A puzzling aspect of man’s personality is that by just closing his eyes he may on occasion envision correctly what is happening on the other side of the world or what will occur in the future. The available evidence from experiments and spontaneous cases seems to justify the assumption that ESP can reach around the globe (5, 21, 26, 31, 32, 38). This apparent sovereignty of ESP over space and time has, over the centuries, spurred the creative imagination of researchers and philosophers (9, 13, 20, 37, 54, 55). Many able individuals have rightly considered the spatial and temporal properties of ESP to be among the most important keys to the riddle of the nature of human personality, For them, the “reach of the mind,” with its virtual omniscience of far lands and the distant past and future, gives * Tam very grateful to Dr. Malcolm E. Turner, Jr., of University, who my conventional ideas about ESP over distances so ins this survey. His help as consultant and collaborator in the evaluations is hi y valued. The late Dr. E. B, Foster, Mrs. M. J. Mason, Mrs. J. E. Nteser, and Mr. J, M. Smith also gave valuable assistance which is deeply appreciated. i ESP Over Distance: Survey of Experiments 23 ck to man some of the dignity and grandeur lost in modern ific concepts of personality. I believe that this line of thought und and constructive, but only up to a point. » space-time problem has proved to be one of the most danger- pitfalls in parapsychology because it has tem researchers ulate far beyond what our factual knowledge of ESP war- We know for certain only a few facts about ESP as related ‘to space and time, and even these facts are entangled with a de of other variables inseparably interwoven in the data. Let re of constructing gigantic castles with just these few solid of knowledge, added to a whole warehouse of building mate- om philosophies of the past. There is no reason why we n't speculate if we admit that this is what we are doing; wer, it becomes dangerously confusing if strong empirical nce which simply does not exist is claimed as a basis for these castles am fully aware of the controversial nature of this topic and, ness to the reader, I want to state my own bias. When, as a z student, I wrote my dissertation on the hypothesis of ESP, eved that the independence of ESP from distance in space time was an established fact. I took it as the cornerstone of thesis. After coming to the United States, the original sources me available to me; 1 checked there the validity of claims for cut empirical evidence for ESP’s independence of space. my horror I found none. ¢ power of our cultural heritage, or Zeitgeist, to distort ‘ing in parapsychology is clear in the case of the distance problem. I believe, therefore, that we definitely need clarification of the available empirical facts about ESP over distances; this is ‘the purpose of the present ae In order to escape the of the Zeitgeist, I went straight to the primary data in the published (periments sampled and by-passed the opinions of the experi- menters. This was necessary in order to clear the ground for new ive theoretical thinking and better future experimentation. ; DELINEATION OF THE PROBLEM The ESP communication system clearly has three parts. At one end is the internal functioning of the percipient, who acquires information by ESP. At the other end is the stimulus; for example, an ESP card, an accident, or an idea in a friend’s mind. The system’s third part, to which this paper is devoted, is the unknown factor which brings both ends of the communication system togeth- er. In my opinion no one yet knows what this factor is. Among other possibilities, it might be visualized as analogous to the light

24 Journal of the American Society for Psychical Research rays which connect our eye with the object seen, or to the sound waves which connect the tones of a violin and the listener. Modern communication theories are accustomed to calling this mediating factor a “channel.” I think we can adopt the term for ESP also. Let us bear in mind that this channel does not have any spatial properties, such as those of a pipe or canal. Radio waves are call- ed a channel in spite of their being distributed around the globe. This paper is concerned only with the channel in the ESP com- munication system, whatever it might turn out to be, and not with the stimuli or the subjective, internal factors. Further, as I have already said, I will survey and evaluate only the experi- mental data pertinent to the channel factor, and not the hypotheses and speculations which have been based upon these data. Theoret- ical discussions are very necessary, of course, but I will wait to develop some of my own in another paper. PRocEDURE For relevant empirical material on distance in ESP the para- psychological literature in English was surveyed: this included research journals and books. The literature in other languages was touched upon only lightly because of linguistic limitations. In ESP research two basic types of stimuli are used: (a) alter- native-choice, ¢.g., ESP cards, in which the subject knows the alternative stimuli, but must ascertain the order of presentation; and (b) free stimulus material, e.g., drawings, in which the sub- ject does not know the stimulus population and must indicate the nature of the stimuli. These two types of research are reviewed separately, with primary emphasis on alternative-choice methods and visual stimuli. Experiments with ESP cards (or other targets with one-fifth probability for guessing correctly) form the backbone of the survey, and special attention is given to research in which the same subject did ESP tests involving different distances within the same experiment. A few distance experiments were found where playing cards, numbers, letters, and clock cards were used, but these will be mentioned only rather briefly since there were too few instances to permit grouping in distance classes, Some of the istance experiments using free stimulus material are reviewed, but with emphasis only on those in which quantitative evaluation was carried out. Excluded from the survey are all experiments not designed for distance comparisons and in which little-known response methods (ESP in dreams, map dowsing, hypnosis induced by ESP, etc.) Were used. Mediumistic and so-called psychometry experiments were also excluded since their results are open to other interpre- a ESP Over Distance: Survey of Experiments 25 tation: ESP over distances. Broadcast experiments were samplec but not included. in the tabulations since none in which fistances were evaluated were found to have been performed under eceptable conditions. Experiments in which distance ESP was oupled with precognition or psychokinesis were not deemed suitable for inclusion because of the obvious confounding of varia- Finally, no tabulation of spontaneous ESP experiences was apted; this would be a herculean task in itself and would in case involve great difficulties of interpretation in terms of the ‘distance factor. : ; Having decided on the of research material to be included n the ier. another sont arose. What