CIA-RDP96-00789R003100120001-4

TeApproved.For Release 2000/08/08AtA-RDP96-00789R00810012000104
Target and Sender Dependencies
in
Anomalous Cognition
Prepared by:
Edwin C. May, Ph.D. and Nevin D. Lantz
2 December 1991
SAIC
®
Science Applications International Corporation
An Employee-Owned Company
Presented to:
The Scientific Oversight Committee
Submitted by:
Science Applications International Corporation
Cognitive Sciences Laboratory
1010 El Camino Real, Suite 330
Menlo Park, California 94025
1010 El Camino Real, Suite 330, P.O. Box 1412, Menlo Park, CA 94025 · (415) 325-8292
961895 817 Pak Alta Seattle Tucson
"Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
UNEDITED DRAFT

Tel:ppicelvedtdoor: Regeasel 2000:08:08nd01ARDR06:00789R006600120001-4
DRAFT
TABLE OF CONTENTS
LIST OF FIGURES
ii
LIST OF TABLES
iii
I
OBJECTIVE
1
II
INTRODUCTION
2
III
APPROACH
4
1.
Target-pool Selection
4
4
2.
Target Perparation
3.
Target Selection
8
4.
Receiver Selection
8
5.
Sender Selection
9
6.
Session Protocol
9
7.
Analysis
10
8.
Hypotheses
10
IV
DISCUSSIONS AND CONCLUSIONS
12
1.
Null Result
12
2.
Significant Deviations
12
V
GLOSSARY
14
REFERENCES
15
APPENDIX
16
UNEDITED DRAFT
Approved For Release 2000708/08 : CIA-RDP96-00789R003100120001-4

Tcp piroMedtocor: Releasel 2000:08/08ndeHA.RDR86607898908100120001-4
DRAFT
LIST OF FIGURES
1.
City with a Mosque
5
2.
Green Intensity Distribution for the City Target (Macro-pixel, 3,3)
6
3.
City with Mosque (|AS | = 1.98 bits/0.25 in2)
6
4.
Pacific Islands (|4S | = 1.35 bits/0.25 in2)
7
5.
Zener Target Cards (Average | AS | = 0.15 bits/0.25 in2)
7
UNEDITED DRAFT
Approved For Release 2000708708 : CIA-RDP96-00789R003100120001-4

TAqbpicoverdtFoul: Ræteased2000/0 8/08nda AieR DF96.00789R963600120001-4
DRAFT
LIST OF TABLES
1. Effect Size as a Function of Target Type
3
2.
Potential Correlation of AS with Effect Size
8
3.
Experiment Conditions
9
UNEDITED DRAFT
Approved For Release 2000708708 : CIA-RDP96-00789R003100120001-4

TAURICAMS Atb&o: Tarter and Sender Dependencies In Anomalous Cognition
lease 2000/08/08 : CIA-RDP96-00789R003100120001-4
DRAFT
I. OBJECTIVE
There are two objectives of this pilot study:
(1) Explore the effects of target properties on AC quality.
(2) Determine the degree to which anomalous cognition (AC) quality depends upon a sender .*
* Definitions of terms can be found in Section V (i.c., Glossary) on page 14.
Approved For Release 2900/08/08 CJA-RPP96-00789R003100120001-4

TEURRESVed.For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
tocol: Target and Sender Dependencies in Anomalous Cognition
DRAFT
II. INTRODUCTION
The field of parapsychology has been interested in improving the quality of responses to target material
since the 1930's when J. B. Rhine first began systematic laboratory studies of extra sensory perception.
Since that time, much of the field's effort has been oriented toward psychological factors that may influ-
ence anomalous cognition (AC). In this section, we review the pertinent literature that describes at-
tempts to improve the quality of AC by categorizing target content.
At a recent conference, Delanoy reported on a survey of the literature for successful AC experiments,
and, she categorized the target material according to perceptual, psychological and physical character-
istics.1* Except for trends related to dynamic, multi-sensory targets, she was unable to observe system-
atic correlations of AC quality with her target categories.
Watt examined the AC-target question from a theoretical perspective.2 She concluded that the "best"
AC targets are those that are meaningful, have emotional impact, and contain human interest; those
targets that have physical features that stand out from their backgrounds or contain movement, novelty,
and incongruity are also good targets.
The difficulty with either the survey of the experimental literature or the psychologically oriented
theoretical approach is that understanding the sources of the variation in AC quality is problematical.
Using a vision analogy, sources of visual material are easily understood (i.e., photons); yet, the percept
of vision is not well understood. Psychological and possibly physiological factors influence what we
"see." In AC research, the same difficulty arises. Until we understand the influence of these factors on
the AC percept, results of systematic studies of AC are difficult to interpret.
Yet, in a few cases, some progress has been realized. In 1990, Honorton et al. conducted a careful meta-
analysis of the experimental Ganzfeld literature.3 In Gansfeld experiments, receivers are placed in a
state of mild sensory isolation and asked to describe their mental imagery. After each trial, the analysis
was performed by the receiver, who was asked to rank order four pre-defined targets, which include the
actual target and three decoys; the chance first-place rank hitting rate was 0.25. In 355 trails collected
from 241 different receivers, Honorton et al. found a hitting rate of 0.31 (z = 3.89, p ≤5 x 10-5) for an
effect size of 0.20. In addition, he found that AC quality was significantly enhanced when the targets
were video clips from popular movies (i.e., dynamic) as opposed to static photographs (i.e., effect sizes
of 0.32 and 0.05, respectively). All trials were conduced with a sender.
In a carefully conducted meta-analysis, Honorton and Ferrari report significant hitting in forced-
choice, precognition experiments.41 They analyzed 53 years of experiments conducted by 62 different
investigators using a limited set of symbols (i.e., called Zener cards) as target material. Fifty thousand
* References may be found at the end of the document.
t Forced-choice means targets are randomly chosen from a known and limited set of possibilities (e.g., red or black playing
cards). Precognition means that the target is generated randomly after the guess has been registered.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
UNEDITED DRAFT

DRAFT
subjects contributed a total of approximately 2 × 106 individual trials. The overall effect size was 0.020
corresponding to a p-value of 6.3×10-25. Similarly, in an earlier review article, Honorton analyzed
7.5×105 forced-choice Zener card trials that were collected from 1934 to 1939 and found a significant
overall effect size of 0.016±0.001.5
Puthoff and Targ publish the results of 39 AC real-time trials where the targets were natural scenes in
the San Francisco Bay area.6 The effect size for the 39 trials was 1.15.
Table 1 summarizes these results for each target type:
Table 1.
Effect Size as a Function of Target Type
Target Type
Trials
Effect Size
Symbols (Real-Time)
7.5 ×105
0.016 ± 0.001
Symbols (Precognitive)
2.0 × 106
0.020 ± 0.001
Static Photographs
165
0.05 ± 0.08
Dynamic Photographs
190
0.32 ± 0.07
Static Natural Scenes
39
1.15 ±0.16
The effect sizes shown in Table 1 are qualitatively monotonically related to target "complexity;" yet an
appropriate quantitative description for target type is currently unknown. Yet, target "complexity" was
one of the experimentally observed and theoretically conceived target concepts found by Delanoy and
Watt, respectively.
A number of confounds exist, however, in this database for the effect-size measures. For example, in all
but the Puthoff and Targ study (i.e., targets were natural scenes), the receivers were unselected. That is,
they did not participate in the various experiments on the basis of their known ability as receivers. So, is
the large effect size for the Puthoff and Targ study because of the accomplished receivers, the natural-
scene targets, or some combination of both? While there are a number of other exceptions, the prepon-
derance of the data were from unselected individuals. In many of the trials, a sender was concentrating
on the target material, and as in most perception experiments, psychological factors and boredom con-
tribute to the variance in the effect sizes.
In this pilot experiment, we will apply one physical measure to static and dynamic photographs to quan-
tify the relationship between target type and AC quality. By careful selection of target content, we will
minimize the psychological factors in perception. In addition, we will minimize individual differences
by conducting many trials with each receiver and by only choosing receivers who have previously dem-
onstrated excellent AC skill.
Because the previous database included trials with and without senders, we will explore the effects of a
sender on AC quality, as well.
Approved For Release 2000
08/08 : CJA-RDP96-00789R003100120001-4
INEDITED DRAFT

TeARKAved Far-Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
et and Sender Dependencies In Anomalous Cognition
DRAFT
III. APPROACH
1. Target-pool Selection
The static target material for this pilot study will be an existing set of 100 National Geographic magazine
photographs. This set has been divided into 20 sets of five photographs that were determined to be
visually dissimilar by a fuzzy set analysis.7 The dynamic target material will be approximately 50, 60 to
90 second clips from popular video movies. These clips will be selected because they:
. are thematically coherent,
· contain obvious geometric elements (e.g., wings of air craft), and
· are emotionally neutral.
The intent of these selection criteria is to control for cognitive surprise, to provide target elements that
are easily sketched, and to control for psychological factors such as perceptual defensiveness.
The video segments will be drawn from a variety of themes including adventure, documentary, and fantasy.
2. Target Preparation
The target variable that we will consider in this experiment is the total change of entropy per unit area,
per unit time. We have chosen this quantity because it is qualitatively related to the "complexity" of
target type shown in Table 1, and because it represents a potential physical variable that is important in
the detection of traditional sensory stimuli. In the case of image data, the entropy is defined as:
N1 -1
j=0
Sk = -
E Piklo82(Pin).
where Pjk is the probability of finding image intensity j of color k. In a standard, digitized, true color
image, each pixel (i.e., picture element) contains eight binary bits of red, green, and blue intensity, re-
spectively. That is, Nk is 256 (i.e., 28) for each k, k = r, g, b. The total change of the entropy in differential
form is given by:
dS+ = VSk . dr + Sr dt .
(1)
We must specify the spatial and temporal resolution before we can compute the total change of entropy
for a real image. Henceforth, we drop the color index, k, and assume that all quantities are computed
for each color and summed.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
UNEDITED DRAFT
4

Tennedvedcor: Release 3000108698 -CIA-RPR96-007895003100120001-4
Cognition
DRAFT
2.1 Static Photographs
Each target from the pool of 100 National Geographic magazine photographs will be scanned at 100 dots per
inch (dpi) for eight bits of information of red, green, and blue intensity. At 0.25 inch spatial resolution, for
example, this scanning density provides 625 pixels for each 0.25 x 0.25 in2 patch to compute the pj.
For a specified resolution, the target photograph is divided into an integral number of macro-pixels ex-
cluding a thin border, if necessary. The entropy for the (i,j) macro-pixel is computed as:
S ;= - = p; log2(p),
N-1
j=0
where pj is computed empirically from the pixels in the (i, j) macro-pixel only. For example, consider the
target photograph shown in Figure 1.
Figure 1. City with a Mosque
Figure 2 shows the probability density for green macro-pixel (3,3), which is shown as a white square in
the upper left hand corner of Figure 1 .* The probability density and the photograph indicates that most
of the intensity in this patch is near zero value (i.e., no intensity of green in this case). In a similar fash-
ion, Sij are calculated for the entire scene. For the photograph shown in Figure 1, i ranges from zero to
43, and j ranges from zero to 32 for a total of 1,452 macro-pixels.
* The original photograph was 8.5 x 11 inches, and we have standardized on 0.25 inch resolution.
Approved For Release 2000/08/08 - CIA-RDP96-00789R003100120001-4
UNEDITED DRAFT

DRAFT
0.4
Probability Density
0.2
0.0
=
0
20
40
60
80
100
Intensity (j)
Figure 2. Green Intensity Distribution for the City Target (Macro-pixel 3,3).
We will use a standard algorithm to compute the 2-dimensional spatial gradient of these 1,452 values of
the entropy. Figure 3 shows contours of constant change of entropy (calculated from Equation 1) for
the city target. The total change per unit area is 1.98 bits/0.25 in2 .*
2
1
Figure 3. City with Mosque (|AS| = 1.98 bits/0.25 in2).
* In this formalism, entropy is in units of bits and the maximum entropy is 24 bits.
Approved For Release 2PAPERTED DRAFT
08/08 : CIA-RDP96-00789R003100120001-4

TARREAved cor: Rele
SAsifa Sender BeBendencies in Anomalous Cognition
DRAFT
108-CJA-RDP96-00789R003100120001-4
The city target was chosen as an example because it was known (qualitatively) to be a "good" static
photograph for AC trials in earlier research. Figure 4 shows contours of constant change of entropy for
a photograph that was known not to be a "good" AC target.
Figure 4. Pacific Islands (|AS | = 1.35 bits/0.25 in2).
For comparison, we show in, Figure 5, the traditional Zener card set, which was used in most of the
forced-choice experiments shown in Table 1 and described above.
O
Figure 5. Zener Target Cards (Average | AS | = 0.15 bits/0.25 in2).
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
UNEDITED DRAFT
7

TeApprovedfor Release 2000/08/98naGIA-RPRA6-A07898993100120001-4
DRAFT
In Table 2, we modify Table 1 to show the values of AS (0.25 in)-2 for two of four target types.
Table 2.
Potential Correlation of AS with Effect Size
Target Type
AS (0.25 in)-2
Symbols
0.15
Static Photographs
1.35
Dynamic Photographs
?
Static Natural Scenes
?
We illustrate in this table the intent of this pilot study. We will compute AS for all the static and dynamic
targets and, using accomplished receivers, measure their associated AC effect sizes.
2.2 Dynamic Photographs
The total change of entropy for the dynamic targets will be calculated in much the same way. The video
target will be digitized at approximately one frame per second. The spatial term of Equation 1 will be
computed exactly as it was for the static targets. The second term, however, will be computed from dif-
ferences between adjacent frames. Or,
AS1;(t)
Si;(t + At) - Stj(t)
At
(2)
where At is the one over the digitizing frame rate. We can see immediately that the dynamic targets will
have a larger 4S than do the static ones becuase Equation 2 is identically zero for all static targets.
2.3 Cluster Analysis
As a result of the above calculations, the static and dynamic target sets will have associated sets of 4S.
Using standard cluster analysis, each set will be grouped into relatively orthoginal clusters of relatively
constant 4S. Inspection and fuzzy set analysis will be used to construct packets of five visually dissimilar
targets from within each cluster. Since we do not yet know how to assign entropy to an AC response, the
AC analysis must be performed on the basis of visual discrimination.
3. Target Selection
For a specified target type (e.g., static photographs), a target pack will be selected randomly and one
target of the five within the that pack will also be chosen randomly.
4. Receiver Selection
Six experienced receivers, who have produced significant AC effect sizes in previous investigations, will
contribute 40 AC trials each. Each receiver will contribute ten trials in each of the conditions shown in
Table 3.
Approved For Release 290008-08 CJA-RPP96-00789R00310012000124

TAnncavedcor: Release 3900108698 -CIA-RPR96-00789R003100120001-4
DRAFT
Table 3
Experiment Conditions
Condition
Target Type
Sender
1
Static
Yes
2
Static
No
3
Dynamic
Yes
4
Dynamic
No
5. Sender Selection
The sender for all trials will be the principal investigator (PI).
6. Session Protocol
Before the pilot experiment begins, the experiment coordinator will generate, randomly, a counter bal-
anced set of 20 dynamic and 20 static targets and, within each target type, generate randomly a counter
balanced set of sender/no sender conditions. Each of the six receivers will have their own individual set
of targets/conditions. For each receiver, the experiment coordinator will prepare 40 sealed envelopes
containing the target number and condition for each trial. For the no-sender condition, the target num-
ber will be sealed in a smaller internal envelope so that the PI will remain blind to the target choice, but
in the sender condition, the target number is visible in the outer envelope. The receivers will be notified
about the dates and times of day when their individual targets are available.
For each trial and for each receiver, the PI will perform the following tasks:
· Determine from the above list, the target and sender condition.
. In the sender condition, study the selected target and attempt to "transmit" it to the intended receiver.
In the no-sender condition, do nothing
. At the conclusion of the 15 minute trial period and after the receipt of the receiver's response by fac-
simile, send a copy of the target material (i.e., either a photograph or video tape) to the receiver by
over night mail.
During each trial, the receiver will perform the following tasks:
. At a prearranged time, the receiver will find a quiet and lighted room in his or her home and sit at a desk.
. For a period lasting no longer than 15 minutes, the receiver will write and draw his or her impressions
of the intended target material, which will be located in Lititz, PA.
. At the end of the AC trial, the subject will send the response by facsimile to the principal investigator (PI).
. By overnight mail, the subject will receive a copy of the actual target as feedback for the trial.
We will not provide specific instructions beyond logistical information to the receivers, because they are
all experienced in this type of task.
For each receiver, the 40 trials will occur at a rate of three per week (i.e., one every other day) during a
five-month period beginning in January 1992. There will be significant breaks during this period for
Approved For Release 2000/08/08 - CIA-RDP96-00789R003100120001-4
UNEDITED DRAFT

DRAFT
holidays and to allow the receiver to participate in other experiments. The PI will maintain frequent
phone contact with them during the experiment.
At the end of the study, the PI will remove the receiver's name, date, and time from each response; ran-
domize the order within a receiver set; and provide an analyst with a set of responses and associated
target packs. The indented target within each pack will not be disclosed.
7. Analysis
For each trial, there is a single response and its associated target pack (i.e., either static or dynamic).
During the first part of the analysis, a judge, who is blind to the condition and target for the trial, will be
asked to rank-order the targets within the given pack. This is a forced rank, so regardless of the quality
of match between the response and targets within the pack, the judge must assign a first place match to
the response, a second place match to the response, and so on for each of the five targets. The output
from this part of the analysis is a rank-order number (i.e., one to five, one corresponding to a first place
match) for the correct target.
For each receiver, target type, and condition there are 10 such rank-order numbers that constitute a
block of data. A rank-order effect size will be computed for a block as:
Rij - Ro
V
IN2-1
12
(3)
where Rij is the average rank for target type i and sender condition j, and Ro is the expected average
rank, which for this study is equal to three for all cases. In Equation 3, N is the number of possible ranks
and is equal to five throughout this study. Thus, Equation 3 reduces to:
Rj - 3
During the second part of the analysis, a two-way analysis of variance (ANOVA) will be computed for
each receiver. The main effects are target type and sender condition.
In this part of the analysis, we do not plan to combine data across receivers.
In the third part of the analysis we will construct a scatter diagram of rank-order number versus 4S.
Using a logistic transformation on the rank-order numbers, we will compute a linear correlation coeffi-
cient to determine the degree to which AC quality linearly depends upon 4S. By inspection of the scat-
ter diagram, we will determine if higher-order correlations should be calculated.
8. Hypotheses
8.1 Null Hypothesis
The overall null hypothesis is that aj will not be significantly different from zero. Even with only 10 trials
in each condition and given that the historical effect size of many of the receivers is approximately 0.8,
there is an 80% chance of observing a significant effect size for a given block of data.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
UNEDITED DRAFT
10

R003100120001-4
DRAFT
8.2 Sender and Target Condition
Using an F-test we will test the hypothesis that the quality of AC does not depend upon a sender regard-
less of target type. Similarly, we will use an F-test to test the hypothesis that the quality of AC does not
depend upon target type regardless of the sender condition.
The interaction terms in the ANOVA will test the hypothesis that a sender might improve AC quality for
only a specific target type.
8.3 Target Entropy
The AC quality of each trial is assessed within a given target type and as closely as possible with similar
4S. Thus, a significant correlation between target 4S and AC quality will be a valid indication of the
primary hypothesis that they are linearly related.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
UNEDITED DA
RAFT

DRAFT
IV. DISCUSSIONS AND CONCLUSIONS
In this pilot investigation we will study the degree to which the change of target entropy affects the qual-
ity of anomalous cognition, and we will explore the relationship of a sender to the AC process. There
are a number of potential outcomes to this investigation and a number of post hoc analyses that could
yield productive insight. We discuss these outcomes and analyses below.
1. Null Result
At the 95% confidence level, no statistically significant deviations are observed for any of the block ef-
fect sizes, Ej. If a X2 test for homogeneity of effect sizes across receivers demonstrates that the data are
homogeneous (i.e., p(X2) > 0.05), then we conclude that the experiment failed to demonstrate signifi-
cant AC functioning. In this case we will recommend that a replication be conducted with more trials,
because there is a 20% chance the the data produced by a single receiver would not reach statistical
significance even if an alternate hypotheses was true. That is, the Type II error is 20%.
If, however, the effect sizes across receivers is not homogeneous (i.e., p(X2) ≤ 0.05), then the data for
each receiver will be examined individually. Depending upon available resources and the advice of the
SOC, the receivers who may have demonstrated individually significant results might be asked to con-
tribute additional data.
2. Significant Deviations
The are a number of different ways, in accordance with the analysis described above, that significant
deviations could be observed.
2.1 Dependency on Target Type
Suppose that the ANOVA demonstrates significant effects for the target type regardless of the sender
condition. Suppose further that we observe a significant correlation between 4S and AC quality. In this
case, we would consider that the primary hypothesis (i.e., the change of target entropy is sensed by AC)
has been confirmed. We would recommend that we extend the study to include natural scenes as target
material. To do this properly, however, we must estimate the potential change of thermodynamic entro-
py for real locations.
2.2 Dependency on Sender
If the ANOVA demonstrates significant effects in support of a sender regardless of target type and
there is no significant interaction terms between target type and sender condition, we would conclude
that a sender can significantly improve the quality of AC. Furthermore, we would conclude, as Delanoy
before us, that we still do not understand what constitutes an AC target.
Approved For Release 2000/08/08 : CJA-RDP96-00789R003100120001-4
JEDITED DRAFT

TAppcavedtFor: Release 2090608/08delAeBAR9S-AA789R9A3100120001-4
DRAFT
We would recommend, therefore, that a post hoc analysis be conducted to search for target systematics
in the existing database. If any were found, we would formulate hypotheses to be tested in later studies.
2.3 Other Post Hoc Analyses
Depending upon time and resources, we will re-analyze the AC data. Decoy targets for the blind rank-
ing would be selected not on the basis of constant 4S, but rather on a visual basis alone; this is the tradi-
tional method usually employed in AC studies. Depending upon the content of the targets, there might
be other dimensions that could be used to construct decoy targets (e.g., function, physical proximity of
target elements).
There has been some indication in the literature that AC quality depends weakly upon the noise in the
geomagnetic field. Since we routinely record the time, date, and location of each trial, we will add the
results from this experiment to that analysis.
Approved For Release 2990 88TD Di
A-RDP96-00789R003100120001-4

DRAFT
V. GLOSSARY
Not all the terms defined below are germane to the MEG study, but they are included here for com-
pleteness. In a typical anomalous mental phenomena (AMP) task, we define:
· Anomalous Cognition-A form of information transfer in which all known sensorial stimuli are ab-
sent. That is, some individuals are able to gain access, by as yet an unknown process, to information
that is not available to the known sensorial channels.
· Receiver-An individual who attempts to perceive and report information about a target.
· Agent-An individual who attempts to influence a target system.
· Target-An item that is the focus of an AMP task (e.g., person, place, thing, event).
. Target Designation-A method by which a specific target, against the backdrop of all other possible
targets, is identified to the receiver (e.g., geographical coordinates).
· Sender/Beacon-An individual who, while receiving direct sensorial stimuli from an intended target,
acts as a putative transmitter to the receiver.
· Monitor -- An individual who monitors an AC session to facilitate data collection.
· Session-A time period during which AC data is collected.
· Protocol -- A template for conducting a structured data collection session.
. Response-Material that is produced during an AC session in response to the intended target.
· Feedback-After a response has been secured, information about the intended target is displayed to
the receiver.
· Analyst-An individual who provides a quantitative measure of AC.
· Speciality-A given receiver's ability to be particularly successful with a given class of targets (e.g.,
people as opposed to buildings).
Approved For Release 20028802 GA-RPP96-00789R003100120001-4
14

DRAFT
REFERENCES
1. D. L. Delanoy, "Characteristics of Successful Free-Response Targets: Experimental Findings and
Observations," Proceedings of Presented Papers of the Parapsychological Association 31st
Annual Convention, pp. 230-246, Montreal, Canada (August 1988).
2. C. Watt, "Characteristics of Successful Free-Response Targets: Theoretical Considerations,"
Proceedings of Presented Papers of the Parapsychological Association 31st Annual Convention,
pp. 247-263, Montreal, Canada (August 1988).
3. C. Honorton, R. E. Berger, M. P. Varvoglis, M. Quant, P. Derr, E. I. Schechter, and D. C. Ferrari,
"PSI Communication in the Ganzfeld," Journal of Parapsychology, Vol. 54, pp. 99-139 (June 1990).
4. C. Honorton and D. C. Ferrari, "Future Telling:' A Meta-analysis of Forced-choice Precognition
Experiments, 1935-1987," Journal of Parapsychology, Vol. 53, pp. 282-308 (December 1989).
5. C. Honorton, "Error Some Place!" Journal of Communication, pp. 103-116, (Winter, 1975).
6. H. E. Puthoff and R. Targ, "A Perceptual Channel for Information Transfer over Kilometer
Distances: Historical Perspective and Recent Research," Proceedings of the IEEE, Vol. 64, No. 3,
pp. 329-354, (March, 1976).
7. E. C. May, J. M. Utts, B. S. Humphrey, W. L. W. Luke, T. J. Frivold, and V. V. Trask, "Advances in
Remote-Viewing Analysis," Journal of Parapsychology, Vol. 54, pp. 194-228, (September, 1990).
Approved For Release 29ROOSIPECUARPP96-00789R003100120001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Technical Protocol: Target and Sender Dependencies in Anomalous Cognition
DRAFT
APPENDIX
This appendix contains the full reprints of the following seven papers:
(1) Characteristics of Successful Free-Response Targets: Experimental Findings and Observations
(2) Characteristics of Successful Free-Response Targets: Theoretical Considerations
(3) PSI Communication in the Ganzfeld
(4) "Future Telling:" A Meta-analysis of Forced-choice Precognition Experiments, 1935-1987
(5) Error Some Place!
(6) A Perceptual Channel for Information Transfer over Kilometer Distances: Historical Perspective
and Recent Research
(7) Advances in Remote-Viewing Analysis
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
UNEDITED DRAFT

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CHARACTERISTICS OF SUCCESSFUL FREE-RESPONSE TARGETS:
EXPERIMENTAL FINDINGS AND OBSERVATIONS
Deborah L. Delanoy
Psychology Department
University of Edinburgh
7 George Square
Edinburgh EH8 9JZ
Scotland, U.K.
.
Abstract
This paper reviews experimental findings and observations concerning
characteristics of successful free-response targets. Information relevant
to the following categories of target characteristics was examined:
colour/black and white; complex/simple; novel/familiar; abstract/
concrete; dynamic/static; form/idea and meaning; emotion; and theme/
content.
Very few conclusions could be drawn from the data base,
although a tentative finding related dynamic, multi-sensory targets to ESP
success. Other suggestive findings were reported for novel and abstract
characteristics. The discussion considers possible reasons for the general
lack of findings and presents a possible avenue for future research.
ACKNOWLEDGEMENTS: Ms. Caroline Watt and Professor James Crandall
contributed substantially to the research for this paper, for which I am
most grateful. My thanks also to Dr. Julie Milton and Ms. Watt for
helpful comments on the paper's content and again to Ms. Watt for the
typing of the references.
Approved For Release 2000/08/08/ CIA-RDP96-00789R003100120001-4

CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
This and the following paper, presented by Caroline Watt,
represents the findings of a literature review examining what makes a
successful (in terms of being accurately perceived by the percpient)
and/or unsuccessful free response GESP target. The review was undertaken
to assist the Koestler Lab in constructing a free response target pool for
use in our future research. We thought such a review was necessary as
initial discussions as to what type of targets we should be looking for
revealed that various researchers in our group held differing
opinions/ideas as to what qualities a successful target should have.
These differences were further reinforced when we started discussing
various targets which we had used in our own independent research, and
those of other researchers with whose targets pools we were familiar. An
initial search through some of the major parapsychology journals and
source
books revealed very little coherently arranged information
regarding free response targets. This review was undertaken in an attempt
to remedy this situation.
To this end,
we examined relevant
parapsychological and psychological experimental findings and theoretical
models, post hoc observations, and lab lore in hopes of discovering some
consensus regarding psi-conducive target qualities and materials. This
first paper will present the findings from parapsychological experimental
findings, including post hoc findings and anecdotal observations.
It should be stressed that this review is not meant to be
exhaustive. We have tried to scout out related information in the main
journals and newsletters (Journal of the American Society for Psychical
Research, Journal of Parapsychology, Journal and Proceedings of the
Society for Psychical Research, European Journal of Parapsychology,
International Journal of Parapsychology, Parapsychology Review, and
Research Letter). We have also examined various conference proceedings
(Parapsychological Association and Parapsychology Foundation), major
parapsychological source books, some of the popular literature regarding
the development of psychic abilities, some of the related psychology
literature, and other prominent books in our field which we thought
likely to contain the information we were seeking. However, it was
obviously impossible to examine all of the possibly related literature.
Our survey of the historical literature was necessarily quite limited (in
fact we examined only two main sources, Warcollier's writings and
Phantasms of the Living, 1886).
Target-related information from forced choice studies has not been
systematically considered here, the primary reason for this omission being
the two reviews of this literature already conducted by Palmer (1978) and
Carpenter (1977).
However,
general findings from these sources
occasionally will be referred to where appropriate in this review.
The most frequent comments regarding targets found in these
sources were generalizations regarding the choice of target material. For
example comments might be made that targets were chosen which were vividly
coloured, intrinsically interesting,
pleasant, and so on. While such
comments may be viewed as conveying the experimenters' perspective of what
constitutes an easy-to-perceive target, to list all such comments would
have been a very tedious task for both the author and her audience.
Furthermore, no comment could be made upon the utility of whatever
parameters were adopted when choosing targets unless one were to attempt a
meta-analysis of the relevant studies, a project which was far beyond the
scope of the present undertaking. Thus, such comments were not included
in this review unless information was provided which related particular
target characteristics to the success or failure of the study, and/or the
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
231

target materials were specifically chosen to be unusual in some respect,
with Approved Fes Release +2000/08/08 fc|AURID BOB.00789R003100120001-4
The initial task in this undertaking required finding some way to
organize the target information in a meaningful and useful manner. This
proved to be quite problematic, as target materials and content are
seldom one-dimensional.
Thus it was required to find a means of
categorizing a diverse range of target materials, such as film clips,
actual geographical sites, agents' experience of some sensory stimulus,
and a large range of assorted pictorial material,
each representing
varying degrees of denotative and connotative complexity.
Indeed, even
defining the target in many studies
was not a straight forward
proposition. For example, in telepathic designs, is the target the
agent's experience of the target material or the target material itself?
In approaching this task it was thought that the target information
could perhaps be divided according to the type of target material used
(e.g. art prints, film clips, geographical locations, etc. ). However,
this approach was rejected as in many cases there was not enough available
information about a specific target material
to allow sensible
generalizations to be made. Also explored were various ways of trying to
represent and categorize the obtained target
information in a
multi-dimensional manner,
taking into account both denotative and
connotative meaning. To this end, attempts were made to apply to the
data various three dimensional conceptualizations of the sort obtained
from the semantic differential. Thus, we sought to find one scale
which would categorize the obtained target information taking into
consideration various connotative components such as evaluation (does the
information convey something which is good-bad, clean dirty, sacred-
profane, etc.), potency (weak-strong, powerless-powerful, light-heavy,
etc.), and activity (fast-slow, active-passive, sharp-dull, etc.).
This approach of organizing the data was rejected as there was not enough
information about most targets to justify a post hoc fitting of the
obtained information into such a model. Thus, in the end the task was
necessarily defined by the type of information obtained in the literature
search.
Looking through the data obtained,
it was decided that the
information could best be organized according to the following target
characteristics: colour / black and white; complex / simple; novel /
familiar; abstract / concrete; dynamic / static;
form / idea and
meaning; emotion; and theme / content. The "working definitions" of
these categories will be delineated in the following appropriate sections
of this paper. There were many instances where the same data fitted into
several different categorizations. For instance, in Krippner, Ullman, et
al. (1972) the target consisted of a randomly chosen word, an art print
which portrayed the word, and then a multi-sensory (auditory, gustatory,
olfactory, tactile and kinesthetic) environment relating to the
word/picture was created for the agent. Such a target could easily be
classified as complex, novel, dynamic, emotional, and as having a strong
theme. In such situations, the author has attempted to refer to the
information in all the relevant categories, but has only provided
details of the study in the category where it was first mentioned.
Colour / Black and White
The colour category referred to all target materials which were
coloured, as opposed to black and white. A telepathic dream study by
Krippner and Zeichner (1974) obtained a significant degree (p < . 002) of
psi-hitting using 74 art prints as the targets. A descriptive analysis of
232
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGH

CPYRGHT
Approved FertRelease-2000/08/08 - CIA-RDP96-00789R003100120001-4
using an adapation of Gough and
Heilbrun's Adjective Check List. Three judges evaluated each of the art
prints using this list. If two judges checked the same adjective for any
picture, that adjective was deemed to describe the particular print.
This analysis revealed
that
a higher percentage of hits
were
associated with targets which had blue in them, where targets containing
orange and yellow were associated with more misses (whether results were
significant is not reported). Puthoff and Targ (1979), in an anecdotal
comment upon their remote viewing studies stated that most hits were
associated with various nonanalytic aspects of a target, such as colour.
However, in another remote viewing study (Targ, Targ and Lichtarge, 1986)
where colour was superimposed over black and white slides of locations,
it was found that the viewers were unable to perceive the colour. In
commenting upon these results the authors speculated that the lack of
colour perception may have been due to the restricted number of colour
choices which resulted in making the colour perception a more analytic
task than the free-response perception of possible target sites. Much
of Warcollier's (1938) work used simple black and white line drawings
as targets. However, he observed informally that when colour was in the
target, it appeared to be perceived as frequently as was the form of the
drawing.
A non-psi study by Braud, Davis, and Opella (1985) examined the
frequency of occurrence of different types of imagery in dreaming and
ganzfeld states. As this study used no targets, the results could
indicate what types of imagery have an a priori probability of being
mentioned more often than others. In relation to this category, they
found that dreaming and ganzfeld imagery contain a predominance of colour
(among other things). These results could be pertinent to the findings
discussed in this paper, in that some of these findings could be due to a
simple predominance of certain naturally occurring types of imagery as
opposed to reflecting actual transmission of target-related content. It is
possible that the higher frequency of colour imagery in general could lead
to spurious observations of success with colour targets unless formally
examined. This should be borne in mind when considering anecdotal
observations.
Many studies have been conducted using black and white targets, most
notably those experiments where the target consisted of simple line
drawings. However, we found no free-response work which compared the
effectiveness of black and white to coloured targets. McMahan and
Rhine (1947) conducted a forced-choice study using both coloured and black
and white Zener cards. They found a higher average score with the
coloured cards than with the black and white, but the difference was not
significant.
The findings from this category do not indicate any clear-cut
differences between the success-rate of colour and black and white target
materials. As both have a long track record of obtaining significant psi
outcomes, research specifically aimed at comparing the two in a
free-response setting would be needed before any conclusions regarding the
superiority of one over the other could be made.
Complex / Simple
Information included in the "complex" category referred to comments
and findings about target materials, most commonly pictorial, which
were complex and/or rich in content. Krippner and Zeichner (1974) found a
higher percentage of misses with more complex targets (whether the
Finding was significant was not reported). Stuart (1946b) stated that
Approved For Release 2000/08/08 : CIA-RD1396-00789R003100120001-4

CPYRGHT
reasonably complex
target material might mislead subjects. krippher
Release 2000/08/98exCIA-RDP96-00789RG0840012000h4
(1970)AP%
APRO
subjects but the judges, as complex targets could make the evaluation
procedure overly problematic,
with the creative judge finding numerous
correspondences between many dream sequences and complex, detailed
pictures. On the other hand, significant results have been obtained with
very complex target material such as film clips (Psychophysical Research
Laboratory, 1985) and the multi-sensory target environment of Krippner,
Ullman, et. al. (1972) described in the introduction.
Information classified as "simple" included references to targets
composed of clear, unequivocally definable, common objects and symbols.
Most frequently these targets were simple line drawings. Both Carington
(1940) and Stuart (1946a) recommended the use of simple, as opposed to
compound, drawings so as not to confuse the subject. Warcollier (1963)
noted that even though his targets were simple, percipients' responses
still showed considerable distortion. As above, Krippner and Zeichner
(1974) found a higher percentage of hits associated with more simple
targets as measured by the number of adjectives used to describe the
target (again, whether this finding was significant is not reported).
Several forced choice studies
have
examined the
use of
multiple aspect targets. Generally these targets would be considered to
be 'simple' by free response standards.
However, being multi-aspect by
definition, they would represent more complex material than many
forced-choice targets. Palmer (1978) in reviewing this work concluded
that when multiple-aspect targets were used subjects tended "to score at
least as high or higher on the total target than on any of its primary
attributes. Such results suggest either that such targets are perceived
holistically (even if the overt responses are fragmentary) or that a
correct guess on one attribute somehow facilitates correct guesses on
other attributes."(Palmer, 1978, p.88)
In a review of six studies
utilizing dual-aspect targets, Kennedy
(1980) examined whether
complex target information was treated as a gestalt or whether the
individual parts of the information appeared to be processed separately.
No support for or against either mode of information processing was
obtained.
The above findings do not merit any clear conclusions. Before
such conclusions could be drawn direct comparison within studies of
carplex target material is needed.
Novel / Familiar
Information relating to unexpected,
unfamiliar, unusual and/or
incongruous target material was included in the novel category. Cavanna
and Servadio (1964) conducted a pilot study to investigate suitable
methodologies for studying the occurrence of ESP during states induced by
taking hallucinogenic drugs. Their targets were photographs consisting of
very incongruous elements, for example an upside-down foot, balancing an
artificial eye between the toes. The results were non-significant,
although this outcome could have been due to the difficulties involved in
attending to a test situation when under the influence of an
hallucinogenic drug. Krippner and Zeichner (1974) obtained a higher
percentage (whether or not significant was not reported) of hits when
targets were described as imaginative and interesting (qualities which
could be construed as novel). Ullman and Krippner (1973) ran a four
subject dream study in which the same target was used for half of the
testing nights and a different target used for each REM period for the
other half.
They observed that the the four participants preferred the
Approved For Release 2000/08/082361A-RDP96-00789R003100120001-4

CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
use of different targets for
every dream against a single target.
The authors thought this indicative of the dreamers' attention being more
engaged by novel ESP stimuli. In another of the dream studies (female
subjects, eight nights ESP, eight of control, no significant scoring)
Ullman and Krippner (1973) commented that the
subjects felt that the
target material should be as unusual as possible.
Roll and Harary (1976)
found that "interesting responses" (hits) were obtained when spontaneous,
unexpected changes were made in the experiment. Two examples they
provided of this involved last minute changes being made to the target
material.
Several forced-choice studies have considered the effect of novelty
of task and/or target material upon ESP performance. In reviewing these
studies
Carpenter (1977) concluded that novelty could facilitate
psi-hitting for most subjects, but could be counter productive for star
subjects used to a specific routine.
Information classified as "familiar" included references to targets
which held varying degrees of recognition for the percipients. Many
studies have been conducted using targets of emotional significance to the
subject and with which the subject would have been also necessarily
familiar. However, as emotional significance was usually deemed the more
important aspect of such targets, these studies will be considered under
that section.
Irwin (1982) conducted a study examining the influence of subjects'
Familiarity with the targets. Half of the targets (Maimonides slides)
were exposed to the subjects prior to testing, and half were not. This
manipulation had no significant effect upon the study's outcome.
Harcollier's (1938) research lead him to anecdotally conclude that only
elements of a target familiar to both the subject and agent could be
successfully transmitted. Targ, Puthoff and May (1979) have commented on
the basis of informal observations of their own research that use of
either repetitive target sequences and/or use of target pools of which the
subject had prior knowledge would inhibit remote viewing success.
The few findings reported in this category do not support the
drawing of any firm conclusions. There is some anecdotal support for the
utility of using a different target, with which the subject is not
familiar, for each testing of that subject. Also, the Krippner and
teichner (1974) findings offer some support for the use of imaginative and
interesting targets.
Abstract / Concrete
Abstract information included references to targets which portrayed
a potentially realistic scene or object in either an abstract and/or
unrealistic manner (to varying degrees) or in a not readily recognizable
fashion. Krippner and Zeichner (1974) found a greater percentage of
misses with targets which were described as unrealistic (whether
this finding was significant was not reported). Ullman and Krippner
1973) in the series of dream studies with Erwin , reported that purely
abstract pictures which lacked human figures gave poorer results than
targets which contained human figures engaged in activity.
Information included in the concrete category would be references
to target material which presented an object or scene in an immediately
recognizable, undistorted manner. While a great number of studies have
used targets which could be characterized as being concrete, we found no
specific reference regarding the utility of this characteristic in the
free response studies.
Although Krippner and Zeichner's (1974) finding and Ullman and
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
235

CPYRGHT
Approved For Release 2000/08/08 ; CTA-RDP96-00789R003100120001-4
Krippner's (1973) observation suggest that abstract targets
conducive to psi-hitting,
more
research is needed before firm
conclusions can be drawn.
Dynamic / Static
The dynamic categorization was used to refer to informations about
targets which portrayed and/or conveyed movement, a sense of movement,
and/or
gustatory, olfactory, auditory, tactile, and/or kinesthetic
stimulation. Thus a wide diversity of target materials fell into this
category including pictorial material (showing movement), film clips
(containing movement), and a variety of non-visual target material such
as music excerpts, the taste of a food, etc. In considering this large
category perhaps it should first be noted that Braud, Davis, and Opella
(1985) in their non-psi, no target study,
found a predominance of
activity contained in ganzfeld and dreaming imagery. Gurney, Myers and
Podmore (1886) reporting on the findings of the Society for Psychical
Research's Census of Hallucinations found that in cases of apparent
GESP of literal reproductions of the agent's bodily sensation (pain,
smell, touch, etc. ) were rarely transmitted. They noted from their own
experience that while taste was perceived in experimental situations,
they received no accounts of such in the spontaneous reports. The
spontaneous cases seldom contained reports of touch, and when it was
reported it was normally associated with auditory and/or visual
impressions. Music and other auditory stimuli were frequently reported.
Warcollier (1963) informally observed that moving objects or the ability
of the target to suggest movement seemed to be perceived by the subject.
Warcollier (1938) also expressed the belief that kinesthetic sensations
should be easily transmitted, but admitted to having little data to back
this up. Reporting on an Esalen Meeting on Psi Research, Schlitz (1984)
reported general agreement among the participants that kinesthetic,
auditory and olfactory images were as important,
if not more so, as
visual images in conveying psi information.
Honorton and Schechter (1987), reporting on the significant (p =
0.027, 1-t) outcome of 187 automated testing ganzfeld sessions, found
that sessions using dynamic targets (video segments and other "lifelike"
material) were independently significant (p = 0.007, 1-t), while those
using static targets (defined as "still pictures") were at chance. The
difference between the two was suggestive, but not significant (p =
0.079, 2-t). Likewise, Krippner and Zeichner (1974) found more hits
associated
with
targets having dynamic content (whether this
finding was significant was not reported).
-
Altom and Braud (1976) ran a pilot study aimed at exploring the idea
that right-hemisphere brain activity may be conducive to psi. They used
four different excerpts of music as targets, which it was thought might
encourage right-hemisphere activity. They obtained a significant level of
psi scoring (p = 0.05). Kesner and Morris (1978) conducted a guided
imagery, precognition study using music from records and their album
covers as targets. The subjects' imagery was rated by an independent
judge who individually rated subjects' visual and auditory imagery.
Neither the results from the visual or the auditory ratings were
independently significant, however the two combined were (p < 0.02),
suggesting that the more senses involved in a target, the better.
Several dream studies have been conducted using dynamic target
material. Krippner, Honorton, and Ullman (1972) obtained significant
results (p < . 001) usings thematically related slides, accompanied by an
Approved For Release 2000/08/08 21A RDP96 00789R003100120001-4
may

.-
CPYRGHT
ApprovedFor Release 2000/08/08 - CIA-RDP96-00789R003100120001-4
same type of target material Krippner,
Honorton, et. al. (1972) again elicited a significant level of psi-hitting
(p = . 004). An even higher level of significant scoring (p = . 0002) was
obtained by Krippner, Ullman, et. al. (1972)
using the multi-sensory
target environment described in the introduction of this paper. As
previously mentioned, Ullman and Krippner (1973) found that paintings of
humans engaged in activity seemed to be more successful than abstract
paintings in the Erwin series.
The second
Erwin study, which again
obtained a significant degree of psi-hitting (reported effects "on the
order of a thousand to one" p.116), used art prints together with
associated objects and activities on the part of the agent.
Dunne and Bisaha (1979), reviewing seven remote viewing series,
noted that dynamic targets were perceived as readily as stationary ones.
Yet, Puthoff and Targ (1979) commenting upon their remote viewing work
said that motion was very rarely reported, even when it was an important
component of the scene. Although, Targ, Puthoff, and May (1979) stated
"that real-time activities at the target site are often perceived" (p.94).
These authors also noted that "in addition to visually observable detail,
subjects sometimes report sounds, smells, electromagnetic fields, and so
forth, which can be verified as existing at target locations" (p.95). It
should be noted that the above three observations were all ancedotal.
Two studies
made specific comparisons between
static
and
dynamic target characteristics. Honorton and Schechter (1987) obtained
highly significant psi effects with dynamic targets, while static targets
obtained chance results.
Krippner and Zeichner (1974) found more hits
associated with dynamic targets. The findings of Kesner
and Morris
(1978) and those of the reviewed dream studies further
suggest the
possible benefits of using multi-sensory target materials.
Form / Meaning and Idea
Comments related to the importance of the shape or form of the target
or some of its components are included in this category. Puthoff and Targ
(1979), in discussing their remote viewing work stated "most of the
correct information that subjects relate is of a nonanalytic nature
pertaining to shape, form, colour, and material rather than to function or
name" (p. 65). Barrington (1983), reviewing past work with the medium
Stefan Ossowiecki, found many examples where the form of the target had
been correctly identified but not the meaning, a situation which she
labelled as "incomprehending clairvoyance". Similarly, Warcollier (1938
& 1963) observed that frequently the shape of a target would be perceived
without reference to the target's meaning or idea, although he also notes
that meaning and idea may also be perceived without specific reference to
shape. Warcollier (1938) also discusses the work of Richonnet
(no
reference provided) noting that Richonnet thought that form was both
easier to perceive than meaning and would be perceived prior to perception
of the identity (idea) of the ESP target.
The "meaning and idea" categorization includes information referring
to situations where the meaning, idea and or identity were perceived,
without reference to the shape or physical appearance of the target.
Carington (1940) believed that the idea of a target, not the form, was
what would come through to the subject. Gurney, Myers and Podmore (1886)
received reports which indicated that meaning and idea were the important
aspects of the target. The example they provide of this is where a word
in one language is received in another, having been suitably translated.
Marsh (1960), in a study using simple line drawings as targets,
carmented that subjects tended to reproduce the concept of the target
'Approved For Release 2000/08/08 : CIA RBn96-00789R003100120001-4
-

