Pk Tests With Repeated Efforts On Prerecorded Targets

ABSTRACT: A study was conducted with the aim of deciding how two different PK efforts on the same prerecorded target would compare with a single PK effort. In part of the runs, the two PK efforts were aimed in the same direction, and in another part, in opposite directions. Unfortunately we could not detect any significant PK effect, so no conclusions with regard to the underlying question could be drawn.

Central to the interpretation of PK effects on prerecorded random events (Schmidt, 1976) is the question of what happens when consecutive PK efforts are made on the same prerecorded event. Can only the first effort have an effect, making the result immune to further PK efforts, or do subsequent PK efforts compound, so that even the experimenter who checks the data, or perhaps even the reader of the final report in a journal, has an effect on the outcome? The truth may lie somewhere in between. A possible experimenter effect on the outcome of psi experiments has long been realized, but most researchers are hesitant to consider a retroactive PK effect exerted by the readers of the published report.

Some theoretical understanding of PK effects on prerecorded events and the possible effects of future observers was attempted by what is now called "observational theories" (Houtkooper, J. H., 1983; Schmidt, H., 1975, 1982), and a more recent attempt has even been published in the respected Physical Review (Stapp, H., 1994), but we are still far from a satisfactory theory.

The present study compares PK action on prerecorded events under three conditions:

1. Where a single PK effort is made on the prerecorded events

2. Where two subsequent efforts are made

3. Where one PK effort is followed by a second effort in the opposite direction.

The question is whether a first PK effort closes the situation so that a second PK effort has no further effect, or whether a second effort can still strengthen the effect in Case 2 and reduce the effect in Case 3.

We were aware that, due to experimenter effects, no single study might conclusively answer the question, but we hoped to contribute to a final solution that might result from several such efforts.

In this study, however, we could not find any significant PK effect so that no such useful contribution could be made. Therefore, here we will only briefly sketch out procedures and results.

THE PK TARGET AND THE REPEATED PRESENTATION OF SOME RUNS

The Basic Task

The basic task was always the same: to obtain high numbers in repeated "throws of a 16-sided electronic die." One test run contained 32 such "die throws;" that is, for each run a sequence of 32 random numbers in the range from 0 to 15 was generated, and the score was defined as the sum of these numbers minus the chance expectation value of 260. Thus the subject's aim was always a high positive score.

Repeated Presentation of Some Runs

The experiments were structured into sets of 120 runs. Each set was based on a smaller subset of 80 independent runs, with each run represented by a sequence of 32 random numbers in the range from 0 to 15. From this, a full set of 120 runs was formed by combining (a) the first 40 runs of the subset (Type A runs), (b) two copies of each of the next 20 runs of the subset (Type B runs), and (c) one copy of the last 20 runs and one copy of these runs in inverted form (Type C runs).

Here an inverted run is obtained by replacing each of its random numbers n by (15 - n).

A PK effort towards high scores made on each of the 120 runs, was then equivalent to (a) a single PK effort on the first 40 independent runs, (b) a double PK effort on the next 20 independent runs, and (c) two PK efforts in opposite directions on the last 20 independent runs.

Shuffling of the Runs before Presentation to the Subject

One would want to shuffle the runs so that a subject might not guess the type of a current run and so that, for example, not all C-type runs are at the end. If one would simply randomly phuffle the order of the 120 runs, one would end up with a set in which the runs with the second PK effort are over-represented towards the end. Then already the often-observed decline effect from the beginning to the end of a set would make the second effort appear less effective. In an attempt to avoid such a possible artifact, we performed the shuffling within smaller units of only 12 runs. For the first unit, for example, we took the first four runs from Group A, added the first two runs from Group B twice, and added the first two runs from Group C once in original and once in inverted form. Then we randomly shuffled the order of these 12 runs with the restriction that an inverted run should never appear before the original run.