is the minimum dis- tance between subject and target in an experiment. which will nt its being called a “distance experiment” ? Subjects usually rience little change in attitude if they are separated from the et by only a few yards; but they seem to consider it a “real rc tance” if the cards are placed in another building. Therefore, ‘in order to minimize the possible effect of the attitude factor, a hundred yards—the distance ee to separate two buildings— a an arbi lower limit. ; ; : The Pistence within: the feet mile seems to be rather interesting ‘as a factor in the transmission of known energies. Therefore, the distance from a hundred yards to one mile was used for the first “distance class.” ‘The range of the other classes is a thousand ‘miles, If the original source reported the data within a distance ‘tange of, say, 500-1400 miles, the middle of the range—in this case 950 miles—was taken as the basis for classification in the present Distances were measured as around the globe; that is, on the Biccumference. Of course, the distance through the globe, the - diameter, is somewhat smaller, but as the known energies used for ‘communication, ¢.g., radio waves, bounce around the globe, it “was decided to use this same distance measure for ESP in order to i it direct comparison. ae PeThe va ie ength of experiments accepted for evaluation in the survey is twenty runs (500 guesses), However, this rule was liberalized for some experiments which are mentioned, but not included in the main evaluations. ‘a No special measures were taken for selection in terms of se of the experiments as I wanted to include all the data which the claims of no lawful relationship between ESP success ego ‘his ail give rather inaccurate informe because of limitations in the data. Approximately half of the ulated material involves multiple calling for the same set of targets,

26 Journal of the American Society for Psychical Research and this could not be taken into account in our evaluations. Experi- ments | and 2 (in Table 1 below) are i iently described in the original publications. Some experiments involved hi selected However, vie ‘ettaply'cha: ot dea qaipiaiaey better t3: stun anv evaluating claims of evidence for the independence of ESP from REsuLts Card Experiments with Five Alternative Choices Experiments in which cards with five alternative choices were used as stimuli are summarized in Table 1. The levels according to the distance classes are distributed in definite non- AVERAGE PER RUN : classes i with chance fashion. When the distance increases, the class average seg consistently from class to class without exception (Figure ). Most decline is registered between the first two classes; that is, 4 between the 100-yards-to-l-mile class and the 1.1-mile-to-1000- 8 ’SP Over Distance: Survey of Experiments VI 7001-8000 mi. Avg. 6.17 24 | 3.83 48 | 5.00 Runs 24 5.22 353 | 5.07 1000 | 5.09 244 100 | 5.24 Rane | Ave. | 100 yd.-1 mi.| 1.1-1000 mi. |1001-2000 mi,!2001-3000 mi./3001-4000 m u I

28 Journal of the American Society for Psychical Research mile class, Actually, the first of these classes ranges only from one hundred yards to a quarter of a mile, as there are no date for the rest of the range. Dr. Malcolm E. Turner, Jr., Professor of Statistics at Emory University, has developed a statistical model for evaluation of the problem of ESP over spatial distances (see Appendix). He is the sole author of this model. The basic idea of the model is to express in mathematical terms any lawful relationship that ESP might have to distance. The model utilizes modern information theory, as well as the latest advances in statistics. The gist of the procedure in over-simplified, non-mathematical terms is this: ESP scores are transformed by means of appropriate Statistical operations into amounts of information transmitted, as expressed in “natural units” or “nits.” This is necessary in order to provide a basis for calculating statistics which describe important trends in the relationship of ESP to distance. One of these statistics is *, which Turner defines in the Appendix: “7 is a constant we may term the ‘extrasensory perceptivity’ and is a measure of the ‘power’ of the Sa ag i system in a particular situation.” It is necessary to calculate this “‘extrasensory perceptivity” in order to describe, by another statistic, 4, precuate the ee in which ESP declines over distances, if indeed it does. This statistic 5, or “proportionality constant,” is Turner’s basic tool for pinning down mathematically any distance law. With it one can sweep across all the major ibilities of a distance law to find precisely that one which dts e data best. For example, if 8 is found to be 0, ESP is not related to distance; if 5 is equal to -2, the result is the inverse square law characteristic of known energies when radiating in all directions. Other values of 8 would show, for example, whether ESP decreases over the distance rapidly or slowly; uniformly over any distance; with a rapid decline at first, sti to a slight decline, etc. Dr. Turner kindly volunteered to evaluate the survey data repre- sented in Table 1. The evaluations were done in accordance with his model, and the steps are outlined in the Appendix. He found that the following trends occurred: 1. ESP scores are significantly related to distance. 2. “Extrasenso perceptivity or “power constant” 7# = 0.608. 3. “Proportional i Ege which “indexes the class of distance ws,” 5 = —0.398. 4- Some chance variation (error variance) could, of course, affect the magnitude of 8, Its true value most likely is within the usual statistical confidence limits, At the 95% confidence limit, the true 8 should be somewhere between the following values: P{—409 < 8 < —.387} = .95. ESP Over Distance: Survey of Experiments 29 At the 99°, confidence limit, which is more conservative, the finding is: P{—.416 < 8 < —.380} ~ .99. : j : 5. Turner calculated on the basis of statistics the way in which the average run score would with distance. The predicted mean run score, as a function of distance x, is: hy = 5e%8=;"", where ¢ is the base of the natural logarithm. Turner concludes that 5 = —.4 = —2/5 seems to be as good a choice as any for describing the ESP distance law; in other words, ESP obeys the inverse two-fifth’s law in the data of Table 1. Let us sum up: In the data evaluated, there is a significant relationship between ESP scores and the distance between subject and stimulus. ‘The relationship indicated is not of the usual inverse “square type, but more like that of an inverse two-fifth’s. In other words, in these experiments the ESP scores decreased somewhat less than by the inverse square root of the distance. _ The crucial question arises: Can we generalize from the results of this survey sample that ESP effects are inversely related to dis- tance? The limitations and weaknesses of the published data have been pointed out above. In my opinion the data are too scant and ive to warrant safe conclusions as to ESP in general. These same data have been the basis for the Zeitgeist interpre- ‘tation that ESP is independent of distance