CPYRGHT
rather
than
the
shape.
Lodge
apparently shared these beliefs as,
transmitted than a drawing (i.e. form). As noted above, Puthoff and Targ
(1979)
believed that most correct information provided by subjects
pertained to the nonanalytic aspects of targets such as form, shape and
colour. Indeed, they thought that errors could arise when the subject
tried to make sense (i.e. label according to name and function) of such
nonanalytical target components.
This category presents some conflicting observations and opinions,
all of which are anecdotal in nature, regarding the utility of form, as
opposed to meaning and idea, in conveying psirrelated information.
Given this state of affairs, the only conclusion that can be drawn is
that research aimed at resolving this question is needed.
Emotion
Any comments having to do with the emotional content of or emotional
reactions to target materials were included in this category.
Same
researchers have also made comments about specific target themes/content
which could be interpreted as having a strong emotional component
(e.g. war scenes, erotic scenes, religious themes, etc.). However,
whether these themes would be regarded as positive or negative would
probably vary greatly from subject to subject. Therefore, these findings
will not be referred to in this section unless the author specifies that
the emotionality of the target was an important factor in the study's
success or failure.
Gurney, Myers, and Podmore (1886) observed that in spontaneous
cases
emotions were frequently received, often with the receiver having
no idea why they were experiencing certain feelings. However, the
emotion experienced by the percipient was later found to be appropriate to
the event which was taking place at the time, unknown to the percipient
(e.g. feeling sadness over the death of a close friend). Warcollier
(1938) also comments that in spontaneous cases, the message is almost
always emotional.
Williams and Duke (1979) conducted a study specifically examining
various target qualities and their relationship to psi performance. They
devised a 39-item Target Evaluation Rating which measured various target
qualities,
including overall emotional impact and positive and negative
emotional dimensions, upon which each of 152 targets were rated. They
then looked at data, gathered from 174 subjects, from other free response
studies which had used these targets.
For the purposes of their
analysis, they excluded any target which had not been randomly chosen as a
target at least three times in the previous studies. This criterion
provided 22 targets, and ESP data from 91 subjects (overall
significant psi-hitting was obtained, p < . 047, 2-t). The individual psi
scores obtained for each of these 22 targets were averaged to provide a
composite psi score for each target. The composite psi scores were divided
into good psi targets and poor psi targets resulting in 12 high
psi-scoring targets and 10 low psi scoring targets.
Comparing these
targets to the total emotion score (the mean of the positive and negative
emotion ratings) from the Target Evaluation Rating, they found that
targets containing a stronger emotional content were significantly better
(i.e. high psi-scoring targets) than non-emotional targets (p < . 001).
Sondow, Braud and Barker (1981) conducted a ganzfeld study also aimed
at investigating target qualities,
which obtained a significant outcome
using a sum of ranks (p < . 04 1-t),
but did not reach significance using
direct hits as a measurement.
Using the Target Evaluation Rating,
238
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGHT
relatively high emotion rating and five having a low emotion rating.
Each high emotion pack consisted of two positive and two negative emotion
pictures; the low emotion packs consisted of two natural scenes and two
pictures of material objects. This complex study involved many different
measurements and analyses, of which only those most relevant to this
paper will be reported. The neutral (low) emotion target packs showed
more psi-hitting than the high emotion pictures, with the difference
approaching significance (p = . 052, 2-t). Using a scale they devised to
measure emotion which both the subjects and agents completed, they found
that when a high emotion picture was the target, receivers would feel
more total emotion whilst in the ganzfeld than did receivers with a low
emotion target pack (p < . 04, 2-t). Also receivers felt more emotion when
senders felt more emotion while sending (p < . 04, 1-t). However,
Stanford (1984) has pointed out that this latter finding could be
artifactual due to commonalities of experience between subjects and
agents (e.g. the weather that day). Using Osgood 's Semantic Differential
to measure the components of the target pictures, they found there more
hits when the receivers' and senders' evaluation of the targets were in
close agreement than when their categorizations widely differed. Of
twenty targets where agreement was close, nine were direct hits (p = . 04,
1-t).
Both Williams and Duke (1979) and Sondow, Braud and Barker (1981)
found significant outcomes in various analyses examining how well their
subjects liked (emotionally preferred) the target. Williams and Duke
(1979), comparing subjects' ratings of target preference for hit and
missed targets for two different groups of subjects (with the rating
being made prior to obtaining feedback as to the target identity),
found the first group of 101 subjects significantly preferred targets
with which they had obtained a hit (p < . 035, 2-t), as did the second
group of 80 subjects (p < . 0038, 2:t). A similar finding was reported in
the Sondow et al. (1981) study, where a comparison between liking for
psi-hit and for psi-missed targets again yielded a significant outcome (p
< . 0096, 2-t). Another analysis in this study showed that pictures
received a significantly higher liking rank (p < . 0094, 2-t) when they
were the target than when they were a control. Braud and Loewenstern
(1982) also found that psi-hitters liked their targets significantly
better than psi-missers (p < . 025, 1-t). Two other significant target
preference findings were presented in Braud and Boston (1986). The
authors replicated the preference effect (p < . 036, 1-t), and also
reported similar results from Braud, Ackles & Kyles (p < . 045, 1-t).
However,
these findings may be contaminated due to response bias
problems.
To quote Stanford (1984)
"these findings could be
artifactual ;... Because of their desire for success, subjects may tend to
like pictures which correspond to their ganzfeld mentation, and such
correspondence tends to be greater and more detailed when ESP has actually
occurred. Thus such pictures may be liked appreciably more." (p. 107).
Many forced choice studies have examined the role of target preference.
These findings have been reviewed by Carpenter (1977) and Palmer (1978).
In drawing some conclusions about these findings Palmer comments that
while a preferential effect has been found most often "with respect to
response type rather than target type, it (the preference hypothesis)
offers our best hope to date of intergrating a very messy and inconsistent
body of data concerning the effect of target type on ESP scoring in
forced-choice experiments." (p. 87).
Krippner, Honorton, et al.
239
(1972) considered their targets
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGHT
( thenapproved For Release. 2000/08/08 : CIA-RDP96-00789R003100120001-4
and
appropriate sound
tracks )
to be
emotionally arousing, and thought that their significant results provided
support for the use of such material.
Ullman and Krippner (1973) also
felt
"that an important ingredient in the success of experiments in
dream telepathy over waking telepathy ... is the use of potent, vivid,
emotionally impressive human interest pictures to which both agent and
subject can relate." (p. 210).
Moss (1968; also see: Moss,
1969; and Moss & Gengerelli, 1968)
described the evolution of her experimental methodology over a series
of six experiments. Emphasizing the importance of using emotionally
arousing targets, her targets evolved to consist of slides accompanied
by appropriate sound effects paired so as to present contrasting
emotions. The results from these studies were very sketchily presented,
although significant outcomes were described for some of the studies.
However, no comparison was made between either emotionally arousing
targets and neutral ones, or between the effectiveness of the different
contrasting emotions. In a series of studies Preiser (1986) found that
ESP performance was highly dependent on the emotional loading of the
target material. The information about this study is limited as it was
obtained from an abstract. However, while no overall significance was
obtained, one part of the series did get a significant ESP outcome.
Cavanna and Servadio (1964) stressed the careful choosing of targets which
they considered to have definite emotional significance. While they did
not obtain significant psi-scoring, they did express the belief that
their future targets should be chosen to be even stronger, emotionally.
Some studies utilizing physiological
measurements have used
targets chosen to have specific emotional significance for individual
subjects. Esser, Etter, and Chamberlain (1967) used plethysmographic
responses to personalized target material,
devised from initial
interviews with the participants. The resulting targets, designed to
have greater emotional significance for either the percipient or the
agent, were either names of importance to the subject or sentences or
quotes describing a emotional conflict of relevance to the participants.
No significant outcomes were obtained, but the results were suggestive in
that there was some correspondence between onset of the sending period and
plesthysmograph responses.
Dean (1971) contrasted plethysmograph
recordings of vasoconstriction examining the reaction of subjects to
targets consisting of either a blank card or a card upon which was written
a name of a person who has emotional significance to the subject. He
found larger vasoconstrictions (i.e. more emotional arousal) for the names
than for the blanks. This study also had a group of control subjects for
whom the names would have had no special relevance. Interestingly, he
found that the control subjects displayed a greater level of reaction to
the names than did the subjects for whom the names had emotional
Haraldsson (1983) again used names
of emotional
significance to the participants as the target in a study using a
significance.
plesthysmograph. No overall significant results were obtained, however,
he did obtain a significant outcome in the first 20 sessions of the study
p < . 003), with results declining later.
Several studies have compared targets having positive emotional
qualities to those having negative emotional characteristics. Williams
and Duke (1979), comparing good psi targets to poor psi targets, found
that targets which contained a positive emotion were significantly better
targets (p < . 02) than those which did not and that targets which
contained negative emotion were significantly worse (p < . 047) than
those which did not. Sondow, Braud and Barker (1981) found no significant
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
difference between positive and negative emotional targets. Eisenberg and
Donderi (1979) used 7 emotionally stimulating sound films as targets in a
study incoporating both forced-choice and free response conditions. They
obtained a significant degree of psi-hitting (forced choice condition: p <
.02; free-response condition:
p
<
.001).
The film clips were
classified as conveying either positive
or negative emotions,
although no significant difference was found between the scoring on the
positive and negative emotional targets. Krippner and Zeichner (1974)
found more misses when the target was described as pleasant and more hits
when the target was described as unpleasant (whether these findings were
significant is not reported).
One forced-choice
study
which
specifically
addresses the
positive/negative issue was conducted by Johnson (1971) who asked subjects
to provide two words, one having an exceedingly pleasant meaning for the
subject and the other having a very unpleasant meaning,
from which he
created targets of associated words/concepts.
These concepts (secondary
targets) were paired with a digit from one to five (primary targets),
although 20 per cent of the primary targets were left unpaired as a
control (emotionally neutral targets). The subjects in this precognitive
study were to guess what number would be selected as the target.
Johnson
compared performance on positive, neutral and negative emotions. No
significant overall scoring was obtained, the positive targets showed a
non-significant degree of psi-hitting, the negative targets significantly
psi-missed (p = . 0094, 1-t), and the neutral targets scored at chance.
The difference between the positive and negative targets was significant
(p < . 005, 1-t).
The anecdotal observations in this category reveal that many
researchers believe emotional targets to be superior to non-emotional
ones.
However, only two studies (Williams & Duke, 1979; and Sondow et
al., 1981) explicitly examined this assumption and they obtained
conflicting results. One analysis in Sondow et al. (1981) found that the
percipient would experience more emotion with a high emotion target, but
as this study also obtained a greater degree of psi-hitting with low
emotion targets, this result could be seen as arguing against the use of
high emotion targets. Nor can the physiological studies be readily
interpreted as providing support for the utility of using target material
chosen to have specific emotional significance for individual subjects.
Aside from the general lack of significant outcomes of these studies, the
Dean (1971) study actually obtained a greater response from his control
subjects to whom the target material should have had no special
relevance. The studies comparing positive emotional targets to those
with negative emotive qualities also obtained conflicting results. Thus,
again more research is needed before any conclusions can be drawn
regarding the psi-conducive effects of emotional targets.
theme / Content
This category includes all references which associate the specific
content or theme of individual targets with the success/failure of these
targets. Williams and Duke (1979) found that "most of the psi-hitting
targets were natural, while the missing targets were material
objects-metal, concrete, man-made, and mechanical. "(p. 8) A post hoc
analysis revealed this difference to be significant (p < . 02). Dunne,
Jahn, and Nelson (1983), reporting on several
remote viewing studies,
noted that there was no difference in effectiveness between the following
site characteristics: natural vs. man-made; permanent vs. transient; and
indoor vs.
outdoor. The Psychophysical Research . Laboratory (1985)
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
241

CPYRGHT
carparkd provinsForReleases2009/08/08etCIA-RER96-00789 R003100120091.4
successful than others. The category of "disasters" obtained significant
psi-hitting (p = . 014, 2-t). Sexual
themes were associated with
significant psi-missing (p = . 008, 2-t). Non-significant scoring in the
psi-hitting direction was obtained by (listed in descending order of
strength of effect) the categories of religion, sports/hunting, locales,
and animals.
Non-significant scoring in the psi-missing direction was
obtained by the racing and fighting/warfare categories. A post hoc
analysis by Sondow (1979) found that targets were chosen and non-targets
avoided significantly often when the pictures showed horses (p < . 01),
water (p < . 02), fire (p < . 03), and flying-leaping-swinging (p < . 04).
Such effects were not found with the target categories of food, war and
famine, and music. Ullman and Krippner (1973) observed that the art
prints containing/portraying religion, colour, eating/drinking, emotions,
and people tended to be successful, as did the agent's multi-sensory
involvement with the target. Stuart (1945), using simple line drawings as
targets found that the two most successful targets portrayed a cartoon
character and a candle. The two least successful targets were a book and
a mathematical equation. In another drawing study, Stuart (1947) found
the best target was a church and the worst was a train. Lastly, Braud,
Davis, and Opella (1985) found a predominance of human characters and
architectural content contained in ganzfeld and dreaming imagery. Less
frequent were mythical characters, animals, food, and unconnected body
parts.
These findings could contribute
to
spurious anecdotal
observations.
Examining these diverse content categories it was discoved that
religion was mentioned three times as a generally successful target topic.
Warfare was twice mentioned as being less successful. Williams and Duke
(1979) found that natural targets were associated with psi-hitting, and
the categories specified as successful by Sondow (1979) could also be
classified as natural. However, given the wide diversity of actual
targets which these findings represent, these similarities should be
viewed at most as possible trends which require further reseach for
confirmation.
Discussion
The most consistent category findings of this paper relate to the
possible advantages of using dynamic, multi-sensory targets. However,
these findings are based on the outcome of relatively few studies and
thus should be treated with caution pending further confirmation. The
novel category provided some tentative support for the use of new targets
with which the subject is not familiar for each trial with that subject,
and also suggested possible benefits of using imaginative and interesting
targets. But again these finding are derived from very few studies.
The two findings relevant to the abstract categorization both found
abstract targets to be associated with poorer results. The emotionality
of targets, often quoted in the literature as one of the yardsticks by
which targets are chosen, has not been shown to be reliably associated
with psi-hitting. Nor have any of the other categories investigated
herein.
In short, this review has not succeeded in shedding a great deal
light upon what qualities/characteristics might discriminate successful
from unsuccessful free-response targets.
Indeed, the outcome of this
paper could be viewed as demonstrating how very little we actually know
about successful versus unsuccessful target characteristics.
However, another interpretation of these findings could be that
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGHT
target_ characteristics,
n
not be
DP96
Insistent !: 3 16datobop_successful
Approved For Release 2000/08/x8 1
Subject differences.
Many
years
ago
Warcollier (1963) commented that "No two subjects respond alike to the
same target.
No two targets seem to affect the same subject in the
same
way."(p. 56). Indeed,
a great deal of experimentation has
examined and revealed interactions
between
various trait factors and
psi performance (for reviews of this literature see Palmer, 1978; or
Carpenter, 1977). Other variables such as state, setting, response
method, and so on, may also influence the particular type of target
which is successful in any given situation.
Future research could
profitably examine the effects of such variables. In addition, the
development of a descriptive set of scales, such as the three-dimensional
scale discussed in the introduction of this paper, which could be used on
an inter-laboratory basis, could forward our knowledge of target success
considerably. The development of such scales will be the focus of future
research at the Edinburgh Lab.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
243

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
REFERENCES
ALTOM,
K. &
BRAUD, W. G. (1976). Clairvoyant and
telepathic
impressions of musical targets. In J. D.
Morris & W. G. Roll (Eds. ), Research In Parapsychology
1975
(pp.
171-174).
Scarecrow Press, Metuchen,
New
Jersey.
BARRINGTON , M. R. (1983) . The mediumship of Stefan
Ossowiecki. In W. G. Roll, J. Beloff
& R. A. White
(Eds. ), Research in Parapsychology 1982 (pp. 72-75).
Scarecrow Press, Metuchen, New Jersey.
BRAUD, L. W. & BOSTON, D. A. (1986). Target preference
and clairvoyance in selected subjects following relaxation
induction. In D. H. Weiner & D. I. Radin ( Eds. ), Research
in Parapsychology 1985 (pp. 25-28).
Scarecrow Press,
Metuchen, New Jersey.
BRAUD, L. W. & LOEWENSTERN, K. (1982). Creativity and pai.
In W. G. Roll, R. L. Morris & R. A. White (Eds. ), Research
in Parapsychology 1981 (pp. 111-115).
Scarecrow Press,
Metuchen, New Jersey.
BRAUD, W. , DAVIS, G. & OPELLA, J. ( 1985) .
Electrodermal
and imagery
concomitants of
ganzfeld
stimulation.
Parapsychology Review, Vol. 16, No. 4, 1-5.
CARINGTON, W. (1940).
Experiments on
the paranormal
cognition of drawings.
Proceedings of the Society for
Paychical
Research,
Vol.
46,.
No.
164,
34-151.
CARPENTER, J. C. (1977).
Intrasubject & subject-agent
effects in ESP experiments.
In B. B.
Wolman (Bd. )
Handbook of Parapsychology, Van Nostrand Reinhold & Co.,
New York, 202-272.
CAVANNA, R. & SERVADIO, E. (1964).
ESP experiments with
LSD
25
and
Psilocybin:
A
methodological
approach.
Parapsychological
Monographa
No.
5,
Parapsychology
Foundation, New York.
DEAN,
B.
D.
(1971).
Long-distance
plethysnograph
telepathy
with
agent
under
water.
Parapsychology
Association Convention Proceedings, 1969, No.6, 41-42.
KESNER, J. & MORRIS, R. L. (1978).
A precognition test
using guided imagery. In W. G. Roll (Ed. ), Research in
Parapsychology 1977,
(pp.
48-52).
DUNNE, B. J. & BISAHA, J. P. (1979). Precognitive remote
perception:
a critical overview of the experimental
progran.
Parapsychology Convention Proceedings 1979,
22nd. Annual Conference, Morago, California.
DUNNE, B. J., JAHN, R. G.
NELSON, R. D. (1983).
Precognitive remote perception.
Engineering Anomalies
Research Technical Note PEAR 83003, Princeton University,
School of Engineering/Applied Science.
EISENBERG, H. & DONDERI, D. C. (1979).
Telepathic
transfer of emotional information in humans.
Journal of
Paychology, Vol 103, 19-43.
ESSER, A. H., ETTER, T. L. & CHAMBERLAIN, W. B. (1967).
Preliminary report:
physiological
concomitanta
of
"communication" between isolated subjects. International
Journal of Parapsychology, Vol. 9, No. 1, 53-56.
GURNEY, B. , MYERS, P. W. H. & PODMORE,
F. (1886).
Phantasms of the Living. Trubner & Co., London.
HARALDSSON, E. (1983). Vasomotor reactions as indicators
of
extrasensory
perception.
. In R. A. White (Ed. )
Perapsychology Abatraota International, Vol. 1, No. 1,
25-26.
HONORTON, C. & SCHECHTER, E. I. (1987). Ganzfeld target
retrieval with an automated testing system: A model for
initial ganzfeld success. In D. H. Weiner & R. D. Nelson
(Eds.), Research in Parapaychology 1986 (pp. 36-39),
Scarecrow Press, Metuchen, New Jersey.'
IRWIN, C. P. (1982). The roll of memory in free-response
ESP studies: is target familiarity reflected in the
scores?
Journal of the American Society for Psychical
Research, Vol. 76, No. 1, 1-22.
JOHNSON, M. (1971). An attempt to effect scoring behavior
in a group test of precognition by means of manipulation
of motivation and by the use of individually assigned
emotionally loaded target material. Research Letter of
the Parapsychological Division
of
the
Paychological
Laboratory, University of Utrecht, 15-32.
KENNEDY, J. R. (1980). Information processing in ESP: A
survey of forced-choice Experiments using multiple-aspect
targets. Journal of Parapsychology, Vol. 44,
No. 1,
9-34.
Scarecrow
Press,
Metuchen, New Jersey.
KRIPPNER, S. (1970). Commenting upon a paper by T. Moss,
"Telepathy in the waking state: an experimental design".
In R. Cavanna (Ed. ). Proceedings of an International
Conference on Methodology in Psi Research, 1968.
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

D
Parapsychology Foundation, New York, 136-137.
KRIPPNER, S. & ZEICHNER, S. (1974). Descriptive analysis
0
"art prints telepathically transmitted during sleep. In
G. Roll, R. L. Morris & J. D. Morris ( Eds. ), Research
- Parapsychology 1973, (pp. 27-28) .
Scarecrow Press,
M
Luchen, New Jersey.
KRIPPNER, S., HONORTON, C. & ULLMAN, M. (1972). A second
pgecognitive dream study with a selected subject.
In
W
FG.
Roll, R. L. Morris & J. D. Morris (Eds. ),
Parapsychological Association Proceedings, 1971, No. 8,
71079.
KRIPPNER, S., HONORTON, C., ULLMAN, M., MASTERS, R. &
HOUSTON, J. (1972). A long-distance "sensory bombardment"
E&D-dream study. In W. G. Roll. R. L. Morris & J. D.
Mejris (Eds. ), Parapsychological Association Proceedings,
1900, No. 7, 49-51.
KROPPNER, S., ULLMAN, M.,
HONORTON,
C., HUGHES, W.,
GOODMAN, G.
HARRIS, R. (1972).
An eight-night study of
pfocognitive dreams using EEG-EOG techniques. In W. G.
RAJ1,
R.
L. Morris
&
J.
D.
Morris
( Eds. ),
Parapsychological Association Proceedings, .1970, No.
27-28.
?,
MARSH, M. C. (1960). The Rhodes experiment: linkage in
extra-sensory perception.
Journal of the Society for
Paschical Research, Vol. 40, 219-239.
O
MONAHAN, B. A. &
RHINE,
J. B. (1947).
A
second
Zereb-Durham ESP experiment.
Journal of Parapsychology,
Vel. 11, No. 4, 244-253.
0
MOS, T. (1969). ESP effects in "artists" contrasted with
"Con-artists". Journal of Parapsychology, Vol. 33, No. 1,
5569.
0
MOSS, T. (1970).
Telepathy in the waking state:
an
experimental design. In R. Cavanna (Ed. ), Proceedings of
go International
Conference on Methodology in Pai
R&Search, 1968.
Parapsychology Foundation,
New York,
121-142.
A
MOSS,
T.,
GENGERELLI,
J.
A.
(1968).
B8P
effects
generated by affective states.
Journal of Parapsychology,
Vol. 32, No. 2, 90-100.
PALMER,
J. (1978).
Extrasensory perception: research
findings.
In
8.
Krippner
(Ed. )
Advancen
in
Parapsychological Research 2: Extrasensory Perception,
Plenum Press, New York, 59-244.
PREISER,
8.
(1986) .
Emotion
versus
information:
methodological
critique.
In
R.
A.
White
(Ed. )
Parapsychology Abstracts International, Vol. 4, No. 1.
PSYCHOPHYSICAL
RESEARCH
LABORATORY
(1985).
1985
Psychophysical Research Laboratory Annual Report.
PUTHOFF, H. E. & TARG, R. (1979) .
A perceptual channel
for
information
transfer
over
kilometer
distances:
historical perspective and recent
research.
In C. T.
Tart, H. E. Puthoff & R. Targ (Eds. ) Mind at Large,
Praeger Publishers, New York, New York.
ROLL, W. G. & HARARY, K. (1976). Target reponses during
out-of-body experiences. In "The 1976 SERPA Convention -
a report by Gerald Solfvin", Parapsychology Review, Vol.
7, No. 3, 1-7.
SCHLITZ, M. (1984).
Esalen meeting
pai
research.
Parapsychology Review, Vol. 15, No. 6, 10-12.
SONDOW, N. (1979). Effects of association and feedback on
psi in the ganzfeld: is there more than meets the judge's
eye ?. Journal of the American Society for Psychical
Research, Vol. 73, No. 2, 123-150.
SONDOW, N., BRAUD, L. & BARKER,
P.
(1981) .
Target
qualities and affect measures
in
an
exploratory psi
ganzfeld.
Parapsychological
Association
Convention
Proceedings, 1981.
STANFORD, R. G. (1984). Recent ganzfeld-ESP research:
survey and critical analysis.
In R. G. Stanford & S.
Krippner (Eds. ) Advances in Parapsychological Research 1.
McFarland & Co., Jefferson, North Carolina, 83-111.
STUART, C. B. (1945). A classroom ESP experiment with the
free-response method.
Journal of Parapsychology, Vol. 9,
No. 2, 92-105.
STUART,
C. E.
(1946a) .
free-response method.
Journal of Parapsychology, Vol. 10,
GESP
experiments
with
the
No. 1, 21-35.
STUART, C. E. (1946b). An interest inventory relation to
ESP scores.
Journal of Parapsychology, Vol. 10, No. 3,
154-161.
STUART, C. E., HUMPHREY, B. M. & McMAHAN, E. (1947).
Personality
measurements and ESP tests with cards and
drawings.
Journal of Parapsychology, Vol.
11,
No.
1,
118-146.
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
--.

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
TARG, E., TARG, R. & LICHTARGE, O. (1986).
Realtime
clairvoyance: a study of remote viewing without feedback.
In D. H. Weiner & D.
I.
Radin
(Eds.) Research in
Parapsychology, 1985
(pp.
36-39).
Scarecrow
Press,
Metuchen, New Jersey.
TARG, R., PUTHOFF, H. E. & MAY, E. C. (1979).
Direct
perception of remote geographical locations.
In C. T.
Tart, H. E. Puthoff
& Targ, R.
(Eds. ) Mind at Large,
Praeger, New York.
ULLMAN,
M.
&
KRIPPNER,
S. (1973).
Dream Telepathy.
Penguin, Baltimore.
WARCOLLIER, R. (1938).
Experimental Telepathy.
Boston
Society for Psychical Research, Boston.
WARCOLLIER, R. (1963).
Mind to Mind.
Collier Books, New
York, New York.
WILLIAMS, L.
B.
&
DUKE, M.
( 1979) .
Qualities
of
free-response
targets
and
their relationship
to psi
performance.
Parapsychology Association 22nd.
Annual
Conference Proceedings, 1979, Moraga, California.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
246

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CHARACTERISTICS OF SUCCESSFUL FREE-RESPONSE TARGETS:
THEORETICAL CONSIDERATIONS
by
Caroline Watt
Psychology Department
University of Edinburgh
ABSTRACT
This paper describes theoretical ideas from a variety of
sources as to what might be expected to make a
successful free-response GESP target. Popular "how to
be psychic" literature, analyses of the characteristics of
spontaneous cases, and theoretical suggestions from
psychology and parapsychology show considerable
consistency in their suggestions about the likely features of
a good target. Two main recommendations appear to
emerge from these sources - good GESP targets should
be psychologically salient and physically salient: 1. targets
in parapsychological research should be meaningful, have
emotional impact and human interest - this may make
them salient in the minds of our experimental participants;
and, 2. targets should also be physically salient by
standing out from their backgrounds - properties such as
movement, novelty, brightness and contrast tend to make
stimuli physically salient.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGHT
CHARACTERISTICS +R 0008ESSHARPPIEE RESPONSE140AGETS:
THEORETICAL CONSIDERATIONS,
INTRODUCTION
Deborah Delanoy (1988) examined the observations from some free-response
literature on what makes a good GESP target. Despite the flaws and contradictory
findings seen in this literature, it was possible to make a few general statements
about what experimenters believe constitutes a good GESP target. This paper can
be seen as forming the second half of our observations and thoughts about targets in
parapsychological research. Delanoy described what is currently believed about the
characteristics of successful GESP targets, concentrating on relatively formal
free-response experiments in parapsychology. In contrast this paper describes
theoretical suggestions as to what might be expected to make good targets, roaming
more widely (and consequently with less depth) over some varied literature which has
something relevant to say on this question.
As stressed by Delanoy, our combined efforts are far from comprehensive, being
primarily aimed at getting some idea of what kind of targets we should use in our
research in Edinburgh. To do this, we looked through some parapsychological
journals (JASPR, JP, JSPR, EJP, IJP), parapsychological abstracts, PA and PF
convention proceedings, RIP, Parapsychology Review, certain "relevant" books held
in the Koestler Chair library, and I have also examined some psychological research
which I consider relevant to the target question. Particular attention was given to
cases where authors made specific comments about the characteristics of successful
GESP targets.
Firstly, this paper briefly considers so-called "Airport Project" books [named after
some research by Professor Robert Morris and his students using the kind of "how to
be psychic" books which can be found in airport bookshops (Morris, 1977)].
Secondly, the paper examines (again briefly) the kind of "target" information which
seems to be transmitted in people's spontaneous psychic experiences. Thirdly, this
paper considers some theoretical suggestions by parapsychologists as to what might
be expected to make a good GESP target. Then I make some suggestions of my
own on possible characteristics of a successful GESP target, derived from some of
the psychological literature on human-environment interactions, curiosity, attention,
and attributions of causality. The paper ends with a summary and conclusions.
1 I would like to thank Prof. Jim Crandall, Dr. Deborah Delanoy, Dr. Julie Milton, Prof.
Robert Morris and Mr. Robin Taylor for their valuable criticisms of and contributions to
this paper.
248
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGHApproved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
1. "AIRPORT PROJECT" BOOKS
A skim through the 21 "how to be psychic" books which form part of the Koestler
Chair library, and which I felt might have some comments to make about targets,
found only 6 authors who made recommendations on what might make a good target
when training psychic powers. Even then, the authors invariably failed to define their
terms or write more than a sentence on the subject. These recommendations should
therefore be treated with caution, as they do not represent the findings of careful
scientific experimentation. On the other hand, they may have something to suggest
about popular ideas of what makes a good GESP target, and these ideas may be
based on some grain of truth.
Boswell (1969) recommended the use of "mentally stimulating" targets. Also, he felt
that physical sensation and especially emotion were easily transmitted, and that
colour was picked up better than black and white. Edwards (no date) suggests that
faces and pictures make good targets. Denning & Phillips (1981) recommend trying
to transmit a message of emotional significance to the receiver. Likewise, Sherman
(1960) says that it is crucial to have some emotional content to the target. A related
area of interest is psychometry, where an object is used to provide further information
about its owner. Powell (1979) recommends using as a token object metal or leather
which has been close to the skin for a long time and therefore has had a chance to
build up some personal association with the owner. Finally, Burns (1981) feels the
following make good practice targets for developing GESP: pictures (rather than
words); something experienced vividly by the agent; flavours; body position of the
agent, or whether the agent is sitting in the light or dark; and sizes and weights of
objects.
There do seem to be some common themes in these authors' suggestions, though
the small sample covered here means that any patterns could be illusory: emotional
impact seems to be important (though little is said about whether the specific
emotions should be positive or negative ones); and targets conveying information
about events happening to humans seem popular.
2. SPONTANEOUS CASES
There is a considerable literature concerning the sort of information conveyed in
spontaneous cases of ESP, and so as a necessary constraint this section is limited to
observations from Sybo Schouten's (1979b, 1982) examination of two great
collections of spontaneous cases - Phantasms of the Living and the Louisa Rhine
collection.
Schouten made a quantitative analysis of these collections with a view to finding
patterns and relationships which might stimulate further experimental research. As he
pointed out, the two collections covered quite different cultures and eras, and were
gathered for different purposes. The collectors of the "Phantasms" cases took great
pains to investigate and verify their cases, and had a special interest in receiving
apparition reports as they felt these might lend support to their hypothesis that
Release 2000/08/08- CIA-RDP96-00789R003100120001-4
249

CPYRGHT
persons. In contrast, the Rhine collection took cases more or less at face value, with
the idea that inaccuracies would cancel each other out over a large number of cases,
and the reports were gathered with the aim of providing suggestions for future
laboratory research (Schouten, 1986).
Excluding 150 of the cases (for reasons outlined in Schouten 1979b), Schouten
analysed the remaining "Phantasms" cases according to 32 previously-defined
categories (Schouten, 1979a) and found that about 75% of the cases involved death,
illness or injury to the target person, though a tendency to remember serious events
for longer than trivial events accounted for some of this pattern. Only 1.4% of cases
conveyed information about positive experiences of the target person.
Table 1 (from Schouten, 1979b, p.432)
Situation of target person at time of experience
death
66.7%
serious illness
12.5%
slight injuries
8.7%
serious material
.5%
slight material
.2%
trivial
10.0%
positive
1.4%
It is interesting to note that slight personal injuries were more often the topic of
spontaneous experiences (8.7%) than serious material damage (for example, a
building on fire, considerable financial loss) (0.5%). This suggests that negative
events related to humans are particularly strong targets in spontaneous cases.
Similar patterns are observed in Schouten's (1982) study of the Rhine collection,
where he analysed a representative sample (15%) of cases (excluding PK). About
75% of the sample concerned negative events such as death, Injury and accident
while almost no cases concerned material damage. As with the Phantasms study, a
tendency to remember and report serious events more often than non-serious events
accounts for some of this pattern. However, the distribution of negative events in the
Rhine collection differs from the Phantasms collection, with the former having fewer
cases involving death of the target person (37.7% compared with 66.7%), but more
cases involving serious accidents and slight injuries. As Schouten points out, part of
this difference may be due to the Phantasms collectors' preference for apparition
cases.
In summary, Schouten's analyses of spontaneous case collections suggest that
negative events related to humans feature predominantly as "targets", although this
observation may be partly due to a reporting bias. It is significant that both the Rhine
and the Phantasms cases share this pattern despite the very different methods used
Approved For Release 2000/08/08 : 2504-RDP96-00789R003100120001-4

CPYRGAJproved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
to gather these collections. Evidently parapsychologists cannot inflict physical injury
on their experimental participants in order to simulate real-life spontaneous cases.
However, negative physical events are likely to have a negative emotional impact
both on the target person and on the percipient (especially if they are emotionally
close). Possibly, therefore, targets which have some strong negative emotional
impact on a person may have more success in a free-response experimental setting
than trivial or impersonal targets. Further, it might be expected that any emotional
impact is better than none, and so positive emotional targets could perhaps be
successfully used in experimental research - this might circumvent any researcher's
concern about the ethics of exposing experimental participants to unpleasant targets.
3. THEORETICAL SUGGESTIONS BY PARAPSYCHOLOGISTS
Although this is not a comprehensive review, I have tried to cover instances where
authors have made specific comments about likely successful targets.
Their
suggestions range from post hoc inferences based on the kinds of targets which were
successful in experimental studies to observations of what makes a good target in
areas of research related to parapsychology.
Le Shan (1977) criticizes parapsychologists for often neglecting to consider the
theoretical assumptions underpinning their research. There has been little discussion,
he feels, of what kind of information psi transmits even though there seems to be
wide agreement that psi does transmit information. As an example of how theorising
on this issue might influence our experimental design and choice of target material,
Le Shan considers the possibility that psi might depend on individual differences,
being better adapted for one purpose with one person and another purpose in a
different person. In this case, he suggests we should "customize" our targets by
examining experimental participants for their personal interests, philosophies,
preferred sensory modalities, and so on.
One of the few studies specifically to examine how target characteristics relate to psi
performance was conducted by Williams & Duke (1979), who go on to discuss
theoretical suggestions derived from their observations. Taking an evolutionary
perspective and asking what sort of information might have been most crucial to
communicate before language evolved in humans, they conclude that targets
reflecting "emotion, sex, survival, nature, food and other basic concerns might be
psychically perceived better than other types of targets" (p.15)
In a similar vein, a theoretical paper by Nash (1980) on the characteristics of psi
communication considers that, to be effective, psi communication must convey
"meaningful information". Also, one of the Maimonides experimental participants, in a
letter to Ullman and Krippner, gave her overall impressions of a dream telepathy
series in which she had recently participated. She felt that the more "potent and
unusual" the target material the better, because with subjects who might be
subconsciously afraid of telepathy this kind of target might be less likely to be "kept
out" (Ullman & Krippner, 1973). Perhaps unfortunately, it is very rare to find any
published opinions from the experimental participants who play a crucial part in
parapsychological reseptbase 2000/08/08 : CIA-RDP96-00789R003100120001-4
251

CPYRGHT
Approved For Release.2000/08/08 : CIA-RDP96-0078950031001200fa Much of
William Braud (1982) questioned the assumption wi
our research - that psi involves redundancy with our other known senses. For
instance, most of our experimentation involves primarily visual targets such as art
prints. Braud suggests that it would be useful if psi provided information which is not
immediately evident to our other known senses. Such non-evident information could
concern the larger relationships in which a target participates, for example its history.
Similarly, Gertrude Schmeidler in her 1971 PA Presidential Address stated that the
ESP target is not the physical stimulus variables, but the "meaning" of the target or
an "informational pattern" (Schmeidler, 1972). Braud conducted a pilot experiment to
test the theory of non-evident psi, where subjects were confronted with five identical
boxes containing, respectively, three control objects and two samples of hair cut from
one person's head. The hair samples were therefore related to each other, while the
control objects had no long-term association to a particular person. Subjects were
told which box was the "key" (one of the two boxes containing a hair sample) and,
while remaining unaware of the contents of all the boxes, were asked to rank the
remaining four boxes according to how "related" their contents were to the contents of
the key box. This study failed to achieve significant results, but this may still be an
idea worth further investigation.
The 1986 Esalen Conference discussed techniques to improve the reliable practical
use of psi abilities. Targ (1987) recommended that experimenters look for common
elements in the "psychic appearance" of targets (i.e. in mentations), and that they
should compose a glossary of typical target transformation errors. Tart (1987), at the
same conference, suggested that experimenters create a pool of "hot" targets - ones
that are consistently successful, either because they are correctly described or are
described in a recognisable fashion. In other words, what makes a good target would
be defined operationally.
So far, this section has considered research purely within parapsychology. Some
parapsychologists have taken a more interdisciplinary approach, however, and have
related the findings from other areas of research back to the question of what makes
a good GESP target.
Tart (1982) looked at how responses to targets are measured in conventional
psychophysiology, and asked what were the characteristics of a successful target in
this field of research: what kind of stimuli are most readily responded to, and easiest
to analyse. To be successful, a target stimulus in psychophysiology should stand
out from its background. For targets in parapsychological research, this may be
achieved by having the target stimulus occur suddenly, be discrete in time, and have
what Tart calls "psychic intensity" - the sense that the target is Important and
meaningful within the experimental context. Tart suggests that we could instruct our
experimental participants on the significance of the target in order to give it the
required meaningfulness. Psychic intensity could also reflect an intense event
happening to an agent - a methodology which Tart finds attractive. The idea that a
good target should stand out from its surroundings is strongly supported by the
psychological literature on human attention which I will be introducing later.
252
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGApproved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
A second area of research which has had some heuristic value for parapsychological
research concerns subliminal perception, or preconscious processing (Dixon, 1981).
Comparisons of psi and subliminal perception have noted that "right hemisphere"
processing facilitates subliminal perception (Roney-Dougal, 1981, 1986) - a
suggestion which has also been made for psi perception (e.g. Braud, 1975). This
could suggest that "right hemisphere targets" such as music, pictures and other
non-analytic targets might be preferable to "left-hemisphere" targets such as words
and numbers. Another parallel between psi and subliminal perception is that emotive
stimuli can evoke clear
autonomic responses in the percipient in both cases
(Roney-Dougal, 1986).
Serena Roney-Dougal feels that the use of negative emotional targets is both morally
and methodologically unsound, partly because some of her subjects reported
unpleasant experiences while receiving target impressions and might psi-miss with
this kind of target, and also because of the perceptual defence phenomenon seen in
subliminal perception. Sondow, Braud & Barker (1981) considered that "defensive"
subjects might be likely to psi-miss with unpleasant targets, and devised an
"Openness Questionnaire" to identify such subjects. They found no significant
difference between the "openness" of receivers who psi-hit and those who psi-missed
in a ganzfeld study. Unfortunately, no extensive description is made of the format of
the questionnaire, or of whether or not it measures perceptual defensiveness as seen
in subliminal perception or some other, unspecified, form of defensiveness.
In perceptual defence, a person may raise his or her recognition threshold for a
threatening or unpleasant stimulus - in other words, they perceive it less clearly.
Roney-Dougal interprets this as being due to the person's desire or motivation not to
perceive the threatening stimulus, a motivation which, she feels, may underlie
psi-missing also. However, Dixon reports experiments which suggest that the
perceptual defence effect, rather than representing the motivations of the
experimental participant, is best explained in physiological terms: emotive stimuli
cause changes in a person's arousal level which in turn affect the sensitivity of the
sensory receptors.
Whatever the mechanism of the effect of emotional stimuli on recognition thresholds,
it is clear that this effect is not uni-directional. One aspect of perceptual defence
which, it seems, tends to be overlooked is sometimes called vigilance. While some
people may raise their recognition thresholds to emotional stimuli, others may actually
lower them (Brown, 1961; Dixon, 1981). Without digressing too much on the reasons
for this apparent contradiction, it has been found that there is a correlation between
personality-type and a person's tendency to raise or lower his or her recognition
threshold, with extroverts raising their thresholds, and introverts lowering them
(Brown, 1961; Corcoran, 1965).
This has some interesting implications for
parapsychology. While Roney-Dougal felt that the raised recognition thresholds seen
in perceptual defence might be linked with the psi-missing of her own subjects with
negative emotional targets, other researchers have found the opposite (Delanoy,
1988), and the vigilance effect suggests that some parapsychological subjects could
even psi-hit with unpleasant targets. Donn Byrne (1961, 1963, 1964) has developed
a "repression-sensitization" scale which indicates whether
might, be
Approved For Release 2000/08/08 : CIA-RDP96-00789R0039681
0120001-4
25.3

exngstedet For Refeasiye doxigiant. perhaps parapsychologists could study the
mechanisms of psi-hitting and psi-missing With the Help of filis 29cale-4Crandall,
personal communication, 1988).
Having looked at popular literature, spontaneous cases, and theoretical suggestions
from parapsychologists on what might make a good target, I will now make some
inferences from areas of psychology which I consider to be relevant to this discussion.
(1) EMOTIONAL RESPONSES TO STIMULI
CPYRGHT
Mehrabian and Russell (1974) outline a theoretical approach to environmental
psychology (the study of the impact of the physical and social environment on man's
emotions, attitudes and behaviour). In their own words, "Evidence suggests that
there are three basic emotional responses (pleasure, arousal, and dominance) (the
dominance-submissiveness dimension refers to the degree of control. which
individuals feel they have over a situation or environment), combinations of which can
be used to describe adequately any emotional state (e.g. anxiety). By considering
their impacts on these basic emotional dimensions, the effects of diverse stimulus
components within or across sense modalities (e.g. color, pitch, texture, temperature)
can be readily compared" (preface, Mehrabian & Russell, 1974, [my italics]).
There is evidence of considerable intermodality of human response to stimulation -
that is, stimulation in one sensory modality may affect perception in another. For
instance, people who visualize auditory stimulation tend to agree in associating colour
names and mood adjectives with types of music: "Such persons were found to
visualize exciting music in bright forms or sharp and angular figures, and slow music
in rounder forms" ( p. 11, Mehrabian & Russell, 1974). The three basic emotional
responses to stimuli reported above (pleasure, arousal and dominance) are seen as
providing a measure with which to compare people's varied intermodal responses to
stimuli. This is relevant because it suggests that an additional important aspect to
our consideration of what might be expected to be salient features of a GESP target
is not only the actual physical characteristics of the target, but also the emotional
response (a combination of pleasure, arousal and dominance) which that target elicits
in the percipient.
Further, the theory may provide a methodological framework for the consideration of
the impact of various target characteristics on our experimental participants (Delanoy,
personal communication, 1988). A semantic differential scale is used to measure
people's emotional state in particular settings, or to measure their characteristic
emotions over time. Mehrabian and Russell's scale comprises 18 adjective pairs
describing various aspects of pleasure, arousal and dominance, and their subjects
are asked to mark on the scale the degree to which one or other of the adjective pair
most accurately reflects their feelings. Semantic differential scales have already been
used in parapsychology, though for different purposes than suggested here. McBain
et al (1970) used Osgood's Semantic Differential to find pairs of people with common
affective reactions to the same concept, though, contrary to their expectations, they
found no relation between the degree to which people agreed in their reactions to the
target stimulus and their GESP scores with that stimulus. Sondow, Braud & Barker
254
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
(1981) used Osgood's Semantic Differential as one of several measures of target
picture emotionality. However, it should be possible to make more extensive use of
the semantic differential, and it is planned to investigate further how a scale such as
Mehrabian & Russell's could be adapted to measure the reactions of
parapsychological subjects to targets and to provide a method to standardise
descriptions of successful targets.
The second aspect of Mehrabian and Russell's theory of environmental psychology
which may be relevant to our discussion about targets is their consideration of how
emotional reactions to physical environmental stimuli are related to the concept of
approach-avoidance. This they define broadly as including " ... physical movement
toward, or away from, an environment or stimulus, degree of attention,
exploration ... favourable attitudes such as ... preference or liking ... " (p.96, Mehrabian &
Russell, 1974). Arousal is seen as a mediator of approach-avoidance behaviour. A
literature review suggests that approach-avoidance is an inverted-U-shaped function
of arousal: an organism seeks an optimum level of arousal - whether or not it
approaches or avoids a stimulus depends on how arousing the stimulus is, and
extremely high or low levels of arousal are avoided.
In animals, there is a tendency to explore the unfamiliar. When the stimuli are
fear-inducing, animals repeatedly withdraw and approach the stimuli. Mehrabian and
Russell note that the animals are maintaining an optimum level of arousal with this
behaviour. Similar behaviour is seen in human children and adults (for references
see Mehrabian & Russell, 1974; Berlyne, 1960). Anecdotally, humans do seem to
have a penchant for voluntarily and repeatedly exposing themselves to negative
emotive and fear-inducing stimuli - hair-raising roller coaster rides and horror films, for
example.
The idea of approach-avoidance being mediated by arousal relates to the
consideration above (re perceptual defence and vigilance) of the merits of using
negative emotive targets in parapsychology. It suggests that people might have
some attraction to negative emotive targets insofar as these targets tend to increase
arousal.
Too much arousal, however, will cause people to withdraw from an
unpleasant target. On the other hand the use of neutral and bland GESP targets is
unlikely to arouse our experimental participants at all, consequently failing to elicit
approach. Of course, positive emotive targets would also be expected to influence
the arousal of our subjects and to elicit approach-avoidance behaviour.
A second area of psychological research which may make suggestions relevant to the
question of what makes a good GESP target concerns the characteristics of stimuli
which attract people's attention.
(2) STUDIES OF ATTENTION
While the theory discussed in the preceding section suggested that stimuli could be
described in terms of people's basic emotional responses to them, other research has
examined characteristics of the stimuli themselves, to see what stimulus features tend
lo attract attention.
Approved For Release 20807888 Poker8:00100eze00-14on
255

relevant to the discussion here as it could suggest the kind of target features which
might attract the attention of our experimental percipients in free-response GESP
tasks.
Berlyne (1970) noted the difficulty of even defining what is meant by the word
"attention". In his series of experiments (described in Berlyne, 1960) on curiosity,
conflict and arousal he seems to use an operational definition. These experiments
typically presented the subject simultaneously with several stimuli and observed the
percipient's eye fixation movements - the inference being that attention was given to
the stimulus which attracted most eye fixation (e.g. Berlyne, 1958). Other
experiments used a different measure of attention, allowing subjects to expose
themselves to very brief sights of stimulus pictures as many times as they liked -
presumably attention was attracted by the stimuli which were chosen to be seen most
often by subjects. The characteristics of stimuli which seemed to influence direction
of attention included: intensity; brightness; contrast; colour; novelty; complexity; and
incongruity.
Intensity. Berlyne (1960) states that the intensity of stimulation is seen in "the
frequency of nerve impulses and the number of fibers activated" (p.170) in the
reticular arousal system. Generally, large stimuli are more intense than small
stimuli; "warm" colours (e.g. red) are more intense and arousing than "cold"
colours (e.g. blue); high-frequency sounds are more intense than low
frequency sounds; and (in cats and monkeys) painful stimuli are most intense,
followed by proprioceptive, auditory, and visual stimuli respectively. Berlyne
found that attention was attracted by relatively intense stimuli - for example, to
larger than to smaller circles; to brighter than to dimmer visual stimuli.
Intensity is related to brightness, which also appears to attract attention.
Colour. Infants preferred looking at colour to looking at black and white
stimuli. Adults' attention was attracted more to a coloured stimulus than to a
white one (Berlyne, 1960).
Contrast. It was found that attention was attracted to a lighter stimulus on
black and medium grey backgrounds, and to a darker stimulus on a white
background. So, contrast with the background attracted attention. Above we
saw that brightness also attracts attention. When presenting subjects with
stimuli which differed from their background to equal extents but in different
directions, it was found that subjects were more likely to respond to the lighter
stimulus - that is, in the absence of a contrast difference, brightness was a
secondary determinant of attention (McDonnell, 1968).
Novelty. This can be defined as an unusual combination of parts of various
objects, or a change from the kind of stimulus to which the organism has
recently been exposed (Stotland & Canon, 1972). It has repeatedly been
found that novel stimuli attract more attention than familiar stimuli (e.g. Langer,
Fiske, Taylor & Chanowitz, 1976; Berlyne, 1958), though the effect of novelty
declines over time (perhaps as the subject habituates to the stimulus and
Approved For Release 2000/08/08 : C054RDP96-00789R003100120001-4

CPYRGHTApproved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
arousal drops). Berlyne (1960) considers attention to be most effectively
attracted by a stimulus whose novelty is often renewed. Novelty is related to
change or surprisingness of a stimulus (Stotland & Canon, 1972). On surprise,
Berlyne says "In experiments on learning, orienting behaviour (a set of
psychological and physiological responses through which the organism "sits up
and takes notice" when an aspect of its environment changes) is often found to
be strengthened by an unheralded change in experimental conditions" (p. 98,
Berlyne, 1960, [my italics]). This observation strongly resembles one made
from a parapsychological experiment by Roll & Harary (1972), that "some of
the more interesting results came when unannounced changes in the
experiment were made spontaneously", and similar results occurred "when
there was a last-minute change in the target materials" (p.4).
Complexity. This can be defined as the number of distinguishable parts which
a stimulus possesses, the degree of difference among these parts, and the
difficulty of integrating the parts involved (Stotland & Canon, 1972).
Incongruity, evidently related to both complexity and novelty, was found by
Berlyne (1958) to attract attention. Under examination, the distinction between
complexity and novelty grows blurred, and, as Stotland & Canon point out,
both involve stimulus change. Humans seem compelled to attend to stimulus
change - a response which might be expected to be evolutionarily adaptive.
Infants are attracted to relatively complex visual patterns and the attention of
adults is also determined partly by stimulus complexity (Berlyne, 1960; Jeffrey,
1968).
This research on the determinants of selective attention also states that, consistent
with the discussion earlier of approach-avoidance behaviour, people seek an optimum
level of arousal: either too much or too little arousal is unpleasant for individuals, and
factors such as stimulus novelty, complexity, intensity and incongruity are seen as
contributing to an organism's arousal.
The research outlined above tended to use fairly sterile tachistoscopic stimulus
presentation, however more recent studies of human causal judgement in social
situations have shown that these early findings can generalise to much more realistic
and complex situations. Shelley Taylor and Susan Fiske (1978), reviewing the
literature on the influence of salient stimuli on people's causal judgements, found that
bright, contrasting, moving and novel stimuli all attract attention in social situations
(e.g. Langer et. al., 1976; McArthur & Post, 1977). Movement can be regarded as
simply another aspect of stimulus complexity/novelty, and we have already seen that
stimulus change (a feature of movement) compels attention.
As it is not yet clear whether the process of psi perception is similar to perception
with our known senses it may be argued that the above findings from psychology on
attention-grabbing stimulus characteristics may not generalise to the "psi stimulus".
However, it would seem to be evolutionarily adaptive for any organism to attend to
bright, contrasting, moving and novel stimuli as such features may indicate either food
or threat to the organism. Insofar as psi perception may be an evolved attribute or
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
257
. .

