A NEW TECHNIQUE OF TESTING ESP IN A REAL-LIFE, HIGH-MOTIVATIONAL CONTEXT


By MARTIN JOHNSON


ABSTRACT: A novel, non-laboratory approach was used in three experiments designed to measure the degree to which ESP was used by well-motivated subjects engaged in a cognitive task. University students took a psychology test composed of eight questions. The examination sheets, with spaces for the answers, were attached to the front and back of a large envelope. Inside, underneath the appropriate spaces, were answers to four of the questions (randomly chosen). The subjeets were unaware that the information was there or that this was an ESP test.

In experiments 1 and 2 (the pilots), the information relating to the target questions was correct. In confirmatory experiment 3, it was incorrect, though relevant, and was accompanied by discouraging remarks. The nontarget questions in this experiment were accompanied by encouraging remarks. The results of experiment 3, which were rated independently by two judges, gave highly significant support (P = .005) to the hypothesis that the subjects' scores on the questions provided with incorrect answers and negative psychological reinforcement would be lower than scores on answers to the nontarget questions, which were provided with positive psychological reinforcement. The inter-rater reliability was also significant for this experiment (P << .01).




It is well realized that laboratory testing of ESP, apart from the advantages it offers, suffers from certain serious limitations. The restrictive conditions of the experimental situation are not overly conducive to an optimum and consistent level of motivation in the ESP subjects. This is probably because in the laboratory set-up, the ESP task, after a few initial runs, becomes devoid of personal significance to the subject, who tends to lose interest quickly. Previous studies of the application of psi in practical life situations that retain spontaneity have seemed to be efforts at overcorning such inherent limitations, of the laboratory (Brier & Tyminski, 1970a and 1970b; Cadoret, 1955; Hallett, unpublished).

The present study was one more such attempt at measuring the extent to which a person can benefit from extrasensory sources when he finds himself in a compelling real-life situation that calls for a high degree of motivation. This new approach afforded certain advantages. For one thing, the subjects were completely unaware of the fact that they were being tested for ESP; for another, the relatively "natural" setting of the experiment prevented any existing response preferences of the subject from influencing his potential responses and precluded the possible development of such preference patterns during testing. (The same, of course, could not be said of a comparable laboratory test.) Thus, the present experiment succeeded in avoiding any nonrandom behavior in the subject which might have had a nullifying influence on psi, particularly if the psi happened to be "weak".


METHOD AND PROCEDURE
The present study had two broad aspects, cognitive and extrasensory. The cognitive part was represented by a written university examination and the extrasensory part by a clairvoyance test held concurrently without the subject's knowledge. The two parts were interwoven for the purposes of this study in a manner described in the succeeding paragraphs. Three experiments were conducted: the first two were similar and were pilot attempts; experiment 3 was confirmatory.

The written examination was specially designed so that there were hidden targets consisting of typed information which (in experiments 1 and 2 contained correct answers to the questions in the examination. In experiment 3 they were misleading answers, although they were relevant to the examination questions. These target details are explained further on in the paper.

Special "test sets" were prepared for the purpose of the written examination. They contained the following material. One item in each set was a large opaque envelope with two sheets of paper pasted on either side of it. The questions selected for the written examination were typed on these sheets beforehand. Out of the total eight questions, the first four were typed in serial order on the sheet which was pasted on the front of the envelope and the questions numbered from 5 to 8 were typed serially on the sheet pasted on the back. Space for a relatively brief answer was provided below each question.

In addition to the above arrangement, for each test set a piece of cardboard of a size slightly smaller than that of the envelope was provided. The purpose of this cardboard was to have the target sheets pasted on either side of it before being wrapped in aluminum foil and enclosed in the envelope and sealed.

The preparation of the target sheets varied from experiment to experiment, being the same for the first two experiments and modified for experiment 3. For experiments 1 and 2 they were prepared thus: four examination questions out of the total eight for each set were chosen randomly, with the help of random number tables, by two assistant experimenters. These four questions were provided with target material consisting of appropriate answers accessible to clairvoyance. The answers, typed on strips of paper, were pasted under the appropriate questions on mimeographed copies of the exam. The target sheet for the other side of the cardboard for each set was prepared in a similar manner. The mimeographed sheets with the pastedon strips were then reproduced by a photocopier in order to secure a uniform surface free of sensory cues. Then they were pasted onto the cardboard, wrapped in foil, inserted into the envelopes, and sealed.