Sten Tugner's analysis does not su the independency hy is at all; i ‘anything, what is in Pied is the dependency of ESP on distance. ever, Turner’s evaluations set a challenging example of what can be extracted from the data of better-designed future experi- ments, aimed at the channel factor in ESP. 1 Ss the S Subj Two Different Dis- gh of cores ihe ubject over Two Differ __ In the data presented in Table | variables associated with the subjects (individual differences, attitudes, etc.) are neither i nor statistically controlled. The first distance class is filled exclusively with the scores of some of the best subjects ever known in parapsychology. This factor alone might explain the scores in this class. . we could sufficiently balance factors other than distance influencing ESP scores, and at the same time vary only the distance between subjects and targets, much more dependable information would be obtained. I do not know of any published experiment where this requirement was fully met. The closest approximations to it are the cases in which the same subject has guessed cards at two different distances within one experiment. were seven such cases in the surveyed publications (see

30 Journal of the American Society for Psychical Research ESP Over Distance: Survey of Experiments 31 of kit at ‘Table 2). In cases | and 2, and in the Antwerp Series of 3, near 3 S=8 3 SRS| = and far distances were not used in the same session. In the Cam- addi ¥ oa] & bridgeSeries of case3,and in case 4, the long-distance calls were done t in the favorable first half of the session and the near-distance calls afterwards. In case 1, the procedures are only sketchily des- 8 eribed. In case 3 (Cambridge Series) some synchronization errors z = m have interfered with the distance part of the experiment.* Only in case 3 (Antwerp Series) did the same agent look at the targets at both distances. is kind of comparison clearly indicates the same trend as in distance } previous analyses. In all asthe scores at the loge are than the scores for the near-by — (Table 2 and Figure 2). In five out of seven cases the scoring differences are individually 5 significant (in one case P < .05; in four cases P < .01). The CR of ewe difference of the total is 7.16. The consistency of the trend is “aad ZZA Mest Satemes a Es to Y : - Y y 5 7“ Y i Y : ‘ Y : Y : ] d pt sd yr 2. ESP scores of the subject different distances - _ There is in Table 2 a considerably smaller number of variables that could account for the differences observed than there is in Table 1. Nevertheless, these differences still might be explained in terms of the subjective variables only. The most serious counterhypothesis is raised by J. B. Rhine and ? If we exclude case 3 (Cambridge Series) from the analyses, the findings will not change. et ant at the 0S level. Experiment 1 (16 5 (16) (19) Pratt Series (39) Antwerp Series Haeckel Exp. ( Totals average scores, average differences, and CR¢'s were rounded for sake of convenience in presentation; hence, they will not ibson: Sub Sub Sub: Rhine. No. 1 G n: j 2 ine, : 3 Soal, : series (50) 4 Osis: . (31) £h er vmracts with Give aiteenative cholees ace inctaded: check ** Significant at the .01 level

32 Journal of the American Society for Psychical Research others (36, 50, 59) to the effect that the subject’s attitude toward the distance condition could in itself explain any scoring trends observed. This is a sound hypothesis, but it still lacks empirical validation. For example, if the subject feels challenged and alerted by the task of “overcoming distance” by ESP, he might on this account score best on distance targets. On the other hand, if he is adversely affected by the frustrating situation of distance ESP, the scores might go down. The importance of the attitude factor in distance telepathy was already recognized by E. M. Sidgwick and A. Johnson in 1892. At that time they called it “'... the effect of the distance on the imagination of the agent or pid cu eel (46, p. 547). ¢ obvious way out of this difficulty would be to set up exper- iments in which the variables of distance and attitude are con- trolled. We could ascertain the subject’s attitude and take it into account in statistical assessment; and by the same token we could also find out what the subject habitually does when confronted with a frustrating situation, since distance ESP targets are frustra- ting for most subjects, Alternative-Choice-Tests Not Included in the Main Evaluations In addition to the standard card tests with five alternative choices, playing cards, clock cards, numerals, letters of the alphabet, etc., have been used in distance experiments, with the number of alternatives varying from two to eighty. Scores on these tests are not directly comparable, nor do the tests provide enough data to permit grouping in distance classes. Furthermore, very few of these experiments satisfy the minimum requirement of twenty runs. Nevertheless, I want to mention some of this work even though its relevance to the distance problem is slight. Experiments with separate scores on two p wei provide more useful material for compere we can easily detect the general trend; that is, which of the two scores is lower. However, unified statistical treatment is not possible here because the probabilities of hitting are so heterogeneous and the reported statistics so varied. Therefore, only a simple tabulation of the igher and lower scores is noted. n 1938 R. Shulman (45) performed ESP experiments using sound stimuli with five alternative choices over distances from a hundred to four hundred yards. These were compared with tests using screened color ESP cards, where distance was not involved. Unfortunately, there is very little known about ESP and auditory stimuli, as nearly all ESP experiments have utilized visual stimuli. We cannot include Shulman’s data in the main evaluations because ESP Over Distance: Survey of Experiments 33 stimulus and distance variables were confounded. However, his results follow the general trend of decline already observed in the survey: 345 runs, no distance, average score: 5.17; 223 runs, over distance, average score: 5.08. _ Some experiments with ESP cards (25) and with other cypre sf cards (14, 15, 29) obviously involved distances, but either e data ‘were not separated into distance categories or the distances were not stated. ; F. L. Usher and F. P. Burt (56) in 1907 performed a pioneering i iment with playing cards over distances ranging up to ; ‘The number of trials was very small: 36 trials, in the same ‘room, 9 complete hits; 60 trials, over distance, 4 complete hits. (This involved multiple calling of some cards.) The trend is obvious. J. E. Coover (11) reported in an appendix to his larger work a brief experiment using playing cards with the face cards omitted (P = 1/40). The distances varied from twenty-seven to ninety ‘miles. He obtained 5 hits in 149 trials. The scores on the various distances are not clearly separated in the report. : S. G. Soal (49) performed distance tests involving free material and a few playing card trials in the early 1930's. His