CPYRGHT
stimulus features such as those outlined above.
For parapsychologists, these findings suggest that: 1. stimuli which are likely to
attract the attention of our experimental participants and consequently make
successful GESP targets may possess the following characteristics in some (as yet
unspecified) degree or form: movement, complexity, novelty, incongruity, contrast,
colour, brightness and intensity; and, 2.
these attention-determining target
characteristics must be present at moderate levels - too much and our subjects will
be overwhelmed, too little and they will be bored.
SOME LIMITATIONS OF THIS PAPER
Although this paper may seem to have rambled over a wide range of subjects, it has
mainly been restricted to a consideration of targets' physical features, and has not
examined in any depth the idea that "the target" is in part defined by the experimental
participant's own personal reactions to and interactions with it. Taylor & Fiske (1978)
considered some ways in which the salience of a stimulus may be influenced by
factors independent of the actual physical stimulus characteristics, and the following
table summarizes their findings.
Table 2 (after Taylor & Fiske, 1978)
Determinants of Selective Attention
Properties of Stimuli
Brightness
Contrast
Movement
Novelty
Properties of Situation
Environmental Cues
Instructional Set
Properties of Perceiver
Temporary Need States
Enduring Individual Differences in Traits, Reinforcement Schedules, Schemas
As Table 2 suggests, properties of a situation and properties of the perceiver may
influence what aspects of an individual's environment, or a free-response target,
appear as salient to any individual. For instance, if a person is hungry then food will
become especially salient to that individual. An Individual's cognitive schemata will
play some part in determining the direction of his or her attention (Stotland & Canon,
Approved For Release 2000/08/08 : CIA_RDP96-00789R003100120001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
SAIC
Science Applications International Corporation
An Employee-Owned Company
September 16, 1993
Defense Intelligence Agency
Bolling AFB, DT-5
Washington, D.C. 20340-6150
Attention:
SG1J
Reference:
MDA908-93-C-0004
Data Item No. A001
SG1J
Project Periodic Status Report
Dear
As you know, we are required by the official contract to deliver formal reports. So, here is our Periodic
Status Report deliverable dated 10 September 1993. We are required by contract to distribute the deliver-
ables according to the Contract Data Requirements List. Although it specifies the reports should be mailed
to RSQ-4, I am sending them to you for the final distribution:
DISTRIBUTION
Requiring Office
Final Report Copies
DIA/DT-5A
2
RSQ-4
1
Also, please contact whomever it is that tracks these deliverables to let them know you have received both
the draft and the final reports.
If you have any questions, please do not hesitate to contact me at (415) 325-8292.
Sincerely,
SCIENCE APPLICATIONS INTERNATIONAL CORPORATION
Elrom C. May
Edwin C. May, Ph.D.
Director, Cognitive Sciences Laboratory
cc Tom Albert/w/enclosure
Joe Angelo/w/enclosure
Betty Muzio/w/o/enclosure
file
1010 El Camino Real, Suite 330, P.O. Box 1412, Menlo Park, CA 94025 · (415) 325-8292
"Approved For Release 2000/08/08 '.CIA-RDP96-00789R0031009 20001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
SAIC
R
Science Applications International Corporation
An Employee-Owned Company
April 23, 1993
Defense Intelligence Agency
Bolling AFB, DT-5
Washington, D.C. 20340-6150
Attention:
SG1J
Reference:
MDA908-93-C-0004
SG1J
01-0187-03-3880-XXX
Status and Management Report
Dear
As you know, we are required by the official contract to deliver formal reports. So, here is our Status and
Management Report deliverable dated . We are required by contract to distribute the deliverables accord-
ing to the Contract Data Requirements List. Although it specifies the reports should be mailed to RSQ-4,
I am sending them to you for the final distribution:
DISTRIBUTION
Requiring Office
Final Report Copies
ODT-S
2
DPP-4
1
Also, please contact whomever it is that tracks these deliverables to let them know you have received both
the draft and the final reports.
If you have any questions, please do not hesitate to contact me at (415) 325-8292.
Sincerely,
SCIENCE APPLICATIONS INTERNATIONAL CORPORATION
C. May
Edwin C. May, Ph.D.
Director, Cognitive Sciences Laboratory
cc Tom Albert/w/enclosure
Joe Angelo/w/enclosure
Betty Muzio/w/o/enclosure
file
1010 El Camino Real, Suite 330, P.O. Box 1412, Menlo Park, CA 94025 · (415) 325-8292
Other SAIC Offices: Albuquerque, Boston, Colorado Springs, Dayton, Huntsville. Las Vegas, Los Angeles, McLean, Oak Ridge, Orlando, Palo Alto, Seattle, Tucson
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGHApproved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
1972). If a person has a phobia of spiders, then a picture of a spider will be very
salient to that person, while it may have no impact on another person who has a
phobia about water. If we as researchers instruct our experimental participants to
attend to one aspect of their environment, then that feature will become salient to
them.
So, we see that there are many influences on what makes target
characteristics grab attention, and it is unwise to restrict our view to physical target
characteristics alone. Nevertheless, these conclusions about the salience of physical
target characteristics remain valid so long as it is appreciated that they do not give
the whole picture.
SUMMARY AND CONCLUSIONS
The present paper considered theoretical ideas of what might be expected to make a
successful free-response GESP target.
1.
Popular literature on the training of psychic powers suggested that
emotional impact and human interest content made good targets. A survey of
patterns seen in spontaneous cases seemed to support these observations:
the bulk of the information transmitted concerned negative events related to
humans, though reporting bias accounted for some of this pattern. While
parapsychologists could not physically harm their subjects, it was suggested
that the emotional impact seen in spontaneous cases could be incorporated
into target material for experimental research, as observations from
spontaneous cases suggested that such targets might be expected to have
more success in an experimental setting than trivial or impersonal targets.
2. Varied theoretical suggestions by parapsychologists on what might make a
good target suggested that meaningful, emotional and potent targets could be
expected to be successful in GESP research. Studies of characteristics of
good targets in conventional psychophysiology suggested that targets in
parapsychology should stand out from their background.
This might be
achieved by having the target event occur suddenly, be discrete in time and be
"important" to the percipient.
Several parallels were noted between subliminal and psi perception. From
perceptual defence and vigilance effects seen in subliminal perception it was
suggested that, paradoxically, while some parapsychological subjects might be
expected to psi-miss with negative emotional targets, others might psi-hit with
such targets. It was suggested that the Repression-Sensitization Scale,
diagnostic of an individual's tendency to be defensive or vigilant, might be
useful to parapsychologists wishing to pursue these ideas.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
259

described.
Firstly, from environmental psychology it was suggested that
greater attention should be given to the subject's emotional response to the
target stimuli, and that, from the connection
between arousal and
approach-avoidance, the use of negative emotive stimuli could on the whole be
more likely to arouse our experimental participants and attract their attention
than neutral or bland stimuli. Secondly, research on attention found that
attention was attracted by stimuli which were relatively Intense, bright,
contrasting, colourful, novel, complex and incongruous - though only at
moderate levels.
Similarly, social psychology, using more complex and
realistic settings than attention research, found that bright, moving, contrasting
and novel stimuli attracted attention.
4. Some of the limitations of this paper were noted: there was a narrow focus
on physical target characteristics without considering inevitable influences of
properties of the perceiver and the environment on what aspects of the target
stimuli would appear salient to any individual.
Nevertheless, the findings
presented here were valid in their relevance to considerations of the target
question given that this paper does not present a comprehensive and
exhaustive overview of the subject of targets in parapsychological research.
We have seen that there Is some consistency in the suggestions of popular "psychic
training" literature, spontaneous cases, and parapsychologists' theoretical ideas on
the likely characteristics of successful GESP targets. These findings appear to
suggest that our targets should be psychologically sallent and physically sallent:
1. targets in parapsychological research should be meaningful, have emotional
impact and human interest - this may make them salient in the minds of our
experimental participants; 2. targets should also be physically salient by standing
out from their backgrounds - properties such as movement, novelty, complexity,
incongruity, brightness and contrast tend to make stimuli physically salient.
CPYRGHT
Approved For Release 2000/08/08 : CIA-ROP96-00789R003100120001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
REFERENCES
Berlyne, D.E. (1958) The influence of complexity and novelty in visual figures on
orienting responses. Journal of Experimental Psychology, 55, 289-296.
Berlyne, D.E. (1960) Conflict, Arousal, and Curiosity. McGraw-Hill, New York.
Berlyne, D.E. (1970) Attention as a problem in behaviour theory. In D.I. Mostofsky
(Ed.) Attention: Contemporary Theory and Analysis. Appleton-Century-Crofts, New
York.
Boswell, H.A. (1969) Master Guide to Psychism. Lancer, New York.
Braud, W.G. (1975) Psi-conducive states. Journal of Communication, 25, 142-152.
Braud, W.G. (1982) Nonevident psi. Parapsychology Review, 13, 16-18.
Brown, W.P. (1961) Conceptions of perceptual defence. British Journal of
Psychology Monograph, supplement no. 35.
Burns, J. (1981) Your Innate Psychic Powers. Sphere, London.
Byrne, D. (1961) The repression-sensitization scale: rationale, reliability, and validity.
Journal of Personality, 29, 335-349.
Bryne, D. (1963) Relation of the revised Repression-Sensitization Scale to measures
of self-description. Psychological Reports, 13, 323-334.
Byrne, D. (1964) Repression-Sensitization as a dimension of personality. In B.A.
Maher (Ed) Progress in Experimental Personality Research, Volume 1, Academic
Press, New York.
Corcoran, D.W.J. (1965) Personality and the inverted-U relation. British Journal of
Psychology, 56, 267-274.
Crandall, J. Conversation held in Psychology Department, University of Edinburgh,
spring 1988.
Delanoy, D. Conversation held in Psychology Department, University of Edinburgh,
spring 1988.
Delanoy, D. (1988) Characteristics of successful free-response targets: experimental
findings and observations. Paper submitted for presentation at the 1988 Annual
Convention of the Parapsychological Association
Approved For Release 2000/08/08 . CIA-RDP96-00789R003100120001-4
261

CPYRGHT
8
St. Paul, Minnesota.
Dixon, N. (1981) Preconscious Processing. John Wiley & Sons, Chichester.
Edwards, H. (no date) A Guide for the Development of Mediumship. Spiritualist
Association of Great Britain, London.
Jeffrey, W.E. (1968) The orienting reflex and attention in cognitive development.
Psychological Review, 75, 323-334.
Langer, E.J., Fiske, S., Taylor, S.E. & Chanowitz, B. (1976) Stigma, staring, and
discomfort: a novel-stimulus hypothesis. Journal of Experimental Social Psychology,
12, 451-463.
Le Shan, L. (1977) The purpose of psi. Journal of the Society for Psychical
Research, 49, 637-643.
McArthur, L.Z. & Post, D.L. (1977) Figural emphasis and person perception. Journal
of Experimental Social Psychology, 13, 520-535.
McBain, W.N., Fox, W., Kimura, S., Nakanishi, M., and Tirado, J. (1970)
Quasi-sensory communication: An investigation using semantic matching and
accentuated affect. Journal of Personality and Social Psychology, 14, 281-291.
McDonnell, P. (1968) Effects of intensity, contrast and novelty, on selective attention
and free-choice reaction time. Unpublished doctoral thesis, University of Toronto.
Cited in Berlyne, D.E. (1970) Attention as a problem in behavior theory. In D.I.
Mostofsky
(Ed.)
Attention:
Contemporary
Theory
and
Analysis.
Appleton-Century-Crofts, New York.
Mehrabian, A. & Russell, J.A. (1974) An Approach to Environmental Psychology.
MIT Press, Cambridge, Massachusetts.
Morris, R.L. (1977) The Airport Project: A survey of the techniques for psychic
development advocated by popular books. In J.D. Morris, W.G. Roll & R.L. Morris
(Eds) Research in Parapsychology 1976, Scarecrow Press, Metuchen, N.J.
Nash, C.B. (1980) Characteristics of psi communication.
Parapsychology Review,
11, 17-22.
Powell, I. (1979) How to be More Psychic. Sphere, London.
Roll, W.G. & Harary, K. (1976) Target responses during out-of-body experiences. In
report by G. Solfvin of 1976 SERPA Conference, Parapsychology Review, 7, 1-7.
Roney-Dougal, S.M. (1981) The interface between psi and subliminal perception.
Parapsychology Review, 12, 12-18.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
262

CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Roney-Dougal, S.M. (1986) Subliminal and psi perception: a review of the literature.
Journal of the Society for Psychical Research, 53, 405-434.
Schmeidler, G. (1972) Respice, Adspice, Prospice. In W.G. Roll, R.L. Morris & J.D.
Morris (Eds)
Proceedings of the Parapsychological Association (1971), 8,
Parapsychological Association, Durham, North Carolina.
Schouten, S.A. (1979a)
Analysis of spontaneous cases. Research Letter, 9,
Parapsychology Laboratory, University of Utrecht, 55-62.
Schouten, S.A. (1979b) Analysis of spontaneous cases as reported in 'Phantasms of
the Living'. European Journal of Parapsychology, 2, 408-455.
Schouten, S.A. (1982) Analysing spontaneous cases: a replication based on the
Rhine collection. European Journal of Parapsychology, 4, 113-158.
Schouten, S.A. (1986) A different approach for studying psi. In B. Shapin & L. Coly
(Eds) Current Trends in Psi Research (1984), Parapsychology Foundation, New York.
Sherman, H. (1960) Know Your Own Mind. Anthony, New York.
Sondow, N., Braud, L. & Barker, P. (1981) Target qualities and affect measures in an
exploratory psi ganzfeld.
Proceedings, 24th Annual Convention of the
Parapsychological Association.
Stotland, E. & Canon, L.K. (1972) Social Psychology: A Cognitive Approach. W.B.
Saunders, Philadelphia.
Targ, R. (1987) 1986 Esalen Conference. Parapsychology Review, 18, 6-8.
Tart, C.T. (1982) Physiological correlates of psi reception: some methodological
considerations. In Proceedings, PA & SPR Combined Jubilee and Centenary
Conference.
Tart, C.T. (1987) cited in Targ, R. 1986 Esalen Conference.
Parapsychology
Review, 18, 6-8.
Taylor, S.E. & Fiske, S.T. (1978) Salience, attention, and attribution: top of the head
phenomena. Advances in Experimental Social Psychology, 11, 249-288.
Ullman, M. & Krippner, S. (1973) Dream Telepathy. Penguin, Baltimore.
Williams, L.B. & Duke, M. (1979)
Qualities of free-response targets and their
relationship to psi performance. Proceedings, 22nd Annual Convention of the
Parapsychological Association .
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
263

.----
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
PSI COMMUNICATION IN THE GANZFELD
EXPERIMENTS WITH AN AUTOMATED TESTING SYSTEM
AND A COMPARISON WITH A META-ANALYSIS
OF EARLIER STUDIES
BY CHARLES HONORTON, RICK E. BERGER, MARIO P. VARVOGLIS,
MARTA QUANT, PATRICIA DERR, EPHRAIM I. SCHECHTER, AND
DIANE C. FERRARI
ABSTRACT: A computer-controlled testing system was used in 11 experiments on
ganzfeld psi communication. The automated ganzfeld system controls target selection
and presentation, subjects' blind-judging, and data recording and storage. Video-
taped targets included video segments (dynamic targets) as well as single images
(static targets). Two hundred and forty-one volunteer subjects completed 355 psi
ganzfeld sessions. The subjects, on a blind basis, correctly identified randomly se-
lected and remotely viewed targets to a statistically significant degree, z = 3.89, p =
.00005. Study outcomes were homogeneous across the 11 series and eight different
experimenters. Performance on dynamic targets was highly significant, z = 4.62, p
= . 0000019, as was the difference between dynamic and static targets, p = . 002.
Suggestively stronger performance occurred with friends than with unacquainted
sender/receiver pairs, p = . 0635. The automated ganzfeld study outcomes are com-
pared with a meta-analysis of 28 earlier ganzfeld studies. The two data sets are con-
sistent on four dimensions: overall success rate, impact of dynamic and static targets,
effect of sender/receiver acquaintance, and prior ganzfeld experience. The combined
z for all 39 studies is 7.53, p = 9 x 10-44.
Research on psi communication in the ganzfeld developed as the
result of earlier research suggesting that psi functioning is fre-
quently associated with internal attention states brought about
This work was supported by the James S. McDonnell Foundation of St. Louis,
Missouri, and by the john E. Fetzer Foundation of Kalamazoo, Michigan.
We wish to thank Marilyn J. Schlitz, Peter Rojcewicz, and Rosemarie Pilkington
for their help in recruiting participants; Daryl J. Bem of Cornell University and
Donald McCarthy of St. Johns University for helpful comments on an carlier draft
of this paper; Edwin C. May of SRI International for performing the audio spectrum
analysis; and Robert Rosenthal of Harvard University for suggestions concerning
data analysis. We also wish to thank several PRL colleagues who contributed in var-
ious ways to the work reported here: Nancy Sondow for assistance in the preparation
relaxation exercise and instruction tape that was used throughout, and George
Hansen and Linda Moore who served frequently as lab senders. Hansen also pro-
vided technical assistance and conducted a data audit resulting in the correction of
several minor errors that appeared in a version of this report presented at the 32nd
Annual Convention of the Parapsychological Association. Finally, we thank the 241
volunteer participants for providing us with such interesting data.
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

100
The Journal of Parapsychology
Approved For Release 2000/08/08 _CJARDP 96-00789R
through dreaming, hypnosis, meditation, and similar naturally oc-
curring or artificially induced states (Braud, 1978; Honorton, 1977).
This generalization, based on converging evidence from sponta-
geous case studies, clinical observations, and experimental studies,
Bd to the development of a low-level descriptive model of psi func-
Boning, according to which, internal attention states facilitate psi de-
@ction by attenuating sensory and somatic stimuli that normally
Mask weaker psi input (Honorton, 1977, 1978). This "noise-reduc-
Son" model thus identified sensory deprivation as a key to the fre-
quent association between psi communication and internal attention
Pates, and the ganzfeld procedure was developed specifically to test
the impact of perceptual isolation on psi performance.
Fifteen years have passed since the initial reports of psi com-
Nunication in the ganzfeld (Braud, Wood, & Braud, 1975;
Bonorton & Harper, 1974; Parker, 1975). Dozens of additional psi
Onzfeld studies have appeared since then, and the success of the
gradigm has triggered substantial critical interest. Indeed, there is
least one critical review or commentary for every ganzfeld study
"porting significant evidence of psi communication (Akers, 1984;
Alcock, 1986; Blackmore, 1980, 1987; Child, 1986; Druckman &
Swets, 1988; Harley & Matthews, 1987; Harris & Rosenthal, 1988;
Honorton, 1979, 1983, 1985; Hovelmann, 1986; Hyman, 1983,
(85, 1988; Hyman & Honorton, 1986; Kennedy, 1979; McClenon,
1086; Palmer, 1986; Palmer, Honorton, & Utts, 1989; Parker &
@klund, 1987; Rosenthal, 1986; Sargent, 1987; Scott, 1986;
Stanford, 1984, 1986; Stokes, 1986; Utts, 1986).
Of the many controversies spanning the history of parapsycholog-
ical inquiry, the psi ganzfeld domain is unique in three respects.
Bujst, the central issue involves the replicability of a theoretically
boed technique rather than the special abilities of exceptional in-
dasiduals (Honorton, 1977). Second, meta-analytic techniques have
ben used to assess statistical significance, effect size, and potential
tleats to validity (Harris & Rosenthal, 1988; Honorton, 1985;
Homan, 1985, 1988; Rosenthal, 1986). Third, investigators and crit-
ihave agreed on specific guidelines for the conduct and evaluation
offuture psi ganzfeld research (Hyman & Honorton, 1986).
The Automated Ganzfeld Testing System
Psi ganzfeld experiments typically involve four participants. The
subject (or receiver, R) attempts to gain target-relevant imagery
while in the ganzfeld; following the ganzfeld/imagery period, R
Psi Communication in the Ganzfeld
101
tries-on a blind basis-to identify the actual target from among
four possibilities. A physically isolated sender (Se) views the target
and attempts to communicate salient aspects of it to R. Two exper-
imenters (Es) are usually required. One E manages R, elicits R's ver-
bal report of ganzfeld imagery (mentation), and supervises R's blind
judging of the target and decoys; a second E supervises Se, and ran-
domly selects and records the target.
We developed an automated ganzfeld testing system ("autoganz-
feld") to eliminate potential methodological problems that were
identified in earlier ganzfeld studies (Honorton, 1979; Hyman &
Honorton, 1986; Kennedy, 1979) and to explore factors associated
with successful performance. The system provides computer control
of target selection and presentation, blind judging, subject feedback,
and data recording and storage (Berger & Honorton, 1986). A com-
puter-controlled videocassette recorder (VCR) accesses and auto-
matically presents target stimuli to Se. A second E is required only
for assistance in target selection The system includes an experimen-
tal design module through which E specifies the sample size and
status of a new series.
The system was designed to enable further assessment of factors
identified with successful performance in earlier ganzfeld studies.
Differences in target type and sender/receiver acquaintance seem to
be particularly important. Significantly better performance occurred
in studies using dynamic rather than static targets. Dynamic targets
contain multiple images reinforcing a central theme, whereas static
targets contain a single image. Also, studies permitting subjects to
have friends as their senders yielded significantly superior perfor-
mance compared to those requiring subjects to work, with laboratory
senders. (See "Comparison of Study Outcomes with Ganzfeld Meta-
Analysis" in the Results section.)
The autoganzfeld system uses both dynamic and static targets.
The dynamic targets are excerpts from films; static targets include
art work and photographs. Receivers may, if they choose, bring
friends or family members to serve as their senders; a session setup
module registers the sender type and other session information.
In this report, we present the results of the 11 autoganzfeld
series conducted between the inauguration of the experiments in
February, 1983, and September, 1989, when funding problems.
required suspension of the PRL research program.1 We focus on
' This article conforms to the reporting guidelines recommended by Hyman and
Honorton (1986). Because of the size of this database, however, it is not practical to
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789F
O
120001-4
1

102
The Journal of Parapsychology
(1) evidence for psi in the autoganzfeld situation, (2) the impact of
dynamic versus static targets, (3) the effects of sender/receiver ac-
quaintance, (4) the impact of prior psi ganzfeld experience, and
(5) a comparison of these four factors with the outcomes of earlier
nonautomated psi ganzfeld experiments. Our findings on demo-
graghic, psychological, and target factors will be presented in later
reports.
Subjects
The participants are 100 men and 141 women ranging in age
from 17 to 74 years (mean = 37.3, SD = 11.8). This is a well-
educated group; the mean formal education is 15.6 years (SD =
2.0₽
Our primary sources of recruitment include referrals from col-
leagues (24%), media presentations concerning PRL research (23%),
friends or acquaintances of PRL staff (20%), and referrals from
othe participants (18%).
Belief in psi is strong in this population. On a seven-point scale
what "1" indicates strong disbelief and "7" indicates strong belief
in jas, the mean is 6.20 (SD = 1.03); only two participants rated
their belief in psi below the midpoint of the scale. Personal experi-
enc& suggestive of psi were reported by 88% of the subjects; 80%
repeated ostensible telepathic experiences. Eighty percent of the
par Ripants have had some training in meditation or other tech-
niques involving internal focus of attention.
Participant Orientation
CA-RDP96:00789R003
stial contact. New participants receive an information pack be-
fore Their first session. The information pack includes a 55-item per-
sona history survey (Participant Information Form [PIF]; Psycho-
phySal Research Laboratories, 1983), Form F of the Myers-Briggs
Type Indicator (MBTI; Briggs & Myers, 1957), general information
abono the research program, and directions for reaching PRL. Par-
ticipants usually return the completed questionnaires before their
firstSession. However, if new participants are scheduled on short
noti@, they either complete the questionnaires at PRL or, in a few
cases, at home after the session.
include the data in an appendix to the report. Instead, we will supply the data to
qualified investigators in a Lotus-compatible, MS-DOS computer disk file. There is a
small fee to cover materials and mailing. Address inquiries to the Journal.
Psi Communication in the Ganzfeld
103
Whenever possible, new participants are encouraged to come in
for a preliminary orientation session, prior to their first PRL ganz-
feld session. The orientation serves as a "get acquainted" session for
participants and the PRL staff, and introduces participants to the
PRL program and facility. Participants who avail themselves of this
option generally complete the MBTI and PIF questionnaires during
the orientation session. We inform new participants that they may
bring a friend or family member to serve as their sender. When a
participant chooses not to do so, a PRL staff member serves as
sender. We encourage participants to reschedule their session rather
than feel they must come in to "fulfill an obligation" if they are not
feeling well.
Session orientation. We greet participants at the door when they
arrive and attempt to create a friendly and informal social atmos-
phere. Coffee, tea, and soft drinks are available. E and other staff
members engage in conversation with R during this period. When
a laboratory sender is used, time is taken for sender and receiver to
become acquainted.
If the participant is a novice, we describe the rationale and back-
ground of the ganzfeld research, and we seek to create positive ex-
pectations concerning R's ability to identify the target. This infor-
mation is tailored to our perception of the needs of the individual
participant, but it generally includes four elements: (1) a brief re-
view of experimental, clinical, and spontaneous case trends indicat-
ing that ESP is more readily detected during internal attention states
such as dreaming, hypnosis, and meditation (Honorton, 1977),
(2) the notion that these states all involve physical relaxation and
functional sensory deprivation, suggesting that weak ESP impres-
sions may be more readily detected when perceptual and somatic
noise is reduced, (3) the development of the ganzfeld technique to
test this noise-reduction hypothesis, and (4) the long-term success of
the ganzfeld technique as a means of facilitating psi communication
in unselected subjects.
We encourage "goal orientation" and discourage excessive "task
orientation" during the session; this is especially emphasized with
participants who appear to be anxious or overly concerned about
their ability to succeed in the ganzfeld task. We discourage partici-
pants from analyzing their mentation during the session, and tell
them that they will have an opportunity to analyze their mentation
during the judging procedure. They are encouraged to adopt the
role of an outside observer of their mental processes during the
ganzfeld. Again, this is emphasized with those who appear anxious
CPYRGHT
Approved For Relea
/08/08 : CIA-RDP96-00789R003
00120001-4
:

104
The Journal of Parapsychology
about their performance; they are advised to relax, follow the taped
instructions, and to simply allow the procedure to work. We inform
participants that they may experience various types of correspond-
en between their mentation and the target; they are told that they
magexperience direct, literal correspondences to the target, but that
they should also be prepared for correspondences involving distor-
tions or transformations of the target content, cognitive associations,
andysimilarities in emotional tone. Finally, we orient new partici-
pas to where Se and E will be located during the session.
Approved For Release 2000/98/08 : Gt
METHOD
Layout and Equipment
and Se are sequestered in nonadjacent, sound-isolated and
ele@ically shielded rooms. Both rooms are copper-screened, and
are
4 ft apart on opposite sides of E's monitoring room, which pro-
vid@ the only access. R and Se remain isolated in their respective
rooms until R completes the blind-judging procedure.
2's room is an Industrial Acoustics Corp., IAC 1205A Sound-
Isolation Room, consisting of two 4-inch sheetrock-filled steel
pane's. The two panels are separated by a 4-inch air space, for a
total thickness of one foot.
She inside walls and ceiling of Se's room are covered with 4-inch
Sonex® acoustical material, similar to that used in commercial
bro@Icast studios. A free-standing Sonex-covered plywood barrier
(5 foowide by 8 ft high) positioned inside the sender's room, between
Se'schair and the acoustical door, blocks sound transmission
through the door frame. Figure I shows the floor plan of the ex-
peromental rooms.
E occupies a console housing the computer system and other
equRment. The computer is an Apple II Plus with two disk drives,
a printer, and an expansion chassis. The computer peripherals in-
cluc a real-time clock, a noise-based random number generator
(RN), a Cavri Interactive Video Interfaces», an Apple game pad-
dle, land a fan. Other equipment includes a color TV monitor, the
VCR used to access and display targets, and three electrically iso-
lated audiocassette recorders. One audiocassette recorder presents
audio stimuli (prerecorded relaxation exercises, session instructions,
and white noise). Another plays background music during the ex-
perimental setup. The third records R's ganzfeld mentation and
Psi Communication in the Ganzfeld
105
Industrial Acoustics
Corp. 120SA Sound
Isolation Room
Double wall with 4"
-Sonex Acoustical
Padding and acousti-
cal door
RECEIVER
SENDER
E's cquipment
console
EXPERIMENTER
SCALE
1
$ ft
Figure 1. Floor plan of experimental suite.
judging period associations. There is two-way intercom communi-
cation between E and R. One-way audio communication from R to
Se allows Se to listen to R's ganzfeld mentation.
Receiver Preparation
R sits in a comfortable reclining chair in the IAC room. Se keeps
R company while E prepares R for visual and auditory ganzfeld
stimulation. Translucent hemispheres are taped over R's eyes with
Micropore® tape. Headphones are placed over R's ears. A clip-on
microphone is fastened to R's collar. A 600-watt red-filtered flood-
light, located approximately 6 ft in front of R's face, is adjusted in
intensity until R reports a comfortable, shadow-free, homogeneous
visual field. White noise level is similarly adjusted; R is informed
that the white noise should be as loud as possible without being an-
noying or uncomfortable. The ganzfeld light and white noise inten-
sity are adjusted from E's console after R and Se are sequestered in
their respective rooms.
Sender Preparation
Se sits in a comfortable reclining chair in the sender's room. Se
faces a color, TV monitor, wearing headphones. During the session,
Se can hear R's mentation report through one headphone; if dy-
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

IVO
i ne journal of rarapsychology
namic targets are used, Se hears the target audio channel through
the other headphone.
Series Manager Setup Procedures
Apprayed
E accesses the autoganzfeld computer program through the Se-
Es Manager software. Series Manager is a password-protected, menu-
given control program. It provides the only means through which
experimenter may specify parameters for the series design, reg-
ister new participants in the series, set up a session, and run a ses-
son. The Series Manager menu is accessed through entry of a private
(and nonechoing) password.
Release 2000/08/08 : GIA RDR
D Series design. A valid series design must exist before sessions can
b@ run in an experimental series. This is done through the Series
Manager "design" module. The design module prompts E to specify
ff'e type of series (pilot, screening, or formal), the number of
participants, the maximum number of trials per participant, the
tal number of trials per series, and the series name. There is no
povision for changing the series design once it is accepted by E.
esign parameters are saved in a disk file; they are passed to the
eoperimental program at the beginning of the session.
Participant registration. When R is new to a series, E accesses
Articipant Registration" from the Series Manager menu before the
sassion. E is prompted to enter R's name and identification number.
e module verifies that the maximum number of participants
specified in the design is not exceeded. (An error message appears
ifan attempt is made to register more participants than are speci-
fill in the design; then, control is returned to the Series Manager
n@nu.)
00 Session setup. E then selects "Session Setup" from the Series Man-
aggr menu. E is prompted to enter R's name and the program ver-
iftes that R has not already completed the maximum number of
troals specified in the design module. (An error message appears if
a participant has completed the number of sessions allowed for the
spies or has not been properly registered; control is then returned
tothe Series Manager menu.) E enters Se's name and the sender
type: lab, lab friend, or friend. Lab senders are PRL staff members
W
Rose acquaintance with the participant is limited to the experi-
ment. Lab friend refers to PRI. staff senders who have some social
acquaintance with R outside the laboratory. Friend senders are friends
or family members of the participant. Finally, E enters the ganzfeld
light and noise intensity levels and his or her initials. E then leaves
Psi Communication in the Ganzfeld
107
the monitoring room while another PRL staff person supervises tar-
get selection.
Targets
The system uses short video segments (dynamic targets) and still
pictures (static targets) as targets. Dynamic targets include excerpts
from motion pictures, documentaries, and cartoons. Static targets
include art prints, photographs, and magazine advertisements.
There are 160 targets, arranged in judging sets of four dynamic
or four static targets. The sets were constructed to minimize simi-
larities among targets within a set. The targets are recorded on four
one-half-inch VHS format videocassettes; each videocassette con-
tains 10 target sets (5 dynamic and 5 static). A signal recorded on
an audio track of each videocassette allows computer access of the
targets. Target display time-to Se during each sending period and
to R during the judging period-is approximately one minute;
blank space added to briefer targets insures that the VCR remains
in play mode for the same length of time for all targets.
Preview packs. The video display format of the autoganzfeld tar-
gets does not permit simultaneous viewing of the entire target set
during the judging procedure as is done in many nonautomated
ganzfeld studies. Each target set is therefore accompanied by a pre-
view pack containing brief excerpts of all four targets in the set; this
gives R a general impression of the range of target possibilities. R
views the preview pack at the beginning of the judging procedure;
it runs approximately 30 sec.
Target Selection
The target selector (TS) is a PRL staff member who has no con-
tact with either E or R until after the blind-judging procedure. TS
is needed to load the videocassette containing the target into the
VCR. TS is informed which of the four videocassettes contains the
target, but remains blind to the target's identity. If Se is a staff
member, Se serves this role; otherwise, a staff member not involved
in the session serves as TS. (In the latter case, Se and R are seques-
tered in their respective rooms before TS enters the monitoring
room.)
The Series Manager program prompts TS to press a key on the
computer keyboard. A program call to the hardware RNG obtains
the target value (a number between 1 and 160) and stores it in, com-
CPYRGHT
Approved For Release 2000/08/08 : CTA-RDP96-00789R003100120001-4

108
The Journal of Parapsychology
puter memory.2 The program determines the target set and video-
cassette number from the target value. The videocassette number is
displayed on the monitor, and TS is prompted to insert it into the
CR. The program verifies that the correct videocassette has been
iserted and clears the monitor screen; if the videocassette is not
correct, an error message prompts TS to insert the correct video-
Pssette.
TS places a cardboard cover over the VCR's front panel to con-
coil the digital counters and VU meters. Finally, TS leaves the mon-
aring room with the three remaining videocassettes, knocking
three times on the monitoring room door as a signal for E to return.
R
Haxation Exercises and Ganzfeld Instructions
D
ER and Se undergo a 14-min prerecorded relaxation exercise be-
fge the mentation/sending period. This provides a unique shared
experience for R and Se before the ESP task. The relaxation exer-
ciste includes progressive relaxation exercises and autogenic phrases
(Jobson, 1929; Shultz, 1950). Ganzfeld instructions are recorded
after the relaxation exercise. The instructions and relaxation exer-
cil are delivered in a slow, soothing but confident manner with
ocean sounds in the background. The style of presentation is similar
too hypnotic induction procedure. The ganzfeld instructions to R,
which are also heard by Se, are as follows:
DP96-00789R003100120004-4
During this experiment we want you to think out loud. Report all of the
Omages, thoughts, and feelings that pass through your mind. Do not
Cling to any of them. Just observe them as they go by. At some point
coluring the session, we will send you the target information. Do not try
o anticipate or conjure up this information. Just give yourself the sug-
gestion, right now-in the form of making a wish-that the information
Will appear in consciousness at the appropriate time. Keep your eyes
open as much as possible during the session and allow your conscious-
Bess to flow through the sound you will hear through the headphones.
One of us will be monitoring you in the other room. Now get as com-
fortable as possible, release all conscious hold of your body, and allow
to relax completely. As soon as you begin observing your mental proc-
Asses, start thinking out loud. Continue to share your thoughts, images,
and feelings with us throughout the session.
2 An exception occurs in the two target comparison series (Series 301 and 302).
See pp. 112-113.
Psi Communication in the Ganzfeld
109
Mentation/Sending Procedures
Receiver mentation report. After the relaxation exercise and in-
structions, R listens to the white noise through headphones for 30
minutes. R reports whatever thoughts, images, and feelings occur in
the ganzfeld. The mentation report is monitored by E and Se from
their respective rooms. The mentation report is tape recorded, and
E takes detailed notes for review from R prior to judging.
Target presentation and sender procedures. A Cavri Video Interface
automates computer access and control of targets from a JVC BR-
6400U VCR. An electronic video switcher selectively routes the
video output (VCR or computer text mode) to three color TV mon-
itors, one each for E, R, and Se. E's and R's monitors remain in
computer text mode until the judging period. During each of the
six sending periods, Se's TV monitor is switched from computer
text to VCR mode.
At the beginning of each sending period, Se's monitor displays
the prompt, "Silently communicate the contents and meaning of the
target to [R's first name]." Se views the target and attempts to com-
municate its contents to R. Se mentally reinforces R for- target-
related associations and mentally discourages R when the mentation
is unrelated to the target.
Judging Procedure
After the mentation period, E turns off the ganzfeld light and
reads back R's mentation from the session notes. R remains in ganz-
feld during the mentation review to minimize any abrupt shift in
state. E's and R's TV monitors are switched into VCR mode by the
computer, which also prompts Se to "Silently direct [R's first name]
to select the target that you saw." Se's TV monitor remains blank
(computer mode) during this period.
R removes the eye covers and views the preview pack. From
their respective rooms, R and E then view the four potential targets
(the actual target and three decoys), which are presented in one of
four random sequences. R, viewing each candidate, associates to the
item as though it were the actual target, describing perceived simi-
larities between the item and the ganzfeld mentation. While R as-
sociates to each candidate, E points out potential correspondences
that R may have overlooked.' R views any of the target candidates
as often as desired before proceeding to the judging task.
' This applies to Pilot Series 3, Novice Series 103-105, and to Experienced Series
CPYRGHT
Approved For Release 2000/08/08 . CIA-RDP
6-00789R003100120001-4

Tur Jourmm o Turapsycnowgy
A 40-point rating scale then appears on R's TV monitor. The
scale is labelled 0% on the left and 100% on the right. Using a com-
puter-game paddle to move a pointer horizontally across the rating
scale, R indicates the degree of similarity between his ganzfeld men-
tation and each potential target. E and Se view R's ratings on their
monitors. The program checks for ties, and, if they occur, R re-rates
the four candidates to obtain unique ratings for each. The program
then converts R's ratings into ranks. A rank of 1 is assigned to the
candidate R believes has the strongest similarity to his ganzfeld men-
tacion; a rank of 4 is given to the candidate R believes is least like
hisoganzfeld experience.
Feedback and Post-Session Procedures
After R finishes judging, Se leaves the sender's room and enters
R'Sroom with E. Se reveals the actual target, which the computer
augmatically displays on R's TV monitor. The session data are writ-
terRo a floppy disk file.
Following feedback, E is prompted to backup the series data
di@ The target videocassette is then automatically wound to a po-
sittgh near the center of the videocassette (frame 50,000). E. selects
"Analysis" from the Series Manager menu and obtains a hardcopy
pr @out of the session data file. The printout includes: the file
name, R's name and ID number, series type, session number, Se's
naque, E's initials, date and start time, target number, target position
in The set, R's target ranking, the standardized target rating (z
sco:), target judging sequence, target name, target type and set
nu ?? ber, sender type, light and white noise levels, finish time, and
op@nal experimenter's comments. The printout is attached to E's
notes on R's mentation and placed in a ring binder containing all
such information for the series. The audio tape of the session is sim-
ila filed.
Experimenters
100789R003:00+20001-4
on
ight Es contributed to the autoganzfeld database. Honorton,
of the originators of the psi ganzfeld technique, has conducted
psi
ganzfeld experiments over a 16-year period. Derr and Varvoglis
2014and 302. It does not apply to the earlier series (Pilot Series 1-2; Novice Series
101-102; or Experienced Series 301). This practice was initiated because participants
frequently failed to identify obvious correspondences between their mentation and
target elements.
Psi Communication in the Ganzfeld
111
worked with Honorton at Maimonides Medical Center and were
trained by him. Berger is primarily responsible for the technical im-
plementation of the autoganzfeld system. He trained Honorton,
Derr, Varvoglis, and Schechter in its use. Honorton trained Quant,
Ferrari, and Schlitz in the use of the autoganzfeld system.
Experimental Series
Altogether, 241 participants contributed 355 sessions in 11 se-
ries. To fully address the issue of selective reporting, we included
every session completed from the inauguration of the experiments,
in February, 1983, to September, 1989, when the PRL facility wasq
closed. Thus, this database has no "file-drawer" problem (Rosenthal,
1984).
The studies include three pilot series and eight formal series.
Five of the formal series were single-session studies with novice par-
ticipants. The remaining three formal series involved experienced
participants.
Pilot Series
Series 1. This initial pilot series was conducted during the devel
opment and testing of the autoganzfeld system. It served to test sys-
tem operation, to detect and correct programming errors, and to
fine-tune session timing functions. Nineteen subjects contributed 220
sessions as Rs. Seven, including PRI. staff members, had prior ex
perience as Rs in nonautomated ganzfeld studies at Maimonide
Medical Center. The remaining 12 Rs were novices with no prior
ganzfeld experience. Series sample size was not specified in advance;
the series continued until we were satisfied that the system was op-
erating reliably.
Series 2. This pilot series was designed by Berger in an attempt
to avert potential displacement effects and subject judging problem
by having E rather than R serve as judge. R received feedback only
to the actual target. Four participants contributed to this series
Nine of the planned 50 sessions were completed before Berger's de
parture from PRL when this series was discontinued.
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
'Berger, Schechter, and Varvoglis have doctorate degrees in psychology. Quant
holds a masters degree in counselling psychology, and Ferrari has a bachelors degree
in psychology. Schlitz has conducted independent ganzfeld and remote-viewing re-
search in other laboratories and has a masters degree in anthropology.
Ed For Release 2000/08/08

112
The Journal of Parapsychology
Series 3. This pilot series was a practice series for participants
who completed the allotted number of sessions in ongoing formal
series but who wanted additional ganzfeld experience. This series
also includes several demonstration sessions when TV film crews
were present and provided receiver experience for new PRL staff.
The sample size was not preset.
Nogice ("First-Timers") Series
„The identification of characteristics associated with successful in-
itial performance was a major goal of the PRL ganzfeld project
(Hbnorton & Schechter, 1987). Except for Series 105, each novice
seffles includes 50 ganzfeld novices, that is, participants with no
prior ganzfeld experience. Each novice contributed a single ganz-
fe@ session. Most novices had not participated in any psi experiment
pror to the novice series.
-Series 101. This is the first novice series.
Series 102. Beginning with this series, R was prompted after the
meAtation period to estimate the number of minutes since the end
of the relaxation/instructions tape.
Series 103. Starting with this series, Rs were given the option of
hating no sender (i.e., "clairvoyance" condition). Only four partici-
pauts opted to have no sender.
Deries 104. A visiting scientist (Marilyn Schlitz) served as E in
seon sessions and as Se in six sessions with subjects from The Juil-
lia) School in New York.
006/08/08 : CIA RDP96-00789R093100$200012-4
-Series 105. This series was started to accommodate the overflow
of juilliard students from Series 104. The sample size was set to 25.
Si Osessions were completed at the time the PRL program was sus-
pooded. (There were 20 Juilliard students altogether. Sixteen were
in
Beries 104 and four were in Series 105.)
Experienced Subjects Series
Series 201. This series involved especially promising subjects.
The number of trials was set to 20. Seven sessions by three Rs were
completed at the time the PRL program was suspended.
Series 301. This series compared dynamic and static targets.
Sample size was set to 50 sessions. Twenty-five experienced subjects
each contributed two sessions. The autoganzfeld program was mod-
ified for this series so that each R would have one session with dy-
Psi Communication in the Ganzfeld
113
namic targets and one session with static targets. Subjects were in-
formed of this only after completing both sessions.
Series 302. This series used a single dynamic target set (Set 20)
In earlier series, Target 77 ("Tidal Wave Engulfing Ancient City"
had an especially strong success rate while Target 79 ("High-Speed
Sex Trio") had never been correctly identified. We made two pro-
gram modifications for this series. The target selection ("RandomO
ize") routine was modified to select only targets in Set 20, and then
VCR tape-centering routine was modified to wind the videotape to-
a randomly selected position between frame numbers 85,000 and0
95,000. The second modification insured that E could not be cued;
perhaps unconsciously, by the time required to wind the tape from
its initial position to the target location.
The study involved experienced Rs who had no prior experience
with Set 20. Each R contributed one session. Participants were una
aware of the purpose of the study or that it was limited to one targets
set. The design called for the series to continue until 15 sessions
were completed with each of the two targets of interest. Twenty-five
sessions were completed when the PRL program was suspended.
Statistical Analysis
Except for two pilot series, series sample sizes were specified in
advance. Our primary hypothesis was that the observed successo
rate-the proportion of correctly identified targets-would reliably
exceed the null hypothesis expectation of .25. To test this hypoth-
esis, we calculated the exact binomial probability for the observed
number of direct hits (ranks of 1) with p = . 25 and q = . 75. Once
the basis of the overwhelmingly positive outcomes of earlier studies,
we preset alpha to .05, one-tailed.
We also tested two secondary hypotheses, based on patterns of.
success in earlier psi ganzfeld research. These are: (1) that dynamica
targets are significantly superior to static targets, and (2) that per-
formance is significantly enhanced when the sender is a friend of R,
compared to when R and Se are not acquainted. We initially
planned to test these hypotheses by chi-square tests, a trial-based
analysis. However, a consultant (Dr. Robert Rosenthal) suggested-
that a t test using the series as the unit would be a more powerful
test of these hypotheses, and we have followed his recommendation.
The remaining analyses are exploratory.5
" The statistical analyses in this report were performed using SYSTAT (Wilkin-
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