THE RANDOM NUMBER GENERATION

For a set of 80 independent runs, a sequence for 80 runs (2560 random numbers in the range from 0 to 15) was prepared with the help of a random number generator based on radioactive decays. As a safeguard against imaginable (but not observed) malfunctions of this random generator, the initially generated numbers were combined (via the logic XOR operation) with quasirandom numbers to provide the final random numbers.

THREE TEST SETTINGS AND THEIR RESULTS

The study used three different test settings, with different test machines and externally different tasks. In all cases, the machines were controlled by a microprocessor (DS5000 from Dallas Semiconductor) that had stored in its memory the random number sequence for the 120 runs of one section; the internal goal was always a high positive run score. The subjects were selected and tested by K. D., with the exception of a few sections in which H. S. acted as a subject. Part of the subjects were fully aware that the experiment was based on prerecorded random events, but this did not appear as an obstacle to their efforts to get good scores. Let us now briefly discuss the test arrangements and results for the three settings.

1. The Multi-PK-Test

The subject faced a test machine with a circle of 15 lamps and a speaker for audio feedback. The subject could freely choose from many feedback options with the light running in different directions or swinging with different amplitudes. The feedback sound could also be selected, ranging from up- or down-going tone scales to simple high or low tones as the target. In one case, for example, the machine displayed each of the 32 random numbers in a run by circular light motion: For a random number n the light moved n -- 7 steps in clockwise direction for n [greater than] 7 and 8 -- n steps in counterclockwise direction for n [less than] 8. The subject aimed at more clockwise than counterclockwise motion, and the success rate was indicated by the display of a score at the end of a run. Depending on the subject's choice for the operating speed, a run lasted between 30 seconds and 1 minute.

With several test machines available, the subjects could work alone at home. But the large majority of the runs were done in the presence of K. D. The sessions were mostly kept brief, often containing only a few runs. After some initial exploration, the work focused on 42 subjects: 28 Californian college students with particular interest in art and music, and 14 students of a Parapsychology Summer Study Program offered at the

Rhine Institute in Durham. For testing a possible correlation between PK performance and psychological measures, the first group was evaluated by an Emotional Stroop Test, and the second group by the NEO Personal Inventory.

A total of seven sections were completed. After the first three sections had provided positive results, we decided to complete four more sections as a replication. The main results listed in Table 1 do not suggest the presence of any PK effects. The decline from the positive first three scores to the negative last four scores, however, might perhaps be interpreted by a change in attitude from a more playful approach to a situation where it seemed vital to get good scores. The correlation coefficients between psychological measures and PK performance turned out to be insignificant. We looked also at the results from smaller units of the total tests: Taking the run as unit, we found no significant correlation between the scores from the first half and the second half of each run. Also the run score variances (for Cases A, B, C) were not significantly different from their expectation values. Similarly, considering 12-run units, we found no significant correlations between the scores from the runs of Type A, B, and C.

2. The Meditation-PK-Test

After the Multi-PK-Tests had produced no appreciable PK effects, we changed to a different setting. The participants were meditators, and their task was to listen attentively to tones that were heard at random time intervals at an average rate of 60 tones per 20-minute session. Previous experiments with nonmeditators had indicated that such expectant attention to random signals might shorten the average waiting periods between signals. Each participant contributed one section, which required about 8 sessions.

In order to derive random time intervals of approximately exponential distribution (like the signals from a Geiger Counter exposed to weak radiation), we divided the initial sequence of random numbers into blocks of 8 numbers and then formed the sum s of the 8 numbers in each block. Each of these sums s was to provide one random interval. To this end we transformed the binomial distribution of s by an appropriate monotonically declining function F(s) into a time interval t = F(s) so that t had--in good approximation--the exponential distribution p(t) = exp(-t/to), with to = 20 sec.

In this manner, larger s-values generated shorter waiting intervals so that a decrease of the waiting intervals corresponded generally to larger s-scores.

One section of 32 x 120 = 3840 random numbers provided a total of 480 waiting intervals with each interval terminated by a gong tone. Each session was automatically terminated after 20 minutes, that is, after an average of 60 gong signals. While the lengths of the intervals provided, in principle, information on the score of each run, these scores were not displayed explicitly.