evaluations ‘are much too complex for this review, but he concluded that in his data there was no evidence that ESP functions at any distances. In 1932 'T’. Besterman (2) published the results of an experiment between London and Athens. The stimulus material consisted of pictures with twenty alternative choices. There were 674 guesses: deviation was —4.7. No distance comparisons were made. C. B. Nash (30) in 1961 followed up the star subjects in Chese- brough-Pond’s ESP television contest with distance tests using the numbers 0-9 as stimuli. In 180 runs he obtained a deviation of —41, associated with P = .04. Further, twenty subjects at a dis- tance ranging from 6 to 520 miles had a deviation of —16, and nineteen subjects at a distance from 680 to 2400 miles had a devia- tion of —25. A very slight non-significant decline effect seems to be present, but the numerically larger negative deviation is associated with the greater distance. Nash’s test was meant to be a test of GESP; however, the instructions to the subjects were like those devised for precognition tests, emphasizing future comparison. D. Michie and D. J. West (27) performed in 1957 a mass tele- vision ESP test, and followed up their star subjects. Clock cards were used as stimuli in the follow-up. One high-scoring subject, Mr. Downey, was extensively retested. His ‘mean di ce score deviation” over a distance of approximately a hundr miles was 0.8125; at no distance, 0.7627. The difference is extremely small and in the opposite direction from the general trend of this survey. In 1963 A. E. H. Bleksley (3) reported an experiment on long-

34 Journal of the American Society for Psychical Research distance ESP during sleep. This involved the subject’s awakening each night at a certain time randomly selected by the experi- menter in Johannesburg, South Africa. During the first series the subject was traveling from South Africa to Holland; distance varied and were quite great, up to five thousand miles. During th main series the subject remained in South Africa at a distance o nearly a thousand miles. Bleksley’s evaluations are too involved to describe in this survey, but more ESP success was reported at the constant, but shorter distance. Bleksley gives plausible psy- chological explanations for this difference: He met the subject only after the series at the longer distances was completed. G. L. Mangan (21) performed a distance experiment over approx- imately 2150 miles. He used dual-aspect targets with 1/5 probability of hitting each aspect and 1/25 for hitting the whole card. The subject was instructed to make one guess a day for a hundred days. Ninety-five trials were scored and resulted in 13 complete hits, with 3.8 expected by chance. This result is associated with P= .000002. Partial hits, as well as displacements, were also evaluated. The number of trials is too small for inclusion in the main evalua- tions. However, the rate of ESP scoring over so long a distance is very high indeed and this small experiment stands out as very impressive evidence for long-distance ESP. Experiments Using Free Stimulus Material Distance experiments with free material are numerous, but difficult to evaluate. Most reports do not include quantitative evaluations. In cases where quantitative evaluations are given, subjective judgment on the part of the judges is always involved, leaving room for an alternative interpretation—the judge’s own ESP exerted during the judging process. Of course, the judge’s ESP influence is less likely to be a factor in experiments where complete similarity between response and stimulus is demanded, as in M. C. Marsh’s main evaluations (22, 23). Distance tests with free stimulus material were attempted as early as the 1880's by S.P.R. researchers, and F. Podmore’s book (33), published in 1902, contains an entire chapter summarizing the results of many attempts to produce “telepathic effects at a dis- tance.”” The Miles-Ramsden series (28) was a later long-range exploration of the same . R. Warcollier’s experiments (58) in the 1920's and 30’s with free material made many useful contri- butions to the field of parapsychology, but their lack of quantitative assessment makes them useless for distance comparisons. The same is true for the Sinclair experiments (34, 47). Methods of of evaluating free stimulus material did not reach ESP Over Distance: Survey of Experiments 35 maturity until the 1940's (5, 6, 7, 8, 51, 52). The first distance eat y of sophisticated design (drawings as stimuli) was pub- by W. Carington in 1940 (5) and further similar research in el as papers (6, 7, 8, 10). The statistical assessments yielded highly significant measures. The distances between target “material and subjects varied widely from rather small to transatlantic. Carington did not seem to be particularly interested in the ‘distance factor and evaluated it only in a small, highly-selected portion of his data, Duke Group, Series IV-A (5). The twelve Duke ‘subjects (longer distance) scored significantly better than the rest of the 105 participants, located in England and Holland. This result certainly falls in line with the Zeitgeist, but it does not reflect on the experiment as a whole because we do not know what dis- tance trends may be hidden in the data. Moreover, Carington’s methods, particularly his catalogue technique, have been severely criticized by 5S. G. Soal and F. Bateman (50) and by the Board of Review of the Journal of Parapsychology (4). In the face of these ‘difficulties, I refrain from stating my personal position and t ‘Carington’s own conclusion: “Distance makes no difference” (9, 69). 4 Carington’s experiments were repeated twice by A.S.P.R. ri- menters. The first repetition, by E. Taves, G. Murphy, and L. A. Dale (53), a large-scale experiment with 272 subjects and 8723 ‘responses, did not yield significant results on the main measures taken. No distance breakdown was performed. G. R. Schmeidler and L. W. Allison (44) undertook a second trepetition in which they followed Carington’s method more closely. His catalogue method, which has not withstood later criticism (50), was used. The experiment as a whole gave significant results (P = .012). When the agents and subjects were in the same build- ing, the average score was .484; when the subjects were in “the New York area,”’ the average score was .125; and when the subjects were scattered over the U.S.A. (including some in New York), the average score dropped to .062. J. H. Rush and A. Jensen (42) performed a reciprocal drawing experiment of fifty trials over distances of 200-500 miles, Assess- ‘ment gave P = .002. No distance comparisons were made. M. C, Marsh developed a well-designed experiment with draw- ings as stimuli, and used a distance of 470 miles. This is documented in his Ph.D. dissertation (22) and also in a brief published report (23). Marsh’s 371 subjects achieved “133 raw A hits on the targets used in the experiments, as against 37 hits on the unused control ” (23, p. 8). Marsh used several statistical assessments which proved to be highly significant. No distance comparisons are possible.