The journal of tarapsychology
TABLE. I
OUTCOME BY SERIES
Series
Hits
Effect size
type
N
subjects
N
trials
N
%
(h)
Z
Pilot
19
22
8
36
.25
.99
2
Pilot
4
9
3
33
.18
.25
3
Pilot
25
36
10
28
.07
.22
101
Novice
50
50
12
24
.02
-. 30
102
Novice
50
50
18
36
.24
1.60
103
Novice
50
50
15
30
.11
.67
104
Novice
50
50
18
36
.24
1.60
105
Novice
6
6
4
67
.87
1.78
201
Experienced
3
7
3
43
.38
.69
301
Experienced
25
50
15
30
.11
.67
302
Experienced
25
25
16
64
.81
3.93
Overall
241
355
122
34
.20
3.89
NoteThe z scores are based on the exact binomial probability with p = . 25
and
.75.
RESULTS
Overall Success Rate
Ganzfeld hit rate. There were 241 participants, who contributed
355 auitoganzfeld sessions. The 122 direct hits (34.4%) yield an exact
binogsial p of .00005 (z = 3.89). The effect size, Cohen's h (Cohen,
1977 is .20. The 95% confidence interval (CI) is a hit rate from
30%-0 39%. Because this level of accuracy would occur about one
timeg3 20,000 by chance, we reject the null hypothesis. (See Table
1.) 0
Access rate by series. Of the 11 series, 10 yield positive outcomes.
TheChean series effect size is .29, SD = . 29, t (10) = 3.32.
Homogeneity of effect sizes. Traditionally, psi investigators have
been preoccupied by whether there is a significant nonzero effect.
An equally important issue, however, is the size of the effect. There
is a growing tendency among behavioral scientists to define replic-
abilit in terms of the homogeneity of effect sizes (Hedges, 1987;
son, 1988). When t tests are reported on samples with unequal variances, they are
calculated using the separate variances within groups for the error and degrees of
freedom following Brownlee (1965). Combined zs are based on Stouffer's method
(Rosenthal, 1984). Unless otherwise specified, p levels are one-tailed.
Psi Communication in the Ganzfeld
TABLE 2
OUTCOME BY EXPERIMENTER
Hits
Experimenter
N
trials
N
%
Effect
size (h)
Quant
106
38
36
.24
Honorton
72
27
38
.29
Berger
53
18
34
.20
Derr
45
12
27
.05
Varvoglis
43
=
26
.03
Schechter
14
5
36
.23
Ferrari
15
9
60
.72
Schlitz
7
2
29
.08
Rosenthal, 1986; Utts, 1986). Two or more studies are replicates of
one another if their effect sizes are homogeneous. We assess the
homogeneity of effect sizes across the 11 series by performing a chi-
square homogeneity test comparing the effect size for each series
with the weighted mean effect size (Hedges, 1981; Rosenthal, 1984).
The formula is:
x2(k - 1) = EN,(h; - h)2,
1 =
where k is the number of studies, N, is the sample size of the ith
study, and the weighted mean effect size is:
*
Σ ΝΑ.
*
ΣΝ.
i = 1
The test shows that the series effect sizes are not significantly non-
homogeneous: x2 = 16.25, 10 df, p = . 093.
Homogeneity of Outcome by Experimenter
Eight Es contributed to the autoganzfeld database. (See Table 2.)
All eight experimenters have positive effect sizes. A chi-square ho-
mogeneity test, using the mean effect sizes for each E weighted by
sample size, indicates that the results are homogeneous across ex-
perimenters: x2 = 7.13, 7 df, p = . 415.
CPYRGHT
115
Approved For Release 2000/08/08 : CJA-RDR98
Approved For Release 2000/08/08 . CIA-RDP96-00789R003100120001-4

116
The Journal of Parapsychology
TABLE. 3
GANZFELD SUCCESS IN RELATION TO NUMBER OF SESSIONS
No. of sessions as receiver
1
2
3
4 +
2
183
23
24
11
Z
183
79
5.1
53
19
31
19
% Hits
29
41
43
35
Effot size (h)
.09
.34
38
.22
Subject-Based Analysis
Seventy-six percent of the participants (N = 183) contributed a
singe session as R. Fifty-eight Rs contributed multiple sessions. Par-
ticiPants with multiple sessions either had direct hits or strongly
suggestive target mentation correspondences in their first session.
(SesTable 3.)
Success rate by subjects. To test the consistency of ganzfeld perfor-
mange across participants, we use the standardized ratings of the
target and decoys (Stanford's z scores; Stanford & Sargent, 1983) as
thendependent variable. Stanford zs are averaged for participants
wit multiple sessions. Direct hits and Stanford zs are highly corre-
lated. In this database, N (353) is .776. The mean Stanford z for the
24 participants is .21 (SD = 1.04), and t (240) = 3.22 (p = . 00073).
Th&95% CI is a Stanford z from .08 to .35. The effect size (Cohen's
d; hen, 1977) is .21. (The effect size for subjects is nearly identical
to de trial-based effect size, h = . 20.) Thus, there is a general ten-
dengy for participants to give higher ratings to the actual target
tha to the decoys, and the significance of these experiments is not
attroutable to exceptional performance by a few outstanding sub-
jec
Dynamic Versus Static Targets
89R003100120001-4 2
The success rate for dynamic targets is highly significant. There
are1190 dynamic target sessions and 77 direct hits (10%, 4 = 32;
exact binomial p = 1.9 x 10-", z = 4.62). The hit rate for static
targets is not significant (165 trials, 45 hits, 27%, h = . 05, p = . 276,
z = . 59). Using the series effect size as the outcome variable and
target type as the predictor variable, the point-biserial correlation
(rp) between ganzfeld performance and target type is .663, t (17) =
Psi Communication in the Ganzfeld
117
TABLE 4
SENDER/RECEIVER PAIRING
Sender as:
Lab
Lab
friend
Friend
N trials
140
66
145
N hits
46
24
52
% Hits
33
36
36
Effect size (h)
.18
.24
.24
Z
2.01
1.93
2.83
p
.023
.026
.0023
3.65, p = . 002." The 95% CI for dynamic targets is a hit rate from
34% to 47%. The CI for static targets is from 21% to 34%. Thus,
our hypothesis concerning the superiority of dynamic targets is
strongly supported.
Sender/Receiver Pairing
.
Receivers are more successful with friends than with laboratory
senders, although the difference is not statistically significant. The
number of sessions in this analysis is 351 because four subjects
opted to have no sender. The best performance occurs with friend
senders. Sessions with laboratory senders, although significant, have
the lowest success rate. (See Table 4.)
Using series effect sizes as the unit of analysis and sender type
as the predictor variable (combining lab friend and friends), ry, is
.363, 1 (17) = 1.61, 0 = . 0635.7 The 95% CI for sessions with
friends is a hit rate from 33.3% to 47%. For lab senders, the CI is
from 18.3% to 41.8%. Thus, although the effect of sender type is
not statistically significant, there is a trend toward better results with
friends.
" Separate effect sizes were obtained for the dynamic and static target sessions of
each series. Since Series 302 used dynamic targets only, the analysis is based on 11
dynamic target effect sizes and 8 static target effect sizes; two static target series (105
and 201) had extremely small sample sizes (2 and 3 sessions, respectively). A similar
procedure is used in the analyses of sender/receiver pairing and experienced versus
novice subjects.
1 Three series involving laboratory senders were eliminated from this analysis be-
cause of extremely small sample sizes. These include Series 2 (n = 2), Series 105 (n
= 2), and Series 201 (n = 1). Thus, the point biserial correlation is based on 11
series with friends and 8 series with laboratory senders.
CPYRGHTApproved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Approved For Release 2000/08/08 : CIA-RDP96

118
The journal of Parapsychology
Psi Communication in the Ganzfeld
119
Ganzfeld Experience
Two hundred and eighteen participants had their first experi-
ence as ganzfeld receivers in the autoganzfeld series. (This includes
the >> Novice Series 101-105 and 12 novices in Series 1.) For all but
24 [{1%), their initial autoganzfeld session provided their first ex-
perignce as participant in any parapsychological research. Of the
218 govices, 71 (32.5%, h = . 17) correctly identified their target (ex-
act nomial p = . 0073, z = 2.44).
Participants with some ganzfeld experience contributed 137
trial@and 51 hits (37%, h = . 26, p = . 001, z = 3.09). When series
effect sizes are used as the unit of analysis and prior ganzfeld ex-
peri@ice is used as the predictor variable, ry, is .078, { (10) = 0.25,
p = 241. The 95% CI for novices is a hit rate from 25.5% to 49.5%.
TheUCI for experienced participants is from 29% to 50%.
Part&pation by PRL Laboratory Staff
H&r completeness, we report the contribution of laboratory staff
as subjects in this database. PRL staff members contributed 12 ses-
sionPas R. These sessions yield 3 hits (exact binomial p = . 50; h =
.00)
lease 2000/08/08 : CIA-RDF96-00789003100120001-4-
White Noise and Ganzfeld Illumination Levels
Tge mean white noise level (in arbitrary units of 0-7.5) is 2.97
(SD @> 1.77). As measured from the headphones, the mean noise
levelOs approximately 68 dB. The mean light intensity (arbitrary
units of 0-100) is 73.8 (SD = 26.1). Preferred noise and light in-
tensig; levels are highly correlated: r = . 569, t (353) = 12.99.
Neither noise nor light intensity is significantly related to ganz-
feld gerformance. The point-biserial correlation between hits and
noise level is - . 026, t (353) = - 0.48, p = . 631, two tailed. For light
intensity, rp is -. 040, t (353) = - 0.76, p = . 449, two tailed.
RANDOMNESS TESTS
The adequacy of randomization was a major source of disagree-
ment in two meta-analytic reviews of earlier psi ganzfeld research
(Honorton, 1985; Hyman, 1985). In this section we document the
adequacy of our randomization procedure according to guidelines
agreed on by Hyman and Honorton (1986).
Global Tests of Random Number Generator
Full-range frequency analysis. As described earlier, autoganzfeld
targets are selected through a program call to the RNG for values
within the target range (1-160). The number of experimental ses-
sions (N = 355) is too small to assess the RNG output distribution
for the full range, so we performed a large-scale control series to
test the distribution of values. Twelve control samples were col-
lected. These included five samples with 156,000 trials, six samples
with 1,560 trials, and one sample of 1,560,000 trials. The 12 result-
ing chi-square values were compared to a chi-square distribution
with 155 df, using the Kolmogorov-Smirnov (KS) one-sample test.
The KS test yields a two-tailed p = . 577, indicating that the RNG
used in these experiments provides a uniform distribution of values
throughout the full target range."
Test of frequency distribution for Set 20. We used a single target set
(Set 20) in Series 302. We repeated the frequency analysis in a
40,000-trial control sample, restricting target selection to the four
target values within Set 20 (Targets 77-80). A chi-square test of the
distribution of targets within Set 20 shows that the RNG produces
a uniform distribution of the target values within the set: x2 = 3.19,
3 df, p = . 363.
Tests of the Experimental RNG Usage
Each autoganzfeld session required two RNG calls. An RNG call
at the beginning of the session determined the target; another,
made before the judging procedure, determined the order in which
the target and decoys were presented for judging.
Distribution of targets in the experiment. A chi-square test of the dis-
tribution of values within the target sets shows that the targets were
selected uniformly from among the four possibilities within each set;
x2 with 3 df is 0.86, p = . 835.
Distribution of judging order. A chi-square test of the judging order
indicates that the targets were uniformly distributed among the four
possible judging sequences: the x2 with 3 df is 1.85, p = . 604.
" One of the preview pack clements for Set 6, containing Targets 21-24, was
damaged. This required filtering the RNG calls in the experiment and control tests
to bypass the damaged portion of the videotape, leaving the targets in Pool 6 unused.
Thus, for the full-range analyses reported here, there are 155 df rather than 159.
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

120
The Journal of Parapsychology
Psi Communication in the Ganzfeld
121
Summary
The randomness tests demonstrate that the RNG used for target
selecion in these experiments provides an adequate source of ran-
dong numbers and was functioning properly during the experi-
mer
Ps.
EXAMPLES OF TARGET-MENTATION CORRESPONDENCES
this section, we present some examples of correspondences
betwpen targets and ganzfeld mentation. Although conclusions can-
not Be drawn from qualitative data, this material should not be ig-
nored. It constitutes the raw data on which the objective statistical
evidence is based, and may provide important insights concerning
the giderlying process. These examples are excerpts from sessions
of subjects' ganzfeld mentation reports, identified by them during
the Wind judging procedure as providing their basis for rating the
target
Target 90, Static: Dali's "Christ Crucified."
SerieDI. Participant II): 77. Rank = 1. z score = 1.67.
G. I think of guides, like spirit guides, leading me and I come into like
a Court with a king. It's quiet .... It's like heaven. The king is something
18 Jesus. Woman. Now I'm just sort of summersaulting through
heaven .... Brooding .... Aztecs, the Sun God .... High priest ....
ler ..
..
Graves. Woman. Prayer .... Funeral .... Dark. Death ....
Seuils.
.. Ten Commandments. Moses.
...
Tare 77, Dynamic: Tidal wave engulfing ancient city. From "The Clash
of theTitans," a film based on Greek mythology. A huge tidal wave crashes
into Re shore. The scene shifts to a center courtyard of an ancient Greek
city; Dere is a statue in the center, and buildings with Greek columns around
the periphery. People are running to escape consumption by the tidal wave.
WateBrushes through the buildings, destroying the columns and the statue;
peop scurry through a stone tunnel, just ahead of the engulfing water;
debris floats through the water.
Series. 1. Participant ID: 87. Rank = 1. z score = 1.42.
" ... The city of Bath comes to mind. The Romans. The reconstruction
of the baths through archaeology. The Parthenon. Also getting sort of
buildings like Stonehenge but sort of a cross between Stonehenge and
the Parthenon. The Byzantine Empire. The Gates of Thunder. The
Holy See. Tables floating about. ... The number 7 very clearly. That just
popped out of nowhere. It reminds me a bit of one of the first Clash
albums, however. The Clash, "Two Sevens" I think it was called, I'm not
sure .... " [The target was number 77.]
Series 302. Participant ID: 267. Rank = 1. z score = 2.00.
" ... A big storm over New York City. I'm assuming it's New York City.
No, it's San Francisco .... A big storm and danger. It looks so beautiful
but I'm getting the sense of danger from it .... It's a storm. An earth-
quake.
Target 63, Dynamic: Horses. From the film, "The Lathe of Heaven." An
overhead view of five horses galloping in a snow storm. The camera zooms
in on the horses as they gallop through the snow. The scene shifts to a close-
up of a single horse trotting in a grassy meadow, first at normal speed, then
in slow-motion. The scene shifts again; the same horse trotting slowly
through empty city streets.
Series: 101. Participant ID: 92. Rank = 1. z score = 1.25.
.. I keep going to the mountains .... It's snowing .... Moving again,
this time to the left, spinning to the left .... Spinning. Like on a carousel,
horses. Horses on a carousel, a circus. .
.. "
Target 46, Dynamic: Collapsing Bridge. Newsreel footage of the collapse of
a bridge the 1940s. The bridge is swaying back and forth and up and down.
Light posts are swaying. The bridge collapses from the center into the water.
Series: 101. Participant ID: 135. Rank = 1. z score = 1.94.
" ... Something, some vertical object bending or swaying, almost some-
thing swaying in the wind. ... Some thin, vertical object, bending to the
left .... Some kind of ladder-like structure but it seems to be almost
blowing in the wind. Almost like a ladder-like bridge over some kind of
chasm that's waving in the wind. This is not vertical this is horizon-
tal .... A bridge, a drawbridge over something. It's like one of those old
English type bridges that opens up from either side. The middle part
comes up. I see it opening. It's opening. There was a flash of an old
English stone bridge but then back to this one that's opening. The
bridge is lifting, both sides now. Now both sides are straight up. Now
it's closing again. It's closing, it's coming down, it's closed. Arc, images
of arcs, arcs, bridges. Passageways, many arcs. Bridges with many
arcs .. ..
Target 137, Static: "Working on a Watermelon Farm." This painting shows
a black man with his back to the picture; his suspenders form a V-shape
CPYRGHT Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

12Z
i ne journal of rarapsychology
around his shoulders. A dog is in front of the man; there are watermelons
between the dog and the man. The man faces a dirt path with watermelon
patches on either side. On the left side, another man pushes a wheelbarrow
filled with huge watermelons.
Series 101. Participant ID: 105. Rank = 2. z score = 0.98.
.. a small lamb, very soft, outside. Small, playful .... I see a 'V'
sbipe .... An apple .... I see a kitchen towel with a picture on it. Apple
seeds or a fruit cut in half showing the seeds. A tomato or an apple.
The fruit was red on the outside .... I thought of watermelon as in a
wmlermelon basket. Thinking of kids playing on a beach. Little kids
playing with balls that are bigger than they are and buckets that are
three-quarters their size .... I had a thought of going through a tunnel,
not the kind of tunnel you see on Earth but the type of tunnel described
when someone dies."
Targy 64, Dynamic: 1920s Car Sinking. From the film "Ghost Story." The
scenesdepicts the murder of a young blonde woman by three young men in
the
820s. The men are all wearing suits; one of the men is wearing a
fedo@ hat that is turned up in the back. The men push an old car into a
lake.Os he camera shifts between close-ups of their facial expressions, and the
car,
" it slowly sinks into the water. The woman's face and hand appear in
the car's large rectangular rear window; she silently screams out for help.
The Cor disappears beneath the water as the sequence ends.
Series 102. Participant ID: 154. Rank = 1. z score = 1.45.
2000/08/083 CIA-RDP96-00289R005100120001 4
. Girl with a haircut. . .. Blond hair. ... A car .... The back of some-
(me's head .... Someone running to the right .... Someone on the right
ico a brown suit ... and a fedora hat turned up very much in the
Rtk .... Fedora, trench coat, dark tie .... A tire of a car. The car's going
" the left. An old movie .... I'm picturing an Edward G. Robinson
movie .... Big roundish car like 1940's. Those scenes from the back win-
Bw. Bumping once in a while up and down looking through the back
window you could see that it was probably a big screen in back of the
and the car's standing still actually .... I think it's a movie I saw.
They're being shot at and shooting at the window and then the girl gets
$Et .... Girl with the blonde haircut .... Someone walking in a suit,
(@own suit .... It's the 1940's again, 30's maybe. Except it looks like it's
iru color. Something red, blood ... blood on someone's lap .... A dead
Brson all of a sudden .... A big mouth opened. Yelling, but no
bind .... Two people running near a train .... Dressed in 1920 type
Fits with balloony pants, like knickers .... A big, old-fashioned white car
ith a flat top. 1920's, 30's .... "
Target 107, Static: Stained-Glass Madonna with Child. This is a stained-
glass window depicting the Virgin Mary and Christ child.
123
Psi Communication in the Ganzfeld
Series: 102. Participant ID: 183. Rank = 2. z score = 0.61.
"Some kind of a house, structure .... Some kind of wall or building.
Something with the sky in the background. Thinking of a bell. A bell
structure. Something with a hole with the light coming through the
hole .... Like a stained glass window like you see in churches."
Target 19, Static: Flying Eagle. An eagle with outstretched wings is about
to land on a perch; its claws are extended. The eagle's head is white and its
wings and body are black.
Series: 104. Participant ID: 316. Rank = 1. z score = 2.00.
" ... A black bird. I see a dark shape of a black bird with a very pointed
beak with his wings down .... Almost needle-like beak .... Something
that would fly or is flying ... like a big parrot with long feathers on a
perch. Lots of feathers, tail feathers, long, long, long .... Flying, a big
huge, huge eagle. The wings of an eagle spread out .... The head of an
eagle. White head and dark feathers .... The bottom of a bird .... "
Target 144, Dynamic: Hell. From the film "Altered States." This sequence
depicts a psychedelic experience. Everything is tinted red. The rapidly shifting
scenes include: A man screaming; many people in the midst of fire and
smoke; a man screaming in an isolation tank; people in agony; a large sun
with a corona around it; a mass crucifixion; people jumping off a precipice,
in the midst of fire, smoke, and molten lava; spiraling crucifixes. There is a
close-up of a lizard's head, slowly opening its mouth, at the end of the se-
quence.
Series: 104. Participant ID: 321. Rank = 1. z score = 1.49.
" ... I just see a big 'X'. A big 'X' .... I see a tunnel in front of me. It's
like a tunnel of smog or a tunnel of smoke. I'm going down it .... I'm
going down it at a pretty fast speed. ... I still see the color red, red, red,
red, red, red, red, red .... Ah, suddenly the sun. ... The kind of cartoon
sun you see when you can see each pointy spike around the sphere .... I
stepped on a piece of glass and there's a bit of blood coming out of my
foot .... A lizard, with a big, big, big head .... "
Target 148, Static. Three Unusual Planes. Three small aircraft flying in
formation. The planes are white and have swept-back wings; their landing-
gear is extended. A winding road is visible below.
Series: 104. Participant ID: 322. Rank = 2. z score = 0.39.
" ... A jet plane .... A 747 on the way to Greece. Blue skies. Sounds like
it's going higher. ... I think I'm back on the plane again. I never used
to be afraid of flying until recently .... They need better insulated jets,
soundproof like these rooms. They could use these comfortable seats,
too. And the leg room. The service isn't bad either .... Still can't get the
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

124
The Journal of Parapsychology
Approved For Release 2000/08/08 : CIA-RDP96-00
feeling of being in an airplane out of my mind. Flying over Greenland
and Iceland when I went to England .... Feels like we're going higher
and higher .... Descending. It seems we're descending .... Big airplanes
flying over with people like me staring down. ... Flying around in a
piece of tin. ... Feel like I'm getting a G-force. Maybe I am taking off.
Sure feels like it. Feels like we're going straight up .... I always feel like
when I'm on the plane going home, { just hope that plane makes it past
the Rocky Mountains. .
Torget 10, Static: Santa and Coke. This is a Coca-Cola Christmas ad from
1950s, showing Santa Claus holding a Coke bottle in his left hand; three
batons are visible on Santa's suit. Behind Santa and to his left, is a large
b&tle cap with the Coca-Cola logo leaning against an ornamented Christmas
Spries: 104. Participant ID: 332. Rank = 1. 2 score = 1.14.
" ... There's a man with a dark beard and he's got a sharp face ....
There's another man with a beard. Now there's green and white and
ohe's in bushes and he's sort of colonial. He looks like Robin Hood and
She's wearing a hat. ... I can see him from behind. I can see his hat and
he has a sack over his shoulder .... Window ledge is looking down and
there's a billboard that says 'Coca-Cola' on it .... There's a snowman
"again and it's got a carrot for a nose and three black buttons coming
down the front .... There's a white beard again. There's a man with a
"white beard .... There's an old man with a beard ....
Teoget 70, Dynamic: Dancing in NY City Streets. From the film "The Wiz."
TR span of yellow-paved bridge over a body of water and automobile traffic
is Bisible in the opening scene; the New York City skyline is in the back-
groend. A hot-air balloon flies overhead. The scene shifts as Dorothy (Diana
Reos), her dog Toto, the Lion, Tin Man, and Scarecrow dance along the
brage; one of the bridge's supporting arches is behind them. The Chrysler
Balding is in the background. At the end of the sequence, the characters
dance in front of a painted backdrop of an old-fashioned building.
S@es: 105. Participant ID: 336. Rank = 1. z score = 1.40.
"Big colorful hot air balloons .... White brick wall .... Ocean .... People
walking before my eyes. Several people .... A dog. Hot air balloon ....
a nightclub singer .... Back of a woman's head, short curly hair ....
Water .... Balloon, big balloon. ... Yellow .... Very tall building. Look-
ing down at a city. Leaving a city, going up .... Faces. An arc ....
Water .... A woman's face .... Cars, freeway .... A rock-n-roll star
chanting .... Architecture. A jester's hat .... geometrical figures, designs.
... Yellow chocolate bar. Water. Going down into water, deep down ....
Man with long golden hair and sun glasses .... The Bay, San Francisco
Psi Communication in the Ganzfeld
125
Bay. A lion .... Highways .... Lion, see a lion .... Tornado ....
Bal-
loon .... Face mask. ... City .... Leaning Tower of Pisa. ... Long hall-
way, doorway .... Long road. Long, long desert road .... "
Target 22, Dynamic: Spiders. From the documentary "Life on Earth." A
spider is weaving its web. The spider's long legs spring up and down re-
peatedly, weaving strands of the web. The body of the spider is constantly in
motion, and bounces up and down. A close-up shows one of the veins of the
web being stretched out by the spider. Various views of the web.
Series: 301. Participant ID: 146. Rank = 2. z score = 0.65.
" ... Now visual patterns more like a spider web and the color. And then
like the form of the veins of a windmill .... Something like a spider web
again. A spider web. A pattern that instead of a spider web it looks like
basket weaving .... An image of the way some children were able to do
something like flying when I was a child though I never had one. It was
a-forgotten what it was called-a pogo stick or a jump stick, something
in which you jumped up and down and you could hop quite a distance
by doing so .... I have kinesthetic images all over as in vigorous motion
expressed in flying or jumping on this sort of spring stick that I men-
tioned .... Vigorous motion. It's as though I were trying to combine re-
laxation with participating in an image of something very vigorous .... I
really feel carried away by these images of vigorous activity without
being able to localize this activity as to what it is .... "
Target 108, Static: Two fire eaters. A young fire eater, in the foreground,
facing to the right of the picture, blows a huge flame out of his mouth. In
the background there is another fire cater. A group of people are watching
on the left side of the picture.
Series: 301. Participant ID: 146. Rank = 1. z score = 1.71.
" ... I keep having images of flames now and then .... The sound re-
minds me of flames too. ... I find flames again .... In these new images
the fire takes on a very menacing meaning .... Rather mountainous
sticking up of bare rocks just as though they had come from a recently
formed volcano. Volcanos of course get back to the fire, extreme heat.
I had an image of a volcano with molten lava inside the crater. Molten
lava running down the side of the volcano .... Cold. Written out there
behind the visual field and thinking how it contrasts with my images of
Hames. Although my images of flames didn't actually include much real
feeling of heat. I didn't have any imagery of heat in connection with the
flames. Just abstract thought of flames .... Now I think of the water as
a way of putting out flames. Suddenly, I was biting my lip. Biting my
lip as though lips had something to do with the imagery and I see lips
out in front of me .... And the lips I see are bright red, reminding me
of the flame imagery earlier. And then a bright heart such as Valentine's
CPYRGHT Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

126
The Journal of Parapsychology
candy in the shape of a heart. The cinnamon favored candies that I
remember as a child having at Valentine's. Red color .... This red as in
the cinnamon candy is a deep very intense red. And similarly for the
flames. And now I see the word 'red'.
Tapet 94, Dynamic: Hang Gliders. The sequence shows a skier on a V-
shapsd hang glider. The skies somos high up above snow covered mountains
and pine forest. At the end, the skier lands on a mountain slope and skis
aweg. The sequence is accompanied by Pachelbel's Canon.
Seres: 301. Participant ID: 188. Rank = 1. z score = 1.26.
" .. Some kind of 'V' shape, like an open book .... I get some moun-
tin .... Some kind of bird with a long wing. ... The shape of an upside
wn 'V' .... Ski, something about skiing came to me .... Some kind of
body like an oval shape of a body with wings on top of it in a 'V'
pape. Another 'V' like a wing shape .... Something with wings ....
Again the shape of an umbrella came into my mind. A butterfly
gape.
Tart 80, Dynamic: Bugs Bunny in Space. In this cartoon, there is a close-
up of the lower part of a cigar-shaped rocketship and the supports holding
it ugo The rocket assembly slides over to the launching pad, directly above
Bu
Bunny's underground patch. The scene shifts to the underground
patch, as Bugs Bunny climbs up the ladder leading out of his patch. Un-
knowingly, he climbs up through the interior of the rocketship. The rocket's
supports pull away and then it takes off into space. The rocket's nose cone
spinas Bugs Bunny appears through the top and he sees the Earth recede
rapay in the distance. As the sequence ends, Bugs Bunny is hit in the belly
by aDomet.
Seres: 302. Participant ID: 292. Rank = 1. z score = 1.48.
.. Space craft. ... The solar system. The underside of a helicopter or
Isubmarine or some kind of fish that you're seeing from under-
Bath. ... Sort of being underneath it. Sort of being underneath it .... A
Dry strange image like a cartoon character, animated character. With
Es mouth open kind of .... Like a hypodermic needle or a candle or
@is shaft like thing with the a pointed top again .... missiles
Bing .... An aerial perspective .... I'm just kind of editing here I think.
An really hoping all this rocketship kind of imagery isn't because of the
puise. I feel like I'm in a rocketship or something. ... That image of the
Bip going into the belly of the mother ship .... "
P96-00789R003100120001-4
COMPARISON OF STUDY OUTCOMES WITH
GANZFELD META-ANALYSIS
In this section, we compare the automated ganzfeld study out-
comes with the results of earlier ganzfeld studies, summarized in'a
Psi Communication in the Ganzfeld
127
TABLE. 5
COMPARISON OF OVERALL. PERFORMANCE IN AUTOMATED GANZFELD AND
META-ANALYSIS DATA SETS
Outcome
N
variable
Database
studies
Mean
ST)
1
df
p
z scores
Meta-analysis
28
1.25
1.57
0.33
25
.748
Effect sizes (h)
Meta-analysis
28
.28
.46
Autoganzfeld
11
.29
.29
0.14
28
.892
Note. The p values are two-tailed.
meta-analysis (Honorton, 1985). We compare the two databases on
four dimensions: (1) overall success rate, (2) dynamic versus static
targets, (3) sender/receiver pairing. and (4) novice versus experi-
enced subjects.
Overall Success Rate
To assess the consistency of results, we compare the ll auto-
ganzfeld series to the 28 studies in a meta-analysis of earlier ganz-
feld studies (Honorton, 1985, Table Al, p. 84), using direct hits as
the dependent variable. The outcomes of the two data sets are con-
sistent. Both display a predominance of positive outcomes: 23 of the
28 studies in the meta-analysis (82%) and 10 of the 11 autoganzfeld
series (91%) yield positive z scores. The mean autoganzfeld z scores
and effect sizes are very similar to those in the meta-analysis. (See
Table 5.)
Combined Estimates of Ganzfeld Success Rate
Because the z scores and effect sizes for the automated ganzfeld
are consistent with the original set of 28 studies in the meta-analysis,
a better estimate of their true population values may be obtained by
combining them. Positive outcomes were obtained in 33 of the 39
studies (85%); the 95% Cl is from 69% to 99%. Table 6 shows a
stem-and-leaf frequency plot of the z scores (Tukey, 1977). Unlike
other methods of displaying frequency distributions, the stem-and-
leaf plot retains the numerical data precisely. (Turned on its side,
the stem-and-leaf plot becomes a conventional histogram.) Each
number includes a stem and one or more leaves. For example, the
stem 1 is followed by leaves of 6,6,6,7,7,7, representing z scores of
1.6,1.6,1.6,1.7,1.7,1.7. In the display, the letter "H" identifies the
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Autoganzfeld
11
1.10
1.14
Approved ForRelease/2000/08/085 6

128
The Journal of Parapsychology
TABLE 6
DISTRIBUTION OF Z SCORES
Stem
Leaf
Minimum z
=
- 1.97
- 1.
97
Lower hinge
0.25
85
Median z
=
0.92
0.
33
Mean z
=
1.28
0.
H 222224
Upper hinge
=
2.08
0.
M 6667777999
Maximum z
=
4.02
666777
SD
=
1.44
2.
H 011
Skewness (g.)
=
0.05
2 .
8
Kurtosis (g2)
=
-0.37
3.
9
4.
0
per and lower hinges of the distribution, and "M" identifies its
Bedian. The z's range from - 1.97 to 4.02 (mean z = 1.21, SD =
95), and the 95% Cl is a z from .76 to 1.66.
The combined z for the 39 studies is 7.53 (p = 9 x 10-1).
osenthal's (1984) file-drawer statistic indicates that 778 additional
udies with z scores averaging zero would be required to reduce the
ignificance of the combined ganzfeld database to nonsignificance;
Icht is a ratio of 19 unknown studies for every known study.
A stem-and-leaf display of the effect sizes is shown in Table 7.
Bie effect sizes range from - . 93 to 1.44 (mean k = . 28, SD - . 41).
The two most extreme values on both sides of the distribution are
Guliers. The 95% Cl is an h between .15 and .41; the equivalent hit
De is from 31.5% to 44.5%.
89 R008100120901 4년
The use of video sequences as targets is a novel feature of the
aoganzfeld database. However, a comparable difference in target
cpe exists in the earlier ganzfeld studies. Of the 28 direct hits stud-
in the meta-analysis, 9 studies (by three independent investiga-
tors) used View Master stereoscopic slide reels as targets
(Honorton, 1985, Studies 7-8, 16-19, 21, 38-39). Static targets
(single pictures or slides) were used in the remaining 19 studies by
seven independent investigators (Studies 1, 2, 4, 10-13, 23-31, 33-
34, 41-42). Like the autoganzfeld video sequences, View Master tar-
gets present a variety of images reinforcing a central target theme.
Psi Communication in the Ganzfeld
129
TABLE. 7
DISTRIBUTION OF EFFECT SIZES (COHEN'S h)
Stem
Leaf
- 9
3
-. 4
0
OUTSIDE VALUES
Minimum h
=
-0.939
-. 3
1
Lower hinge
=
0.101
-. 1
0
Median h
=
0.25℃
-. 0
51
Mean h
=
0.28
.0
7779
Upper hinge
=
0.44
.1
H 002888
Maximum h
=
1.44D
M 1334
SD
=
0.4 IN
.3
11144777
Skewness (g.)
=
.4
H 01113
Kurtosis (g2)
=
5
7
7
3
.8
17
OUTSIDE VALUES
1.3
3
1.4
4
To compare the relative impact of dynamic and static targets in
the autoganzfeld and meta-analysis, we obtained point-biserial cor-
relations for each data set using target type (static or dynamic) as
the predictor variable and the series effect size, Cohen's h, as the
outcome variable. We test the difference between the two correla-
tions using Cohen's q (Cohen, 1977). Dynamic targets yield signifi-
cantty farger effect sizes in both data sets. For the meta-analysis, r
is .409, t (26) = 2.28, p = . 015; and for the autoganzfeld, as rec
ported above, ry, is .663. The two correlations are not significantly
different (q = . 36; z = 1.14). Therefore, we combine the-two data
sets to obtain a better estimate of the relationship between effect size
and target type: 1p = . 439, ( (45) = 3.28, p = . 002. The 95% CIsc
are 24% to 36% for static targets and 38% to 55% for dynamic tar-
gets. Thus, the cumulative evidence strongly indicates that dynamick
targets are more accurately retrieved than static targets.
Sender/Receiver Pairing
A similar analysis compares the effects of sender/receiver pairing
in the two databases. Studies in the meta-analysis did not routinely
CPYRGHT
Approved F
1.
3.
01124
Combined (Stouffer) z =
7.53
Approved For Release 2000/08/08 : CIA RDE96-1
Dynamic Versus Static Targets
2000/08/08 . CIA-RDP96-00
89R003100120001-4

101
provide detailed breakdowns regarding sender/receiver pairing.
Sender/receiver pairing in the meta-analysis can only be coded ac-
cording to whether subjects could bring friends to serve as their
sender os were restricted to laboratory sendees. In 17 studies, by six
independent investigators, subjects were free to bring friends
Slonorton, 1985, Studies 1-2, 4, 7-8, 16, 23-28, 30, 33-34, 38-
g9). Laboratory-assigned senders were used exclusively in the re-
gaining 8 studies, by four independent investigators (Studies 10-
22, 18-19, 21, 29, 41). (Three studies using clairvoyance proce-
Bures and no senders are excluded from this analysis.) For the au-
mganzfeld studies, we calculated separate effect sizes for each series
&y sender type (combining lab friend and friend for comparability
yoth the meta-analysis). In the meta-analysis, r,, (23) is .403; larger
Affect sizes occurred in studies where friends could serve as sender
0= 2.11, p = . 023). For the autoganzfeld, as reported above, r, is
(363, in the same direction. The two correlations are very similar (q
.05; z = 0.14) and are combined to give a better estimate of the
lationship between sender/receiver pairing and ganzfeld study
Ruitcome: r,. = . 38, 7 (42) = 2.66. p = . 0055. The 95% CIs are 20%
6 34% for unacquainted sender/receiver pairs and 34.1% to 49.2%
Or friends. Thus, the sender/receiver relationship does have a sig-
uficant impact on performance.
Ffect of Prior Ganzfeld Experience
The meta-analysis includes 14 studies, by nine independent in-
Ostigators, in which novices are used exclusively (Honorton, 1985,
fudies 2, 4, 8, 10-12, 16-18, 23-24, 31, 41-42). Experienced or
Dixed samples of novice and experienced subjects are used in the
omaining 14 studies, by four different investigators (Studies 1, 7,
9, 21, 25-30, 33-34, 38-39). Studies using experienced subjects
core more successful than those limited to novices; the point-biserial
Brrelation between level of experience and effect size is .229, t (26)
$ 1.20, p = . 12. For the autoganzfeld studies, as reported above,
is .078. The two correlations do not differ significantly (q = . 155;
0.40), and the combined rp, is .194, t (38) = 1.22, p = . 105. The
¿spective 95% CIs are 24.5% to 44.5% for novices and 35.5% to
8% for experienced subjects.
¿The 95% Cls for these comparative analyses are shown graphi-
cally in Figure 2. The bottom two rows are CIs for the overall hit
rates in the meta-analysis and autoganzfeld, respectively. The next
Auto-Exper -
Mela:Exper
Auto:Novice
-
MeLa:Novice
Auto:SR=Fr
-
Data set and condition
Meln:SK=Fr
Auto:SR=Lab
-
Mela:SR=Lab
Auto:TGT=Sta
Meta:TGT=Sta
Aulo:TGT=Dyn
Mela:TGT=Dyn
Autoganz.
-
Meta-anal. -
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 .0.8 1.0
Effect size (h)
Figure 2. Comparison of autoganzfeld and meta-analysis 95% confidence
limits. Abbreviations are defined as follows: Meta = meta-analysis studies,
Auto = automated ganzfeld studies, Dyn = dynamic targets, Sta = static
targets, Lab = laboratory senders, Fr = sender is friend or acquaintance
of receiver, Novice = no prior ganzfeld experience, Exper = prior ganz-
feld experience.
two rows give the CIs for dynamic targets in the two data sets, and
SO On.
DISCUSSION
We now consider various rival hypotheses that might account for
the experimental outcomes, and the degree to which the automated
ganzfeld experiments, viewed in conjunction with the earlier psi
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
se 2000/08/082 CIA-RDP96-08789R003100120001-4 5
....

132
The Journal of Parapsychology
ganzfeld studies, constitute evidence for psi communication. Finally,
we consider directions for future research suggested by these find-
Days.
Bival Hypotheses
2
Sensory Cues. Only Se knows the identity of the target until R
finishes the automated judging procedure. If Se is not a PRL staff
gember, a staff member not otherwise involved in the session su-
pærvises target selection. In either case, the target selector knows
Only which videocassette contains the target. The target selector
Baves the monitoring room with the remaining three target tapes
Efter knocking three times on the monitoring room door, signalling
to return. Since the target selector only knows the videocassette
@amber, variations in knocking cannot communicate any useful in-
ormation to E. The cardboard cover over the VCR eliminates any
isual cues to E regarding the position of the videotape or the activ-
y of the VU meters (which are active when the target is dynamic
oad has a soundtrack).
-2608/08/08 : CIA-RDP96-00789R003100120061
Sensory transmission from Se to R during the ganzfeld session is
minated by having R and Se in separate, sound-attenuated rooms.
fi either participant leaves their room before R's ratings have been
Egistered in the computer, the session is unconditionally aborted.
The videotape target display system prevents potential handling
ges during the judging procedure. Computer registration of R's
Erget ratings and automated feedback after the session prevents the
Possibility of cheating by Se during feedback, raised by Hyman
985).
After about 80% of the sessions were completed, it was becoming
Dear that our hypothesis concerning the superiority of dynamic tar-
gets over static targets was receiving substantial confirmation. Be-
use dynamic targets contain auditory as well as visual information,
Re conducted a supplementary test to assess the possibility of audi-
rory leakage from the VCR soundtrack to R. With the VCR audio
&t to normal amplification, no auditory signal could be detected
Prough R's headphones, with or without white noise. When an ex-
Jernal amplifier was added between the VCR and R's headphones
and with the white noise turned completely off, the soundtrack
could sometimes be faintly detected. It is unlikely that subjects could
have detected any target audio signal with the normal VCR ampli-
fication and white noise; as we have reported, there is no correlation
between ganzfeld success rate and white noise level in these exper-
Psi Communication in the Ganzfeld
133
iments. Nevertheless, to totally exclude any possibility of subliminal
cueing, we modified the equipment. Additional testing confirmed
that this modification effectively eliminated all leakage. This was
formally confirmed by an audio spectrum analysis, covering the fre-
quency domain between 475 Hz and 15.2 kHz. The critical question,
of course, is whether performance on dynamic targets diminished
after this modification. The answer is no; in fact, performance im-
proved. Before the modification, the direct hit rate on dynamic tar-
gets was 38% (150 trials, 57 hits, h = . 28, exact binomial p =
.00029, z = 3.44); the 95% CI was from 31% to 45%. Following the
modification, the direct hit rate was 50% (40 trials, 20 hits, h = . 52,
exact binomial p = . 00057, z = 3.25) with a 95% CI from 37% to
63%. The direct hit rate for all targets-static and dynamic-after
the modification was 44% (64 trials, 28 hits, h = . 39, exact binomial
b = . 00082, z = 3.15).
Randomization. As Hyman and Honorton (1986, p. 357) have
pointed out, "Because ganzfeld experiments involve only one target
selection per session ... , the ganzfeld investigator can restrict his or
her attention to a frequency analysis allowing assessment of the de-
gree to which targets occur with equal probability." We have docu-
mented both the general adequacy of the RNG used for target se-
lection and its proper functioning during the experiment.
Data selection. Except for two pilot studies, the number of partic-
ipants and trials were specified in advance for each series. The pilot
or formal status of each series was similarly specified in advance and
recorded on disk before beginning the series. We have reported all
trials, including pilot and ongoing series, using the automated ganz-
feld system. Thus, there is no "file-drawer" problem in this data-
base.
Psi ganzfeld success rate is similar for pilot and formal sessions.
The proportion of hits for the 66 pilot sessions is .32 (h = . 16, p =
.129, z = 1.13). For the 289 formal sessions, the proportion correct
is .35 (h = . 22, p = . 0001, z = 3.71). The difference is not signifi-
cant: x2 = 0.11, 1 df, p = . 734.
If we assume that the remaining trials in the three unfinished
series would yield only chance results, these series would still be sta-
tistically significant (exact binomial p = . 009, z = 2.36). This would
reduce the overall z for all 11 series from 3.89 to 3.61. Thus, inclu-
sion of the three incomplete studies does not pose an optional stop-
ping problem. '
Multiple analysis. Informal examination of recent issues of several
American Psychological Association journals suggests that correction
CPYRGHT Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

Psi Communication in the Ganzfeld
135
for multiple comparisons is not a common practice in more conven-
tional areas of psychological inquiry. Nevertheless, half of Ilyman's
(1985) 50-page critique of carlier psi ganzfeld research focused on
issues related to multiple testing. In the present case, advance spec-
ification of the primary hypothesis and method of analysis prevents
Problems involving multiple analysis of multiple indices in our test
& the overall psi ganzfeld effect. Our direct hits analysis is actually
lo significant than either the sum of ranks method (z = 4.04, p =
8 x 10 ") or Stanford's z scores (t = 4.53, 354 df, p = 4.1 x
').
TiIn addition to the primary hypothesis, however, we also tested
b) secondary hypotheses concerning the impact of target type and
Ader/receiver pairing on psi performance, and we have presented
segeral purely exploratory analyses as well. Our Results section in-
(Wides 15 significance tests involving psi performance as the depen-
dent variable, and the p values cited are not adjusted for multiple
comparisons. Of the 15 significance tests, 9 are associated with p <
The Bonferroni multiple comparisons procedure provides a
c&servative method of adjusting the alpha level when several si-
ngltaneous tests of significance are performed (Holland & Copen-
høer, 1988; Hyman & Honorton, 1986; Rosenthal & Rubin, 1984).
When the Bonferroni adjustment is applied, six of the nine individ-
ually significant outcomes remain significant; these are: the overall
hlgrate, the subject-based analysis using Stanford z scores, the dif-
ference between dynamic and static targets, the dynamic target hit
rau, and the hit rate for experienced subjects.
2000/08/08 : CIA-RDP966
DAlthough the relationship between psi performance and sender
type is not independently significant in the autoganzfeld, the cor-
ration coefficient of .363 is close to that observed in the meta-
analysis (r = . 403), and the combined result is significant. The cu-
negative evidence, therefore, does support the conclusion that the
stoler/receiver relationship is a significant moderator of ganzfeld
p&performance.
00789R0031003200
Security. Given the large number of subjects and the significance
ofthe outcome using subjects as the unit of analysis, subject decep-
tion is not a plausible explanation. The automated ganzfeld protocol
h
been examined by several dozen parapsychologists and behav-
iogl researchers from other fields, including well-known critics of
parapsychology. Many have participated as subjects, senders, or ob-
servers. All have expressed satisfaction with our handling of security
issues and controls.
In addition, two experts on the simulation of psi ability have ex-
amined the autoganzfeld system and protocol. Ford Kross has been
.
a professional mentalist for over 20 years. He is the author of many
articles in mentalist periodicals and has served as Secretary/ Treas-
urer of the Psychic Entertainers Association. Mr. Kross has provided
us with the following statement: "In my professional capacity as a
mentalist. I have reviewed Psychophysical Research Laboratories'
automated ganzfeld system and found it to provide excellent secu-
rity against deception by subjects" (personal communication, May,
1989). We have received similar comments from Daryl Bem, Pro-
fessor of Psychology at Cornell University. Professor Bem is well
known for his research in social and personality psychology. He is
also a member of the Psychic Entertainers Association and has per-
formed for many years as a mentalist. He visited PRI. for several
days and was a subject in Series 101.
The issue of investigator integrity can only be conclusively ad-
dressed through independent replications. It is, however, worth
drawing attention to the 13 sessions in which a visiting scientist,
Marilyn J. Schlitz, served as either experimenter (N = 7, 29% hits,
h = . 08) or sender (N = 6, 67% hits, h = . 36). Altogether, these
sessions yielded 6 direct hits (N = 13, 46.2% hits, h = . 45). This
effect size is more than twice as large as that for the' database as a
whole.
Status of the Evidence for Psi Communication in the Ganzfeld
The automated ganzfeld studies satisfy' the methodological
guidelines recommended by Hyman and Honorton (1986). The re-
sults are statistically significant. The effect size is homogeneous
across 11 experimental series and eight different experimenters.
Moreover, the autoganzfeld results are consistent with the outcomes
of the earlier, nonautomated ganzfeld studies; the combined z of
7.53 would be expected to arise by chance less than one time in 9
trillion.
We have shown that, contrary to the assertions of certain critics
(Druckman & Swets, 1988, p. 175), the ganzfeld psi effect exhibits
"consistent and lawful patterns of covariation found in other areas
of inquiry." The automated ganzfeld studies display the same pat-
terns of relationships between psi performance and target type,
sender/receiver acquaintance, and prior testing experience found in
earlier ganzfeld studies, and the magnitude of these relationships is
consistent across the two data sets. The impact of target type and
sender/receiver acquaintance is also consistent with patterns in spon-
taneous case studies, linking ostensible psi experiences to emotion-
ally significant events and persons. These findings cannot be ex-
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Approved For Release