The sheets were arranged in such a way that once the target sheet was pasted onto the cardboard and enclosed in the envelope, the target answers inside occupied spaces that were exactly beneath the spaces on the outer sheets on which the subject was to write his answers. It should be noted that the location of the four targets varied from test set to test set, and was different from subject to subject, having been assigned randomly as described before. This arrangement of providing targets for only half of the total number of questions in the examination allowed each subject to serve as his own control.

While in the first two experiments the concealed targets contained relevant and correct answers to the examination questions they were associated with, in experiment 3 more than one type of target was used. These may be called primary targets (which had a negative function) and secondary targets (which had either a negative or positive function depending on whether they were associated with a primary target or one of the nontarget examination questions).

The primary targets were those lines of typed information on the concealed target sheets which, by psi, should provide relevant but incorrect answers to the questions they were associated with. Thus they were aimed at producing confusion and conflict in the subject and had a negative function. The secondary targets, on the other hand, contained information which had no relevance to the examination questions proper, but were brief remarks aimed either at giving psychological encouragement or discouragement to the subject. Whenever a secondary target was paired with a primary target, it contained such a demoralizing remark as, "You are too stupid to pass this examination," or a remark aimed at reinforcing the negative influence of the primary target. However, when a secondary target appeared alone, associated with a nontarget question, it had a positive function. In such a case, the secondary target was aimed at encouraging the subject psychologically, as, for instance, "You will certainly pass this exam."

Written instructions, which were similar for all the three experiments, were given to the subjects before the start of the examination to explain the use of the special test material. This was done in an attempt to stem the excessive curiosity which the unusual test sets might arouse in the subjects. They were informed that a new method of testing that would facilitate speedy evaluation of the students' performance was being tried out, for which purpose the special tests were introduced. The students were also told not to open any of the test-set envelopes.

All the three written examinations for the three experiments in this study formed part of a psychology course in "clinical interview," and the questions pertained to the topic. The three experiments were carried out at the department of psychology in Lund University, Sweden, where the author was in charge of the course. Experiment 1 was carried out in the spring semester of 1971; experiment 2 in the fall of 1971; and experiment 3 in the spring of 1972.


Table 1
Inter-rater reliability in experiments 1 and 3
  Experiment 1 Experiment 3
  Targets    rs = .29
  (n.s.)		        rs = .90
       t = 11.86; P<< .01
  Nontargets   rs .43
  (n.s.) 		       rs = .84
       t = 8.99; P<<.01


Note. Since only one rater was present in experiment 2, no reliability coefficients could be worked out.


SCORING
A special manual for scoring was prepared in the case of each experiment in order to have a reliable scoring procedure. A seven point scale ranging between 0 and 3 was adopted to rate the answers to the examination for the first two experiments; and a five-point scale with limits of 0 and 2 was used for the third examination.

In experiment 1 the answers were rated by the author and a corater. The help of a co-rater was sought with a view to obtaining a measure of the degree of intersubjectivity of the ratings. In experiment 2 the author was the only rater.

Experiment 3, which was to some degree considered confirmatory in purpose, coming after the first two pilot experiments, was provided with a more elaborate scoring procedure in order to enhance the reliability of the method. A more detailed manual was provided. The ratings made independently by the author in Utrecht and the co-rater (B.N.) in Sweden (to whom copies of the outer sheets of the envelope of every answered test set were sent along with a copy of the scoring manual soon after the completion of the examination) were mailed to a third person (P.A.V.). The envelopes of test sets, which were still sealed, were opened by P.A.V., who made a list after identifying targets and nontargets within each test set.



Table 2
Summary of results in all three experiments
  Experiment 1 Experiment 2 Experiment 3
N    37    35    35
Average ratings on targets for each subject 8.72 8.57 4.42
Average ratings on nontargets for each subject 8.00 7.87 5.55
Z a 2.01 b 1.70 c 2.57 b
P .022 .045 .005


a Z values in experiment 3 for the chief rater and co-rater worked out separately are 2.22 and 2.32, respectively.
b Z based on sign test.
c Z based on Wilcoxon matched-pairs test.


RESULTS
As mentioned before, in each experiment there were equal numbers of questions that were targets and nontargets for each subject. The subjects' answers were scored blindly by the raters. The author was the chief rater in all the experiments, and a co-rater was present in the first and third experiments. The inter-rater reliability was worked out, and the correlation coefficients are given in Table 1.

In the first experiment, out of 37 subjects, only 15 were selected for the co-rater's ratings. The correlation between the ratings of the two raters for targets was .29, which is not significant. Similarly, for the nontargets, the reliability coefficient was positive, but not significant (rs = .43). On the other hand, the agreement between the judges in experiment 3 was very high. In this experiment, all 35 subjects were rated by both judges. The obtained correlation for targets was .90, which gives a t of 11.86, significant well beyond the .01 level. On the nontargets, too, the correlation was highly significant (rs = .84; t = 8.99, P < < .01 ).