36 Yournal of the American Society for Psychical Research On the whole, in distance experiments with free stimulus material the quality of the evaluation methods is very uneven; in many of them evaluation is inadequate by modern standards or lacking altogether. Therefore, this area of research is not a particularly fruitful one for probing the distance problem, although it does seem to give us some clues. Trends in the Data Not Included in the Main Evaluations The experiments not included in the main evaluations (Tables | and 2), but which have been briefly described, are so h c- neous that unified quantitative analyses seem impossible. For that reason this section was nicknamed the “‘Tidbit Basket!’’ However, the experiments in which two Ringe rel were used erent Sarena to comparison by a ranking m ; that is, it is possible to deter- mine which polis at which distance within each experiment ranks higher. Table 3 summarizes the ranking efforts. Of the Table 3 Exvesiments Usinc Two Distances, Not Inctupep iy Mar Evatuations No. of Higher Score ‘ Experiment Alternatives Qualifying Comments Near Far 2 En eet Shulman (45) 5 x Sound stimuli for far tar- fat visual for near Usher, Burt (56) 52 * trials only Nash (30) 10 * Negative deviation numeri- y larger for far subjects ie, W 1 12 x i between suc- mod ae Coept Ciak) and unsuc- cessful (West) experimen- ters i Personal contact with sub- an ©) Poe x ject only after far series Carington (5) free Slanted selection of data: material | x only one group (Duke) in one series evaluated for distance Schmeidler, free Use of Carington catalogue | x method Allison (44) material seven experiments included in this table, five yielded higher scores on the near-distance targets, and two gave higher scores on the long-distance targets. Difficulties of interpretation are summar- ized in qualifying statements in the right-hand column of the table. ESP Over Distance: Survey of Experiments 37 Discussion In a careful analysis of the relevant experimental data it was noted that ESP scores seem to decrease quite uniformly with in- creased distances. On the other hand, in sampling opinions about the relationship between ESP and distance, we find that most researchers express a view opposite to the direction the data take. W. Carington’s statement is quite representative of published opinion: “Distance makes no difference. All workers seem to be agreed on this” (9, p. 69). A few leading experimenters, however, such as S. G. Soal (50) and D. J. West (59), insert cautious clauses to the effect that it is likely, but not definitely established, that distance makes no difference. In the French school there have been dissident voices, among them those of R. Kherumian (19) and R. Amadou (1), who said: “Our conclusion would, then, be simple and prudent. It has not been proven whether distance affects or does not affect the intensity or the intelligibility of a telepathic - message ...” (1, p. 223). ; What is the reason for this discrepancy between the trends in the data and the majority of opinions? In part, as I pointed out above, it seems to be caused by our cultural inheritance, or Zeit- geist. Philosophers such as Kant, Schopenhauer, von Hartmann, and others have stressed the idea that ESP goes beyond the sub- jective forms which we call space and time. In part, the confusion ‘seems to arise from our expectation that a decrease in ESP scores is an either/or situation: that a decrease in scores over distances either will be tremendous or will not occur at all. The logical possi- bility that a decrease in ESP efficiency could, in fact, be quite ‘moderate has been overlooked. G. R. Schmeidler clearly perceived the di of detecting such a decrease: “If, for exam- ple, psi is really proportional to distance or time raised to some } power, where the value of the fraction is small, very ‘wide experimental variations will be necessary to detect any effect" (43, p. 111). Another possible root of confusion may lie in unsystematic interpretations of the data. A logical way to deal with the results ‘would be to ascertain carefully the trends in the data when distance between the stimulus and the percipient is introduced, and only thereafter to look into the causative factors responsible for the trends. For example, if the scores do decline with distance, we cannot tell immediately whether to ascribe this decline to subjective factors, such as attitudes, beliefs, feeling “‘on the spot,” etc., or to the channel factor—the hypothetical energy which connects stimuli and percipient. In practice, we tend to jump quickly to this

38 Journal of the American Society for Psychical Research second level of interpretation without carefully and thoroughly investigating at the first level. Very little work, if any, has been done on the third level of interpretation; that is, evaluating jointly the subjective factors and the channel factor, and their interactions; nor has much effort been made to analyze in a reliable way either the components of the subjective factors or the possible components of the nel factor. Some quite sharp theoretical analysis of the channel factor has been done, such as that of H. A. C. Dobbs (12), B. Hoffmann (17), J. H. Rush (41), and L. L. Vasiliev (57). Their hypotheses, however, have not been followed up experimentally, except in an often sporadic way by Vasiliev. e need for clearly restating the problem and devising sharply- delineated hypotheses and modern means for testing them is great indeed. Turner’s model (see Appendix), upon which the main evaluations in this paper are based, is one attempt. Let us remember that Turner's