136
The Journal of Parapsychology
plained by conventional theories of coincidence (Diaconis &
Mosteller, 1989).
Hyman and Honorton (1986) have stated,
Approved For Release 2803 198103 .2 CIA-R
... the best way to resolve the [ganzfeld] controversy ... is to await the
outcome of future ganzfeld experiments. These experiments, ideally,
will be carried out in such a way as to circumvent the file-drawer prob-
lem, problems of multiple analysis, and the various defects in random-
ization, statistical application, and documentation pointed out by
Hyman. If a variety of parapsychologists and other investigators con-
tinue to obtain significant results under these conditions, then the exis-
tence of a genuine communications anomaly will have been demon-
strated. (pp. 353-354)
We have presented a series of experiments that satisfy these
gidelines. Although no single investigator or laboratory can satisfy
a requirement of independent replication, the automated ganzfeld
dies are quite consistent with the earlier studies. On the basis of
cumulative evidence, we conclude that the ganzfeld effect rep-
Rents a genuine communications anomaly. This conclusion will
esther be strengthened or weakened by additional independent rep-
lications, but there is no longer any justification for the claim made
some critics that the existing evidence does not warrant serious
aftention by the scientific community.
Recommendations for Future Research
[P96-09 78SRD@3100$200984
Recent psi ganzfeld research has necessarily focused on meth-
Cological issues arising from the ganzfeld controversy. It is essen-
that future studies comply with the methodological standards
agreed on by researchers and critics. Yet it is equally imperative that
serious attention be given to conditions associated with successful
Rtcomes.
Small to medium effect sizes characterize many research findings
Sthe biomedical and social sciences (e.g., Cohen, 1977; Rosenthal,
34). Rosenthal (1986) and Utts (1986) make a strong case for
nore careful consideration of the magnitude of effect in the design
od analysis of future ganzfeld studies. The automated ganzfeld
studies show a success rate slightly in excess of 34%. Utts's (1986)
power analysis shows that for an effect of this size, the investigator
has only about one chance in three of obtaining a statistically signif-
icant result in a 50-trial experiment. Even with 100 trials-an unu-
sually large sample size in ganzfeld research-the probability of a
significant outcome is only about .5.
Psi Communication in the Ganzfeld
137
We urge ganzfeld investigators to use dynamic targets and to de-
sign their studies to allow subjects to have the option to have friends
or acquaintances as their senders. The similarity of the autoganzfeld
and meta-analysis data sets strongly indicates that these factors are
important moderators of psi ganzfeld performance. If our estimate
of the impact of dynamic and static targets is accurate, a 50-session
series using dynamic targets has approximately an 84% chance of
yielding a significant outcome. A comparable series with static tar-
gets has only about one chance in five of achieving significance.
REFERENCES
ALCOCK, J. E. (1986). Comments on the Hyman-Honorton ganzfeld contro-
versy. Journal of Parapsychology, 50, 345-348.
AKERS, C. (1984). Methodological criticisms of parapsychology. In S. Krippner
(Ed.), Advances in parapsychological research, Vol. 4 (pp. 112-164). Jeffer-
son, NC: McFarland.
BERGER, R. E., & HONORTON, C. (1986). An automated psi ganzfeld testing
system. In D. H. Weiner & D. I. Radin (Eds.). Research in parapsychology
1985 (pp. 85-88). Metuchen, NJ: Scarecrow Press.
BLACKMORE, S. (1980). The extent of selective reporting of ESP. ganzfeld
studies. European Journal of Parapsychology, 3, 213-219.
BLACKMORE, S. (1987). A report of a visit to Carl Sargent's laboratory. Journal
of the Society for Psychical Research, 54, 186-198.
BRAUD, W. G. (1978). Psi conducive conditions: Explorations and interpre-
tations. In B. Shapin & L. Coly (Eds.), Psi and states of awareness (pp. 1-
34). New York: Parapsychology Foundation, Inc.
BRAUD, W. G., WOOD, R., & BRAUD, L. W. (1975). Free-response GESP per-
formance during an experimental hypnagogic state induced by visual
and acoustic ganzfeld techniques: A replication and extension. Journal of
the American Society for Psychical Research, 69, 105-113.
BRIGGS, K. C., & MYERS, I. B. (1957). Myers-Briggs Type Indicator Form F. Palo
Alto, CA: Consulting Psychologists Press, Inc.
BROWNLEE, K. A. (1965). Statistical theory and methodology in science and engi-
neering. New York: John Wiley & Sons, Inc.
CHILD, I. L. (1986). Comments on the ganzfeld controversy. Journal of Para-
psychology, 50, 337-344.
COHEN, J. (1977). Statistical power analysis for the behavioral sciences (rev. ed.).
New York: Academic Press.
DIACONIS, P., & MOSTELLER, F. (1989). Methods for studying coincidences.
Journal of the American Statistical Association, 84, 853-861.
DRUCKMAN, D., & SWETS, J. (1988). Enhancing human performance: Issues, the-
ories, and techniques. Washington, DC: National Academy Press.
HARLEY, T., & MATTHEWS, G. (1987). Cheating, psi, and the appliance of
science: A'reply to Blackmore. Journal of the Society for Psychical Research,
54, 199-207.
CPYRGHT
-
Approved For Release 2000/08/08 - CIA-RDP96-00789R003100120001-4

Journal of Parapsychology, Vol. 53, December 1989
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
"FUTURE TELLING": A META-ANALYSIS OF
FORCED-CHOICE PRECOGNITION
EXPERIMENTS, 1935-1987
BY CHARLES HONORTON AND DIANE C. FERRARI
ABSTRACT: We report a meta-analysis of forced-choice precognition experiments
published in the English-language parapsychological literature between 1935 and
1987. These studies involve attempts by subjects to predict the identity of target
stimuli selected randomly over intervals ranging from several hundred milli-
seconds to one year following the subjects' responses. We retrieved 309 studies
reported by 62 investigators. Nearly two million individual trials were contributed
by more than 50,000 subjects. Study outcomes are assessed by overall level of sta-
tistical significance and effect size. There is a small, but reliable overall effect (z
= 11.41, p = 6.3 x 10-25). Thirty percent of the studies (by 40 investigators) are
significant at the 5% significance level. Assessment of vulnerability to selective re-
porting indicates that a ratio of 46 unreported studies averaging null results would
be required for each reported study in order to reduce the overall result to nonsig-
nificance. No systematic relationship was found between study outcomes and eight
indices of research quality. Effect size has remained essentially constant over the
survey period, whereas research quality has improved substantially. Four moder-
ating variables appear to covary significantly with study outcome: Studies using
subjects selected on the basis of prior testing performance show significantly larger
effects than studies using unselected subjects. Subjects tested individually by an
experimenter show significantly larger effects than those tested in groups. Studies
in which subjects are given trial-by-trial or run-score feedback have significantly
larger effects than those with delayed or no subject feedback. Studies with brief
intervals between subjects' responses and target generation show significantly
stronger effects than studies involving longer intervals. The combined impact of
these moderating variables appears to be very strong. Independently significant
outcomes are observed in seven of the eight studies using selected subjects, who
were tested individually and received trial-by-trial feedback.
CPYRGHT
Precognition refers to the noninferential prediction of future
events. Anecdotal claims of "future telling" have occurred through-
out human history in virtually every culture and period. Today such
This work was funded by SRI International and the John E. Fetzer Foundation.
We wish to thank our PRL colleague George P. Hansen, who is primarily responsible
for retrieving the studies used in the meta-analysis. We are grateful to Edwin C. May,
Jessica Uuts, and to five anonymous reviewers at SRI for valuable comments on an
earlier draft of this report. Valuable comments were also made by Ephraim Schechter
and by three anonymous referees. The division of authorship responsibility is as fol-
lows: Honorton is responsible for the design of the meta-analysis, definition of study
coding criteria, the actual analyses, and the report itself. Ferrari coded the individual
research reports in consultation with Honorton and/or Hansen.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
.

202
The Journal of Parapsychology
claims are generally believed to be based on factors such as delusion.
irrationality, and superstitious thinking. The concept of precogni-
tion Quns counter to accepted notions of causality and appears to
confout with current scientific theory. Nevertheless, over the past
half-century a substantial number of experiments have been re-
porigi claiming empirical support for the hypothesis of precogni-
tion. Subjects in forced-choice experiments, according to many re-
porto have correctly predicted to a statistically significant degree the
identity (or order) of target stimuli randomly selected at a later
time@
W@performed a meta-analysis of forced-choice precognition ex-
perigents published in the English-language research literature be-
tweeR 1935 and 1987. Four major questions were addressed
through this meta-analysis: (1) Is there overall evidence for accurate
targgidentification (above-chance hitting) in experimental precog-
nitiogstudies? (2) What is the magnitude of the overall precognition
effec@ (3) Is the observed effect related to variations in methodo-
logic8 quality that could allow a more conventional explanation? (4)
Does .precognition performance vary systematically with potential
mod Quing variables, such as differences in subject populations,
stimusis conditions, experimental setting, knowledge of results, and
time Interval between subject response and target generation?
DELINEATING THE DOMAIN
RetriOal of Studies
CIA-RDP96-00789R00310012000
Pagapsychological research is still academically taboo, and it is
unlikely that there have been many dissertations and theses in this
area Bat have escaped publication. Our retrieval of studies for this
metaanalysis is therefore based on the published literature. The
studi@ include all forced-choice precognition experiments appear-
ing in the peer-reviewed English-language parapsychology journals:
Journey of Parapsychology, Journal (and Proceedings) of the Society for
PsychOl Research, Journal of the American Society for Psychical Research,
European Journal of Parapsychology (including the Research Letter of
the Utrecht University Parapsychology Laboratory), and abstracts of
peer-reviewed papers presented at Parapsychological Association
meetings published in Research in Parapsychology.
Criteria for Inclusion
Our review is restricted to fixed-length studies in which signifi-
cance levels and effect sizes based on direct hitting can be calcu-
A Meta-Analysis of Forced-Choice Precognition Experiments
lated. Studies using outcome variables other than direct hitting, such
as run-score variance and displacement effects, are included only
the report provides relevant information on direct hits (i.e., numbe
of trials, hits, and probability of a hit). Finally, we exclude studie
conducted by two investigators, S. G. Soal and Walter J. Levy, whom
work has been unreliable.
Many published reports contain more than one experiment
experimental unit. In experiments involving multiple conditions,
significance levels and effect sizes are calculated for each conditions
Outcome Measures
Significance level. Significance levels (z scores) were calculated fds
each study from the reported number of trials, hits, and probabilito
of success using the normal approximation to the binomial distra
bution with continuity correction. Positive z scores indicate aboveo
chance scoring, and negative z scores reflect below-chance scoring.
Effect size. Because most parapsychological experiments, partic.
ularly those in the older literature, have used the trial rather than
the subject as the sampling unit, we use a trial-based estimator
effect size. The effect size (ES) for each study is the z score divided
by the square root of the number of trials in the study.1
General Characteristics of the Domain
We located 309 studies in 113 separate publications. These stud
ies were contributed by 62 different senior authors and were pub
lished over a 53-year period, between 1935 and 1987. Considering
the half-century time-span over which the precognition experiments
were conducted, it is not surprising that the studies are very diverse
The database comprises nearly two million individual trials and
more than 50,000 subjects. Study sample sizes range from 25
297,060 trials (median = 1,194). The number of subjects range
from 1 to 29,706 (median = 16). The studies use a variety of meth
odologies, ranging from guessing ESP cards and other card symbol
to automated random number generator experiments. The domai
encompasses diverse subject populations: the most frequently used
' Elsewhere (Honorton, 1985), we have used the effect size index Cohen's h
(Cohen, 1977), and one referee has asked that we explain why we are now using
z/N1/2. The answer is that h and z/N12 yield virtually identical results, and z/N12 is.
computationally simpler. For the present sample of 309 precognition studies, the
mean difference between the two indices is .00047, and the standard deviation of the
difference is .026: ((308) = 0.312, p = . 756, two-tailed. The correlation between the
Iwn indirec ic 07
Approved For Release 2000/08/08: CIA-RDP96-00789R.003100120001-
CPYRGHT
283
Approved For Releas

284
The Journal of Parapsychology
A Meta-Analysis of Forced-Choice Precognition Experiments
285
O
0
Investigator
0
O
0
0
0
0
O
0
O
0
T
-0.2
-0.1
-0.0
0.1
0.2
0.3
0 .- 1
0.5
Mean effect size
Figure 1. Mean effect size by investigator. N = 62 investigators.
Replication Across Investigators
Virtually the same picture emerges when the cumulation is byd
investigator rather than study as the unit of analysis; the combined
z is 12.13, and 23 of the 62 investigators (37%) have overall out-
comes significant at the 5% level. The mean (investigator) effect sized
is 0.033 (SD = . 093).
There is a significant difference in the mean ES across investi-
gators, but it is surprisingly small: Kruskal-Wallis one-way ANOVA
by ranks, x2(61) = 82.71, p = . 034. The effect is clearly not due to
a few major contributors. If investigators contributing more thand
three studies are eliminated, leaving 33 investigators, the combined
z is still 6,00 (p = 1.25 x 10-9) and the mean ES is .028 (SD =
.091). Figure 1 shows the mean effect sizes by investigator.
These results indicate substantial cross-investigator replicability
and directly contradict the claim of critics such as Akers (1987) that
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003
00120001-4
Mean
0.65
0.020
OD
2.68
0.100
Dower 95% confidence estimate
0.40
0.011
Combined z = 11.41, p = 6.3 x 10-25
"Fail-safe N" = 14,268
t(ES) = 3.51, 308 df, p = . 00025
Population is students (in approximately 40% of the studies); the
Rast frequently used populations are the experimenters themselves
and animals (each used in about 5% of the studies).
Though a few studies tested subjects through the mail, more typ-
Billy subjects were tested in person, either individually or in groups.
@arget selection methods included no randomization at all (studies
Bing "quasi-random" naturalistic events), informal methods includ-
ing manual card-shuffling or dice-throwing, and formal methods,
Dimarily random number tables or random number generators.
The time interval between the subjects' responses and target gen-
gation varied from less than one second to one year.
OVERALL CUMULATION
Evidence for an overall effect is strong. As shown in the top part
Table 1, the overall results are highly significant.2 Lower bound
ne-tailed) 95% confidence estimates of the mean z score and ES
Be displayed in the bottom portion of Table 1.
Ninety-two studies (30%) show significant hitting at the 5% level,
and significant outcomes are contributed by 40 different investiga-
Qrs. The z scores correlate significantly with sample size: r(307) =
No56, p = . 003. The mean number of trials for significant studies is
3% larger than the mean number of trials for nonsignificant stud-
2 The statistical analyses presented here were performed using SYSTAT (Wilk-
inson, 1988). When t tests are reported on samples with unequal variances, they are
calculated using the separate variances within groups for the error and degrees of
freedom following Brownlee (1965). Unless otherwise specified, p levels are one-
tailed. Combined z's are based on Stouffer's method (Rosenthal, 1984).
TABLE I
OVERALL SIGNIFICANCE LEVEL AND EFFECT SIZE
Z
ES
.

286
The Journal of Parapsychology
successful parapsychological outcomes are achieved by only a few
investigators.
5.0
The Filedrawer Problem
$ Approved For Retease 2000/08/08 : CIA RDB96-00789R00310012000
well-known reporting bias exists throughout the behavioral
sciences favoring publication of "significant" studies (e.g., Sterling,
195
). The extreme view of this "filedrawer problem" is that "the
journals are filled with the 5% of the studies that show Type I er-
rorsOwhile the filedrawers back at the lab are filled with the 95% of
the Studies that show nonsignificance ... " (Rosenthal, 1984, p. 108).
Recognizing the importance of this problem, the Parapsychological
Assesiation in 1975 adopted an official policy against selective re-
porting of positive results.3 Examination of the parapsychological lit-
eratge shows that nonsignificant results are frequently published,
andOn the precognition database, 70% of the studies have reported
non &mificant results. Nevertheless, 75% of the precognition studies
were published before 1975, and we must ask to what extent selec-
tive Publication bias could account for the cumulative effects we ob-
serve.
ge central section of Table I uses Rosenthal's (1984) "fail-safe
N" statistic to estimate the number of unreported studies with z
scoreaveraging zero that would be necessary to reduce the known
database to nonsignificance. The filedrawer estimate indicates that
over 6 unreported studies must exist for each reported study to
reduce the cumulative outcome to a nonsignificant level.
Adifferent approach to the filedrawer problem is described by
Daw@9, Landman, and Williams (1984; personal communication
from Dawes to Honorton, July 14, 1988). Their truncated normal
curvyanalysis, like Rosenthal's "fail-safe N," is based on normal
curveassumptions. Their null hypothesis is that z scores above some
criticas level (e.g., z = 1.65, 1.96, etc.) are randomly sampled from
N(0, IRabove that critical level. The alternative to the null hypothesis
is that because there is some real effect, the distribution of z's is
shifted to the right of 0 and the z's will be larger than predicted by
the nol. For a critical level of z = 1.65, the expected mean z is 2.06
and the variance is .14. In the precognition database, there are 92
studies with z's > 1.65. Their average is 3.61, not 2.06 as predicted
A Meta-Analysis of Forced-Choice Precognition Experiments
287
by the null hypothesis. Since the variance of the normal truncated
above 1.65 is .14, the test z (using the Central Limit Theorem) com-
paring 3.61 to 2.06 is 39.84 [1.55 divided by (.14/92)1/2]. Here, p is
virtually zero. Similar results are found with cut points of 1.96, 2.33,
and 2.58.
On the basis of these analyses, we conclude that the cumulative
significance of the precognition studies cannot satisfactorily be ex-
plained by selective reporting.
OUTLIER REDUCTION
Although the overall z scores and effect sizes cannot reasonably
be attributed to chance, inspection of the standard deviations in
Table 1 indicates that the study outcomes are extremely heteroge-
neous. Given the diversity of methods, subject populations, and
other study features that characterize this research domain, this is
not surprising.
The study outcomes are in fact extremely heterogeneous. Al-
though a major objective of this meta-analysis is to account for the
variability across studies by blocking on differences in study quality,
procedural features, and sampling characteristics, the database
clearly contains extreme outliers. The z scores range from - 5.1 to
19.6, a 25-sigma spread! The standardized index of kurtosis (g2) is
9.47, suggesting that the tails of the distribution are much too long
for a normal distribution.
We eliminated the extreme outliers by performing a "10 percent
trim" on the study z scores (Barnett & Lewis, 1978). This involves
eliminating studies with z scores in the upper and lower 10% of the
distribution, and results in an adjusted sample of 248 studies. The
trimmed z scores range from -2.24 to 3.21 (g) = - 1.1). The re-
vised z scores and effect sizes are presented in Table 2.
Elimination of extreme outliers reduces the combined z scores by
approximately one half, but the outcomes remain highly significant.
Twenty-five percent of the studies (62/248) show overall significant
hitting at the 5% level. Lower bound confidence estimates show that
the mean z's and effect sizes are above 0 at the 95% confidence level.
Elimination of outliers reduces the total number of investigators
from 62 to 57, but the results remain basically the same when the
analyses are based on investigators rather than studies. The com-
bined z is 6.84; 18 of the 57 investigators (31.6%) have overall. sig-
" Analyses indicate no significant difference in the magnitude of reported study
outcomes before and after 1975. The mean ES for studies prior to 1975 is 0.021 (SD)
= . 099), and for studies reported thereafter the mean is 0.017 (SD = . 106): {(307)
= 0.28, p = . 782, two-tailed.
CPYRGHApproved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

288
The Journal of Parapsychology
A Meta-Analysis of Forced-Choice Precognition Experiments
289
O
O
0
O
O
O
Investigator
O
O
O
O
0
0
O
0
O
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
Mean effect size
Figure 2. Mean effect size by investigator for trimmed sample. N = 57 in-
vestigators.
practice, this is usually not feasible, particularly when, as in the pres-
ent case, large numbers of studies are involved. For our analysis of
study quality, statistical and methodological variables are defined
and coded in terms of procedural descriptions (or their absence) in
the research reports. This approach was used in an earlier meta-
analysis of psi ganzfeld research (Honorton, 1985), and it led to
study quality ratings that were generally in agreement, r(26) = . 766,
p = 10-6, with independent "flaw" ratings by an outside critic (Hy-
man, 1985).
One point is given (or withheld) for each of the following eight
criteria:
Specification of sample size. Does the investigator preplan the num-
ber of trials to be included in the study or is the study vulnerable
to the possibility of optional stopping? Credit is given to reports that
explicitly specify the sample size. Studies involving group testing, in
which it is not feasible to specify the sample size precisely, are also
Approved For Release 2000/081
Z
ES
Mean
0.38
0.012
SD
1.45
0.065
Lower 95% confidence estimate
0.23
0.005
Combined z = 6.02, p = 1.1 x 10-9
((ES) = 2.90, 247 df, p = . 002
nifica@ outcomes at the 5% level. The mean (investigator) ES is
0.020@SD = . 05).
Fan the trimmed sample, the difference in ES across investiga-
ors is not significant: Kruskal-Wallis one-way ANOVA by ranks,
"(56)= 59.34, p = . 355. If investigators contributing more than
hree Sudies are eliminated, leaving 37 investigators, the combined
z is s 5.00 (p = 3.0 x 10-7) and the mean ES is 0.022 (SD =
056).@figure 2 shows the mean effect size by investigator.
TIgs, elimination of the outliers does not substantially affect the
conclusions drawn from our analysis of the database as a whole.
Therclearly is a nonchance effect. In the remainder of this report,
ve use the trimmed sample to examine covariations in effect size
and aVariety of methodological and other study features.
CIA-RDP96-00789R00310012009154
STUDY QUALITY
Because target stimuli in precognition experiments are selected
only agger the subjects' responses have been registered, precognition
tudie!
Pare usually not vulnerable to sensory leakage problems.
Other @potential threats to validity must, however, be considered.
The poblem of variations in research quality remains a source of
ontrowersy in meta-analysis. Some meta-analysts advocate eliminat-
ng low quality studies whereas others recommend empirically ac-
essingThe impact of variations in quality on study outcome. Rosen-
hal (284) points out that the practice of discarding studies is
quivant to assigning them weights of zero, and he recommends
weighting study z scores in relation to ratings of research quality.
Study Quality Criteria
Ideally, the assessment of study quality should be performed by
nowledgeable specialists who are blind to the study outcomes. In
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
TABLE 2
SIGNIFICANCE LEVEL AND EFFECT SIZE FOR TRIMMED SAMPLE

290
The Journal of Parapsychology
given credit. No credit is given to studies in which the sample size
is either not preplanned or not addressed in the experimental re-
Approved
Preplanned analysis. Is the method of statistical analysis, including
e outcome (dependent variable) measure, preplanned? Credit is
gwen to studies explicitly specifying the form of analysis and the
gtcome measure. No credit is given to those not explicitly stating
the form of the analysis or those in which the analysis is clearly post
DRandomization method. Credit is given for use of random number
tales, random number generators, and mechanical shufflers. No
cadit is given for failure to randomize (i.e., use of "quasi-random
nguralistic events") or for informal methods such as hand-shuffling,
die-casting, and drawing lots.
Controls. Credit is given to studies reporting randomness control
csecks, such as random number generator (RNG) control series and
e
pirical cross-check controls.
SRecording. One point is allotted for automated recording of tar-
"B and responses, and another for duplicate recording.
g
Checking. One point is allotted for automated checking of
matches between target and response, and another for duplicate
checking of hits.
Study Quality Analysis
Each study received a quality weight between 0 and 8 (mean =
3.3J SD = 1.8). We find no significant relationship between study
quality and ES: +(246) = . 081, p = . 202, two-tailed. This tendency
fogstudy outcomes to correlate positively with study quatity has the
co@equence that the quality-weighted z score of 6.26 is slightly
larger than the unweighted z of 6.02. Table 3 shows the correlations
begeen effect size and each of the eight individual quality meas-
urg. The mean effect sizes by quality level are displayed graphi-
cal in Figure 3.
"The correlation between ES and study quality is also nonsignificant for the un-
tringned sample of 309 studies: r(307) = -. 060, p = . 289. The quality-weighted z
score is 7.38: p = 2.32 x 10-13. However, three of the individual quality measures
are significantly related to performance. Controls and duplicate checking correlate
significantly positively with ES, and randomization correlates significantly negatively
with ES. These correlations appear to be due to a few studies with z scores that are
extreme outliers (z > 7). When the 10 studies with z > 7 are eliminated, the signifi-
cant correlations between quality and ES disappear.
A Meta-Analysis of Forced-Choice Precognition Experiments
291
TABLE 3
CORRELATIONS BETWEEN EFFECT SIZE AND QUALITY MEASURES
Quality measure
r(246)
Sample size specified in advance
- . 100
Preplanned analysis
- . 001
Randomization
.011
Controls
.058
Automated recording
.169
Duplicate recording
.047
Automated checking
.136
Duplicate checking
.078
Quality Extremes
Is there a tendency for extremely weak studies to show larger
effects than exceptionally "good" studies? Analysis on the extremes
of the quality ratings indicates that this is not the case.
This analysis, based on the untrimmed sample of 309 studies,
uses studies with quality ratings outside the interquartile range of
the rating distribution (median = 4, Q, = 2, Q3 = 5). There are 56
"low-quality" studies (ratings of 0-1) and 35 "high-quality" studies
(ratings of 6-8). The high-quality studies have effect sizes that are
not significantly lower than the low-quality studies; the ES means
are 0.017 (SD = 0.063) and 0.037 (SD = 0.137), for the low- and
high-quality studies, respectively: t(82) = -. 92, p = . 358, two-
tailed.
Quality Variation in Publication Sources
Precognition ES is not significantly related to source of publica-
tion: Kruskal-Wallis one-way ANOVA; x2(4). = 0.78, p = . 942.
However, the sources of publication differ significantly in study
quality: Kruskal-Wallis one-way ANOVA, x2(4) = 17.19, p = . 002.
This is due largely to the lower quality of studies published in the
Journal of the Society for Psychical Research and in Research in Parapsy-
chology.
Study Quality in Relation to Year of Publication
Precognition effect size has remained constant over a half-cen-
tury of research, even though the methodological quality of the re-
CPYRGHT Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
ar Release. 2000/08/08: CIA-RDP96-00789R003900120001-4

2
The Journal of Parapsychology
8
" Approved For Release 2000/08/08 : CIA-RDP96
6
-4 5
13
1 7
4 45
Study quality
63
-
35
41
IH 31
0
8
-0.10
-0.05
0.00
0.05
0.10
Mean effect size
ure 3.@recognition effect size in relation to study quality, with 95% con-
ence legits. N = 248 studies.
arch les improved significantly during this period. The correla-
n betpten ES and year of publication is -. 071: t(307) = - 1.25,
.2130two-tailed. Study quality and year of publication are, how-
r, pogively and significantly correlated: r(246) = . 282, p = 2 x
7
, twotailed.
Crities of parapsychology have long believed that evidence for
apsy ological effects disappears as the methodological rigor in-
ases. The precognition database does not support this belief.
"REAL-TIME" ALTERNATIVES TO PRECOGNITION
Investigators have long been aware of the possibility that precog-
ion effects could be modeled without assuming either time rever-
or backward causality. For example, outcomes from studies with
A Meta-Analysis of Forced-Choice Precognition Experiments
293
targets based on indeterminate random number generators (RNGs)
could be due to a causal influence on the RNG-a psychokinetic
(PK) effect-rather than information acquisition concerning its fu-
ture state. In experiments with targets based on prepared tables of
random numbers, the possibility exists that the experimenter or
other randomizer may be the actual psi source, unconsciously using
"real-time" ESP combined with PK to choose an entry point in the
random number sequence that will significantly match the "sub-
ject's" responses. While the latter possibility may seem far-fetched,
it cannot be logically eliminated if one accepts the existing evidence
for contemporaneous ESP and PK, and it has been argued that it is
less far-fetched than the alternative of "true" precognition.
Morris (1982) discusses models of experimental precognition
based on "real-time" psi alternatives and methods for testing "true"
precognition. In general terms, these methods constrain the selec-
tion of the target sequence so as to eliminate nonprecognitive psi
intervention. In the most common procedure, attributed to Mangan
(1955), dice are thrown to generate a set of numbers that are math-
ematically manipulated to obtain an entry point in the random num-
ber table. This procedure is sufficiently complex "as to be appar-
ently beyond the capacities of the human brain, thus ruling out PK
because the 'PKer' would not know what to do even via ESP" (Mor-
ris, 1982, p. 329).
Two features of precognition study target determination proce-
dures were coded to assess "real-time" psi alternatives to precogni-
tion: method of determining random number table entry point and
use of Mangan's method.
.
Methods of eliminating "real-time" psi alternatives have not been
used in studies with random number generators and have only been
used in a small number of studies involving randomization by hand-
shuffling. These analyses are therefore restricted to studies using
random number tables (N = 138).
Method of Determining RNT Entry Point
The reports describe six different methods of obtaining entry.
points in random number tables. If the study outcomes were due to
subjects' precognitive functioning rather than to alternative psi ;
modes on the part of the experimenter or the experimenter's as-
sistants, there should be no difference in mean effect size across the
various methods used to determine the entry point. Indeed, our
analysis indicates that the study effect sizes do not vary systemati-
CPYRGHT Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

294
The Journal of Parapsychology
cally as a function of method of determining the entry point: Krus-
kal-Wallis one-way ANOVA by ranks: x"(5) = 7.32, p = . 198.
De of Mangan's Method
We find no significant difference in ES between studies using
Omplex calculations of the type introduced by Mangan to fix the
hdom number table entry point and those that do not use such
calculations: (45) = 0.38, p = . 370, two-tailed.
MODERATING VARIABLES
" The stability of precognition study outcomes over a 50-year pe-
od, which we described earlier, is also bad news. It shows that in-
Btigators in this area have yet to develop sufficient understanding
othe conditions underlying the occurrence (or detection) of these
Dects to reliably increase their magnitude. We have identified four
variables that appear to covary systematically with precognition ES:
(b) selected versus unselected subjects, (2) individual versus group
testing, (3) feedback level, and (4) time interval between subject re-
sponse and target generation.
OThe analyses use the raw study z scores and effect sizes; we
found that this results in uniformly more conservative estimates of
roationships with moderating variables than when the analyses are
b@ed on quality-weighted z scores and effect sizes.
Saicted Versus Unselected Subjects
Our meta-analysis identifies eight subject populations: unspeci-
fied subject populations, mixtures of several different populations,
annals, students, children, "volunteers," experimenter(s), and se-
lected subjects.
Effect size magnitude does not vary significantly across these
eight subject populations: Kruskal-Wallis one-way ANOVA, x2(7) =
10.90, p = . 143. Effect sizes by subject population are displayed in
Figure 4.
However, studies using subjects selected on the basis of prior
performance in experiments or pilot tests show significantly larger
effects than studies using unselected subjects. As shown in Table 4,
60% of the studies with selected subjects are significant at the 5%
level. The mean z score for these studies is 1.39 (SD = 1.40). The
ES is significantly higher for selected-subjects studies than for sind-
A Meta-Analysis of Forced-Choice Precognition Experiments
295
Selected
25
Exptr
12
Volunteer
.
26
Subject population
Children -
31
Students
108
Animals
- 10
Mixed
28
Unspec -
-
8
-0.06
-0.04
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
Mean effect size
Figure 4. Precognition effect size by subject population, with 95% confi-
dence limits. N = 248 studies.
ies with unselected subjects. The t test of the difference in mean ES
is equivalent to a point-biserial correlation of .198.
Does this difference result from less stringent controls in studies
with selected subjects? The answer appears to be "No." The average
quality of studies with selected subjects is higher than studies using
TABLE 4
SELECTED VERSUS UNSELECTED SUBJECTS
Selected
Unselected
N studies
25
223
Combined z
6.89
4.04
Studies with p. < . 05
60%
21%
Mean ES
.051
.008
SDES
.075
.063
((246) = 3.16 ₺ = 001
CPYRGHApproved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
EApproved For Release 2090108/08
DP 96-00789R003100320001-4
.

296
The Journal of Parapsychology
A Meta-Analysis of Forced-Choice Precognition Experiments
297
TABLE 5
INDIVIDUAL VERSUS GROUP TESTING
Approved For Release 2000/08/08 5 C
Individual
Group
& studies
97
105
ombined z
6.64
1.29
Dudies with p < . 05
30%
19%
Lean ES
.021
.004
.060
.066
t (200) = 1.89, p = . 03
omselected subjects: t(27) = 1.51, p = . 142, two-tailed. This result
pears to reflect a general tendency toward increased rigor and
pore detailed reporting in studies with selected subjects.
olividual Versus Group Testing
Subjects were tested in groups, individually, or through the mail.
Studies in which subjects were tested individually by an experimen-
have a significantly larger mean ES than studies involving group
testing (Table 5).
U The t test of the difference is equivalent to a point-biserial cor-
Hation of .132, favoring individual testing. Of the studies with sub-
ets tested individually, 30% are significant at the 5% level.
The methodological quality of studies with subjects tested indi-
Blually is significantly higher than that of studies involving group
ting: ((137) = 3.08, p = . 003, two-tailed. This result is consistent
«oh the conjecture that group experiments are frequently con-
gcted as "targets of opportunity" and may often be carried out
stily in an afternoon without the preparation and planning that
into a study with individual subjects that may be conducted over
period of weeks or months.
Thirty-five studies were conducted through the mail. In these
Didies, subjects completed the task at their leisure and mailed their
sponses to the investigator. These correspondence studies yield
atcomes similar to those involving individual testing. The com-
Jmed z score is 2.66, with a mean ES of 0.018 (SD = . 082). Ten
correspondence studies (25.7%) are significant at the 5% level.
Eleven studies are unclassifiable with regard to experimental set-
ting.
TABLE 6
FEEDBACK RECEIVED BY SUBJECTS
Feedback of Results
None
Delayed
Run score Trial-by-trial
N studies
15
21
21
47
Combined z
- 1.30
2.11
4.74
6.98
Studies with p < . 05
0.0%
19.0%
33.3%
42.6%
Mean ES
-. 001
.009
.023
.035
SDES
.028
.036
.048
.072
Feedback
A significant positive relationship exists between the degree of
feedback subjects receive about their performance and precognitive
effect size (Table 6).
Subject feedback information is available for 104 studies. These
studies fall into four feedback categories: no feedback, delayed
feedback (usually notification by mail), run-score feedback, and
trial-by-trial feedback. We gave these categories numerical values
between 0 and 3. Precognition effect size correlates .231 with feed-
back level (102 df, p = . 009). Of the 47 studies involving trial-by-
trial feedback, 20 (42.6%) are significant at the 5% level. None of
the studies without subject feedback are significant.
Feedback level correlates positively though not significantly with
research quality: r(102) = . 173, p = . 082, two-tailed. Inadequate
randomization is the most plausible source of potential artifacts in
studies with trial-by-trial feedback. We performed a separate analy-
sis on the 47 studies in this group. Studies using formal methods of
randomization do not differ significantly in mean ES from those
with informal randomization: t(15) = 0.67, p = . 590, two-tailed.
Similarly, studies reporting randomness control data do not differ
significantly in ES from those not including randomness controls:
t(42) = 0.79, p = . 436, two-tailed.
Time Interval
The interval between the subject's response and target selection
ranges from less than one second to one year. Information about
the time interval is available for 144 studies. This information, how-
A-RDP.96-09789R003100120001-4
.
CPYRGHT Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

298
The Journal of Parapsychology
.
Months
-
-
7
Weeks
17
Days
-
15
Hours
35
Minutes
19
Seconds
1-
1
20
Millisec
31
- 0.06
. - 0.04
- 0.02
0.00
0.02
1.0-1
0,06
0.08
0.10
Mean effect size
Bure 5. Effect size by precognition interval, with 95% confidence limits.
F 144 studies.
eter, is often imprecise. Our analysis of the relationship between
Decognitive ES and time interval is therefore limited to seven broad
Rerval categories: milliseconds, seconds, minutes, hours, days,
eks, and months. (Effect sizes by precognition interval are dis-
wyed in Figure 5.)
Although it is confounded with degree of feedback, there is a
sgnificant decline in precognition ES over increasing temporal dis-
tance: r(142) = -. 199, p = . 017, two-tailed. The largest effects oc-
car over the millisecond interval: N = 31 studies, combined z =
6.03, mean ES = 0.045, SD = . 073. The smallest effects occur over
periods ranging from a month to a year: N = 7, combined z =
0.53, mean ES = 0.001, SD = . 049.
Interestingly, the decline of precognition performance over in-
creasing temporal distances results entirely from studies using, un-
A Meta-Analysis of Forced-Choice Precognition Experiments
299
selected subjects: r(122) = - . 235. p = . 009, two-tailed. Studies with
selected subjects show a nonsignificant positive relationship between
ES and time interval: r(18) = . 077. p = . 745, two-tailed. Although
the difference between these two correlations is not significant (z =
1.24), this suggests that the origin of the decline over time may be
motivational rather than the result of some intrinsic physical bound-
ary condition. The relationship between precognition ES and feed-
back also supports this conjecture. Nevertheless, any finding sug-
gesting potential boundary conditions on the phenomenon should
be vigorously pursued.
Influence of Moderating Variables in Combination
The above analyses examine the impact of each moderating var-
iable in isolation. In this final set of analyses, we explore their joint
influence on precognition performance. For this purpose, we iden-
tify two subgroups of studies. One subgroup is characterized by the
use of selected subjects tested individually with trial-by-trial feed-
back. We refer to this as the Optimal group (N = 8 studies). The
second group is characterized by the use of unselected subjects
tested in groups with no feedback. We refer to this as the Suboptimal
group (N = 9 studies).
The Optimal studies are contributed by four independent inves-
tigators and the Suboptimal studies are contributed by two of the
same four investigators. All of the Optimal studies involve short pre-
cognition time intervals (millisecond interval); the Suboptimal stud-
ies involve longer intervals (intervals of weeks or months). All of the
Optimal studies and 5 of the 9 Suboptimal studies use RNG meth-
odology. The two groups do not differ significantly in average sam-
ple size. The mean study quality for the Optimal group is signifi-
cantly higher than that of the Suboptimal studies: Optimal mean =
6.63, SD = 0.92; Suboptimal mean = 3.44, SD = 0.53; t(10) =
8.63, p = 3.3 × 10-6, two-tailed.
The combined impact of the moderating variables appears to be
quite strong (Table 7). Seven of the 8 Optimal studies (87.5%) are
independently significant at the 5% level, whereas none of the Sub-
optimal studies are statistically significant. All four investigators con-
tributing studies to the Optimal group have significant outcomes.5
" In the untrimmed sample of 309 studies, there are a total of 17 Optimal studies.
The mean ES is 0.117 (SD = . 154), and the combined z is 15.84. The percentage of
independently significant studies is virtually the same as it is in the trimmed sample:
15 of the 17 studies (88.2%) are significant.
Precognition interval
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHApproved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

300
The Journal of Parapsychology
TABLE 7
IMPACT OF MODERATORS IN COMBINATION
"Optimal" studies
"Suboptimal" studies
V studies
8
9
Combined z
6.14
~1.29
Studies with p < . 05
87.5%
0.0%
Mean ES
.055
.005
.045
.035
((15) = 2.61, p = . 01
r = . 559
These results are quite striking and suggest that future studies
combining these moderators should yield especially reliable effects.
SUMMARY AND CONCLUSIONS
Our meta-analysis of forced-choice precognition experiments
sonfirms the existence of a small but highly significant precognition
effect. The effect appears to be replicable; significant outcomes are
reported by 40 investigators using a variety of methodological par-
gdigms and subject populations.
The precognition effect is statistically very robust: it remains
highly significant despite elimination of studies with z scores in the
upper and lower 10% of the z-score distribution and when a third
of the remaining investigators-the major contributors of precog-
Bition studies-are eliminated.
Estimates of the "filedrawer" problem and consideration of para-
&sychological publication practices indicate that the precognition ef-
fect cannot plausibly be explained on the basis of selective publica-
Dion bias. Analyses of precognition effect sizes in relation to eight
measures of research quality fail to support the hypothesis that the
Observed effect is driven to any appreciable extent by methodolog-
Scal flaws; indeed, several analyses indicate that methodologically su-
perior studies yield stronger effects than methodologically weaker
istudies.
Analyses of parapsychological alternatives to precognition, al-
though limited to the subset of studies using random number tables,
provide no support for the hypothesis that the effect results from
A Meta-Analysis of Forced-Choice Precognition Experiments
301
the operation of contemporaneous ESP and PK at the time of ran-
domization.
Although the overall precognition effect size is small, this does
not imply that it has no practical consequences. It is, for example,
of the same order of magnitude as effect sizes leading to the early
ermination of several major medical research studies. In 1981, the
National Heart, Lung, and Blood Institute discontinued its study of
propranolol because the results were so favorable to the propranolol
reatment that it would be unethical to continue placebo treatment
Kolata, 1981); the effect size was 0.04. More recently, The Steering
Committee of the Physicians' Health Study Research Group (1988),
n a widely publicized report, terminated its study of the effects of
ispirin in the prevention of heart attacks for the same reason. The
ispirin group suffered significantly fewer heart attacks than a pla-
:ebo control group; the associated effect size was 0.03.
The most important outcome of the meta-analysis is the identi-
ication of several moderating variables that appear to covary sys-
ematically with precognition performance. The largest effects are
observed in studies using subjects selected on the basis of prior test
performance, who are tested individually, and who receive trial-by-
rial feedback. The outcomes of studies combining these factors con-
rast sharply with the null outcomes associated with the combination
if group testing, unselected subjects, and no feedback of results. Be-
ause the two groups of studies were conducted by a subset of the
ame investigators, it is unlikely that the observed difference in per-
ormance is due to experimenter effects. Indeed, these outcomes
inderscore the importance of carefully examining differences in
ubject populations, test setting, and so forth, before resorting to
acile "explanations" based on psi-mediated experimenter effects or
he "elusiveness of psi."
The identification of these moderating variables has important
nplications for our understanding of the phenomena and provides
clear direction for future research. The existence of moderating
ariables indicates that the precognition effect is not merely an
nexplained departure from a theoretical chance baseline, but
: ather is an effect that covaries with factors known to influence
hore familiar aspects of human performance. It should now be pos-
ble to exploit these moderating factors to increase the magnitude
nd reliability of precognition effects in new studies.
REFERENCES
KERS, C. (1987). Parapsychology is science, but its findings are inconclusive.
Behavioral and Brain Sciences, 10, 566-568.
Approved For Release 2000/08/08 : CIA-RDP96-60789R003160120009-4
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
.