As an alternate form of feedback, a vibrator could be connected that briefly activated for each signal. It was suggested that the vibrator might be placed in the solar plexus area as a good focal point for meditation. Only few sections (marked by "V" in Table 2) used this form of feedback.

Table 2 gives the total results for each of the 20 sections in the same format as used for Table 1.

It is seen that the total scores obtained under the conditions A, B, and C in Table 1 are not significant.

Considering the possibility that individual subjects had either a tendency to shorten or perhaps to extend the time intervals, we looked first at the correlation between the scores obtained under single or double efforts listed in columns A and B which we would expect to tend in the same direction. The corresponding correlation coefficient between A and B, [gama](A,B) = 0.311, is not statistically significant (z = 1.4).

We also looked at a possibly increased variance of the scores in column A and B. Both these variances, however, lie even below their theoretical chance expectation values.

3. The Click-PK-Test

In our last attempt to observe PK effects, the subjects were instructed to close their eyes and to listen through headphones to two types of clicks that appeared in random sequence. For example, the subject could select a mixture of clicks that appeared either to originate in the center of the head or more on the sides. By setting a switch, the subjects could select the center clicks or side clicks as the target. This setting encouraged an attitude where the subject did not try to reach Out to affect some external event, but could focus on something happening apparently inside the head.

In order to transform the original random numbers n (range 0,.., 15) into binary events, we translated a number n into a sequence of (n-7) 1's for n[greater than] 7, and into a sequence of (8 - n) 0's for n [less than] 8.

Then the PK task of obtaining large numbers n was equivalent to having many l's and few 0's in the binary sequence. This original sequence was far from random because the 1's and 0's had the tendency to cluster in strings of equal bits. To make the sequence appear more random, the bits in the sequence were shuffled by a simple quasirandom algorithm so that the subject would no longer detect a tendency for strings of equal numbers. Most of the subjects had participated in the previous experiments. After a personal introduction to the test by K. D., the subjects worked at home at a leisurely pace.

Table 3 gives the results in the same format as used in the previous tables. The scores from the 10 subjects, each of which completed one section, show a slight missing tendency with z = -2.11 for the blocks with single effort and z = -0.47 for the blocks with two efforts in the same direction. The combined z-value of z = -1.99 has only suggestive value. There was no significance in the correlation coefficients between the conditions A and B.

CONCLUSION

The goal of this work was to study how two successive PK efforts on the same prerecorded events contribute to the outcome. A clear answer to this question might distinguish between different theoretical psi models, tell us more about the space time independence of psi, and touch on practical questions about the influence of the experimenter who checks the results of psi experiments. Previous work in this direction (Schmidt, 1985) was only suggestive. The absence of any PK effects in the main part of the present study does not allow us even to give a suggestive answer.

The failure to get results was not restricted to the second author (K. D.) who conducted most of the experiments, but also to the performance of the first author as a subject. We can only speculate on the reasons for this failure of previously successful experimenters, because we know too little about the psychodynamics of the interaction between experimenters and subjects that produces psi effects. One could well imagine that initial scores that happened to be low discouraged the experimenters and thereby set the psychological conditions for failure.

The further study of repeated PK efforts seems of central importance. A clear answer will require work from several groups because one needs not only significant PK effects, but also wants researchers with different mental sets such as to overcome the "unreasonable" differential effects which experimenter expectations might induce. Special attention should be given to how the first subject receives feedback. If the subject receives full numeric feedback and remembers the results, this might reduce the effect of a subsequent PK effort much more than the case where the first subject receives only qualitative feedback (like with our random waiting intervals) or if the subject immediately forgets the feedback results.

This work was supported by the Institut fur Grenzgebiete der Psychologie und Psychohygiene. We also want to thank Marilyn Schlitz from the Institute of Noetic Sciences for her help in finding test subjects.

REFERENCES

HOUTKOOPER, J. M. (1983). Observational theory: A research program for paranormal phenomena. Lisse, The Netherlands, Swets & Zeitlinger.