The experimental hypothesis to be tested in the first two experiments was the following: that the ratings on target questions should be higher than those on the nontarget questions. In the third experiment, the hypothesis was that the ratings should be lower on primary target questions than on questions not provided with primary targets. To test these hypotheses, the sign test was used in the first and third experiments and the Wilcoxon matched-pairs test in the second. The results thus obtained are given in Table 2.

In experiment 1 the sign test was applied separately for each rater's ratings between targets and nontargets. However, only the chief rater's ratings showed a significant difference between the two sets of targets. The obtained Z value was 2.01, which has a P of .022. The average rating for each subject on targets was 8.72; for nontargets, it was 8.00, thus confirming the first hypothesis.

In the second experiment there was only one rater, and his ratings on the average for targets were higher than for nontargets (8.57 and 7.87, respectively). The difference between the two sets yields a Z of 1.70 with P= .045. These results also support the first hypothesis.

In the third experiment, the average rating of the two raters for targets was 4.42, and for nontargets, 5.55. The difference tested by Wilcoxon's test gave a Z of 2.57, which is significant at the .005 level. A similar trend was seen when the data were broken down for the two raters separately. The Z value for the chief rater was found to be 2.22, and for the co-rater, 2.32, both significant beyond the .05 level. These results support the hypothesis that the targets further reinforced by negative remarks would obtain lower ratings than those coupled with positive reinforcement.


DISCUSSION
The results from this study can be considered promising. Even though the outcome of experiment 2 was not significant, it was still in the expected direction. Experiment 3, provided with stricter procedures, had an extended hypothesis that was well-substantiated. The results of this confirmatory experiment may be considered strong enough to warrant their replication and extended application. Such newer classroom techniques as programmed instruction, for instance, might be found suitable as a means of effectively camouflaging an ESP test on the lines of the present one. As the student learns progressively he has to anticipate the next piece of information, thus providing for the use of his clairvoyant abilities.

The encouraging aspect of the present design, as the author sees it, lies in the fact that it suceceds in avoiding confounding effects of such factors as subjects' attitudes toward ESP or the type of ESP task, etc. This is because subjects are not aware of the ESP part of the examination. Also, the method tends to do away with the other usual dangers such as "experimenter expectancies," artifacts resulting from subject-experimenter interaction, etc., since in this study there is no experimenter in the conventional sense.

The author is presently engaged in a search for any significant correlations that there may be between the ESP effect obtained in the present study and such factors as degree of preparedness on the students' part, performance on similar examinations, and anxiety proneness. Further extensions of this research might include studies of the relative distance of the target from the subject, the form of target, and similar variables which might be related to ESP.

Coming to the anomalies and drawbacks of this study, with particular reference to experiment 3: subjects who may be so well prepared as to have completely correct answers to all the questions probably will jeopardize the working of the method. Also, the provision of primary and secondary targets together with a question makes it difficult to distinguish which of the two created the supposed psi effect. Although these problems are interesting in themselves and may be overcome in a further study, they do exert their unwelcome influence on the present results. Another chief drawback of the method is that it is laborious and time-consuming.

Nevertheless, the findings of this novel approach should prove interesting and stimulating enough to spur innovative and repetitive efforts by other workers in the future.


A preliminary report of this work first appeared in the Research Letter of the Parapsychological Division of the Psychological Laboratory of the University of Utrecht, November 1972.

The author wishes to express his gratitude to Dr. Bertil Nordbeck for his contribution as co-rater of the results and to Dr. Piet Vroon of the Psychological Laboratory in Utrecht for his kindness in functioning as a controller.



REFERENCES

BRIER, R. M., & TYMINSKI, W. V. Psi application: Part I. A preliminary attempt. Journal of Parapsychology, 1970, 34, 1-25. (a)

BRIER, R. M., & TYMINSKI, W. V. Psi application: Part II. The majority-vote technique - analyses and observations. Journal of Parapsychology, 1970, 34, 26-36. (b)

CADORET, R. J. The reliable application of ESP. Journal of Parapsychology, 1955, 19, 203-27.

HALLETT, S. J. Proposed research on the problem of the practical application of ESP. (Unpublished manuscript in the files of the Institute for Parapsychology.)

Dept. of Psychology
Lund University
Lund, Sweden

Reprinted from The Journal of Parapsychology, vol. 37, No. 3. Sept. 1973.

Martin Johnson held a chair in parapsychology at the University in Utrecht, Holland 1974-1988.


To page 1!


www.sokaren.se/INDEX94.HTML