302
The Journal of Parapsychology
BARNETT, V. & LEWIS, T. (1978). Outliers in statistical data. New York: John
Wiley & Sons.
BROWNLEE, K. A. (1965). Statistical theory and methodology in science and engi-
nerring. New York: John Wiley & Sons.
COHEN, J. (1977). Statistical power analysis for the behavioral sciences. New York:
Academic Press.
AWES, R. M., LANDMAN, J., & WILLIAMS, J. (1984). Reply to Kurosawa. Amer-
ican Psychologist, 39, 74-75.
IONORTON, C. (1985). Meta-analysis of psi ganzfeld research: A response to
20 Hlyman. Journal of Parapsychology, 49, 51-92.
LYMAN, R. (1985). The ganzfeld psi experiment: A critical appraisal. Journal
of Parapsychology, 49, 3-50.
POLATA, G. B. (1981). Drug found to help heart attack survivors. Science, 214,
774-775.
BANGAN, G. L. (1955). Evidence of displacement in a precognition test. Journal
of Parapsychology, 19, 35-44.
SORRIS, R. L. (1982). Assessing experimental support for true precognition.
Journal of Parapsychology, 46, 321-336.
OSENTHAL, R. (1984). Meta-analytic procedures for social research. Beverly Hills,
CA: Sage.
STEERING COMMITTEE OF THE PHYSICIANS' HEALTH STUDY RESEARCH GROUP.
(1988). Preliminary report: Findings from the aspirin component of the
ongoing Physicians' Health Study. New England Journal of Medicine, 318,
262-264.
STERLING, T. D. (1959). Publication decisions and their possible effects on
inferences drawn from tests of significance-or vice versa. Journal of the
American Statistical Association, 54, 30-34.
SHL.KINSON, L .. (1988). SYSTAT: The system for statistics. Evanston, 11 .: SYSTAT.
CHRONOLOGICAL LISTING OF STUDIES IN META-ANALYSIS
ARINGTON, W. (1935). Preliminary experiments in precognitive guessing. Jour-
nal of the Society for Psychical Research, 29, 86-104.
BUINE, J. B. (1938). Experiments bearing on the precognition hypothesis: I.
Pre-shuffling card calling. Journal of Parapsychology, 2, 38-54.
HINE, J. B., SMITH, B. M., & WOODRUFF, J. L. (1938). Experiments bearing
on the precognition hypothesis: II. The role of ESP in the shuffling of
cards. Journal of Parapsychology, 2, 119-131.
TUMPHREY, B. M., & PRATT, J. G. (1941). A comparison of five ESP test
procedures. Journal of Parapsychology, 5, 267-293.
RHINE, J. B. (1941). Experiments bearing upon the precognition hypothesis:
III. Mechanically selected cards. Journal of Parapsychology, 5, 1-57.
STUART, C. E. (1941). An analysis to determine a test predictive of extra-chance
scoring in card-calling tests. Journal of Parapsychology, 5, 99-137.
HUMPHREY, B. M., & RHINE, J. B. (1942). A confirmatory study of salience in
precognition tests. Journal of Parapsychology, 6, 190-219.
A Meta-Analysis of Forced-Choice Precognition Experiments 303
RHINE. J. B. (1942). Evidence of precognition in the covariation of salience
ratios. Journal of Parapsychology, 6, 111-143.
NicoL., J. E., & CARINGTON, W. (1947). Some experiments in willed die-throw-
ing. Proceedings of the Society for Psychical Research, 48, 16-4-175.
THOULESS, R. H. (1949). A comparative study of performance in three psi
tasks. Journal of Parapsychology, 13, 263-273.
BASTIN, E. W., & GREEN, J. M. (1953). Some experiments in precognition.
Journal of Parapsychology, 17, 137-143.
MCMAHAN, E. A., & BATES, E. K. (1954). Report of further Marchesi exper-
iments. Journal of Parapsychology, 18, 82-92.
MANGAN, G. L. (1955). Evidence of displacement in a precognition test. Journal
of Parapsychology, 19, 35-44.
Osis, K. (1955). Precognition over time intervals of one to thirty-three days.
Journal of Parapsychology. 19, 82-91.
NIELSEN, W. (1956). An exploratory precognition experiment. Journal of Para-
psychology, 20, 33-39.
NIELSEN, W. (1956). Mental states associated with success in precognition. Jour-
nal of Parapsychology, 20, 96-109.
FAHLER, J. (1957). ESP card tests with and without hypnosis. Journal of Para-
psychology, 21, 179-185.
MANGAN, G. L. (1957). An ESP experiment with dual-aspect targets involving
one trial a day. Journal of Parapsychology, 21, 273-283.
ANDERSON, M., & WHITE, R. (1958). A survey of work on ESP and teacher-
pupil attitudes. Journal of Parapsychology, 22, 246-268.
NASH, C. B. (1958). Correlation between ESP and religious value. Journal of
Parapsychology, 22, 204-209.
ANDERSON, M. (1959). A precognition experiment comparing time intervals of
a few days and one year. Journal of Parapsychology, 23, 81-89.
ANDERSON, M., & GREGORY, E. (1959). A two-year program of tests for clair-
voyance and precognition with a class of public school pupils. Journal of
Parapsychology, 23, 149-177.
NASH, C. B. (1960). Can precognition occur diametrically? Journal of Parapsy-
chology, 24, 26-32.
FREEMAN, J. A. (1962). An experiment in precognition. Journal of Parapsychology,
26, 123-130.
1
RHINE, J. B. (1962). The precognition of computer numbers in a public test.
Journal of Parapsychology, 26, 244-251.
RYzL., M. (1962). Training the psi faculty by hypnosis. Journal of the Society for
Psychical Research, 41, 234-252.
SANDERS, M. S. (1962). A comparison of verbal and written responses in a
precognition experiment. Journal of Parapsychology, 26, 23-34.
FREEMAN, J. (1963). Boy-girl differences in a group precognition test. Journal
of Parapsychology, 27, 175-181.
RAO, K. R. (1963). Studies in the preferential effect: II. A language ESP test
involving precognition and "intervention." Journal of Parapsychology, 27, 147-
160.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGI Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

904
A Meta-Analysis of Forced-Choice Precognition Experiments
305
:
CARPENTER, J. C. (1969). Further study on a mood adjective check list and
ESP run-score variance. Journal of Parapsychology, 33, 48-56.
DUVAL, P., & MONTREDON, E. (1969). Precognition in mice: A confirmation.
Journal of Parapsychology, 33, 71-72.
FREEMAN, J. A. (1969). The psi-differential effect in a precognition test. Journal
of Parapsychology, 33, 206-212.
FREEMAN, J. A. (1969). A precognition experiment with science teachers. Journal
of Parapsychology, 33, 307-310.
JOHNSON, M. (1969). Attitude and target differences in a group precognition
test. Journal of Parapsychology, 33, 324-325.
MONTREDON, E., & ROBINSON, A. (1969). Further precognition work with mice.
Journal of Parapsychology, 33, 162-163.
SCHMIDT, H. (1969). Precognition of a quantum process. Journal of Parapsy-
chology, 33, 99-108.
BENDER, H. (1970). Differential scoring of an outstanding subject on GESP and
clairvoyance. Journal of Parapsychology, 34, 272-273.
FREEMAN, J. A. (1970). Sex differences in ESP response as shown by the Free-
man picture-figure test. Journal of Parapsychology, 34, 37-46.
FREEMAN, J. A. (1970). Ten-page booklet tests with elementary-school children.
Journal of Parapsychology. 34, 192-196.
FREEMAN, J. (1970). Shift in scoring direction with junior-high-school students:
A summary. Journal of Parapsychology, 34, 275.
FREEMAN, J. A. (1970). Mood, personality, and attitude in precognition tests.
Journal of Parapsychology, 34, 322.
HARALDSSON, E. (1970). Subject selection in a machine precognition test. Journal
of Parapsychology, 34, 182-191.
HARALDSSON, E. (1970). Precognition of a quantum process: A modified rep-
lication. Journal of Parapsychology, 34, 329-330.
NIELSEN, W. (1970). Relationships between precognition scoring level . and
mood. Journal of Parapsychology, 34, 93-116.
SCHMIDT, H. (1970). Precognition test with a high-school group. Journal of
Parapsychology, 34, 70.
BELOFF, J., & BATE, D. (1971). An attempt to replicate the Schmidt findings.
Journal of the Society for Psychical Research, 46, 21-31.
HONORTON, C. (1971). Automated forced-choice precognition tests with a "sen-
sitive." Journal of the American Society for Psychical Research, 65, 476-481.
MITCHELL, E. D. (1971). An ESP test from Apollo 14. Journal of Parapsychology,
35, 89-107.
SCHMIDT, H., & PANTAS, L. (1971). Psi tests with psychologically equivalent
conditions and internally different machines. Journal of Parapsychology, 35,
326-327.
STANFORD, R. G. (1971). Extrasensory effects upon "memory." Journal of the
American Society for Psychical Research, 64, 161 - 186.
STEILBERG, B. J. (1971). Investigation of the paranormal gifts of the Dutch
sensitive Lida T. Journal of Parapsychology, 35, 219-225.
Approved For Release 2006/08/08 &CIA-REP96-00789R003100120601-4
FREEMAN, J. (1964). A precognition test with a high-school science club. Journal
of Parapsychology, 28, 214-221.
FREEMAN, J., & NIELSEN, W. (1964). Precognition score deviations as related
to anxiety levels. Journal of Parapsychology, 28, 239-249.
SCHMEIDLER, G. (1964). An experiment on precognitive clairvoyance: Part I.
The main results. Journal of Parapsychology, 28, 1-14.
FREEMAN, J. A. (1965). Differential response of the sexes to contrasting ar-
rangements of ESP target material. Journal of Parapsychology, 29, 251-258.
OSIs, K., & FAHLER, J. (1965). Space and time variables in ESP. Journal of the
American Society for Psychical Research, 59, 130-145.
MAHLER, J., & Osis, K. (1966). Checking for awareness of hits in a precognition
experiment with hypnotized subjects. Journal of the American Society for
Psychical Research, 60, 340-346.
FREEMAN, J. A. (1966). Sex differences and target arrangement: High-school
booklet tests of precognition. Journal of Parapsychology, 30, 227-235.
ROGERS, D. P. (1966). Negative and positive affect and ESP run-score variance.
Journal of Parapsychology, 30, 151 - 159.
ROGERS, D. P., & CARPENTER, J. C. (1966). The decline of variance of ESP
scores within a testing session. Journal of Parapsychology, 30, 141 - 150.
PRIER, B. (1967). A correspondence ESP experiment with high-I.Q. subjects.
Journal of Parapsychology, 31, 143-1418.
UZBY, D. E. (1967). Subject attitude and score variance in ESP tests. Journal
of Parupsychology, 31, 43-50.
BUZBY, D. E. (1967). Precognition and a test of sensory perception. Journal of
Parapsychology, 31, 135-142.
REEMAN, J. A. (1967). Sex differences, target arrangement, and primary men-
tal abilities. Journal of Parapsychology, 31, 271-279.
ONORTON, C. (1967). Creativity and precognition scoring level. Journal of
Parapsychology, 31, 29-42.
ARPENTER, J. C. (1968). Two related studies on mood and precognition run-
score variance. Journal of Parapsychology, 32, 75-89.
IVAL, P., & MONTREDON, E. (1968). ESP experiments with mice. Journal of
Parapsychology, 32, 153-166.
ATHER, S. R., & BRIER, R. (1968). The possible effect of the checker in
precognition tests. Journal of Parapsychology, 32, 167-175.
BEEMAN, J. A. (1968). Sex differences and primary mental abilities in a group
precognition test. Journal of Parapsychology, 32, 176-182.
FISH, C. S., & NASH, C. B. (1968). Effect of target selection, field dependence,
and body concept on ESP performance. Journal of Parapsychology, 32, 248-
257.
AINE, L. E. (1968). Note on an informal group test of ESP. Journal of Para-
psychology, 32, 47-53.
YZL, M. (1968). Precognition scoring and attitude toward ESP. Journal of Para-
psychology, 32, 1-8.
RYzI., M. (1968). Precognition scoring and attitude. Journal of Parapsychology,
32, 183-189.
CPYRGHT
Approved For Releas
2000/08/08 - CIA-RDP96-00789R003100120001-4

306
The Journal of Parapsychology
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4-
THOULESS, R. H. (1971). Experiments on psi self-training with Dr. Schmidt's
pre-cognitive apparatus. Journal of the Society for Psychical Research, 46, 15-
21.
HONORTON, C. (1972). Reported frequency of dream recall and ESP. Journal
of the American Society for Psychical Research, 66, 369-374.
JOHNSON, M., & NORDBECK, B. (1972). Variation in the scoring behavior of a
"psychic" subject. Journal of Parapsychology, 36, 122-132.
TKELLY, E. F., & KANTHAMANI, B. K. (1972). A subject's efforts toward voluntary
control. Journal of Parapsychology, 36, 185-197.
SCHMIDT, H., & PANTAS, L. (1972). Psi tests with internally different machines.
Journal of Parapsychology, 36, 222-232.
@GRAIG, J. G. (1973). The effect of contingency on precognition in the rat.
Research in Parapsychology 1972, 154-156.
PFREEMAN, J. A. (1973). The psi quiz: A new ESP test. Research in Parapsychology
1972, 132-134.
PARTLEY, B. (1974). Confirmation of the small-rodent precognition work. Journal
of Parapsychology, 38, 238-239.
HARRIS, S., & TERRY, J. (1974). Precognition in a water-deprived Wistar rat.
Journal of Parapsychology, 38, 239.
RANDALL, J. L. (1974). An extended series of ESP and PK tests with three
English schoolboys. Journal of the Society for Psychical Research, 47, 485-494.
EYSENCK, H. J. (1975). Precognition in rats. Journal of Parapsychology, 39, 222-
227.
HARALDSSON, E. (1975). Reported dream recall, precognitive dreams, and ESP.
Research in Parapsychology 1974, 47-48.
HONORTON, C., RAMSEY, M., & GABIBBO, C. (1975). Experimenter effects in
extrasensory perception. Journal of the American Society for Psychical Research,
69, 135-149.
KANTHAMANI, H., & RAO, H. H. (1975). Response tendencies and stimulus
structure. Journal of Parapsychology, 39, 97-105.
LEVIN, J. A. (1975). A series of psi experiments with gerbils. Journal of Para-
psychology, 39, 363-365.
TERRY, J. C., & HARRIS, S. A. (1975). Precognition in water-deprived rats.
Research in Parapsychology 1974, 81.
DAVIS, J. W., & HAIGHT, J. (1976). Psi experiments with rats. Journal of Para-
psychology, 40, 54-55.
ACOBS, J., & BREEDERVELD, H. (1976). Possible influences of birth order on
ESP ability. Research Letter (Parapsychology Laboratory, University of
Utrecht), No. 7, 10-20.
NEVILLE, R. C. (1976). Some aspects of precognition testing. Research in Para-
psychology 1975, 29-31.
DRUCKER, S. A., DREWES, A. A., & RUBIN, L .. (1977). ESP in relation to cognitive
development and IQ in young children. Journal of the American Society for
Psychical Research, 71, 289-298.
A Meta-Analysis of Forced-Choice Precognition Experiments
307
HARALDSSON, E. (1977). ESP and the defense mechanism test (DMT): A further
validation. European Journal of Parapsychology, 2, 104-114.
SARGENT, C. L. (1977). An experiment involving a novel precognition task.
Journal of Parapsychology, 41, 275-293.
BIERMAN, D. J. (1978). Testing the "advanced wave" hypothesis: An attempted
replication. European Journal of Parapsychology, 2, 206-212.
BRAUD, W. (1979). Project Chicken Little: A precognition experiment involving
the SKYLAB space station. European Journal of Parapsychology, 3, 149-165.
HARALDSSON, E., & JOHNSON, M. (1979). ESP and the defense mechanism test
(DMT) Icelandic study No. III: A case of the experimenter effect? European
Journal of Parapsychology, 3, 11-20.
O'BRIEN, J. T. (1979). An examination of the checker effect. Research in Para-
psychology 1978, 153-155.
CLEMENS, D. B., & PHILLIPS, D. T. (1980). Further studies of precognition in
mice. Research in Parapsychology 1979, 156.
HARALDSSON, E. (1980). Scoring in a precognition test as a function of the
frequency of reading on psychical phenomena and belief in ESP. Research
Letter (Parapsychology Laboratory, University of Utrecht), No. 10, 1-8.
SARGENT, C., & HARLEY, T. A. (1981). Three studies using a psi-predictive
trait variable questionnaire. Journal of. Parapsychology, 45, 199-214.
WINKELMAN, M. (1981). The effect of formal education on extrasensory abil-
ities: The Ozolco study. Journal of Parapsychology, 45, 321-336.
NASH, C. B. (1982). ESP of present and future targets. Journal of the Society for
Psychical Research, 51, 374-377.
THALBOURNE, M., BELOFF, J., & DELANOY, D. (1982). A test for the "extra-
verted sheep versus introverted goats" hypothesis. Research in Parapsychology
1981, 155-156.
CRANDALL, J. E., & HITE, D. D. (1983). Psi-missing and displacement: Evidence
for improperly focused psi? Journal of the American Society for Psychical Re-
search, 77, 209-228.
SCHWARTZ, S. A., & DEMATTEI, R. J. (1983). The Mobius Psi-@ test: Prelim-
inary findings. Research in Parapsychology 1982, 103-105.
JOHNSON, M., & HARALDSSON, E. (1984). The Defense Mechanism Test as a
predictor of ESP scores: Icelandic studies IV and V. Journal of Parapsychology,
48, 185-200.
TEDDER, W. (1984). Computer-based long-distance ESP: An exploratory ex-
amination (RB/PS). Research in Parapsychology 1983, 100-101.
HESELTINE, G. L. (1985). PK success during structured and nonstructured
RNG operation. Journal of Parapsychology, 49, 155- 163.
HARALDSSON, E., & JOHNSON, M. (1986). The Defense Mechanism Test (DMT)
as a predictor of ESP performance: Icelandic studies VI and VII. Research
in Parapsychology 1985, 43-44.
VASSY, Z. (1986). Experimental study of complexity dependence in precogni-
tion. Journal of Parapsychology, 50, 235-270.
CPYRGHT
Approved For Releas
2000/08/08 . CIA-RDP96-00789R0031
001
20001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
·
308
The Journal of Parapsychology
HONORTON, C. (1987). Precognition and real-time ESP performance in a com-
puter task with an exceptional subject. Journal of Parapsychology, 51, 291-
320.
Psychophysical Research Laboratories
P. O. Box 569
Plainsboro, NJ 08536
4
CPYRGHT Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

2
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Paranormal Communication
"Error Some Place!"
by Charles Honorton
Review of the ESP controversy
traces debate from statistical
and methodological issues to
the a priori critique and the
paradigm of "normal science."
Asked his opinion of ESP, a skeptical psychologist once retorted, "Error
Some Place!" I believe he was right, but for the wrong reasons. Western
science has always been ambivalent toward the mental side of reality, and
it is perhaps not surprising that the occurrence of "psychic" phenomena is
one of the most controversial topics in the history of science.
CPYRGHT
The first serious effort toward scientific examination of psi claims was
undertaken by the Society for Psychical Research (SPR), founded in London
in 1882 for the purpose of "making an organized and systematic attempt
to investigate the large group of phenomena designated by such terms as
mesmeric, psychical, and spiritualistic." The SPR leadership included many
distinguished scholars of the period, and similar organizations quickly
spread to other countries, including the American Society for Psychical Re-
search, founded in New York in 1885 under the aegis of William James,
who himself took an active role in early investigations of mediumistic
communications.
These turn-of-the-century investigators focused much of their attention
on authenticating individual cases of spontaneous experiences suggestive
of psi communication. While a great deal of provocative material was care-
fully examined and reported (e.g., 13), the limitations inherent in the case
study approach prohibited definitive conclusions. However thoroughly au-
thenticated, spontaneous cases cannot provide adequate assessment of such
potential sources of contamination as chance coincidence, unconscious in-
ference and sensory leakage, retroactive falsification, or deliberate fraud.
Charles Honorton is director of research in the Division of Parapsychology and
Psychophysics, Department of Psychiatry, Maimonides Medical Center, Brooklyn, N.Y.
Approved For Release 2000/08/08 : CIA RDP96 00789R003100120001 4
Id3

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Journal of Communication, Winter 1975
Early experimental approaches primarily involved the "telepathic" repro-
duction of drawings at a distance (62). While often striking correspondences
were obtained, the experimental conditions did not usually provide for
random selection of target (stimulus) material, and were not always totally
adequate with respect to the possibility of sensory leakage, intentional or
otherwise.
Neither the spontaneous case studies nor the early experimental efforts
made much impact upon the scientific community, though they drew critical
comment from prominent period scientists. "Neither the testimony of all
the Fellows of the Royal Society, nor even the evidence of my own senses,"
proclaimed Helmholtz, "would lead me to believe in the transmission of
thought from one person to another independently of the recognized chan-
nels of sense." Thomas Huxley declined an invitation to participate in
some of the early SPR investigations, saying he would sooner listen to the
idle gossip of old women.
The rudiments of an experimental methodology
for testing psi were suggested three
centuries ago by Francis Bacon.
In Sylva Sylvarum, a work published posthumously, Bacon discussed
"experiments in consort, monitory, touching transmission of spirits and
forces of imagination." He suggested that "the motions of shuffling cards,
or casting of dice" could be used to test the "binding of thoughts. . . .
The experiment of binding of thoughts should be diversified and tried to
the full; and you are to note whether it hit for the most part though not
always" (2).
The application of probability theory to the assessment of deviations
from theoretically expected chance outcomes was introduced to psychical
research in 1884 by the French Nobel laureate, Charles Richet, in experi-
ments involving card-guessing. The popularity of card-guessing as an ex-
perimental methodology was greatly influenced by the work of J. B. Rhine
and his associates at Duke University in the early 1930s. Rhine (50) devised
a standard set of procedures around a simplified card deck containing
randomized sequences of five geometric forms (circle, cross, wavy lines,
square, and circle. These "ESP cards" were prepared in packs of 25, and
each "run" through the pack was associated with a constant binomial prob-
ability of 1/5, since subjects were not given trial-by-trial feedback. Provid-
ing the experimental conditions were adequate to eliminate illicit sensory
cubes, recording errors, and rational inference, statistically significant de-
partures from binomial chance expectation were interpreted as indicating
extrasensory communication.
R
S
Initially, "telepathy" tests consisted of having a subject in one room
attempt to identify the order of the cards as they were observed by an
"agent" in another room. In "clairvoyance" tests, the subject attempted to
"guess" the order of the cards directly, as they lay concealed in an opaque
104
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Journal of Communication, Winter 1975
scores were in all cases nonsignificant, with a mean scoring rate of 5.04 (43).
Several critics questioned the applicability of the binomial distribution
as a basis for assessing the statistical significance of ESP card-guessing data.
Willoughby (78) proposed the use of an empirical control series, but later
withdrew the suggestion after comparing the two methods (79). Alternative
methods of deriving the probable error and recommendations for using
the empirical standard deviation were also proposed and later withdrawn
(21, 22). Concern over this issue diminished and was generally abandoned
following the publication of a large chance control series involving half
a million trials and demonstrating close approximation to the binomial
model (12).
Another question arose about whether the binomial model provides
sufficient approximation to the normal distribution to allow use of normal
probability integral tables for determination of significance levels (17).
Stuart and Greenwood (73) showed that when the normal distribution is
used as an approximation to the binomial model, discrepancies are im-
portant only with cases of borderline significance and few trials.
The use of the binomial critical ratio (z) to evaluate the significance of
the ESP card-guessing deviations was generally approved by professional
statisticians (6, 20). Fisher (10), however, commented that high levels of
statistical significance should not be accepted as substitutes for independent
replication. In another vein, Huntington (20) asked, "If mathematics has
successfully disposed of the hypothesis of chance, what has psychology to
say about the hypothesis of ESP?"
The most frequently expressed methodological
concern was the possibility of some
form of "sensory leakage," giving the ESP
subject enough information about the targets
to account for significant, extrachance results.
--
As early as 1895, two Danish psychologists, Hansen and Lehmann
(16), reported that with the aid of parabolic reflectors subjects could detect
digits and other material silently concentrated upon by an agent. In these
experiments, the subject and agent sat with their heads close to the foci
of two concave mirrors. While the agent concentrated on the number, he
made a special effort to keep his lips closed. Under these conditions, the
subjects were frequently successful in identifying the number. These results
were interpreted by Hansen and Lehmann as supporting the hypothesis
of "involuntary whispering." The utilization of subtle sensory cues was
demonstrated in a careful investigation by S. G. Soal of a stage "telepathist"
(66). There were also reports, such as the case of "Ilga K.," a mentally
retarded Latvian child who could read any text, even in a foreign language,
when someone stood behind her, reading "silently." Experiments with
dictaphone recordings revealed that "Ilga" was responding to very slight
auditory cues (3).
106
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Paranormal Communication / "Error Some Place!"
container or in another room, without an agent. "Precognition" tests,
introduced somewhat later (59), required the subject to make anticipatory
guesses of the card order before the pack was shuffled or otherwise random-
ized.
Rhine introduced the term "ESP" in his first major report on the Duke
University work in 1934 (50). He reported a total of 85,724 card-guessing
trials, carried out with a wide variety of subjects and under a wide range
of test conditions. The results as a whole were astronomically significant,
though informal exploratory trials were indiscriminately pooled with those
carried out under more carefully controlled conditions. The best-controlled
work during this period was the Pearce-Pratt distance series of clairvoyance
tests (58), in which the subject, Pearce, located in one building, attempted
to identify the order of the cards as they were handled, but not viewed, by
Pratt, the experimenter, located in another building. The level of accuracy
obtained in this series of 1,850 trials was associated with a probability of
10-22.
As a stimulant to experimental research, Rhine's work had unprece-
dented influence. For the first time a common methodology was adopted
and employed on a large scale by a number of independent and widely
separated investigators. For the first time, also, the scientific community
was confronted with a body of data, collected through conventional meth-
ods, which it could no longer ignore-nor too hastily accept. The wide-
scale adoption of the card-guessing methodology was accompanied by a
plethora of critical articles, challenging almost every aspect of the evalua-
tive techniques and the experimental conditions. During the period be-
tween 1934 and 1940, approximately 60 critical articles by 40 authors ap-
peared, primarily in the psychological literature. While card-guessing is
no longer the primary methodology in experimental parapsychology, the
questions which arose over its use are of equal relevance to the more
sophisticated approaches used today.
The first major issue concerned the
validity of the assumption that the
probability of success in the card-guessing
experiments was actually 1/5.
If chance expectation is other than 1/5, the significance of the observed
deviations would obviously be in doubt. This issue was quickly resolved
by mathematical proof and through empirical "cross-checks," a form of
control series in which responses (guesses) were deliberately compared with
target orders for which they were not intended (e.g., responses on run n,
matched with the target sequence for run n.). Empirical cross-checks were
reported for 24 separate experimental series involving a total of 12,228
runs (305,700 individual trials). While the actual experimental run scores
(e.g., guesses on run n, compared to targets for run n1) were highly sig-
nificant and yielded a mean scoring rate of 7.23/25, the control cross-check
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
105

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Paranormal Communication / "Error Some Place!"
It is clear that at least some of the early exploratory series reported in
Rhine's monograph were open to criticism for inadequate controls against
sensory cues. While Rhine did not base major conclusions on such poorly
controlled data, inclusion of them in his monograph provided a ready target
for critical reviewers and sidetracked discussion away from the better con-
trolled work, such as the Pearce-Pratt series, which was not susceptible to
explanation by sensory cues.
Defects in an early commercial printing of ESP cards were reported by
several investigators (18, 25). It was found that the cards were warped and
could under certain conditions be identified from the back. This discovery
circulated widely for a time as an explanation of all successful (i.e., statis-
tically significant) experimental series. The parapsychologists retorted that
defective cards had not been employed in any of the experiments reported
in the literature and that, in any case, they could not account for results
from studies involving adequate screening with such devices as opaque
envelopes, screens, distance, or work involving the precognition paradigm
in which the target sequences were not generated until after the subject
had made his responses (53, 54, 72).
By 1940 nearly one million experimental trials had been reported under
conditions which precluded sensory leakage. These included five studies
in which the target cards were enclosed in opaque sealed envelopes (41, 45,
46, 54, 59), 16 studies employing opaque screens (7, 8, 11, 19, 33, 34, 35, 38,
41, 42, 44, 45, 46, 59, 71), ten studies involving separation of subjects and
targets in different buildings (50, 51, 52, 53, 34, 32, 8, 77, 61, 60), and two
studies involving precognition tasks (59, 75). These data are summarized in
Table 1. The results were independently significant in 27 of the 33 experi-
ments. By the end of the 1930s there was general agreement that the better-
controlled ESP experiments could not be accounted for on the basis of
sensory leakage.
The hypothesis that significant "extrachance" deviations in ESP experi-
ments might be attributable to motivated scoring errors was investigated in
several studies. In one investigation (26), 28 observers recorded 11,125 mock
ESP trials. Of these, 126 (1.13 percent) were misrecorded. Observers favor-
Table 1: ESP card-guessing experiments (1934-1939) excluding sensory cues'
Method
Studies
N (Trials)
Mean/25
P<
"Clairvoyance" paradigm,
stimuli in sealed, opaque
envelopes
5
129,775
5.21
4.0 × 10-11
"Clairvoyance" paradigm,
stimuli concealed by opaque
screens
16
497,450
5.44
2.0 × 10-12
Distanceb
10
164,475
5.37
18-12
Precognition paradigme
2
115,330
5.15
2.95 × 10 -*
· References given in text.
b Includes work with both "telepathy" and "clairvoyance" paradigms
· Stimuli generated after subjects made their responses
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4197

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Journal of Communication, Winter 1975
able to the ESP hypothesis made 71.5 percent more errors of commission
(increasing ESP scores), while those who were unfavorable to the ESP
hypothesis made 100 percent more errors of omission (decreasing ESP
scores). Murphy ($7) reported an analysis of 175,000 trials from experiments
reporting positive evidence for ESP and found only 175 errors (0.10 per-
cent). Greenwood (12) reported only 90 recording errors in rechecking his
500,000-trial control study, of which 76 were errors of omission.
Some critics also alleged that improper selection of data could account
for experimental successes. This could be done in several ways: (a) selection
of subjects; (b) selection of particular blocks of data out of larger samples;
(c) selection of one of several forms of analysis; and (d) selective reporting
of particular studies. The questions raised have sometimes been stated
cynically in the form, "Parapsychologists must run 100 subjects before
they find one with 'ESP'." As if in defense against this charge, a number
of the reported studies specifically stated that all of the data collected were
included in the analysis (see 43, pp. 118-124, Table 12).
Concerning selection of subjects, Warner (76) suggested two criteria:
first, results of "poor" subjects must be included up to the point when
they are discontinued since it does not matter how many trials a given
subject makes as long as all of the trials (for all subjects) are included;
second, exclude all preliminary trials (for both "good" and "poor" sub-
jects) and use preliminary screening studies to select "good" candidates for
formal work. These criteria were generally endorsed by the chief critics of
the period (e.g., 23).
The question of post hoc selection of analyses was not a point of serious
concern in the period between 1934 and 1940, though it is relevant to the
assessment of some of the process-oriented investigations reported more
recently. The question of whether nonsignificant studies were withheld
from publication involves an issue which is of great concern to the be-
havioral sciences as a whole (70, 81) and one which is difficult to accurately
assess since there is no way of knowing how many studies may have been
withheld from publication because their results failed to disconfirm the
null hypothesis.
Several studies of American Psychological Association publication poli-
cies (4, 70, 81) indicate that experimental studies in general are more likely
to be published if the null hypothesis is rejected at the conventional .05
and .01 alpha levels than if it is not rejected. These studies also indicate
that a negligible proportion of published studies are replications. Bozarth
and Roberts (4), in a survey of 1,334 articles from psychological journals,
found that 94 percent of the articles involving statistical tests of significance
reported rejection of specific null hypotheses; only eight articles (less than
1 percent) involved replications of previously published studies.
With respect to the implications of such selection for the ESP hypothesis,
there are two partial answers. First, considering the degree of critical inter-
est which prevailed in the 1980s, it seems unlikely that nonsignificant find-
ings would have been repressed during this period; second, the high levels
108
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

ApprovedForRelease 2000/08/08 : CIA-RDP96-00789R003100120001-4
Paranormal Communication / "Error Some Place!" .
of significance attached to some of the reported ESP investigations would
necessitate postulating astronomical numbers of "chance" trials in order
to dilute the overall deviations to chance. To take one example, consider
the Pearce-Pratt series of 1,850 trials which yielded p = 10-22. As Soal and
Bateman (66) pointed out, it is difficult to believe that 1010 (ten thousand
million) sets of 1,850 trials could have possibly been carried out between
1984 and 1940 (or, for that matter, since 1940). But, as Soal and Bateman
suggest, ". .. if we posit this absurd estimate as an upper limit [with overall
chance totals], that would still give us odds of 1010 . .. against the supposi-
tion that the Pearce-Pratt results were a run of pure luck."
The possibility of obtaining significant "extrachance" results by stopping
an experimental series at "favorable" points was also raised (9, 31). While
this "optional stopping" hypothesis was generally agreed to be of significance
only in cases of marginally significant results, it led to the adoption of
several procedural modifications: specification of the total number of trials
in advance of data collection, or accumulation of data in blocks of pre-
determined size.
The possibility was raised by several critics that hand-shuffled cards may
display a tendency to stick together or otherwise produce patterns which
could produce spurious results (24, 82). While the cross-check type of con-
trol series, described earlier, failed to reveal any evidence of patterning,
there was a general trend away from hand shuffling in the later published
studies, which utilized tables of prepared random numbers as a basis for
generating target sequences.
There was-and is (e.g., 15)-a rather widespread belief that most of
the evidence supporting the ESP hypothesis originated in the Duke Uni-
versity studies and that most independent replications by other investigators
were nonconfirmatory. A survey of the published literature between 1934
and 1940 fails to support this claim. Table 2 shows all the published experi-
mental reports during this period which provided statistical treatment of
the data. Inspection of this table reveals that a majority (61 percent) of
the outside replications report significant results (p < . 01) and that the
proportion of significant studies was not significantly greater for the Duke
University group (x2 = 1.70, 1 df).
By 1940, the active methodological
controversy was over.
The issues raised were, for the most part, legitimate, and investigators
modified their procedures to safeguard their results from methodological
criticism. The major issues raised since 1940 center on alleged anomalies in
probability theory and the hypothesis of widespread investigator fraud.
Spencer Brown (68, 69) has suggested that statistically significant card-
guessing studies provide evidence, not of extrasensory modes of communica-
tion, but of fundamental defects in probability theory. He makes three
criticisms of random number sequences: (a) published random number
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
109

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Journal of Communication, Winter 1975
Table 2: Breakdown of experimental ESP studies (1934-1939)
N
(Studies)*
N studies reported
significant (p < ,01)
% signif.
Duke group
17
15
88
Non-Duke-
33
20
61
Total
50
35
70
· Includes all English-language studies involving assessment of statistical significance
of data, 1934-1939 inclusive.
xª (Duke vs. non-Duke X significant vs. nonsignificant) = 1.70 (1 df)
sequences have been "doctored" prior to publication in order to remove
certain nonrandom features; this practice, according to Spencer Brown,
makes such sequences nonrandom and invalidates the use of standard
significance tests; (b) the source of some random number sequences involves
randomizing machines which utilize the unpredictability of human be-
havior when examined for microscopic variation; such variation, says
Spencer Brown, may be predictable enough to account for observed anom-
alies in random sequences, as well as some of the significant results reported
in ESP guessing experiments; (c) Spencer Brown produces evidence to show
that anomalous (significant) departures from probability theory can be
obtained by matching columns of random numbers (39).
A detailed examination of these points was undertaken by Scott (64).
With respect to "doctored" sequences, Scott showed that the maximum
error due to rejected (edited) sequences would not affect interpretations
of results which are more than marginally significant and could, in fact,
increase the likelihood of making a Type II error. On the hypothesis that
ESP results are due to some kind of hyper-regularity affecting both the
target sequence and the response (guess) sequence simultaneously and
similarly, Scott makes the point that this would lead to the expectation of
similar results from matching any set of humanly produced random se-
quences. The cross-check type of control series and the Greenwood chance
control series described earlier demonstrate that this is not the case. The
anomalies reported by Spencer Brown (68), obtained by arbitrarily matching
columns of random numbers, have been criticized on the basis of post hoc
selection (40) and illustrate not that there are fundamental defects in prob-
ability theory, but rather that significant deviations from chance can
occur in any data where hypotheses and analyses are not specified in advance.
The most recent phase of the ESP controversy centers on the hypothesis
of investigator fraud. This argument was most forcefully presented in a
lead article in Science, entitled "Science and the Supernatural," by G. R.
Price (47), who began with the following observations:
Believers in psychic phenomena . . . appear to have won a decisive
victory and virtually silenced opposition. . . . This victory is the result
of an impressive amount of careful experimentation and intelligent
argumentation. . . . Against all this evidence, almost the only defense
110
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Paranormal Communication / "Error Some Place!"
remaining to the skeptical scientist is ignorance, ignorance concerning
the work itself and concerning its implications. The typical scientist
contents himself with retaining ... some criticism that at most applies
to a small fraction of the published studies. But these findings (which
challenge our very concepts of space and time) are-if valid-of enor-
mous importance . .. so they ought not to be ignored.
Following Hume's argument on miracles, Price asserted that ESP is
"incompatible with current scientific theory," and that it is therefore more
parsimonious to believe that parapsychologists cheat than that ESP is a
real phenomenon. He concluded, "My opinion concerning the findings of
the parapsychologists is that many of them are dependent on clerical and
statistical errors and unintentional use of sensory clues, and that all extra-
chance results not so explicable are dependent on deliberate fraud or mildly
abnormal mental conditions" (47, p. 360). This extraordinary critique and
the ensuing discussion in Science (5, 36, 48, 55, 56, 65) were widely reviewed.
As Meehl and Scriven (46) pointed out, Price's argument rests on two highly
questionable assumptions, namely that contemporary scientific knowledge
is complete, and that ESP necessarily conflicts with it. Seventeen years
later, in an "Apology to Rhine and Soal," Price retracted his accusations
of investigator fraud (49).
Very similar arguments have, however, been made more recently by
the British parapsychological critic C. E. M. Hansel (14, 15), who began
his examination of the ESP hypothesis by suggesting that "the a priori
arguments . .. may even save time and effort in scrutinizing the [ESP]
experiments. ... In view of the a priori arguments against it we know in
advance that telepathy, etc., cannot occur."
Because of the "a priori unlikelihood" of ESP, Hansel's examination
of the literature centered primarily on the possibility of fraud, by subjects
or investigators. He reviewed in depth four experiments which he regarded
as providing the best evidence of ESP: the already-mentioned Pearce-Pratt
distance series (59); the Pratt-Woodruff (44) series, also conducted at Duke;
and Soal's work with Mrs. Stewart and Basil Shackleton (66), as well as
a more recent series by Soal and Bowden (67). Hansel showed, in each
case, how fraud could have been committed (by the experimenters in the
Pratt-Woodruff and Soal-Bateman series, and by the subjects in the Pearce-
Pratt and Soal-Bowden experiments). He gave no direct evidence that fraud
was committed in these experiments but said, "If the result could have
arisen through a trick, the experiment must be considered unsatisfactory
proof of ESP, whether or not it is finally decided that such a trick was in
fact used" (15, p. 18, italics mine).
Hansel's argument is unclear, inasmuch as he quite properly insists that
no single experiment can be conclusive, then proceeds to show that none
is, given the theoretical possibility of fraud by subjects or investigators.
Hansel's only conclusion after more than 250 pages of careful scrutiny
was that these experiments were not "fraud-proof" and therefore not con-
clusive proof of ESP.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
111

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
1
Journal of Communication, Winter 1975
Two recent examples, one involving cancer research (74) and the other
involving parapsychology (57), serve to remind us of the importance of
cross-validation in the assessment of any experimental finding. In both
cases, it should be added, the fraudulent acts were detected in-house, by
the researchers themselves. The point is that in the final analysis an experi-
mental finding is of value and is to be taken seriously only to the extent
that it leads to further inquiry. To regard any experiment as an end in
itself is to remove it from the domain of experimental science. It is obvious
that hypothetical construct, such as ESP, cannot be validated by any
isolated experiment, no matter how well controlled it might be. Inde-
pendent replication is a necessary prerequisite.
The claim that psi phenomena operate
outside the framework of physical probability
has been a major source of a
priori arguments against acceptance of ESP.
It has been suggested that to accept ESP requires the rejection of
physics. This is absurd, and it is worth noting that such arguments have
usually been advanced and defended by psychologists rather than by
physicists.
The debate over the incompatibility of physics and ESP has been
conducted almost exclusively within the framework of nineteenth-century
deterministic physics, wherein the ultimate constituent of physical reality
was still believed to be solid matter. Inasmuch as modern microphysics has
exorcised the material out of matter and deals with processes which on
our macrophysical level of sensory perception are every bit as erratic and
anomalous as ESP, the a priori claim that ESP violates specifiable laws of
physics can no longer be considered to be of more than historical interest.
ESP and other psi phenomena, while no longer incompatible with
physics, are not yet accounted for by physics; but then, neither are the
more familiar processes of memory and conscious experience. Indeed, the
transformation of "raw feels" into conscious experience is no less a problem
for the neurophysiologists of today than it was for the speculative philos-
ophers of classical antiquity. Sir John Eccles, among others, has repeatedly
warned, "We should not pretend that consciousness is not a mystery."
The ESP controversy illustrates several features of the paradigmatic
view of science developed by Thomas Kuhn (28). Normal science, according
to Kuhn, is essentially a clean-up operation, constrained by a broad
theoretical framework, or paradigm, which defines the boundaries of legiti-
mate inquiry. Paradigms are scientific world views which provide coherence
and structure and determine the types of questions to be posed of nature,
as well as the manner in which answers are sought. Normal science is thus
a process of paradigm-articulation, rather than of discovery. Within the
paradigm structure of normal science, observations which conflict with
112
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Paranormal Communication / "Error Some Place!"
the paradigm are seldom made; anomalies are ignored. When the anomalies
become sufficiently persistent that they can no longer be ignored, they
are hotly disputed. Eventually, a new paradigm is tentatively erected
which attracts a group of adherents, and a period of crisis ensues which
Kuhn calls a paradigm clash.
In this review I have focused at some length on the period of the
1930s, not because it provides the best available evidence for ESP or
the best understanding of the processes underlying its operation-it does
neither, but rather because it was during this period that the major
substantive methodological issues were raised and to a large extent con-
sensually resolved. Since 1940, well over 10,000 journal pages devoted to
parapsychological research have been published, and at least 250 experi-
mental studies have been reported. The methodological foundations of
the research have gradually diversified, enlarging and enriching the scope
of inquiry and providing a basis for more sophisticated study. Automated
testing equipment has replaced card-guessing in forced-choice ESP tasks,
and quantitative methods have been developed for the objective assessment
of psi interactions in nonguessing tasks. Psychophysiological techniques,
permitting determination of psi-optimal organismic states, have been in-
troduced and utilized in conjunction with experimental methods more
closely approximating the conditions under which psi interactions occur
in vivo. More important, parapsychological investigators have to a large
extent shifted their attention away from the "proof-oriented" approach,
which can only reaffirm the presence of anomaly, toward systematic at-
tempts to identify the antecedent conditions necessary for the occurrence
and detection of psi interactions, the delineation of positive attributes,
and the study of individual differences. Only through the pursuit of such
"process-oriented" research can we ever hope to achieve the goals of control,
assured replicability (or at least predictability), and eventual understanding.
REFERENCES
1. American Psychological Association. "ESP Symposium at the American Psychological
Association." Journal of Parapsychology 2, 1938, pp. 247-272.
2. Bell. M. "Francis Bacon, Pioneer in Parapsychology." International Journal of Para-
psychology 6, 1964, pp. 199-208.
3. Bender, H. "The Case of Ilga K .: Report of a Phenomenon of Unusual Perception."
Journal of Parapsychology 2, 1938, pp. 5-22.
1. Bozarth, J. D., and R. R. Roberts. "Signifying Significant Significance." American
Psychologist 27. 1972. pp. 774-775.
6. Bridgman, P. W. "Probability, Logic, and ESP." Science 123. 1956, pp. 15-17.
6. Camp. B. H. "Statement under 'Notes.'" Journal of Parapsychology 1, 1937. p. 305.
. Carpenter, C. R., and H. R. Phalen. "An Experiment in Card Guessing." Journal of
Parapsychology 1, 1937, pp. 31-43.
. Crumbaugh. J. C. An Experimental Study of Extra-sensory Perception. M.A. Thesis,
1938, Southern Methodist University Library.
9. Feller, W. "Statistical Aspects of ESP." Journal of Parapsychology 4, 1940, pp. 271-298.
10. Fisher. R. A. Letter to J. B. Rhine, cited in (50), p. 45.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
113

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Journal of Communication, Winter 1975
11. Gibson, E. P. "A Study of Comparative Performance in Several ESP Procedures."
Journal of Parapsychology 1, 1937, pp. 264-275.
12. Greenwood. J. A. Analysis of a Large Chance Control Series of ESP Data." Journal
of Parapsychology 2, 1938, pp. 138-146.
13. Gurney. E., F. Myers, and F. Podmore. Phantasms of the Living (2 vols.). London:
Trubner and Co., 1886.
14. Hansel, C. F. M. "Experiments on Telepathy in Children." British Journal of Statistical
Psychology 13, 1960, pp. 175-178.
15. Hansel, C. E. M. ESP-A Scientific Evaluation. New York: Scribners, 1966.
16. Hansen, F. C. C., and A. Lehmann. "Über Unwillkürliches Flüstern." Philosophical
Studies 17. 1895, pp. 471-530.
17. Heinlein, C. P., and J. H. Heinlein. "Critique of the Premises and Statistical Meth.
odology of Parapsychology." Journal of Psychology 5, 1938, pp. 135-148.
18. Herbert, C. V. C. "Experiment in Extra-sensory Perception." Journal of the Society
for Psychical Research 30, 1938, pp. 215-218.
19. Humphrey, B. M., and J. A. Clark. "A Comparison of Clairvoyant and Chance Match-
ing." Journal of Parapsychology 2. 1938, pp. 31-37.
20. Huntington, E. V. "Is It Chance or ESP?" American Scholar 7, 1938, pp. 201-210.
21. Kellogg. C. E. "Dr. J. B. Rhine and Extra-sensory Perception." Journal of Abnormal
and Social Psychology 31, 1936, pp. 190-193.
22. Kellogg, C. E. "New Evidence (?) for Extra-sensory Perception." Scientific Monthly,
1937. pp .- 331-341.
23. Kellogg, C. E. Critical comments in (43). pp. 228-238.
24. Kennedy, J. L. "A Methodological Review of Extra-sensory Perception. Psychological
Bulletin 36. 1939, pp. 59-103.
25. Kennedy, J. L. "The Visual Cues from the Backs of the ESP Cards." Journal of
Psychology 5, 1938, pp. 149-153.
26. Kennedy, J. L., and H. F. Uphoff. "Experiments on the Nature of Extra-sensory
Perception: III. The Recording Error Criticism of Extra-chance Scores." Journal
of Parapsychology 3. 1939, pp. 226-245.
27. Kennedy, J. L. Critical comments in (43). pp. 218-220.
28. Kuhn, T. S. The Structure of Scientific Revolutions. Chicago: University of Chicago
Press, 1962.
29. Lemmon, V. W. "The Role of Selection in ESP Data," Journal of Parapsychology 3.
1939, pp. 104-106.
30. Lemmon, V. W. Critical comments in (43), pp. 222-228.
31. Leuba. C. "An Experiment to Test the Role of Chance in ESP Research." Journal of
Parapsychology 2, 1938, pp. 217-221.
32. MacFarland, J. D. "Discrimination Shown Between
Experimenters by Subjects."
Journal of Parapsychology 2, 1938, pp. 160-170.
33. MacFarland, J. D., and R. W. George. "Extra-sensory Perception of Normal and
Distorted Symbols." Journal of Parapsychology 1. 1937, pp. 93-101.
34. Martin, D. R. "Chance and Extra-chance Results in Card-matching." Journal of
Parapsychology 1, 1937, pp. 185-190.
35. Martin, D. R., and F. P. Stribic. "Studies in Extra-sensory Perception: I. An Analysis
of 25,000 Trials; and II. An Analysis of a Second Series of 25,000 Trials." Journal
of Parapsychology 2, 1938, pp. 23-30: 287-295.
36. Mechi, P. E., and M. Scriven. "Compatibility of Science and ESP." Science 123, 1956,
pp. 14-15.
37. Murphy, G. "On Limits of Recording Errors." In (1), pp. 262-266.
38. Murphy. G., and E. Taves. "Covariance Methods in the Comparison of Extra-sensory
Tasks." Journal of Parapsychology 3, 1939, pp. 38-78.
39. Oram. A. T. "An Experiment with Random Numbers." Journal of the Society for
Psychical Research 37, 1954, pp. 369-377.
40. Oram, A. T. Correspondence. Journal of the Society for Psychical Research 38, 1955.
pp. 143-144.
114
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
:

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Paranormal Communication / "Error Some Place!"
41. Pratt, J. G. "Clairvoyant Blind Matching." Journal of Parapsychology 1, 1937. Pp.
10-17; and "The Work of Dr. C. Hilton Rice in Extra-sensory Perception, pp. 239-
259.
42. Pratt, J. G., and M. Price. "The Experimenter-subject Relationship in Tests for ESP."
Journal of Parapsychology 2. 1938, pp. 84-94.
43. Pratt, J. G., J. B. Rhine, B. M. Smith, C. E. Stuart, and J. A. Greenwood. Extrasensory
Perception After Sixty Years. Boston: Bruce Humphries, 1966. (Originally published
in 1940 by Henry Holt, New York.)
44. Pratt, J. G., and J. L. Woodruff. "Size of Stimulus Symbols in Extra-sensory Percep-
tion." Journal of Parapsychology 3, 1939, pp. 121-158.
45. Price, M. "A Comparison of Blind and Seeing Subjects in ESP Tests." Journal of
Parapsychology 2. 1938, pp. 273-286.
46. Price, M., and M. Pegram. "Extrasensory Perception Among the Blind." Journal of
Parapsychology 1, 1937, pp. 143-155.
47. Price, G. R. "Science and the Supernatural," Science 122, 1955, pp. 359-367.
48. Price, G. R. "Where is the Definitive Experiment?" Science 123. 1956, pp. 17-18.
49. Price, G. R. "Apology to Rhine and Soal." Science 175, 1972, pp. 359.
50. Rhine. J. B. Extra-sensory Perception. Boston: Bruce Humphries, 1964. (Originally
published in 1934.)
51. Rhine, J. B. "Telepathy and Clairvoyance in the Normal and Trance States of a
Medium." Character and Personality 3, 1934, pp. 91-111.
52. Rhine, J. B. "Some Selected Experiments in Extra-sensory Perception." Journal of
Abnormal and Social Psychology 31, 1936, pp. 216-228.
53. Rhine, J. B. "The Question of Sensory Cues and the Evidence." Journal of Para-
psychology 1, 1937, pp. 276-291.
54. Rhine. J. B. "ESP Tests with Enclosed Cards." Journal of Parapsychology 2. 1938,
pp. 199-216.
55. Rhine, J. B. "Comments on 'Science and the Supernatural.'" Science 123, 1956, pp.
11-14.
56. Rhine, J. B. "The Experiment Should Fit the Hypothesis." Science 123, 1956, p. 19.
57. Rhine, J. B. Comments. Journal of Parapsychology, 1974, p. 38.
58. Rhine, J. B., and J. G. Pratt. "A Review of the Pearce-Pratt Distance Series of ESP
Tests." Journal of Parapsychology 18, 1954, pp. 165-177.
59. Rhine, J. B., B. M. Smith, and J. L. Woodruff. "Experiments Bearing on the Pre-
cognition Hypothesis: II. The Role of ESP in the Shuffling of Cards." Journal of
Parapsychology 2, 1938. pp. 119-131.
60. Riess, B. F. "A Case of High Scores in Card Guessing at a Distance." Journal of
Parapsychology 1, 1937, pp. 260-263.
61. Shulman, R. "A Study of Card-guessing in Psychotic Subjects." Journal of Parapsy-
chology 2, 1938, pp. 96-107.
62. Sinclair, U. Mental Radio, with a Preface by Albert Einstein. Springfield, Ill .: Charles
C. Thomas, 1962. (Originally published in 1930.)
63. Shapiro, A. "Review of ESP-A Scientific Evaluation by C. E. M. Hansel." International
Journal of Clinical and Experimental Hypnosis 16, 1968, pp. 133-134.
64. Scott, C. "G. Spencer Brown and Probability: A Critique." Journal of the Society
for Psychical Research 39, 1958, pp. 217-234.
65. Soal, S. G. "On 'Science and the Supernatural.' " Science 123, 1956, pp. 9-11.
66. Soal, S. G., and F. Bateman. Modern Experiments in Telepathy. New Haven: Yale
University Press, 1954.
67. Soal, S. G., and H. T. Bowden. The Mind Readers: Recent Experiments in Telepathy.
New Haven, Conn .: Yale University Press, 1954.
68. Spencer Brown, G. "Statistical Significance in Psychical Research." Nature 172, 1953,
pp. 154-156.
69. Spencer Brown, G. Probability and Scientific Inference. New York: Longmans, Green,
1957.
70. Sterling, T. C. "Publication Decisions and Their Possible Effects of Inference Drawn
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
115

Approved Far Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Journal of Communication, Winter 1975
from Tests of Significance-Or Vice Versa." Journal of the American Statistical As-
sociation 54, 1959, pp. 30-34.
71. Stuart, C. E. "The Effect of Rate of Movement in Card Matching Tests of Extra-
sensory Perception." Journal of Parapsychology 2, 1938, pp. 171-183.
72. Stuart, C. E. "A Review of Certain Proposed Hypothèses Alternative to Extra-sensory
Perception." Journal of Abnormal and Social Psychology 33, 1938, pp. 57-70.
73. Stuart, C. E., and J. A. Greenwood. "A Review of Criticisms of the Mathematical
Evaluation of ESP Data." Journal of Parapsychology 1, 1937, pp. 295-304.
74. Time, April 29, 1974.
75. Tyrrell, G. N. M. "Some Experiments in Undifferentiated Extra-sensory Perception."
Journal of the Society for Psychical Research 29, 1935, pp. 52-68.
76. Warner, L. "The Role of Luck in ESP Data." Journal of Parapsychology 1, 1937, pp.
84-92.
77. Warner, L. "A Test Case." Journal of Parapsychology 1. 1937, pp. 234-238.
78. Willoughby, R. "Critical Comment: The Use of the Probable Error in Evaluating
Clairvoyance." Character and Personality 4, 1935, pp. 79-80.
79. Willoughby, R. "A Critique of Rhine's 'Extra-sensory Perception.' " Journal of Ab.
normal and Social Psychology 30, 1935, pp. 199-207.
80. Willoughby, R. Critical comments in (43), pp. 220-221.
81. Wolins, L. "Needed: Publication of Negative Results." American Psychologist 14, 1959,
p. 598.
82. Wolfle, D. L. "A Review of the Work on Extra-sensory Perception." American Journal
of Psychiatry, 1938, pp. 945-955.
116
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
- ----

10
329
PROCEEDINGApproved ForRelease 2000/08/08 : CIA-RDP96-00789R003100120001-4
A Perceptual Channel for Information Transfer
over Kilometer Distances: Historical Perspective
and Recent Research
HAROLD E. PUTHOFF, MEMBER, IEEE, AND RUSSELL TARG, SENIOR MEMBER, IEEE
Abstract -For more than 100 years, scientists have attempted to
determine the truth or falsity of claims for the existence of a perceptual
channel whereby certain individuals are able to perceive and describe
remote data not presented to any known sense. This paper presents an
outline of the history of scientific inquiry into such so-called paranor-
mal perception and surveys the current state of the art in parapsycho-
logical research in the United States and abroad. The nature of this
perceptual channel is examined in a series of experiments carried out
in the Electronics and Bioengineering Laboratory of Stanford Research
Institute. The perceptual modality most extensively investigated is the
ability of both experienced subjects and inexperienced volunteers to
view, by innate mental processes, remote geographical or technical
targets including buildings, roads, and laboratory apparatus. The ac-
cumulated data indicate that the phenomenon is not a sensitive func-
tion of distance, and Faraday cage shielding does not in any apparent
way degrade the quality and accuracy of perception. On the basis of
this research, some areas of physics are suggested from which a descrip-
tion or explanation of the phenomenon could be forthcoming.
CPYRGHT
1
INTRODUCTION
CPYRGHT
unification brought about by the work of Ampere, Faraday,
and Maxwell. Since the early work, however, we have seen the
development of information theory, quantum theory, and
neurophysiological research, and these disciplines provide
powerful conceptual tools that appear to bear directly on the
issue. In fact, several physicists (Section V) are now of the
opinion that these phenomena are not at all inconsistent with
the framework of modern physics: the often-held view that
observations of this type are a priori incompatible with known
laws is erroneous in that such a concept is based on the naive
realism prevalent before the development of quantum theory.
In the emerging view, it is accepted that research in this area
can be conducted so as to uncover not just a catalog of inter-
esting events, but rather patterns of cause-effect relationships
of the type that lend themselves to analysis and hypothesis
in the forms with which we are familiar in the physical
solenocu. One hypothesis is that information transfer under
'T IS THE PROVINCE of natural science to investigate
nature, impartially and without prejudice" [1]. Nowhere
in scientific inquiry has this dictum met as great a chal-
conditions of sensory shielding is mediated by extremely
low-frequency (ELF) electromagnetic waves, a proposal that
does not seem to be ruled out by any obvious physical or
lenge as in the area of so-called extrasensory perception (ESP), "biological facts. Further, the development of information
the detection of remote stimuli not mediated by the usual.
theory makes it possible to characterize and quantify the
performance of a communications channel regardless of
the underlying mechanism.
sensory processes. Such phenomena, although under scientific
consideration for over a century, have historically been fraught
with unreliability and controversy, and validation of the phe-
nomena by accepted scientific methodology has been slow in
coming. Even so, a recent survey conducted by the British
publication New Scientist revealed that 67 percent of nearly
1500 responding readers (the majority of whom are working
scientists and technologists) considered ESP to be an estab-
lished fact or a likely possibility, and 88 percent held the
investigation of ESP to be a legitimate scientific undertaking
[2].
A review of the literature reveals that although experiments
by reputable researchers yielding positive results were begun
over a century ago (e.g., Sir William Crookes' study of D. D.
Home, 1860's) [3], many consider the study of these phe-
nomena as only recently emerging from the realm of quasi-
science. One reason for this is that, despite experimental
results, no satisfactory theoretical construct had been advanced
to correlate data or to predict new experimental outcomes.
Consequently, the area in question remained for a long time
in the recipe stage reminiscent of electrodynamics before the
Manuscript received July 25, 1975; revised November 7, 1975. The
submission of this paper was encouraged after review of an advance
proposal. This work was supported by the Foundation for Parasensory
Investigation and the Parapsychology Foundation, New York, NY; the
Institute of Noetic Sciences, Palo Alto, CA; and the National Aero-
nautics and Spage Administ
The authors Approved For Release-2000/08/08 : CIA-RDP96-0078, 5 "was reported
ectronics and BiSengTheching
Stanford Research Institute, Menlo Park, CA 94025.
For the past three years, we have had a program in the
Electronics and Bioengineering Laboratory of the Stan-
ford Research Institute (SRI) to investigate those facets of
human perception that appear to fall outside the range of well-
understood perceptual/processing capabilities. Of particular
interest is a human information-accessing capability that we
call "remote viewing." This phenomenon pertains to the
ability of certain individuals to access and describe, by means
of mental processes, information sources blocked from ordi-
nary perception, and generally accepted as secure against such
access.
In particular, the phenomenon we have investigated most
extensively is the ability of a subject to view remote geograph-
ical locations up to several thousand kilometers distant from
his physical location (given only a known person on whom to
target)." We have carried out more than fifty experiments
under controlled laboratory conditions with several individuals
whose remote perceptual abilities have been developed suf-
ficiently to allow them at times to describe correctly-often in
great detail-geographical or technical material such as build-
ings, roads, laboratory apparatus, and the like.
As observed in the laboratory, the basic phenomenon appears
to cover a range of subjective experiences variously referred to
6-00789R0031001200Qdrie [4], and re-
printed in the IEEE Commun. Soc. Newsletter, vol. 13, Jan. 1975.