SCHMIDT, H. (1975). Toward a mathematical theory of psi. Journal of the American Society for Psychical Research 69, 302-319.

SCHMIDT, H. (1976). PK elfect on prerecorded targets. Journal of the American Society for Psychical Research, 70, 267-291.

SCHMIDT, H. (1982). Collapse of the state vector and psychokinetic effect. Foundations of Physics, 12, 565-581.

STAPP, H. P. (1994) Theoretical model for a purported empirical violation of quantum theory. Physical Review A, 50, 18-22.

                       RESULTS OF THE MULTI-PK-TEST
                                              Z-values
           Score Tallies                     for (A+B)
Type                   A        B       C
(Sigma:              165      117    117)
Section:
         1           336       71      13         2.01
         2            77        0      19         0.38
         3            19       37     242         0.28
         4          -225      -55    -164        -1.39
         5          -157      125       7        -0.16
         6            10      -79     124        -0.34
         7            64     -187      64        -0.61
Total:               124      -88     305
z=                     0.29    -0.29    1.00      0.06

Note: This table shows results from seven sections, where the first three section were considered as a pilot study and the last four as an attempt to confirm the positive scoring. The table gives for each section (A) the total score on the 40 runs with single effort, (B) the total score on the 20 runs with two efforts, and (C) the total score on the 20 runs with opposite efforts. In the last case the score is the one obtained in the first efforts.

The last column lists the resulting z-value when the runs of A and B (for which we expected positive scores) are combined. Sigma gives the standard deviation for each of the seven entries in the corresponding column. The total z-values in the last row are insignificant in all cases A, B, and C.

                     RESULTS OF THE MEDITATION-PK-TEST
                                     Z-values
                                     for
        Score Tallies                (A+B)      Meditation Method
          A            B     C
(Sigma:  165          117    117)
Subject
MS        82          248      56     1.63      Ananda
MS      -171          -47     118    -1.08      Ananda
HS        70          156     109     1.12    V Autogenic Training
HS         4          108      58     0.55    V Autogenic Training
HS       243          189      -6     2.14    V Autogenic Training
LM       208           35     114     1.20      Kriya Yoga
MM        73          116    -179     0.94      Sufi
MM       113          -81      21     0.16      Sufi
KT       -11           65      60     0.27      Sufi and TM
RO      -173          -60    -316    -1.15      Tibetan
BS      -307           35     241    -1.35      Transcendental
SW      -162          -81      73    -1.20      Vipassana
TP      -136           12    -124    -0.61      Vipassana
JE      -183          108     -69    -0.37      Vipassana
BD       145           -7     155     0.68      Vipassana
DB        -4           -1     -43    -0.02      Visualization
GL        91           91      -8     0.90      Witness
LS       114           11     -15     0.62    V Zen
                      -16
DL        37            4     -71    -0.63      Zen
DZ      -181            9     -60    -0.85      Zen
Total   -148          742     114
z =       -0.2          1.42    0.22  0.66

Note. This table shows results from the 20 sections, with each subject completing a whole section or even several sections. H. S. denotes the first author. A "V" indicates the use of vibrator feedback. The four sections from the two subjects who selected this form of feedback gave a total of z = 2.21. Thus further work with subject who like this feedback might be interesting.

                       RESULTS OF THE CLICK-PK-TEST
                                       Z-values
                                         for
         Score Tallies                  (A+B)
               A         B      C
(Sigma:       165       117    117)
Subject:
JU           -138      -126      22     -1.31
LM            -89       -83     -69     -0.85
MS           -146       122     118     -0.12
JG           -177        -2     102     -0.89
BF            -35      -135      57     -0.84
AV            -82       -65     -33     -0.73
SS           -270      -101      27     -1.84
RP            -59       110    -100      0.25
DB           -240        90     148     -0.74
RN            137        16     -43      0.76
Total       -1099      -174     229
z =            -2.11     -0.47    0.62  -1.99

COPYRIGHT 1999 Parapsychology Press
COPYRIGHT 2004 Gale Group