approach keeps strictly within the data description level; that is, the first order of interpretation. So far as I know, he is the first to propose such a detailed and mathemat- ically sound description of data concerning ESP over distances. He has carefully avoided any confounding of the first-order inter- pretation by the other higher-order interpretations mentioned above. I think this is a desirable approach. It is possible to design experiments in which variables are ascer- tained that will permit mathematically precise data descriptions and also will generate second- or third-order interpretations. For example, in my forthcoming paper, to be published in the next issue of this Journal, I describe an experiment incorporating sta- tistical control not only of some of the attitudes and personality variables, but also of some of the components of the channel varia- bles, i.c., distance in space and time and their interactions. This is just a small beginning. Nevertheless, the twin problems of ESP over distance in space and time and of ESP orientation (how to find a particular point in space) present themselves as very challen- ging and promising tasks. A sound and systematic approach to these problems will probably establish clear-cut trends in the ESP channel factor. This would permit a definitive and precise theoret- ical analysis in scientific terms, which is not possible at present. Only when enough bricks are properly made can one build a sound structure. There are breath-taking possibilities for a future theory of the ESP channel factor. For example, will an unknown energy be Le a with the aid of modern communication-theory tools such as precise energy expansion curves, signal-to-noise ratios, etc ? Will a new additive dimension be found in line with the hypo- theses of Chari, Hart, and Smythies? Or will a qualitatively different ESP Over Distance: Survey of Experiments 39 structure-giving principle emerge such as Driesch's Seelenfeld, Jung’s “‘other-valued reality,” or Tyrrell’s “Elsewhere” ? There is ty of room for a variety of approaches to the channel factor. t seems to be a ‘‘many splendored thing.” This survey does not pretend to give final answers to questions concerning ESP and distance, or to the specific question of the relationship between distance and the channel factor in ESP. But it has succeeded, I hope, in its intent to clear the ground and estab- lish the available empirical trends as a basis for new creative think- ing and better future experiments. REFERENCES 1. Amapou, R. La Parapsychologie. Paris: Editions Denéel, 1954. 2. BesterMAN, THEopore. “An Experiment in Long-Distance Telepathy.” Journal S.P.R., Vol. 27, April, 1932, 235-236. 3. Brexsiey, A. E. H. “An 7 Saige? on Long-Distance ESP During Sleep.” Yournal of Parapsychology, Vol. 27, March, 1963, 1-15. 4. Boarp or Review. “Letters to the Editors.” Journal of Para- psychology, Vol. 4, June, 1940, 153-155. Carincton, Wuartety. “Experiments on the Paranormal Cognition of Drawings, I."” Proc. S.P.R., Vol. 46, 1940, 34- 151. ba 6. . “Experiments on the Paranormal Cognition of Draw- ings, II.” Proc. S.P.R., Vol. 46, 1941, 277-344. 7. . “Experiments on the Paranormal Cognition of Draw- ings, III.” Proc. A.S.P.R., Vol. 24, 1944, 3-107. 8. ———. “Experiments on the Paranormal Cognition of Draw- ings, IV." Proc. 5.P.R., Vol. 47, 1944, 155-228. 9. ———. Thought Transference. New York: Creative Age Press, 1946. 10. CartncTron, WuHaTeLy, AND Heywoop, Rosa.inp. ‘Some Positive Results from a Group of Small Experiments.”” Proc. S.P.R., Vol. 47, 1944, 229-236. 11. Coover, Joun E. Experiments in Psychical Research. Palo Alto, California: Stanford University Press, 1917. 12. Dosas, H. A. C. “Review of L. L. Vasiliev’s Experiments in Mental Suggestion.” Journal S.P.R., Vol. 42, March, 1964, 229-248. 13. Dopns, E. R. “Telepathy and Clairvoyance in Classical Antiquity.” Journal of Parapsychology, Vol. 10, December, 1946, 290-309.

40 Journal of the American Society for Psychical Research 14. Fisk, G. W., and West, D, J. “ESP and Mood: Report on a 15. 16. 17. 18. 19. 20. 21. . Mies, Crarissa, and ‘Mass’ Experiment.” Journal S.P.R., Vol. 38, September, 1956, 320-329. . “Towards Accurate Predictions from ESP Data,” Journal S.P.R., Vol. 39, December, 1957, 157-162. Grsson, EpMonp P. “A Study of Comparative Performance in Several ESP Procedures.” Journal of Parapsychology, Vol. 1, December, 1937, 264-275. Horrmann, B. “ESP and the Inverse Square Law.” Journal of Parapsychology, Vol. 4, June, 1940, 149-152. Kaun, S. Davin. “Studies in Extrasensory Perception: Experi- ments Utilizing an Electronic Scoring Device." Proc. A.S.P.R., Vol. 25, 1952, 1-48. ener ie R. “Anticipations.” Revue Métapsychique, 1950, Lupwic, A. F. Geschichte okkultistischen (metapsychischen ) Forschung, 1. Pfullingen: Johannes Baum, 1921. Manoan, G. L. “An ESP Experiment with Dual-Aspect Targets Involving One Trial a Day.” Journal of Parapsychol- 6), Vol. rape ay aw nae —— arsu, M. C, Le e in a- Perception. Unpub- lished Ph.D. thesis. Ceahacapieunnltteath ‘Africa: Rhodes University, 1958. . “Three ESP Experiments Using Drawings as Target gael Publications of the South African S.P.R., No.5, - McManan, Etizasetu A., and Bates, E. K., Jr. “Report of Further Marchesi Experiments.” Journal of Parapsychology, Vol. 18, June, 1954, 82-92. hme os - McManan, Etizapetu A., and Laver, Joan. “Extrasensory Perception of Cards in an Unknown Location.” Journal of Parapsychology, Vol. 12, March, 1948, 47-57. . McManan, Exizaseru A., and Rute, J. B. “A Second Zagreb-Durham ESP Experiment.” Journal of Parapsychology, Vol. 11, December, 1947, 244-253. a4 - Micutz, D., and West, D. J. “A Mass ESP Test Using Tele- vision.” Journal S.P.R., Vol. 39, September, 1957, 113-133. , Hermione. “Experiments in Thought-Transference.” Proc. S.P.R., Vol. 21, 1907, 60-93. . Murpny, Garpner, and Taves, Ernest. “Covariance Methods in the Comparison of Extra-Sensory Tasks.” Journal of Para- chology, Vol. 3, June, 1939, 38-78. ASH, OLL B. “Retest of High Scoring Subjects in the Chesebrough-Pond’s ESP Television Contest.” Yournal A.S.P.R., Vol. 57, April, 1963, 106-110. ESP Over Distance: Survey of Experiments 41 81. 