330
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003F66/2000MINE IEEE, MARCH 1976
VAEROPUERTO DESPITE THE
Zuday + /12/ 73
1:25 -130
ocear
Hanger of
Arpa
wer. ...
BRIGHT
Sand
Fig. 1. Airport in San Andres, Colombia, used as remote-viewing target, along with sketch produced by subject
in California.
in the literature as autoscopy (in the medical literature); exteri-
orization or disassociation (psychological literature); simple
clairvoyance, traveling clairvoyance, or out-of-body experience
(parapsychological literature); or astral projection (occult liter-
ature). We choose the term "remote viewing" as a neutral
descriptive term free from prior associations and bias as to
mechanisms.
The development at SRI of a successful experimental pro-
cedure to elicit this capability has evolved to the point where
persons such as visiting government scientists and contract
monitors, with no previous exposure to such concepts, have
learned to perform well; and subjects who have trained over a
one-year period have performed excellently under a variety of
experimental conditions. Our accumulated data thus indicate
that both specially selected and unselected persons can be
assisted in developing remote perceptual abilities up to a
level of useful information transfer.
In experiments of this type, we have three principal findings.
First, we have established that it is possible to obtain signifi-
cant amounts of accurate descriptive information about remote
locations. Second, an increase in the distance from a few
meters up to 4000 km separating the subject from the scene
to be perceived does not in any apparent way degrade the
quality or accuracy of perception. Finally; the use of Faraday
cage electrical shielding does not prevent high-quality descrip-
tions from being obtained.
To build a coherent theory for the explanation of these
phenomena, it is necessary to have a clear understanding of
what constitutes the phenomena. In this paper, we first briefly
summarize previous efforts in this field in Section II. We then
present ApprovediFor Release20postes'd more
than fifty experiments with nine subjects carried out in our
own laboratory, which represent a sufficiently stable data base
to permit testing of various hypotheses concerning the func-
tioning of this channel. Finally, in Section V, we indicate
those areas of physics and information theory that appear to
be relevant to an understanding of certain aspects of the
phenomena.
First, however, we present an illustrative example generated
in an early pilot experiment. As will be clear from our later
: discussion, this is not a "best-ever" example, but rather a
typical sample of the level of proficiency that can be reached
and that we have come to expect in our research.
Three subjects participated in a long-distance experiment
focusing on a series of targets in Costa Rica. These subjects
said they had never been to Costa Rica. In this experiment,
one of the experimenters (Dr. Puthoff ) spent ten days traveling
through Costa Rica on a combination business/pleasure trip.
This information was all that was known to the subjects about
the traveler's itinerary. The experiment called for Dr. Puthoff
to keep a detailed record of his location and activities, includ-
ing photographs of each of seven target days at 1330 PDT.
A total of twelve daily descriptions were collected before the
traveler's return: six responses from one subject, five from
another, and one from a third.
The third subject who submitted the single response supplied
a drawing for a day in the middle of the series. (The subject's
response, together with the photographs taken at the site, are
shown in Fig. 1). Although Costa Rica is a mountainous
country, the subject unexpectedly perceived the traveler at a
CIA-RDP96-00789R003/100/2000ivi4g, he described an
airport on a sandy beach and an airstrip with the ocean at the

331
PUTHOFF AMbrovea®ForRelease 2006/08/08ATCIATRDP96-00789R00310092060314T
end (correct). An airport building also was drawn, and shown
to have a large rectangular overhang (correct). The traveler
had taken an unplanned one-day side trip to an offshore island
and at the time of the experiment had just disembarked from a
plane at a small island airport as described by the subject
4000 km away. The sole discrepancy was that the subject's
drawing showed a Quonset-hut type of building in place of the
rectangular structure.
The above description was chosen as an example to illustrate
a major point observed a number of times throughout the
program to be described. Contrary to what may be expected,
a subject's description does not necessarily portray what may
reasonably be expected to be correct (an educated or "safe"
guess), but often runs counter even to the subject's own
expectations.
We wish to stress again that a result such as the above is not
unusual. The remaining submissions in this experiment pro-
vided further examples of excellent correspondences between
target and response. (A target period of poolside relaxation
was identified; a drive through a tropical forest at the base of
a truncated volcano was described as a drive through a jungle
below a large bare table mountain; a hotel-room target descrip-
tion, including such details as rug color, was correct; and so
on.) So as to determine whether such matches were simply
fortuitous-that is, could reasonably be expected on the basis
of chance alone-Dr. Puthoff was asked after he had returned
to blind match the twelve descriptions to his seven target
locations. On the basis of this conservative evaluation proce-
dure, which vastly underestimates the statistical significance
of the individual descriptions, five correct matches were ob-
tained. This number of matches is significant at p = 0.02 by
exact binomial calculation.2
The observation of such unexpectedly high-quality descrip-
tions early in our program led to a large-scale study of the
phenomenon at SRI under secure double-blind conditions (i.e.,
target unknown to experimenters as well as subjects), with
independent random target selection and blind judging. The
results, presented in Sections III and IV, provide strong evi-
dence for the robustness of this phenomenon whereby a
human perceptual modality of extreme sensitivity can detect
complex remote stimuli.
II. BACKGROUND
Although we are approaching the study of these phenomena
as physicists, it is not yet possible to separate ourselves entirely
from the language of the nineteenth century when the labora-
tory study of the paranormal was begun. Consequently, we
continue to use terms such as "paranormal," "telepathy," and
the like. However, we intend only to indicate a process of
information transfer under conditions generally accepted as
secure against such transfer and with no prejudice or occult
assumptions as to the mechanisms involved. As in any other
scientific pursuit, the purpose is to collect the observables that
result from experiments and to try to determine the functional
relationships between these observables and the laws of physics
as they are currently understood.
2 The probability of a correct daily match by chance for any given
transcript is p = +. Therefore, the probability of at least five correct
matches by chance out of twelve tries can be calculated from
12
12!
p Approved
i=s
Organized research into so-called psychic functioning began
roughly in the time of J. J. Thomson, Sir Oliver Lodge, and
Sir William Crookes, all of whom took part in the founding of
the Society for Psychical Research (SPR) in 1882 in England.
Crookes, for example, carried out his principal investigations
with D. D. Home, a Scotsman who grew up in America and
returned to England in 1855 [3]. According to the notebooks
and published reports of Crookes, Home had demonstrated
the ability to cause objects to move without touching them.
We should note in passing that, Home, unlike most subjects,
worked only in the light and spoke out in the strongest pos-
sible terms against the darkened seance rooms popular at the
time [5].
Sir William Crookes was a pioneer in the study of electrical
discharge in gases and in the development of vacuum tubes,
some types of which still bear his name. Although everything
Crookes said about electron beams and plasmas was accepted,
nothing he said about the achievements of D. D. Home ever
achieved that status. Many of his colleagues, who had not
observed the experiments with Home, stated publicly that they
thought Crookes had been deceived, to which Crookes angrily
responded:
Will not my critics give me credit for some amount of common
sense? Do they not imagine that the obvious precautions, which
occur to them as soon as they sit down to pick holes in my
experiments, have occurred to me also in the course of my pro-
longed and patient investigation? The answer to this, as to all
other objections is, prove it to be an error, by showing where
the error lies, or if a trick, by showing how the trick is per-
formed. Try the experiment fully and fairly. If then fraud be
found, expose it; if it be a truth, proclaim it. This is the only
scientific procedure, and it is that I propose steadily to pursue
[3].
In the United States, scientific interest in the paranormal
was centered in the universities. In 1912, John Coover [6]
was established in the endowed Chair of Psychical Research at
Stanford University. In the 1920's, Harvard University set up
research programs with George Estabrooks and L. T. Troland
[7], [8]. It was in this framework that, in 1930, William
McDougall invited Dr. J. B. Rhine and Dr. Louisa Rhine to
join the Psychology Department at Duke University [9]. For
more than 30 years, significant work was carried out at Rhine's
Duke University Laboratory. To examine the existence of
paranormal perception, he used the now-famous ESP cards
containing a boldly printed picture of a star, cross, square,
circle, or wavy lines. Subjects were asked to name the order
of these cards in a freshly shuffled deck of twenty-five such
cards. To test for telepathy, an experimenter would look at
the cards one at a time, and a subject suitably separated from
the sender would attempt to determine which card was being
viewed.
Dr. J. B. Rhine together with Dr. J. G. Pratt carried out
thousands of experiments of this type under widely varying
conditions [10]. The statistical results from these experiments
indicated that some individuals did indeed possess a paranor-
mal perceptual ability in that it was possible to obtain an
arbitrarily high degree of improbability by continued testing
of a gifted subject.
The work of Rhine has been challenged on many grounds,
however, including accusations of improper handling of statis-
tics, error, and fraud. With regard to the statistics, the general
consensus of statisticians today is that if fault is to be found
(12-)
Release. 2000/08/08 : 61A-RDP96,00789R00310012000th4n statistical
grounds [11]. With regard to the accusations of fraud, the

332
CPYRGHT
Approved For Release 2000/08/08 : CIA RDP96 00789R003100120001 4
IEEE, MARCH 1976
most celebrated case of criticism of Rhine's work, that of
G. R. Price [12], ended 17 years after it began when the
accusation of fraud was retracted by its author in an article
entitled "Apology to Rhine and Soal," published in the same
journal in which it was first put forward [13]. It should also
be noted that parapsychological researchers themselves re-
cently exposed fraud in their own laboratory when they
encountered it [14].
At the end of the 1940's, Prof. S. G. Soal, an English mathe-
matician working with the SPR, had carried out hundreds of
card guessing experiments involving tens of thousands of calls
[15]. Many of these experiments were carried out over ex-
tended distances. One of the most notable experiments was
conducted with Mrs. Gloria Stewart between London and
Antwerp. This experiment gave results whose probability of
occurring by chance were less than 10-8. With the publication
of Modern Experiments in Telepathy by Soal and Bateman
(both of whom were statisticians), it appeared that card guess-
ing experiments produced significant results, on the average.3
The most severe criticism of all this work, a criticism diffi-
cult to defend against in principle, is that leveled by the well-
known British parapsychological critic C. E. M. Hansel [17],
who began his examination of the ESP hypothesis with the
stated assumption, "In view of the a priori arguments against
it we know in advance that telepathy, etc., cannot occur."
Therefore, based on the "a priori unlikelihood" of ESP,
Hansel's examination of the literature centered primarily on
the possibility of fraud, by subjects or investigators. He
reviewed in depth four experiments which he regarded as
providing the best evidence of ESP: the Pearce-Pratt distance
series [18]; the Pratt-Woodruff [19] series, both conducted
at Duke; and Soal's work with Mrs. Stewart and Basil Shackle-
ton [15], as well as a more recent series by Soal and Bowden
[20]. Hansel showed, in each case, how fraud could have been
committed (by the experimenters in the Pratt-Woodruff and
Soal-Bateman series, or by the subjects in the Pearce-Pratt
and Soal-Bowden experiments). He gave no direct evidence
that fraud was committed in these experiments, but said, "If
the result could have arisen through a trick, the experiment
must be considered unsatisfactory proof of ESP, whether or
not it is finally decided that such a trick was in fact used" [17,
p. 18]. As discussed by Honorton in a review of the field
[21], Hansel's conclusion after 241 pages of careful scrutiny
therefore was that these experiments were not "fraud-proof"
and therefore in principle could not serve as conclusive proof
of ESP.
Even among the supporters of ESP research and its results,
there remained the consistent problem that many successful
subjects eventually lost their ability and their scores gradually
drifted toward chance results. This decline effect in no way
erased their previous astronomical success; but it was a disap-
pointment since if paranormal perception is a natural ability,
one would like to see subjects improving with practice rather
than getting worse.
One of the first successful attempts to overcome the decline
effect was in Czechoslovakia in the work of Dr. Milan Ryzl, a
chemist with the Institute of Biology of the Czechoslovakian
Academy of Science and also an amateur hypnotist [22].
Through the use of hypnosis, together with feedback and
'Recently, some of the early Soal experiments have been criticized
[16]. However, his long-distance experiments cited here were judged
early ex Approved For Release 2000/08/08
reinforcement, he developed several outstanding subjects, one
of whom, Pavel Stepanek, has worked with experimenters
around the world for more than 10 years.
Ryzl's pioneering work came as an answer to the questions
raised by the 1956 CIBA Foundation conference on extra-
sensory perception.
The CIBA Chemical Company has annual
meetings on topics of biological and chemical interest, and
that same year they assembled several prominent parapsy-
chologists to have a state-of-the-art conference on ESP [23].
The conference concluded that little progress would be made
in parapsychology
research until a repeatable experiment
could be found; namely, an experiment that different experi-
menters could repeat at will and that would reliably yield a
statistically significant result.
Ryzl had by 1962 accomplished that goal. His primary con-
tribution was a decision to interact with the subject as a per-
son, to try to build up his confidence and ability. His protocol
depended on "working with" rather than "running" his sub-
jects. Ryzl's star subject, Pavel Stepanek, has produced highly
significant results with many contemporary researchers [24]-
[29]. In these experiments, he was able to tell with 60-percent
reliability whether a hidden card was green side or white
side up, yielding statistics of a million to one with only a
thousand trials.
As significant as such results are statistically, the information
channel is imperfect, containing noise along with the signal.
When considering how best to use such a channel, one is led
to the communication theory concept of the introduction of
redundancy as a means of coding a message to combat the
effects of a noisy channel [30]. A prototype experiment by
Ryzl using such techniques has proved to be successful. Ryzi
had an assistant select randomly five groups of three digits
each. These 15 digits were then encoded into binary form and
translated into a sequence of green and white cards in sealed
envelopes. By means of repeated calling and an elaborate
majority vote protocol, Ryzl was able after 19 350 calls by
Stepanek (averaging 9 s per call) to correctly identify all 15
numbers, a result significant at p = 10-15
The hit rate for
individual calls was 61.9 percent, 11 978 hits, and 7372 misses
[31].
Note Added in Proof: It has been brought to our attention
that a similar procedure was recently used to transmit without
error the word "peace" in International Morse Code (J. C.
Carpenter, "Toward the effective utilization of enhanced
weak-signal ESP effects," presented at the Annual Meeting of
the American Association for the Advancement of Science,
New York, NY, Jan. 27, 1975).
The characteristics of such a channel can be specified in
accordance with the precepts of communication theory. The
bit rate associated with the information channel is calculated
from [30]
R = H(x) - Hy(x)
(1)
where H(x) is the uncertainty of the source message containing
symbols with a priori probability pi:
H(x) =- 2
2
pi log2 Pi
(2)
i=1
and Hy (x) is the conditional entropy based on the a posteriori
probabilities that a received signal was actually transmitted:
H (x) =-
2
(3)
in a double-blind fashion of the type that escaped the criticism of che_RDP96-00789R003,1001200

333
CPYRGHT
For Stepanek's run, with pi = ₺, Pj (j) = 0.619, and an average
time of 9 s per choice, we have a source uncertainty H(x) = 1
bit and a calculated bit rate
R ~ 0.041 bit/symbol
or
R/T = 0.0046 bit/s.
(Since the 15-digit number (49.8 bits) actually was transmitted
at the rate of 2.9 X 10-4 bit/s, an increase in bit rate by a
factor of about 20 could be expected on the basis of a coding
scheme more optimum than that used in the experiments. See,
for example, Appendix A.)
Dr. Charles Tart at the University of California has written
extensively on the so-called decline effect. He considers that
having subjects attempt to guess cards, or perform any other
repetitious task for which they receive no feedback, follows
the classical technique for deconditioning any response. He
thus considers card guessing "a technique for extinguishing
psychic functioning in the laboratory" [32].
Tart's injunctions of the mid-sixties were being heeded at
Maimonides Hospital, Brooklyn, NY, by a team of researchers
that included Dr. Montague Ullman, who was director of
research for the hospital; Dr. Stanley Krippner; and, later,
Charles Honorton. These three worked together for several
years on experiments on the occurrence of telepathy in dreams.
In the course of a half-dozen experimental series, they found
in their week-long sessions a number of subjects who had
dreams that consistently were highly descriptive of pictorial
material that a remote sender was looking at throughout the
night. This work is described in detail in the experimenters'
book Dream Telepathy [33]. Honorton is continuing work
of this free-response type in which the subject has no precon-
ceived idea as to what the target may be.
In his more recent work with subjects in the waking state,
Honorton is providing homogeneous stimulation to the subject
who is to describe color slides viewed by another person in a
remote room. In this new work, the subject listens to white
noise via earphones and views an homogeneous visual field
imposed through the use of Ping-Pong ball halves to cover the
subject's eyes in conjunction with diffuse ambient illumina-
tion. In this so-called Ganzfeld setting, subjects are again able,
now in the waking state, to give correct and often highly
accurate descriptions of the material being viewed by the
sender [34].
In Honorton's work and elsewhere, it apparently has been
the step away from the repetitive forced-choice experiment
that has opened the way for a wide variety of ordinary people
to demonstrate significant functioning in the laboratory, with-
out being bored into a decline effect.
This survey would be incomplete if we did not indicate
certain aspects of the current state of research in the USSR.
It is clear from translated documents and other sources [35]
that many laboratories in the USSR are engaged in paranormal
research.
Since the 1930's, in the laboratory of L. Vasiliev (Leningrad
Institute for Brain Research), there has been an interest in the
use of telepathy as a method of influencing the behavior of a
person at a distance. In Vasiliev's book Experiments in Mental
Suggestion, he makes it very clear that the bulk of his labora-
tory's experiments were aimed at long-distance communica-
tion combined with a form of behavior modification; for
example, Putting Poved For Release 2000/08/08'. CIA
nosis [36] .
Similar behavior modification types of experiments have been
carried out in recent times by I. M. Kogan, Chairman of the
Bioinformation Section of the Moscow Board of the Popov
Society. He is a Soviet engineer who, until 1969, published
extensively on the theory of telepathic communication [37]-
[40]. He was concerned with three principal kinds of experi-
ments: mental suggestion without hypnosis over short dis-
tances, in which the percipient attempts to identify an object;
mental awakening over short distances, in which a subject is
awakened from a hypnotic sleep at the "beamed" suggestion
from the hypnotist; and long-range (intercity) telepathic com-
munication. Kogan's main interest has been to quantify the
channel capacity of the paranormal channel. He finds that the
bit rate decreases from 0.1 bit/s for laboratory experiments
to 0.005 bit/s for his 1000-km intercity experiments.
In the USSR, serious consideration is given to the hypothesis
that telepathy is mediated by extremely low-frequency (ELF)
electromagnetic propagation. (The pros and cons of this
hypothesis are discussed in Section V of this paper.) In
general, the entire field of paranormal research in the USSR
is part of a larger one concerned with the interaction between
electromagnetic fields and living organisms [41], [42]. At
the First International Congress on Parapsychology and
Psychotronics in Prague, Czechoslovakia, in 1973, for example,
Kholodov spoke at length about the susceptibility of living
systems to extremely low-level ac and de fields. He described
conditioning effects on the behavior of fish resulting from the
application of 10 to 100 uW of RF to their tank [43]. The
USSR take these data seriously in that the Soviet safety re-
quirements for steady-state microwave exposure set limits
at: 10 uW/cm2, whereas the United States has set a steady-state
limit of 10 mW/cm2 [44]. Kholodov spoke also about the
nonthermal effects of microwaves on animals' central nervous
systems. His experiments were very carefully carried out and
are characteristic of a new dimension in paranormal research.
The increasing importance of this area in Soviet research was
indicated recently when the Soviet Psychological Association
issued an unprecedented position paper calling on the Soviet
Academy of Sciences to step up efforts in this area [45].
They recommended that the newly formed Psychological
Institute within the Soviet Academy of Sciences and the
Psychological Institute of the Academy of Pedagogical Sciences
review the area and consider the creation of a new laboratory
within one of the institutes to study persons with unusual
abilities. They also recommended a comprehensive evaluation
of experiments and theory by the Academy of Sciences' Insti-
tute of Biophysics and Institute for the Problems of Informa-
tion Transmission.
The Soviet research, along with other behavioristically
oriented work, suggests that in addition to obtaining overt
responses such as verbalizations or key presses from a subject,
it should be possible to obtain objective evidence of informa-
tion transfer by direct measurement of physiological parame-
ters of a subject. Kamiya, Lindsley, Pribram, Silverman,
Walter, and others brought together to discuss physiological
methods to detect ESP functioning, have suggested that a
whole range of electroencephalogram (EEG) responses such as
evoked potentials (EP's), spontaneous EEG, and the contingent
negative variation (CNV) might be sensitive indicators of the
detection of remote stimuli not mediated by usual sensory
processes [46].
Fagy
carried out by
Yougias Deal at the Newark College
Engineering. In his

... . ...... ...:
334Approved For Release 2000/08/08 : CIA-RDP96-00789R003106120001-4HIE IEEE, MARCH 191
.
search for physiological correlates of information transfer, he
used the plethysmograph to measure changes in the blood
volume in a finger, a sensitive indicator of autonomic nervous
system functioning [47]. A plethysmographic measurement
was made on the finger of a subject during telepathy experi-
ments. A sender looked at randomly selected target cards
consisting of names known to the subject, together with names
unknown to him (selected at random from a telephone book).
The names of the known people were contributed by the sub-
ject and were to be of emotional significance to him. Dean
found significant changes in the chart recording of finger
blood volume when the remote sender was looking at those
names known to the subject as compared with those names
randomly chosen.
Three other experiments using the physiological approach
have now been published. The first work by Tart [48], a later
work by Lloyd [49], and most recently the work by the
authors [4] all follow a similar procedure. Basically, a subject
is closeted in an electrically shielded room while his EEG is
recorded. Meanwhile, in another laboratory, a second person
is stimulated from time to time, and the time of that stimulus
is marked on the magnetic-tape recording of the subject's EEG.
The subject does not know when the remote stimulus periods
are as compared with the nonstimulus periods.
With regard to choice of stimulus for our own experimenta-
tion, we noted that in previous work others had attempted,
without success, to detect evoked potential changes in a sub-
ject's EEG in response to a single stroboscopic flash stimulus
observed by another subject [50]. In a discussion of that
experiment, Kamiya suggested that because of the unknown
temporal characteristics of the information channel, it might
be more appropriate to use repetitive bursts of light to increase
the probability of detecting information transfer [51]. There-
fore, in our study we chose to use a stroboscopic flash train of
10-s duration as the remote stimulus.
In the design of the study, we assumed that the application
of the remote stimulus would result in responses similar to
those obtained under conditions of direct stimulation. For
example, when an individual is stimulated with a low-
frequency (< 30 Hz) flashing light, the EEG typically shows
a decrease in the amplitude of the resting rhythm and a
driving of the brain waves at the frequency of the flashes [52].
We hypothesized that if we stimulated one subject in this
manner (a putative sender), the EEG of another subject in a
III. SRI INVESTIGATIONS OF REMOTE VIEWING
remote room with no flash present (a receiver) might show
Experimentation in remote viewing began during studies
carried out to investigate the abilities of a New York artist,
Ingo Swann, when he expressed the opinion that the insights
gained during experiments at SRI had strengthened his ability
(verified in other research before he joined the SRI program)
to view remote locations [54]. To test Mr. Swann's asser-
tion, a pilot study was set up in which a series of targets from
around the globe were supplied by SRI personnel to the ex-
perimenters on a double-blind. basis. Mr. Swann's apparent
ability to describe correctly details of buildings, roads,
bridges, and the like indicated that it may be possible for
a subject by means of mental imagery to access and describe
randomly chosen
geographical sites located several miles
from the subject's position and demarcated by some appro-
priate means. Therefore, we set up a research program to
test the remote-viewing hypothesis under rigidly controlled
scientific conditions.
changes in alpha (9-11 Hz) activity and possibly an EEG
driving similar to that of the sender, or other coupling to the
sender's EEG [53]. The receiver was seated in a visually
opaque, acoustically and electrically shielded, double-walled
steel room about 7 m from the sender. The details of the
experiment, consisting of seven runs of thirty-six 10-s trials
each (twelve periods each for 0-Hz, 6-Hz, and 16-Hz stimuli,
randomly intermixed), are presented in [4]. This experiment
proved to be successful. The receiver's alpha activity (9-11 Hz)
showed a significant reduction in average power (-24 percent,
p < 0.04) and peak power (-28 percent, p < 0.03) during
16-Hz flash stimuli as compared with periods of no-flash
stimulus. [A similar response was observed for 6-Hz stimuli
(-12 percent in average power, -21 percent in peak power),
but the latter result did not reach statistical significance.]
Fig. 2 shows an overlay of three averaged EEG spectra from
one of the subject's 36 trial runs, displaying difference: i_RDP96-00789RUOBroom2000aceptrated on what we
alpha Approved henke SARkis conditions.
lease.2
extensive
control procedures were undertaken to determine if these
VOLTAGE - arbitrary units
0
~6
/16
1
5 Hz
10 Hz
15 Hz
Fig. 2. Occipital EEG frequency spectra, 0-20 Hz, of one subject (H.H
acting as receiver showing amplitude changes in the 9-11-Hz band as
function of strobe frequency. Three cases: 0-, 6-, and 16-Hz flash
(twelve trial averages).
results were produced by system artifacts, electromagneti
pickup (EMI), or subtle cueing; the results were negative [4]
As part of the experimental protocol, the subject was aske
to indicate a conscious assessment for each trial (via telegrap
key) as to the nature of the stimulus; analysis showed thes
guesses to be at chance. Thus arousal as evidenced by signif:
cant alpha blocking occurred only at the noncognitive level o
physiological response. Hence the experiment provided direc
physiological (EEG) evidence of perception of remote stimul
even in the absence of overt cognitive response.
Whereas in our experiments we used a remote light flash as .
stimulus, Tart [48] in his work used an electrical shock to
himself as sender, and Lloyd [49] simply told the sender to
think of a red triangle each time a red warning light wa
illuminated within his view. Lloyd observed a consisten
evoked potential in his subjects; whereas in our experiment
and in Tart's, a reduction in amplitude and a desynchroniza
tion of alpha was observed-an arousal response. (If a subjec
is resting in an alpha-dominant condition and he is ther
stimulated, for example in any direct manner, one will observe
a desynchronization and decrease in alpha power.) We con
sider that these combined results are evidence for the existence
of noncognitive awareness of remote happenings and that they
have a profound implication for paranormal research.
considered to be our principal responsibility-to resolve under
unambiguous conditions the basic issue of whether or not this

335
PUTHOFF AApproved ForRelease 2000/08/08MADIA-RDP96-00789R003R00120001-4
class of paranormal perception phenomenon exists. At all
times, we and others responsible for the overall program took
measures to prevent sensory leakage and subliminal cueing and
to prevent deception, whether intentional or unintentional.
To ensure evaluations independent of belief structures of both
experimenters and judges, all experiments were carried out
under a protocol, described below, in which target selection at
the beginning of experiments and blind judging of results at
the end of experiments were handled independently of the
researchers engaged in carrying out the experiments.
Six subjects, designated S1 through S6, were chosen for the
study. Three were considered as gifted or experienced subjects
(S1 through S3), and three were considered as learners (S4
through S6). The a priori dichotomy between gifted and
learners was based on the experienced group having been
successful in other studies conducted before this program
and the learners group being inexperienced with regard to
paranormal experimentation.
The study consisted of a series of double-blind tests with
local targets in the San Francisco Bay Area so that several in-
dependent judges could visit the sites to establish documenta-
tion. The protocol was to closet the subject with an experi-
menter at SRI and at an agreed-on time to obtain from the
subject a description of an undisclosed remote site being
visited by a target team. In each of the experiments, one of
the six program subjects served as remote-viewing subject,
and SRI experimenters served as a target demarcation team at
the remote location chosen in a double-blind protocol as
follows.
In each experiment, SRI management randomly chose a
target location from a list of targets within a 30-min driving
time from SRI; the target location selected was kept blind to
subject and experimenters. The target pool consisted of more
than 100 target locations chosen from a target-rich environ-
ment. (Before the experimental series began, the Director of
the Information Science and Engineering Division, not other-
wise associated with the experiment, established the set of lo-
cations as the target pool which remained known only to him.
The target locations were printed on cards sealed in envelopes
and kept in the SRI Division office safe. They were available
only with the personal assistance of the Division Director who
issued a single random-number selected target card that con-
stituted the traveling orders for that experiment.)
In detail: To begin the experiment, the subject was closeted
with an experimenter at SRI to wait 30 min before beginning
a narrative description of the remote location. A second ex-
perimenter then obtained from the Division Director a target
location from a set of traveling orders previously prepared and
randomized by the Director and kept under his control. The
target demarcation team, consisting of two to four SRI experi-
menters, then proceeded by automobile directly to the target
without any communication with the subject or experimenter
remaining behind. The experimenter remaining with the sub-
ject at SRI was kept ignorant of both the particular target and
the target pool so as to eliminate the possibility of cueing
(overt or subliminal) and to allow him freedom in questioning
the subject to clarify his descriptions. The demarcation team
remained at the target site for an agreed-on 15-min period
following the 30 min allotted for travel.4 During the observa-
4The first subject (S1) was allowed 30 min for his descriptions, but
it was found Approved For Release.2000/08/08
min. The viewing time was therefore reduced to 15 min for subjects
S2 through S6.
tion period, the remote-viewing subject was asked to describe
his impressions of the target site into a tape recorder and to
make any drawings he thought appropriate. An informal com-
parison was then made when the demarcation team returned,
and the subject was taken to the site to provide feedback.
A. Subject SI: Experienced
To begin the series, Pat Price, a former California police com-
missioner and city councilman, participated as a subject in
nine experiments.
In general, Price's ability to describe
correctly buildings, docks, roads, gardens, and the like, includ-
ing structural materials, color, ambience, and activity-often
in great detail-indicated the functioning of a remote per-
ceptual ability. A Hoover Tower target, for example, was
recognized and named by name. Nonetheless, in general, the
descriptions contained inaccuracies as well as correct state-
ments. A typical example is indicated by the subject's drawing
shown in Fig. 3 in which he correctly described a park-like
area containing two pools of water: one rectangular, 60 by
89 ft (actual dimensions 75 by 100 ft); the other circular,
diameter 120 ft (actual diameter 110 ft). He incorrectly indi-
cated the function, however, as water filtration rather than
recreational swimming. (We often observe essentially correct
descriptions of basic elements and patterns coupled with in-
complete or erroneous analysis of function.) As can be seen
from his drawing, he also included some elements, such as
the tanks shown in the upper right, that are not present at the
target site. We also note an apparent left-right reversal, often
observed in paranormal perception experiments.
To obtain a numerical evaluation of the accuracy of the
remote-viewing experiment, the experimental results were
subjected to independent judging on a blind basis by an SRI
research analyst not otherwise associated with the research.
The subject's response packets, which contained the nine
typed unedited transcripts of the tape-recorded narratives
along with any associated drawings, were unlabeled and pre-
sented in random order. While standing at each target loca-
tion, visited in turn, the judge was required to blind rank order
the nine packets on a scale 1 to 9 (best to worst match). The
statistic of interest is the sum of ranks assigned to the target-
associated transcripts, lower values indicating better matches.
For nine targets, the sum of ranks could range from nine to
eighty-one. The probability that a given sum of ranks s or
less will occur by chance is given by [55]
Pr (s or less) =
1
where s is obtained sum of ranks, N is number of assignable
ranks, n is number of occasions on which rankings were made,
and / takes on values from zero to the least positive integer k
in (i - n)/n. (Table I is a table to enable easy application of
the above formula to those cases in which N = n.) The sum in
this case, which included seven direct hits out of the nine, was
16 (see Table II), a result significant at p = 2.9 × 10-5 by
exact calculation.
In Experiments 3, 4, and 6 through 9, the subject was se-
cured in a double-walled copper-screen Faraday cage. The
Faraday cage provides 120-dB attenuation for plane-wave
radio-frequency radiation over a range of 15 kHz to 1 GHz.
decreases to 3 dB at 60 Hz. The results of rank order judging
(Table II) indicate that the use of Faraday cage electrical

336
CPYRGHTApproved For Release 2000/08/08 : CIA-RDP96-00789R00390012000F-ZE IEEE, MARCH 197
2 person
THOK.
TROY'S
.
i
AVE.
TENNIS COURTS
ELECT.
DIST.
76
110
100
RAINBOW EFFECTS
PLAY
GROUND6
POOL HOUSE
CONCRETE BLOCK
SERVICE
YARD
FIRE
STATION
PICNIC AREA/
RD.
SUNIL
n
(a)
(b)
Fig. 3. Swimming pool complex as remote-viewing target. (a) City map of target location. (b) Drawing by Price (S1).
TABLE I
CRITICAL VALUES OF SUMS OF RANKS FOR PREFERENTIAL MATCHING
Probability (one-tailed) that the Indicated Sum of Ranks or Less Would Occur by Chance
Number of
Assignable
Ranks (N)
0.20
0.10
0.05
0.04
0.025
0.01
0.005
0.002
0.001
0.0005
10~4
10-6
10-6
10-7
4
7
6
5
5
5
4
4
5
11
10
9
8
8
7
6
6
5
5
6
16
15
13
13
12
11
10
9
8
7
6
7
22
20
18
18
17
15
14
12
12
9
8
8
29
27
24
24
22
20
19
17
16
15
13
11
9
8
9
37
34
31
30
29
26
24
22
21
20
17
14
12
10
10
46
42
39
38
36
33
31
29
27
25
22
19
16
13
11
56
51
48
47
45
41
38
36
34
32
28
24
20
17
12
67
61
58
56
54
49
47
43
41
39
35
30
25
22
Note: This table applies only to those special cases in which the number of occasions on which objects
are being ranked (n) is equal to the number of assignable ranks (N). Each entry represents the largest
number that is significant at the indicated p-level. Source: R. L. Morris [55].
shielding does not prevent high-quality descriptions from being
obtained.
As a backup judging procedure, a panel of five additional
SRI scientists not otherwise associated with the research were
asked simply to blind match the unedited typed transcripts
(with associated drawings) generated by the remote viewer
against the nine target locations which they independently
visited in turn. The transcripts were unlabeled and presented
in random order. A correct match consisted of a transcript
of a given date being matched to the target of that date. In-
stead of the expected number of 1 match each per judge, the
number of correct matches obtained by the five judges was 7,
6, 5, 3, and 3, respectively.
from the judges, 24 such
matches were obtained.
B. Subject S4: Learner
This experiment was designed to be a replication of our pre-
vious experiment with Price, the first replication attempted.
The subject for this experiment was Mrs. Hella Hammid, a
gifted professional photographer. She was selected for this
series on the basis of her successful performance as a per-
cipient in the EEG experiment described earlier. Outside of
that interaction, she had no previous experience with apparent
paranormal functioning.
At the time we began working with Mrs. Hammid, she had
no strong feelings about the likelihood of her ability to suc-
ceed in this task. This was in contrast to both Ingo Swann
total Approved For Release 2000/08108 CA-RUP96-00789R00310012000+ 4om a lengthy and
Thus, rather than the
apparently successful series of experiments with Dr. Gertrude
Schmeidler at City College of New York [56] and Pat Price
SOMLE YARD,
SHEEP of WHAT SPRAY

337
PUTHOF
AApproved ForRelease 2000/08/08 ACIA-RDP96-00789R0031061260d-4
PEDESTRIAN OVERPASS TARGET
1
Light
C
CCCC
Fig. 4. Subject Hammid (S4) drawing, described as "some kind of diagonal trough up in the air."
TABLE II
DISTRIBUTION OF RANKINGS ASSIGNED TO TRANSCRIPTS
ASSOCIATED WITH EACH TARGET LOCATION FOR EXPERIENCED
SUBJECT PRICE (SI)
Target Location
Distance
(km)
Rank of
Associated
Transcript
Hoover Tower, Stanford
3.4
1
Baylands Nature Preserve, Palo Alto
6.4
1
Radio telescope, Portola Valley
6.4
1
Marina, Redwood City
6.8
1
Bridge coll plaza, Fremont
14.5
6
Drive-in theater, Palo Alto
5.1
1
Arts and Crafts Plaza, Menlo Park
1.9
1
Catholic Church, Portola Valley
8.5
3
Swimming pool complex, Palo Alto
3.4
1
Total sum of ranks
16
.
(p=2.9x10-5)
scientific rigor, one of our primary tasks as researchers is to
provide an environment in which the subject feels safe to
explore the possibility of paranormal perception. With a new
subject, we also try to stress the nonuniqueness of the ability
because from our experience paranormal functioning appears
to be a latent ability that all subjects can articulate to some
degree.
Because of Mrs. Hammid's artistic background, she was ca-
pable of drawing and describing visual images that she could
not identify in any cognitive or analytic sense. When the target
demarcation team went to a target location which was a
pedestrian overpass, the subject said that she saw "a kind of
trough up in the air," which she indicated in the upper part
of her drawing in Fig. 4. She went on to explain, "If you
stand where they are standing you will see something like
this," indicating the nested squares at the bottom of Fig. 4.
As it turned out, a judge standing where she indicated would
have a view closely resembling what she had drawn, as can be
seen from the accompanying photographs of the target loca-
tion. It needs to be emphasized, however, that judges did not
have access to our photographs of the site, used here for
illustrative purposes only, but rather they proceeded to each
of the target locations by list.
who felt that he used his remote-viewing ability in his every-
day life.
In comparison with the latter two, many people are more
influenced by their environment and are reluctant under
In another experiment, the subject described seeing “an
open barnlike structure with a pitched roof." She also saw
a "kind of slatted side to the structure making light and dark
bars on the wall." Her drawing and a photograph of the
associated bicycle shed target are shown in Fig. 5. (Subjects
are encouraged to make drawings of anything they visualize
public scrutiny to attempt activities that are generally thought
to be impossible. Society often provides inhibition and nega-
tive feedback to the individual who might otherwise have
explored his own nonregular perceptual ability. We all share
an historical tradition of "the stoning of prophets and the
burning of witchesroved For Release 2000/08/08 : CIA-RDP96 00789R008100120004e4drawings they
tion of those who claim to perceive things that the majority do
make are in general more accurate than their verbal description.)
As in the original series with Price, the results of the nine-
not admit to seeing. Therefore, in addition to maintaining

338
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
PROCEEDINGS OF THE IEEE, MARCH 197
BICYCLE SHED TARGET
DETAIL OF BICYCLE SHED
1
1
1
-
Stone
will
T
i
Fig. 5. Subject Hammid (S4) response to bicycle shed target described as an open "barn-like building" with "slats
on the sides" and a "pitched roof."
TABLE III
DISTRIBUTION OF RANKINGS ASSIGNED TO TRANSCRIPTS ASSOCIATED
WITH EACH TARGET LOCATION FOR LEARNER SUBJECT HAMMID (S4)
Target Location
Distance
( km)
Rank of
Associated
Transcript
Methodist Church, Palo Alto
1.9
1
Ness Auditorium, Menlo Park
0.2
1
Merry-go-round, Palo Alto
3.4
1
Parking garage, Mountain View
8.1
2
SRI International Courtyard, Menlo Park
0.2
1
Bicycle shed, Menlo Park
0.1
2
Railroad trestle bridge, Palo Alto
1.3
2
Pumpkin patch, Menlo Park
1.3
1
Pedestrian overpass, Palo Alto
5.0
2
Total sum of ranks
13
(p=1.8x10-6)
Again, as a backup judging procedure, a panel of five addi-
tional judges not otherwise associated with the research
were asked simply to blind match the unedited typed tran-
scripts and associated drawings generated by the remote viewer,
against the nine target locations which they independently
visited in turn. A correct match consisted of a transcript of
a given date being matched to the target of that date.' In-
stead of the expected number of 1 match each per judge,
the number of correct matches obtained by the five judges
was 5, 3, 3, 2, and 2, respectively. Thus, rather than the ex-
pected total number of 5 correct matches from the judges,
15 such matches were obtained.
C. Subjects S2 and S3: Experienced
Having completed a series of 18 remote-viewing experiments,
9 each with experienced subject SI (Price) and learner S4
(Hammid), additional replication experiments, four with each
subject, were carried out with experienced subjects S2 (Elgin)
and S3 (Swann) and learners S5 and S6. To place the judging
on a basis comparable to that used with S1 and S4, the four
transcripts each of experienced subjects S2 and S3 were com-
bined into a group of eight for rank order judging to be com-
pared with the similarly combined results of the learners
S5 and S6.
experiment series were submitted for independent judging on
a blind basis by an SRI research analyst not otherwise associ-
ated with the research. While at each target location, visited
n turn, the judge was required to blind rank order the nine
unedited typed manuscripts of the tape-recorded narratives,
Along with any associated drawings generated by the remote
fiewer, on a scale 1 to 9 (best to worst match). The sum of
The series with S2 (Elgin, an SRI research analyst) provided
a further example of the dichotomy between verbal and draw-
ing responses. (As with medical literature, case histories often
tanks assigned to the target-associated transcripts in this case
vas 13, a result significant at p = 1.8 X 10-6 by exact calcula-
tion (seA Bishoved ForRelease2000/08/08c CIA-RDR96-00789R0031001.20001-4results.) The ex-
and four second ranks (Table III).
periment described here was the third conducted with this
. . ..