45. s$ S$ #238 & £8 8 Osis, Karis. “ESP Tests at Long and Short Distances.’ Journal of Parapsychology, Vol. 20, June, 1956, 81-95. Osis, Karuis, and Prenaar, D. C. “ESP Over a Distance of Seventy-Five Hundred Miles.” Journal of Parapsychology, Vol. 20, December, 1956, 229-232. . Popmore, Frank. Apparitions and Thought-Transference. New York: Charles Scribner’s Sons, 1902. Prince, WALTER FRANKLIN. “The Sinclair Experiments Dem- onstrating Telepathy.” Bulletin 16, Boston Society for Psychic Research, 1932. Rune, J. B. Extra-Sensory Perception. Boston: Bruce Hum- phries, 1934. . “The Effect of Distance in ESP Tests.” Journal of Parapsychology, Vol. 1, September, 1937, 172-184. . New World of the Mind. New York: William Sloane Associates, 1953. Rute, J. B., and Humpnrey, Berry M. “A Transoceanic ESP Experiment.” Journal of Parapsychology, Vol. 6, March, 1942, 52-74. . Rue, J. B., and Pratt, J. G. ‘A Review of the Pearce-Pratt Distance Series of ESP Tests.” Journal of Parapsychology, Vol. 18, September, 1954, 165-177. . Riess, Bernarp F. “A Case of High Scores in Card Guessing at a Distance.” Journal of Parapsychology, Vol. 1, December, 1937, 260-263. . Rusu, J. H. “Some Considerations as to a Physical Basis of ESP.” Journal of Parapsychology, Vol. 7, March, 1943, 44- 4 9, . Rusu, J. H., and Jensen, Ann. “A Reciprocal Distance GESP Test with Drawings.” Journal of Parapsychology, Vol. 13, June, 1949. . Scumempier, Gertrupe R. “Research Projects in Parapsy- chology.” Journal of Parapsychology, Vol. 12, June, 1948, 107-113. . Scumetpter, Gertrupe R., and Auiison, Lypia W. “A Repetition of Carington’s Experiments with Free Drawings.” Journal A.S.P.R., Vol. 42, July, 1948, 97-107. ; Suutman, R. ‘Research Notes: An Experiment in Extra- Sensory Perception with Sounds as Stimuli.” Journal of Parapsychology, Vol. 2, December, 1938, 322-325. 46. Sipcwick, Mrs. H., and Jounson, Miss Atce, “ 47. iments in Thought-Transference.” Proc. S.P.R., Vol. 8, 1892, 536- 596. Sinciair, Upton. Mental Radio. Monrovia, California: Upton Sinclair, 1930.

42 ournal of the American Society for Psychical Research 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. Sxreinsky, Morris. “A Comparison of Names and Symbols in a Distance ESP Test.’’ Journal of Parapsychology, Vol. 14, June, 1950, 140-156. Soar, S. G. “Experiments in Supernormal Perception at a sae sa See — 40, 1932, 165-362. aL, S. G., and Bateman, F. Modern Experiments in Telepathy. New Haven: Yale University Press, 1954. Stuart, C. E, “An ESP Test with Drawings.” Journal of Parapsychology, Vol. 6, March, 1942, 20-43. . “An ESP Experiment with Enclosed Drawings.” Journal of Parapsychology, Vol. 9, December, 1945, 278-295. Taves, Ernest, Murpny, GArpner, and Date, L.A. “American Experiments on the Paranormal Cognition of Drawings.” Journal A.S.P.R., Vol. 39, July, 1945, 144-150. Tiscuner, R. Geschichte okkultistischen (metapsychischen) Forschung, U1. Pfullingen: Johannes Baum, 1924. Tyrrett, G. N. M. The Personality of Man. Baltimore: Pen- guin Books, 1947. Usuer, F. L., and Burr, F. P. ‘Thought Transference.” Annals of Psychical Science, Vol. 8, 1909, 561-600. Vasiuiey, L. L. 7 in Mental Suggestion, Church Crookham, Hampshire, England: Institute for the Study of — Images, 1963. ARCOLLIER, RENE. Experimental Te. . Boston: Boston Society for Psychic Research, 1938. zee West, D. J. Psychical Research Today. London: Gerald Duckworth, 1954. A.S.P.R. 880 Fifth Avenue New York, N. Y. 10021 transmitted units determined by the base of the logarithm used in the definition (1) or (2); here the units are natural units or nits). Solving (2) for the mean scoring level at the i” distance gives Appendix: A Statistical Model for Examining the Relations between ESP and Distance M. E. Turner, Jr. We consider a standard card-guessing experiment ormed at several distances between objects being guessed and guessing subject. We suppose that it is of interest to investigate what relation (if any) holds between success at guessing and distance. We assume that the experiment has been carried out in such a way (the usual practice) that in the case of no ESP the chance of success on a oh trial, p,, is a constant at all distances (with the standard e cards p = 0.2), and the number of successes y, out of n, trials, at the i” distance, has a binomial distribution. Now, if ESP occurs, y, will still have a binomial distribution we will assume) but the probability of success will be changed m p, to p(x,), a function of the distance x, For a binomial variate the mean is given by np. ‘Thus, in the case of ESP we ma say that the mean scoring level is changed from the chance level, (¥) = "Pp, to a new level, M(y,) = mip(*,). We now describe this change in mean scoring level in terms of the change, », in probability entropy. We define (cf. 1, pp. 2-4) » by the relation 1 = —log.[p(*:)/Pe), (1) or, in terms of mean scoring level, 1 = —log.[M.0)/mP-). (2) We note that increasing the scoring level decreases the amount of probability entropy (the negative of the change in probability entropy is sak y referred to as the amount of information an like the probability entropy itself, is measured in MA) = pe. (3) By equations (2) and (3) we can express the change in probability ‘entropy in terms of the mean scoring level, or the mean scoring level in terms of the change in probability entropy, at the # distance. Let us now postulate a class of possible relations between the change in negative entropy and the distance, and hence by (3) between the mean scoring level and the distance. We suppose

44 Journal of the American Society for Psychical Research a the change in negative entropy obeys the “general distance w"": dn, ‘ eer “) where 6 is a proportionality constant and yé is an origin adjustment for the distance scale. The integrated form of (4) is found to be m = — a(x, — ¥5)%; (5) m is a constant we may term the “extrasensory perceptivity”’ and is a measure of the ‘“‘power"’ of the object-subject system in a particular situation. It may be seen that the proportionality constant 8 indexes the class of distance laws. For example, if 5 = —2, (5) becomes the familiar “inverse square law” so useful in describing the decrease in the radiant energy of light. If 5 = 0, no increase in probability entropy occurs (this case corresponds to the state- ment: “ESP is not related to distance”). If 8 = —1, the relation is hyperbolic; if § = +1, the relation is linear; if 5 = +2, the relation is parabolic; and if |5| —> o, the relation is exponential. All intermediate cases are also powelle A discussion of this family of curves has been given by ‘Turner, Monroe, and Lucas (2). Combining (3) and (5) we obtain the general distance law for mean scoring level MAy) = npr. (6) Consider two examples. If 6 =0 (no effect of distance), (6) becomes MAy) = 1h"; (7) that is, the mean scoring level is equal to the mean chance scoring level multiplied by a factor which is the antilog, of the extrasensory perceptivity. If 6 = —1 (the hyperbolic case), (6) becomes My) = mpeier™. (8) Other special cases follow similarly. Now the practical problem of investigating the ESP-distance relation given a set of ESP scores at different distances can be solved using the model (6) and appropriate statistical procedures. Model (6) can be fitted to the data yielding estimates of the index of law & (thus identifying the most plausible law). In addition confidence limits for 5 can be obtained identifying the class of lausible laws. If these limits contain zero the hypothesis of no P-distance relation cannot be rejected. Estimates and confidence limits for the extrasensory perceptivity 7 will also be of interest. Consider N scores y,, Vg, .--, ¥v at N distances x,, X, .... *n- The likelihood function is given by ESP and Distance: A Statisucal Model 45 1 na ag ele That choice of values for y, 5, and + which makes the likelihood function a maximum may be taken as the most plausible choice (likelihood principle). In practice finding such a choice may involve rather tedious computations. A simpler approach will be developed here as an approximation to the use of the likelihood principle. We will restrict attention to the case y = 0 (no distance adjustment needed; exponential law excluded) although this approach can be modified for the case y nonzero. We divide (6) through by m,p,, take logarithms twice and then set y = 0. We obtain loge log (Me) — tog, + 8 log x, (10) a linear form. In (10) we now replace M,(y,) 24 the observed estimate y,, the scoring level at the # distance. The relation (10) is no longer exact so we absorb the discrepancy by a residual element ¢; to give log, log,( =) = log, + Slog x, +<¢,, (11) a linear regression model. Note that (11) is defined only for y, > n,p,. For w sufficiently large and for large samples this will usuall be the case. For other cases the procedure will have to be modifie Let the left side of (11) be denoted by Y, and log x, by X,. Then we rewrite (11) Y, = log, # +8X,+«. (12) Now we will use standard linear regression procedures to fit (12), providing estimates of the slope 6 and the intercept loge = (and hence = itself), and to provide confidence limits for 5. The variance of «, can be approximated using the following “propogation of error” argument. see _«s bas am ~ ¥log(yi/miPe)’ and Vie) = Vo) A "= (MG) log [MAy)/np J}? But the V(y,) =M,(y)[1—M.(9)/m). So V(<,) = ee er (13) * MAy,) (log [M.Ay0)/mibel}

46 Journal of the American Society for Psychical Research Since V(e,) is not constant it is appropriate to perform a weighted regression analysis using as weights 1/V(y,). However, V(¢;) is a function of the unknown constants 7 and 4; so as a further approx- imation we suggest using as weights _ Yillog.(yi/miPe)]? oat worse (14) Now we compute in stepwise fashion: (1) Raw sums = w; = w,X, ‘7 p> W, Y, z= w,X, Py = w xX i Y; = Wy Y? (2) Corrected sums S., = De,X7¥;— (2 w,X,)(2w, ¥;) >> WW, Snesene. See vy i+¢ > w, (3) Estimates of parameters §= Szy/Sez # = exp{[d mY, —§ 2D w,X,)/ Dw) (4) Confidence limits for 5 SSR = 5S, SSE = S y— SSR sg = SSE|((N — 2)Sz2] t.(N — 2) = t = tabular value of “Student’s” ¢ at the a significance level with N-2 degrees of freedom § — tss <8 < § + tsy with 1 — a confidence Then any value of 5 lying within the limits of the above inequality has plausibility as great as or greater than 1 — a. REFERENCES 1. Kurncuin, A. I. Mathematical Foundations of Information Theory. New York: Dover Publications, Inc., 1957. 2. Turner, M.E., Monroe, R. J., and Lucas, H. L. Generalized asymptotic regression and nonlinear path analysis, Biometrics, 17: 120-143, 1961. Cf. especially pp. 124-128. Emory University Atlanta, Georgia PRINTED IN BRUGES, BELGIUM, BY THE ST CATHERINE PRESS LTD.