PUTHOFF AND TXI
Approved For Release 2000/08/08 - CIA-RDP96-00789R003100120001-4
339
TARGET-TENNIS COURTS
Windows
8
1
Q1 28
Fig. 6. Subject Elgin (S2) drawings in response to tennis court target.
subject. It was a demonstration experiment for a government
visitor who had heard of our work and wanted to evaluate our
experimental protocol.
In the laboratory, the subject, holding a bearing compass at
arm's length, began the experiment by indicating the direction
of the target demarcation team correctly to within 5°. (In all
four experiments with this subject, he has always been within
10° of the correct direction in this angular assessment.) The
subject then generated a 15-min tape-recorded description and
the drawings shown in Fig. 6.
In discussing the drawings, Elgin indicated that he was
uncertain as to the action, but had the impression that the
demarcation team was located at a museum (known to him)
in a particular park. In fact, the target was a tennis court lo-
cated in that park about 90 m from the indicated museum.
Once again, we note the characteristic (discussed earlier) of a
resemblance between the target site and certain gestalt ele-
ments of the subject's response, especially in regard to the
drawings, coupled with incomplete or erroneous analysis of
the significances. Nonetheless, when rank ordering transcripts
1 through 8 at the site, the judge ranked this transcript as 2.
This example illustrates a continuing observation that most of
the correct information related to us by subjects is of a non-
analytic nature pertaining to shape, form, color, and material
rather than to function or name.
A second example from this group, generated by S3 (Swann),
indicates the Jexeb of proficiencyella ser 2000/08/08.
practice. In the two years since we first started working with
Swann, he has been studying the problem of separating the ex-
thema signal from the internal anion In allt most recent
experiments, he dictates two lists for us to record. One list
contains objects that he "sees," but does not think are located
at the remote scene. A second list contains objects that he
thinks are at the scene. In our evaluation, he has made much
progress in this most essential ability to separate memory
and imagination from paranormal inputs. This is the key to
bringing the remote-viewing channel to fruition with regard to
its potential usefulness.
The quality of transcript that can be generated by this pro-
cess is evident from the results of our most recent experiment
with Swann. The target location chosen by the usual double-
blind protocol was the Palo Alto City Hall. Swann described a
tall building with vertical columns and "set in" windows. His
sketch, together with the photograph of the site, is shown in
Fig. 7. He said there was a fountain, "but I don't hear it."
At the time the target team was at the City Hall during the
experiment, the fountain was not running. He also made an
effort to draw a replica of the designs in the pavement in front
of the building, and correctly indicated the number of trees
(four) in the sketch.
For the entire series of eight, four each from S2 and S3, the
numerical evaluation 'based on blind rank ordering of tran-
scripts at each site was significant at p = 3.8 X 10-4 and in-
cluded three direct hits and three second ranks for the target-
associated transcripts (see Table IV).
D. Subjects S5 and S6: Learners
CIATROP96-09789P003100120001-4h were carried
out with learner subjects S5 and S6, a man and woman on the
SRI professional staff The results in this case, taken as a

340
Approved For Release 2000/08/08 : CIA-RDP96-00789R0091001200014 IEEE, MARCH 1976
TABLE IV
DISTRIBUTION OF RANKINGS ASSIGNED TO TRANSCRIPTS ASSOCIATED
WITH EACH TARGET LOCATION FOR EXPERIENCED SUBJECTS ELGIN (S2)
AND SWANN (S3)
Subject
Target Location
Distance
(km)
Rank of
Associated
Transcript
S2
BART Station (Transit System), Fremont
16. 1
1
S2
Shielded room, SRI, Menlo Park
0.1
2
S2
Tennis court, Palo Alto
3.4
2
S2
Golf course bridge, Stanford
3.4
2
53
City Hall, Palo Alto
2.0
1
$3
Miniature golf course, Menlo Park
3.0
1
53
Kiosk in park,Menlo Park
0.3
3
S3
Baylands Nature Preserve, Palo Alto
6.4
3
Total sum of: ranks
15
(p=3.8x10-4)
TABLE V
DISTRIBUTION OF RANKINGS ASSIGNED TO TRANSCRIPTS ASSOCIATED
WITH EACH TARGET LOCATION FOR LEARNER SUBJECTS S5 AND S6
Subject
Target Location
Distance
(km)
Rank of
Associated
Transcript
S5
Pedestrian overpass, Palo Alto
5.0
3
S5
Railroad treatle bridge, Palo Alto
1.3
6
S5
Windmill, Portola Valley
8.5
2
55, S6
White Plaza, Stanford (2)
3.8
1
S6
Airport, Palo Alto
5.5
2
S6
Kiosk in Park, Menlo Park
0.3
5
$6
Boathouse, Stanford
4.0
1
Total sum of ranks
20
(p=0.08, NS)
vious experience in remote viewing, began to describe a large
square with a fountain. Four minutes into the experiment,
she recognized the location and correctly identified it by name
(see Fig. 8). (It should be noted that in the area from which
the target locations were drawn there are other fountains
as well, some of which were in the target pool.) As an ex-
ample of the style of the narratives generated during remote
viewing with inexperienced subjects and of the part played by
the experimenter remaining with the subject in such a case,
we have included the entire unedited text of this experiment
as Appendix B.
E. Normal and Paranormal: Use of Unselected Subjects in
Remote Viewing
After more than a year of following the experimental pro-
tocol described above and observing that even inexperienced
subjects generated results better than expected, we initiated a
series of experiments to explore further whether individuals
other than putative "'psychics" can demonstrate the remote-
viewing ability. To test this idea, we have a continuing pro-
gram to carry out additional experiments of the outdoor type
with new subjects whom we have no a priori reason to believe
have paranormal perceptual ability. To date we have collected
data from five experiments with two individuals in this cate-
gory: a man and a woman who were visiting government
scientists interested in observing our experimental protocols.
The motivation for these particular experiments was twofold.
cedure, And proceeded 8 thpenease 2000/08/08 201A-RDP96-00789R003100120001 -4dicate the level of
matician In the computer science laboratory who had no pre-
proficiency that can be expected from unselected volunteers.
. .. ***
Savona
13 No 29.
10:40 am
fach of gren .
foliage - winel
trees?
a coundor & com set .
awake helt the trees
lawns .
anapon fuld.
an enclosed aber ofcourse cost.
a quad .
I long holding
afounten.
but I don't hand .
well thus
lan
wall
cros walker.
1
OR
basket ball anost . fuld .
Sester
CROSS
walks
?
Fig. 7. Subject Swann (S3) response to City Hall target.
group, did not differ significantly from chance. For the series
of eight (judged as a group of seven since one target came up
twice, once for each subject), the numerical evaluation based
on blind rank ordering of transcripts at each site was non-
significant at p = 0.08, even though there were two direct hits
and two second ranks out of the seven (see Table V).
One of the direct hits, which occurred with subject S6 in her
first experiment, provides an example of the "first-time effect"
that has been rigorously explored and is well-known to experi-
menters in the field [57]. The outbound experimenter
obtained, by random protocol from the pool, a target blind
to the experimenter with the subject, as is our standard pro-

Approved For Release 2000/08/08 : CIA-RDP96-00789R003400120001-4
PUTHOFF AND TARG: PERCEPTUAL CHANNEL FOR INFORMATION TRANSFER
341
....
-
1
-
somewhat curry
5-benches
Some landscaping
grass
Small - medium trees
afew flower
Fig. 8. Subject (S6) drawing of White Plaza, Stanford University. Sub-
ject drew what she called "curvy benches" and then announced cor-
rectly that the place was "White Plaza at Stanford."
Second, when an individual observes a successful demonstra-
tion experiment involving another person as subject, it inevi-
tably occurs to him that perhaps chicanery is involved. We
have found the most effective way to settle this issue for the
observer is to have the individual himself act as a subject so as
to obtain personal experience against which our reported
results can be evaluated.
The first visitor (V1) was invited to participate as a subject
in a three-experiment series. All three experiments contained
elements descriptive of the associated target locations; the
quality of response increased with practice. The third re-
sponse is sAoBroved For Release 2000/081
elements. in
the drawing appeared to be a closer match than the subject's
analytic interpretation of the target object as a cupola.
white
some sort of structure of
thing on top of apandage
rod in a general sense lite
Grounded done- type top
lightening rod
TOP VIEW
medium denturo thing
cod black x hoursy
RESPONSES OF VISITING
CIA-RDP96-00789ROQ3108020301-4
Fig. 9. Subject (VI) drawing of merry-go-round target.

342
CPYRCApproved For Release 2000/08/08 : CIA-RDP96-00789R0031001200
PROCEEDINGS 86 1-4
THE IEEE, MARCH 1976
TECHNOLOGY SERIES
TYPEWRITER TARGET
Sans to move into 2 posts
on siting on top of the oder -
a machine in 2 posts .
white on the side .
see the flor now - linge
-
1623
suits of slots
levers
mail
- Ly colored
dak ?
te
0
-0
0
0
10
the lights med be verde
a from crescut.
SUBJECT SWANN (S3) RESPONSE
8
SUBJECT HAMMID (S4) RESPONSE
Fig. 10. Drawings of a typewriter target by two subjects.
The second visitor [V2] participated as a subject in two ex-
periments. In his first experiment, he generated one of the
higher signal-to-noise results we have observed. He began
his narrative, "There is a red A-frame building and next to it
is a large yellow thing [a tree-Editor]. Now further left
there is another A-shape. It looks like a swing-set, but it is
pushed down in a gully so I can't see the swings." [All cor-
rect.] He then went on to describe a lock on the front door
that he said "looks like it's made of laminated steel, so it
must be a Master lock." [Also correct.]
For the series of five-three from the first subject and two
from the second-the numerical evaluation based on blind
rank ordering of the transcripts at each site was significant at
p = 0.017 and included three direct hits and one second rank
TABLE VI
DISTRIBUTION OF RANKINGS ASSIGNED TO TRANSCRIPTS ASSOCIATED WITH
EACH TARGET LOCATION FOR VISITOR SUBJECTS VI AND V2
Subject
Target Locacion
Distance
(km)
Rank of
Associated
Transcript
v1
Bridge over stream, Menlo Park
0.3
L
v1
Baylands Nature Preserve, Palo Alto
6.4
2
v1
Merry-go-round, Palo Alto
3.4
1
V2
Windmill, Portola Valley
8.5
1
V2
Apartment swimming pool, Mountain View
9.1
3
Total sum of ranks
8
(p=0,017)
for the Approved ForReteasy-2000/08/08 : CIA-RDP96-00789R003100120001-4

PUTHOFF AApproved For Release-2000/08/08 :- GIA-RDP96-00789R003100120001-4
343
TARGET LOCATION: XEROX MACHINE
(TECHNOLOGY SERIES)
TO ADD INTEREST TO TARGET
LOCATION EXPERIMENTER WITH
HIS HEAD BEING XEROXED
rolling along
table
it they '
Fund
00000
Fig. 11. Drawings by three subjects (S2, S3, and V3) for Xerox machine target. When asked to describe the square at upper left of response on
the right, subject (V3) said, "There was this predominant light source which might have been a window, and a working surface which might have
been the sill, or a working surface or desk." Earlier the subject had said, "I have the feeling that there is something silhouetted against the
window."
Observations with unselected subjects such as those de-
scribed above indicate that remote viewing may be a latent and
widely distributed perceptual ability.
F. Technology Series: Short-Range Remote Viewing
Because remote viewing is a perceptual ability, we consid-
ered it important to obtain data on its resolution capabilities.
To accomplish this, we turned to the use of indoor techno-
logical targets.
Twelve experiments were carried out with five different sub-
jects, two of whom were visiting government scientists. They
were told that one of the experimenters would be sent by
random protocol to a laboratory within the SRI complex and
that he would interact with the equipment or apparatus at
that location. It was further explained that the experimenter
remaining with the subject was, as usual, kept ignorant of the
contents of the target pool to prevent cueing during question-
ing. (Unknown to subjects, targets in the pool were used with
replacement; one of the goals of this particular experiment was
to obtain multiple responses to a given target to investigate
whether correlation of a number of subject responses would
provide enhancement of the signal-to-noise ratio.) The sub-
ject was asked to describe the target both verbally (tape
recorded) and by means of drawings during a time-synchronized
15-min interval in which the outbound experimenter inter-
acted in an appropriate manner with the equipment in the
target area.
Comparisons of the targets and subject drawings for three of
the multiple-response cases (the typewriter, Xerox machine,
and video terminal) are shown in Figs. 10, 11, and 12. As is
apparent from these illustrations alone, the experiments
provide circumstantial evidence for an information channel
of useful bit rate. This includes experiments in which visit-
ing government scientists participated as subjects (Xerox
machine and video terminal) to observe the protocol. In
general, it appears that use of multiple-subject responses to a
single target provides better signal-to-noise ratio than target
identification by a single individual. This conclusion is borne
out by the judging described below.
Given that in general the drawings constitute the most
accurate portion of a subject's description, in the first judging
procedure a judge was asked simply to blind match only the
drawings (i.e., without tape transcripts) to the targets. Multiple-
subject responses to a given target were stapled together, and
thus seven subject-drawing response packets were to be
matched to the seven different targets for which drawings were
made. The judge did not have access to our photographs of
the target locations, used for illustration purposes only, but
rather proceeded to each of the target locations by list. While
standing at each target location, the judge was required to rank
order the seven subject-drawing response packets (presented in
random order) on a scale 1 to 7 (best to worst match). For
seven targets, the sum of ranks could range from 7 to 49. The
sum in this case, which included 1 direct hit and 4 second
ranks out of the 7 (see Table VII) was 18, a result significant
at p = 0.036.
In the twelve experiments, seven targets were used: a drill
press, Xerox machine, video terminal, chart recorder, four-
state random number generator, machine shop, and type-
In the second more detailed effort at evaluation, a visiting
scientist selected at random one of the 12 data packages (a
dependent analysis to an engineer with a request for an esti-
writer. Three of these were used twice (drill press, video
terminal, andApproved.For Release 2000/08/08p: CIA-RDP96.00789R003100120001-4ted it for in-
three times in our random selection procedure.

344
Approved For Release 2000/08/08 : CIA-RDP96-00789R008400+2000f-ZIE IEEE, MARCH 197
TARGET: VIDEO MONITOR FOR TEXT EDITING (TECHNOLOGY SERIES)
Side
all mat Black
--
Ton
3-
window
TI Slant 700 :
(a)
(b)
Fig. 12. Drawing by two subjects of a video monitor target. (a) Subject (S4) drawing of "box with light coming out of it . .. painted flat black
and in the middle of the room." (b) Second subject (V2) saw a computer terminal with relay racks in the background.
TABLE VII
DISTRIBUTION OF RANKINGS ASSIGNED TO SUBJECT
DRAWINGS ASSOCIATED WITH EACH TARGET LOCATION
Subject
Target
Rank of
Associated
Drawings
$3, 54
Drill press
2
$2, S3, V3
Xerox machine
2
S4, V2
Video terminal
1
S3
Chart recorder
2
S4
Random number generator
6
S4
Machine shop
3
S3, S4
Typewriter
2
Total sum of ranks
18
(p=0.036)
TABLE VIII
SUMMARY: REMOTE VIEWING
Subject
Number of
Experiments
p-Value, Rank
Order Judging
With natural targets
S1 (experienced)
9
2.9 x 10-5
S2 and S3 (experienced)
8
3.8 x 10-4
S4 (learner)
9
1.8 x 10-6
S5 and S6 (learners)
8
0.08 (NS)
VI and V2 (learners/visitors)
5
0.017
With technology cargets
S2, S3, S4, V2, V3
12
0.036
tabulation of the statistical evaluations of these fifty-one ex-
periments with nine subjects is presented in Table VIII. The
overall result, evaluated conservatively on the basis of a
judging procedure that ignores transcript quality beyond that
necessary to rank order the data packets (vastly underestimat-
ing the statistical significance of individual descriptions),
clearly indicates the presence of an information channel of
useful bit rate. Furthermore, it appears that the principal
difference between experienced subjects and inexperienced
volunteers is not that the latter never exhibit the faculty, but
rather that their results are simply less reliable, more sporadic.
Nevertheless, as described earlier, individual transcripts from
the inexperienced group of subjects number among some of
the best obtained.
Such observations indicate a hypothesis
mate as to what was being described. The analyst, blind as to
the target and given only the subject's taped narrative and
drawing (Fig. 13), was able, from the subject's description
alone, to correctly classify the target as a "man-sized vertical
boring machine."
G. Summary of Remote Viewing Results
1) Discussion: The descriptions supplied by the subjects
in the experiments involving remote viewing of natural targets
or laboratory apparatus, although containing inaccuracies,
were sufficiently accurate to permit the judges to differentiate
amonApproved For Release 2000/08/08-CIA-RDP95.00789R00310612060and widely distributed
that remote_viewir
.
5
O
0

PUTHOFF AMApproveckFreruRelease:2000/08/084TOMATRDP96-00789RB03100120001-4
345
Light canvas
RetORY
Wh.T
Black
Fig. 13. Subject (S4) drawing of drill press showing belt drive, stool,
and a "vertical graph that goes up and down."
Thus the primary achievement of the SRI program was the
elicitation of high-quality remote viewing from individuals
who agreed to act as subjects. Criticism of this claim could
in principle be put forward on the basis of three potential
flaws. 1) The study could involve naivete in protocol that
permits various forms of cueing, intentional or unintentional.
2) The experiments discussed could be selected out of a larger
pool of experiments of which many are of poorer quality.
3) Data for the reported experiments could be edited to show
only the matching elements, the nonmatching elements being
discarded.
All three criticisms, however, are invalid. First, with regard
to cueing, the use of double-blind protocols ensures that none
of the persons in contact with the subject can be aware of the
target. Second, selection of experiments for reporting did not
take place; every experiment was entered as performed on a
master log and is included in the statistical evaluations. Third,
data associated with a given experiment remain unedited; all
experiments are tape recorded and all data sodo 08/08":
For Release
edited in the data package to be judged and evaluated.
In the process of judging-attempting to match transcripts
against targets on the basis of the information in the
transcripts-some patterns and regularities in the transcript
descriptions became evident, particularly regarding individual
styles in remote viewing and in the perceptual form of the
descriptions given by the subjects. These patterns and the
judging procedure are discussed below.
a) Styles of response: The fifty-one transcripts were
taken from nine different subjects.
Comparing the tran-
scripts of one subject with those of another revealed that each
pattern tended to focus on certain aspects of the remote
target complex and to exclude others, so that each had an
individual pattern of response, like a signature.
Subject S3, for example, frequently responded with topo-
graphical descriptions, maps, and architectural features of the
target locations. Subject S2 often focused on the behavior of
the remote experimenter or the sequence of actions he carried
out at the target. The transcripts of subject S4, more than
those of other subjects, had descriptions of the feel of the lo-
cation, and experiential or sensory gestalts-for example,
light/dark elements in the scene and indoor/outdoor and
enclosed/open distinctions. Prominent features of S1's tran-
scripts were detailed descriptions of what the target persons
were concretely experiencing, seeing, or doing-for example,
standing on asphalty blacktop overlooking water; looking at
a purple iris.
The range of any individual subject's responses was wide.
Anyone might draw a map or describe the mood of the remote
experimenter, but the consistency of each subject's overall
approach suggests that just as individual descriptions of a
directly viewed scene would differ, so these differences also
occur in remote-viewing processes.
b) Nature of the description: The concrete descriptions
that appear most commonly in transcripts are at the level of
subunits of the overall scene. For example, when the target
was a Xerox copy machine, the responses included (S2) a
rolling object (the moving light) or dials and a cover that is
lifted (S3), but the machine as a whole was not identified by
name or function.
In a few transcripts, the subjects correctly identified and
named the target. In the case of a computer terminal, the
subject (V2) apparently perceived the terminal and the relay
racks behind it. In the case of targets which were Hoover
Tower and White Plaza, the subjects (S1 and S6, respectively)
seemed to identify the locations through analysis of their
initial images of the elements of the target.
There were also occasional incorrect identifications. Gestalts
were incorrectly named; for example, swimming pools in a
park were identified as water storage tanks at a water filtration
plant (S1).
The most common perceptual level was thus an intermediate
one-the individual elements and items that make up the tar -.
get. This is suggestive of a scanning process that takes sample
perceptions from within the overall environment.
When the subjects tried to make sense out of these fragmen-
tary impressions, they often resorted to metaphors or con-
structed an image with a kind of perceptual inference. From
a feeling of the target as an "august" and "solemn" building,
a subject (S4) said it might be a library; it was a church. A
pedestrian overpass above a freeway was described as a conduit
(S4). A rapid transit station, elevated above the countryside,
Yas
ssociated with
6522
These responses
of attempts to process partial informa-
CIA-RDP
6e19-00789R083100120001
..... ...

346
CPYRGUT
PROCEEDINGS OF THE IEEE, MARCH 1976
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
tion: similarly, this occurs in other parapsychological experi-
ments. These observations are compatible with the hypotheses
that information received in a putative remote-viewing mode
is processed piecemeal in pattern form (consistent with a low
bit rate process, but not necessarily requiring it); and the
errors arise in the processes of attempted integration of the
data into larger patterns directed toward verbal labeling.
When the subjects augmented the verbal transcripts with
drawings or sketches, these often expressed the target elements
more accurately than the verbal descriptions. Thus the draw-
ings tended to correspond to the targets more clearly and
precisely than the words of the transcript.
The descriptions given by the subjects sometimes went be-
yond what the remote experimenter experienced, at least con-
sciously. For example, one subject (S4) described and drew
a belt drive at the top of a drill press that was invisible even to
the remote experimenter who was operating the machine;
another subject .(S1) described a number of items behind
shrubbery and thus not visible to members of the demarcation
team at the site.
Curiously, objects in motion at the remote site were rarely
mentioned in the transcript. For example, trains crossing the
railroad trestle target were not described, though the remote
experimenter stood very close to them.
Also in a few cases, the subject descriptions were inaccurate
regarding size of structures. A 20-ft courtyard separating two
buildings was described as 200 ft wide, and a small shed was
expanded to a barn-like structure.
c) Blind judging of transcripts: The judging procedure
entailed examining the transcripts for a given experimental
series and attempting to match the transcripts with the cor-
rect targets on the basis of their correspondences. The tran-
scripts varied from coherent and accurate descriptions to mix-
tures of correspondences and noncorrespondences. Since the
judge did not know a priori which elements of the descriptions
were correct or incorrect, the task was complicated, and tran-
scripts often seemed plausibly to match more than one target.
A confounding factor in these studies is that some target lo-
cations have similarities that seem alike at some level of per-
ception. For example, a radio telescope at the top of a hill,
the observation deck of a tower, and a jetty on the edge of a
bay all match a transcript description of "looking out over a
long distance." A lake, a fountain, and a creek may all result
in an image of water for the subject. Therefore, in several
cases, even correct images may not help in the conservative
differential matching procedure used.
According to the judge, the most successful procedure was a
careful element-by-element comparison that tested each tran-
script against every target and used the transcript descriptions
and drawings as arguments for or against assigning the tran-
script to a particular target. In most cases, this resulted in
either a clear conclusion or at least a ranking of probable
matches; these matches were subjected to the statistical
analyses presented in this paper.
2) Summary: In summary, we do not yet have an under-
standing of the nature of the information-bearing signal that a
subject perceives during remote viewing. The subjects com-
monly report that they perceive the signal visually as though
they were looking at the object or place from a position in its
immediate neighborhood. Furthermore, the subjects' per-
ceptual viewpoint has mobility in that they can shift their
point of view so as to describe elements of a scene that would
not be visible to an observer merely standing at ground leve:
and describing what he sees. (In particular, a subject often
correctly describes elements not visible to the target demarca-
tion team.) Finally, motion is seldom reported; in fact,
moving objects often are unseen even when nearby static
objects are correctly identified.
A comparison of the results of remote viewing (a so-called
free-response task) with results of forced-choice tasks, such as
the selection of one of four choices generated by a random
number generator [58], reveals the following findings. From a
statistical viewpoint. a subject is more likely to describe, with
sufficient accuracy | to permit blind matching, a remote site
chosen at random than he is to select correctly one of four
random numbers. Our experience with these phenomena leads
us to consider that| this difference in task performance may
stem from fundamental signal-to-noise considerations. Two
principal sources of noise in the system apparently are mem-
ory and imagination, both of which can give rise to mental
pictures of greater clarity than the target to be perceived. In
the random number task, a subject can create a perfect mental
picture of each of the four possible outputs in his own imagi-
nation and then attempt to obtain the correct answer by a
mental matching operation. The same is true for card guessing
experiments. On the other hand, the subject in remote view-
ing is apparently more likely to approach the task with a
blank mind as he attempts to perceive pictorial information
from remote locations about which he may have no stored
mental data.
Finally, we observe that most of the correct information
that subjects relate to us is of a nonanalytic nature pertaining
to shape, form, color, and material rather than to function or
name. In consultation with Dr. Robert Ornstein of the Langley-
Porter Neuropsychiatric Institute, San Francisco, CA, and with
Dr. Ralph Kiernan of the Department of Neurology, Stanford
University Medical Center, Stanford, CA, we have formed
the tentative hypothesis that paranormal functioning may
involve specialization characteristic of the brain's right hemi-
sphere. This possibility is derived from a variety of evidence
from clinical and neurosurgical sources which indicate that the
two hemispheres of the human brain are specialized for dif-
ferent cognitive functions. The left hemisphere is predomi-
nantly active in verbal and other analytical functioning and the
right hemisphere predominates in spatial and other holistic
processing [59], [60]. Further research is necessary to
elucidate the relationship between right hemisphere function
and paranormal abilities. Nonetheless, we can say at this
point that the remote-viewing results of the group of subjects
at SRI have characteristics in common with more familiar
performances that require right hemispheric function. The
similarities include the highly schematicized drawings of ob-
jects in a room or of remote scenes. Verbal identification of
these drawings is often highly inaccurate and the drawings
themselves are frequently left-right reversed relative to the
target configuration.
Further, written material generally is
not cognized. These
characteristics have been seen in left
brain-injured patients and in callosal-sectioned patients.
As a result of the above considerations, we have learned to
urge our subjects simply to describe what they see as opposed
to what they think they are looking at. We have learned that
their unanalyzed perceptions are almost always a better guide
to the true target than their interpretations of the perceived
data.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

CPYRGHT
PUTHOFF AApproveck For Release:2000/08/08 41CIA-R.DR96-00789R003100120001-4
347
IV. CONSIDERATIONS CONCERNING TIME
If the authors may be forgiven a personal note, we wish to
express that this section deals with observations that we have
been reluctant to publish because of their striking apparent in-
compatibility with existing concepts. The motivating factor
for presenting the data at this time is the ethical consideration
that theorists endeavoring to develop models for paranormal
functioning should be apprised of all the observable data if
their efforts to arrive at a comprehensive and correct descrip-
tion are to be successful.
During the course of the experimentation in remote viewing
(Section III), subjects occasionally volunteered the informa-
tion that they had been thinking about their forthcoming par-
ticipation in a remote-viewing experiment and had an image
come to them as to what the target location was to be. On
these occasions, the information was given only to the experi-
menter remaining at SRI with the subject and was unknown to
the outbound experimenter until completion of the experi-
ment. Two of these contributions were among the most
accurate descriptions turned in during those experiments.
Since the target location had not yet been selected when the
subject communicated his perceptions about the target, we
found the data difficult to contend with.
We offer these spontaneous occurrences not as proof of pre-
cognitive perception, but rather as the motivation that led us
to do further work in this field. On the basis of this firsthand
evidence, together with the copious literature describing years
of precognition experiments carried out in various other labo-
ratories, we decided to determine whether a subject could per-
form a perceptual task that required both spatial and temporal
remote viewing.
It is well known and recently has been widely discussed that
nothing in the fundamental laws of physics forbids the appar-
ent transmission of information from the future to the present
(discussed further in Section V). Furthermore, there is a gen-
eral dictum that "in physical law, everything that is not forbid-
den, is required" [61]. With this in mind, we set out to con-
duct very well-controlled experiments to determine whether
we could deliberately design and execute experiments for the
sole purpose of observing precognition under laboratory
conditions.
The experimental protocol was identical to that followed in
previous remote-viewing experiments with but one exception.
The exception was that the subject was required to describe
the remote location during a 15-min period beginning 20 min
before the target was selected and 35 min before the outbound
experimenter was to arrive at the target location.
In detail, as shown in Table IX, each day at ten o'clock one
of the experimenters would leave SRI with a stack of ten
sealed envelopes from a larger pool and randomized daily, con-
taining traveling instructions that had been prepared, but that
were unknown to the two experimenters remaining with the
subject. The subject for this experiment was Hella Hammid
(S4) who participated in the nine-experiment series replicating
the original Price work described earlier. The traveling experi-
menter was to drive continuously from 10:00 until 10:30 be-
fore selecting his destination with a random number generator.
(The motivation for continuous motion was our observation
that objects and persons in rapid motion are not generally seen
in the remote-viewing mode of perception, and we wished the
traveler to be a poor target until he reached his target site.) At
the end of 3Approvedng,oreRelease 2000/08/08r CIA-RDP96-00789R003100120001-4
TABLE IX
EXPERIMENTAL PROTOCOL: PRECOGNITIVE REMOTE VIEWING
Time
Schedule
Experimenter/Subject Activity
10:00
Outbound experimenter leaves with 10 envelopes (containing
target locations) and random number generator;
begins half-hour drive
10:10
Experimenters remaining with subject in the laboratory
elicit from subject a description of where outbound
experimenter will be from 10:45-11:00
10:25
Subject response completed, at which time laboratory part
of experiment is over
10:30
Outbound experimenter obtains random number from a random
number generator, counts down to associated envelope, and
proceeds to target location indicated
10:45
Outbound experimenter remains at target location for
15 minutes (10:45-11:00)
Fig. 14. Subject Hammid (S4) described "some kind of congealing tar,
or maybe an area of condensed lava .. . that has oozed out to fill up
some kind of boundaries."
ated a random digit from 0 to 9 with a Texas Instruments
SR-51 random number generator; while still in motion, he
counted down that number of envelopes and proceeded di-
rectly to the target location so as to arrive there by 10:45. He
remained at the target site until 11:00, at which time he re-
turned to the laboratory, showed his chosen target name to a
security guard, and entered the experimental room.
During the same period, the protocol in the laboratory was
as follows. At 10:10, the subject was asked to begin a descrip-
tion of the place to which the experimenter would go 35 min
hence. The subject then generated a tape-recorded description
and associated drawings from 10:10 to 10:25, at which time
her part in the experiment was ended. Her description was
thus entirely concluded 5 min before the beginning of the tar-
get selection procedure.
Four such experiments were carried out. Each of them ap-
peared to be successful, an evaluation later verified in blind
judging without error by three judges. We will briefly sum-
marize the four experiments below.
The first target, the Palo Alto Yacht Harbor, consisted en-
tirely of mud flats because of an extremely low tide (see Fig.
14). Appropriately, the entire transcript of the subject per-
tained to "some kind of congealing tar, or maybe an area of
condensed lava. It looks like the whole area is covered with
some kind of wrinkled elephant skin that has oozed out to fill
up some kind of boundaries where (the outbound experi-
menter) is standing." Because of the lack of water, the dock
where the remote experimenter was standing was in fact rest-

348
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
-
PROCEEDINGS OF THE IEEE, MARCH 1976
Fig. 15. Subject (S4) described a formal garden "very well manicured"
behind a double colonnade.
Note that the subject has learned not to rush into interpreta-
tion as to the nature or purpose of the place. This is a result
of our cautioning based on the observation that such efforts
tend to be purely analytical and in our experience are almost
invariably incorrect. If a subject can limit himself to what he
sees, he is often then able to describe a scene with sufficient
accuracy that an observer can perform the analysis for him and
identify the place.
The second target visited was the fountain at one end of a
large formal garden at Stanford University Hospital (Fig. 15).
The subject gave a lengthy description of a formal garden be-
hind a wall with a "double colonnade" and "very well mani-
cured." When we later took the subject to the location, she
was herself taken aback to find the double colonnaded wall
leading into the garden just as described.
The third target was a children's swing at a small park 4.6 km
from the laboratory (Fig. 16). The subject repeated again and
again that the main focus of attention at the site was a "black
iron triangle that the outbound experimenter had somehow
walked into or was standing on." The triangle was "bigger
than a man," and she heard a "squeak, squeak, about once a
second," which we observe is a match to the black metal swing
that did squeak.
Fig. 16. Subject (S4) saw a "black iron triangle that Hal had somehow
walked into" and heard a "squeak, squeak, about once a second."
Fig. 17. Subject (S4) described a very tall structure located among city
streets and covered with "Tiffany-like glass."
The final target was the Palo Alto City Hall (Fig. 17). The
subject described a very, very tall structure covered with
"Tiffany-like glass." She had it located among city streets and
with little cubes at the base. The building is glass-covered, and
the little cubes are a good match to the small elevator exit
buildings located in the plaza in front of the building.
To obtain a numerical evaluation of the accuracy of the pre-
cognitive viewing, the experimental results were subjected to
independent judging on a blind basis by three SRI scientists
who were not otherwise associated with the experiment. The
judges were asked to match the four locations, which they
visited, against the unedited typed manuscripts of the tape-
recorded narratives, along with the drawings generated by the
remote viewer. The transcripts were presented unlabeled and
in random order and were to be used without replacement. A
correct match required that the transcript of a given experi-
ment be matched with the target of that experiment. All three
judges independently matched the target data to the response
data without error. Under the null hypothesis (no information
channel and a random selection of descriptions without re-
placement), each judge independently obtained a result signifi-
cant at p = (4!)-1 = 0.042.
For reasons we do not as yet understand, the four transcripts
generated in the precognition experiment show exceptional co-
herence and accuracy as evidenced by the fact that all of the
judges were able to match successfully all of the transcripts to
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

PUTHOFF
Approved For Release 2000108708"ACPA-RDP96-00789R003960120001-4
349
the corresponding target locations. A long-range experimental
program devoted to the clarification of these issues and involv-
ing a number of subjects is under way. The above four experi-
ments are the first four carried out under this program.
Currently, we have no precise model of this spatial and tem-
poral remote-viewing phenomenon. However, models of the
universe involving higher order synchronicity or correlation
have been proposed by the physicist Pauli and the psychologist
Carl Jung [62].
ACAUSALITY. If natural law5 were an absolute truth, then of
course there could not possibly be any processes that deviate
from it. But since causality5 is a statistical truth, it holds good
only on average and thus leaves room for exceptions which must
somehow be experienceable, that is to say, real. I try to regard
synchronistic events as acausal exceptions of this kind. They
prove to be relatively independent of space and time; they rela-
tivize space and time insofar as space presents in principle no ob-
stacle to their passage and the sequence of events in time is in-
verted so that it looks as if an event which has not yet occurred
were causing a perception in the present.
We shall see in the next section that such a description,
though poetic, has some basis in modern physical theory.
V. DISCUSSION
It is important to note at the outset that many contempo-
rary physicists are of the view that the phenomena that we
have been discussing are not at all inconsistent with the
framework of physics as currently understood. In this emerg-
ing view, the often-held belief that observations of this type
are incompatible with known laws in principle is erroneous,
such a concept being based on the naive realism prevalent
before the development of modern quantum theory and
information theory.
One hypothesis, put forward by I. M. Kogan of the USSR,
is that information transfer under conditions of sensory
shielding is mediated by extremely low-frequency (ELF)
electromagnetic waves in the 300-1000-km region [37]-
[40]. Experimental support for the hypothesis is claimed
on the basis of slower than inverse square attenuation, com-
patible with source-percipient distances lying in the induc-
tion field range as opposed to the radiation field range; ob-
served low bit rates (0.005-0.1 bit/s) compatible with the
information carrying capacity of ELF waves; apparent ineffec-
tiveness of ordinary electromagnetic shielding as an attenuator;
and standard antenna calculations entailing biologically gener-
ated currents yielding results compatible with observed signal-
to-noise ratios.
M. Persinger, Psychophysiology Laboratory, Laurentian Uni-
versity, Toronto, Canada, has narrowed the ELF hypothesis to
the suggestion that the 7.8-Hz "Shumann waves" and their
harmonics propagating along the earth-ionosphere waveguide
duct may be responsible. Such an hypothesis is compatible
with driving by brain-wave currents and leads to certain other
hypotheses such as asymmetry between east-west and west-
east propagation, preferred experimental times (midnight-4
A.M.), and expected negative correlation between. success
and the U index (a measure of geomagnetic disturbance
throughout the world). Persinger claims initial support for
these factors on the basis of a literature search [63], [64]. .
On the negative side with regard to a straightforward ELF
interpretation as a blanket hypothesis are the following: a) ap-
parent real-time descriptions of remote activities in sufficient
detail to require a channel capacity in all probability greater
than that allowed by a conventional modulation of an ELF
signal; b) lack of a proposed mechanism for coding and decod-
ing the information onto the proposed ELF carrier; and c) ap-
parent precognition data. The hypothesis must nonetheless re-
main open at this stage of research, since it is conceivable that
counterindication a) may eventually be circumvented on the
basis that the apparent high bit rate results from a mixture of
low bit rate input and high bit rate "filling in the blanks" from
imagination; counterindication b) is common to a number of
normal perceptual tasks and may therefore simply reflect a
lack of sophistication on our part with regard to perceptual
functioning [65]; and counterindication c) may be accom-
modated by an ELF hypothesis if advanced waves as well as
retarded waves are admitted [66], [67]. Experimentation to
determine whether the ELF hypothesis is viable can be carried
out by the use of ELF sources as targets, by the study of para-
metric dependence on propagational directions and diurnal
timing, and by the exploration of interference effects caused
by creation of a high-intensity ELF environment during ex-
perimentation, all of which are under consideration in our lab-
oratory and elsewhere.
Some physicists believe that the reconciliation of observed
paranormal functioning with modern theory may take place at
a more fundamental level-namely, at the level of the founda-
tions of quantum theory. There is a continuing dialog, for
example, on the proper interpretation of the effect of an ob-
server (consciousness) on experimental measurement [68],
and there is considerable current interest in the implications
for our notions of ordering in time and space brought on by
the observation [69], [70] of nonlocal correlation or "quan-
tum interconnectedness" (to use Bohm's term [71]) of distant
parts of quantum systems of macroscopic dimensions. The
latter, Bell's theorem [72], emphasizes that "no theory of
reality compatible with quantum theory can require spatially
separated events to be independent" [73], but must permit
interconnectedness of distant events in a manner that is con-
trary to ordinary experience [74]-[75]. This prediction has
been experimentally tested and confirmed in the recent
experiments of, for example, Freedman and Clauser [69],
[70].
E. H. Walker and O. Costa de Beauregard, independently
proposing theories of paranormal functioning based on quan-
tum concepts, argue that observer effects open the door to the
possibility of nontrivial coupling between consciousness and
the environment and that the nonlocality principle permits
such coupling to transcend spatial and temporal barriers [76],
[77].
Apparent "time reversibility"-that is, effects (e.g., observa-
tions) apparently preceding causes (e.g., events)-though con-
ceptually difficult at first glance, may be the easiest of appar-
ent paranormal phenomena to assimilate within the current
theoretical structure of our world view. In addition to the
familiar retarded potential solutions f(t - r/c), it is well known
that the equations of, for example, the electromagnetic field
admit of advanced potential solutions f(t +r/c)-solutions that
would appear to imply a reversal of cause and effect. Such
solutions are conventionally discarded as not corresponding to
any observable physical event. One is cautioned, however, by
statements such as that of Stratton in his basic text on electro-
"Asproved For Release 2000/08/08 : CHAND+5℃66789R003100120001-4
' As usually un

350
CPYRGHT
PROCEEDINGS OF THE IEEE, MARCH 1-97(
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
The reader has doubtless noted that the choice of the function
f(t - r/c) is highly arbitrary, since the field equation admits also
a solution f(t + r/c). This function leads obviously to an advanced
time, implying that the field can be observed before it has been
generated by the source. The familiar chain of cause and effect
is thus reversed and this alternative solution might be discarded
as logically inconceivable. However, the application of "logical"
causality principles offers very insecure footing in matters such
as these and we shall do better to restrict the theory to retarded
action solely on the grounds that this solution alone conforms to
the present physical data.
Such caution is justified by the example in the early 1920's of
Dirac's development of the mathematical description of the
relativistic electron that also yielded a pair of solutions, one of
which was discarded as inapplicable until the discovery of the
positron in 1932.
In an analysis by O. Costa de Beauregard, an argument is put
forward that advanced potentials constitute a convergence
toward "finality" in a manner symmetrical to the divergence
of retarded potentials as a result of causality [77]. Such
phenomena are generally unobservable, however, on the gross
macroscopic scale for statistical reasons. This is codified in
the thermodynamic concept that for an isolated system entropy
(disorder) on the average increases. It is just this requirement
of isolation, however, that has been weakened by the observer
problem in quantum theory, and O. Costa de Beauregard argues
that the finality principle is maximally operative in just those
situations where the intrusion of consciousness as an ordering
phenomenon results in a significant local reversal of entropy
increase. At this point, further discussion of the subtleties of
such considerations, though apropos, would take us far afield,
so we simply note that such advanced waves, if detected, could
in certain cases constitute a carrier of information precognitive
to the event.
The above arguments are not intended to indicate that the
precise nature of the information channel coupling remote
events and human perception is understood. Rather, we in-
tend to show only that modern theory is not without resources
that can be brought to bear on the problems at hand, and we
expect that these problems will, with further work, continue
to yield to analysis and specification.
Furthermore, independent of the mechanisms that may be
involved in remote sensing, observation of the phenomenon
implies the existence of an information channel in the
information-theoretic sense. Since such channels are amenable
to analysis on the basis of communication theory techniques,
as indicated earlier, channel characteristics such as bit rate can
be determined independent of a well-defined physical channel
model in the sense that thermodynamic concepts can be ap-
plied to the analysis of systems independent of underlying
mechanisms. Furthermore, as we have seen from the work of
Ryzl discussed in Section II, it is possible to use such a channel
for error-free transmission of information if redundancy coding
is used. (See also Appendix A.) Therefore, experimentation
involving the collection of data under specified conditions per-
mits headway to be made despite the formidable work that
needs to be done to clarify the underlying bases of the
phenomena.
VI. CONCLUSION
-1
For the past three years we have had a program in the Elec-
tronics and Bioengineering Laboratory of SRI to investigate
those facets of human perception that appear to fall outside
the raApproved For Release 2000/08/08g: GIA-RD .96-0078
bilities. The primary achievement of this program has beer
the elicitation of high-quality "remote viewing"-the ability
of both experienced subjects and inexperienced volunteers
to view, by means of innate mental processes, remote geo-
graphical or technical targets such as roads, buildings, and
laboratory apparatus. Our accumulated data from over fifty
experiments with more than a half-dozen subjects indicate
the following. a) The phenomenon is not a sensitive function
of distance over a range of several kilometers. b) Faraday cage
shielding does not appear to degrade the quality or accuracy of
perception. c) Most of the correct information that subjects
relate is of a nonanalytic nature pertaining to shape, form,
color, and material rather than to function or name. (This
aspect suggests a hypothesis that information transmission
under conditions of sensory shielding may be mediated pri-
marily by the brain's right hemisphere.) d) The principal
difference between experienced subjects and inexperienced
volunteers is not that the latter never exhibit the faculty, but
rather that their results are simply less reliable. (This observa-
tion suggests the hypothesis that remote viewing may be a
latent and widely distributed, though repressed, perceptual
ability.)
Although the precise nature of the information channel cou-
pling remote events and human perception is not yet under-
stood, certain concepts in information theory, quantum
theory, and neurophysiological research appear to bear directly
on the issue. As a result, the working assumption among re-
searchers in the field is that the phenomenon of interest is
consistent with modern scientific thought, and can therefore
be expected to yield to the scientific method. Further, it is
recognized that communication theory provides powerful
techniques, such as the use of redundancy coding to improve
signal-to-noise ratio, which can be employed to pursue special-
purpose application of the remote-sensing channel independent
of an understanding of the underlying mechanisms. We there-
fore consider it important to continue data collection and to
encourage others to do likewise; investigations such as those
reported here need replication and extension under as wide a
variety of rigorously controlled conditions as possible.
APPENDIX A
SIGNAL ENHANCEMENT IN A PARANORMAL
COMMUNICATION CHANNEL BY APPLICATION
OF REDUNDANCY CODING
Independent of the mechanisms that may be involved in
remote sensing, observation of the phenomenon implies the
existence of an information channel in the information-
theoretic sense. As we have seen from the work of Ryzl dis-
cussed in Section II,º it is even possible to use such a (noisy)
channel for error-free transmission of information if suf-
ficient redundancy coding is used [30], [31]. Following is a
general procedure that we have used successfully for signal
enhancement.
We shall assume that the "message" consists of a stream of
binary digits (0,1) of equal probability (e.g., binary sort of
green/white cards as in Ryzl's case, English text encoded as in
Table X and sent long distance by strobe light on/off, and so
on). To combat channel noise, each binary digit to be sent
through the channel requires the addition of redundancy bits
(coding). Efficient coding requires a compromise between the
desire to maximize reliability and the desire to minimize re-
"063'1001200014essful work done by

CPYRGHT
PUTHOFF AADBIOVedRFor Release 2000/08/08/A MIA-RDP96-00789R003100120001-4
TABLE X
5-BIT CODE FOR ALPHANUMERIC
CHARACTERS
E
00000
Y
01000
T
11111
G.J
10111
N
00001
พ
01001
R
11110
V
10110
I
00010
B
01010
0
11101
10101
A
00011
1
01011
S, X, Z
11100
2
10100
D
00100
3
01100
H
11011
4
10011
L
00101
5
01101
C, K, Q
11010
6
10010
F
00110
7
01110
P
11001
8
10001
U
00111
9
01111
M
11000
.
10000
Note: Alphabet characters listed
in order of decreasing frequency
in English text. See, for example,
A. Sinkov [79]. (The low-fre-
quency letters, X, Z, K, Q, and J,
have been grouped with similar
characters to provide space for
numerics in a 5-bit code.) In
consideration of the uneven dis-
tribution of letter frequencies in
English text, this code is chosen
such that 0 and 1 have equal
probability.
dundancy. One efficient coding scheme for such a channel is
obtained by application of a sequential sampling procedure of
the type used in production-line quality control [80]. The
adaptation of such a procedure to paranormal communication
channels, which we now discuss, was considered first by
Taetzsch [81]. The sequential method gives a rule of proce-
dure for making one of three possible decisions following the
receipt of each bit: accept 1 as the bit being transmitted; reject
1 as the bit being transmitted (i.e., accept 0); or continue
transmission of the bit under consideration. The sequential
sampling procedure differs from fixed-length coding in that
the number of bits required to reach a final decision on a
message bit is not fixed before transmission, but depends on
the results accumulated with each transmission. The principal
advantage of the sequential sampling procedure as compared
with the other methods is that, on the average, fewer bits per
final decision are required for an equivalent degree of
reliability.
CUMULATIVE NUMBER OF MATCHES AGAINST
50
"1" TARGET MESSAGE SEQUENCE
40
DECISION 1
Accept "1" as
30
the Bit Being
DECISION 3
Transmitted
Continue
20
Transmission
DECISION 2
Accept "O" as
10
the Bit Being
Transmitted
0
0
10
20
30
40
50
60
70
80
NUMBER OF TRIALS
Fig. 18. Enhancement of signal-to-noise ratio by sequential sampling
procedure (Po = 0.4, p, = 0.6, x = 0.01, 8 = 0.01).
The question to be addressed is whether, after repeated
transmission, a given message bit is labeled a "1" at a low rate
Po commensurate with the hypothesis Ho that the bit in ques-
tion is a "0," or at a higher rate p1 commensurate with the
hypothesis H1 that the bit in question is indeed a "1." The
decision-making process requires the specification of four
parameters.
Po The probability of labeling incorrectly a "O" message bit
as a "1." The probability of labeling correctly a "0" as
a "0" is p = 0.5 + +b = 0.6. Therefore, the probability
of labeling incorrectly a "O" as a "1" is 1 - p = 0.4 = Po.
P: The probability of labeling correctly a "1" message bit
.
as a "1," is given by p1 = 0.5 + V b = 0.6.
α
The probability of rejecting a correct identification for
a "0" (Type I error). We shall take @ = 0.01.
β
The probability of accepting an incorrect identification
for a "1" (Type II error). We shall take $ = 0.01.
With the parameters thus specified, the sequential sampling
procedure provides for construction of a decision graph as
shown in Fig. 18. The equations for the upper and lower limit
lines are
Σ, = d1 +SN
Σ. = - do + SN
Use of the sequential sampling procedure requires the speci-
where
fication of parameters that are determined on the basis of the
following considerations. Assume that a message bit (0 or 1)
log -B
log
do =
1 - a
:
a
d1 =
is being transmitted. In the absence of a priori knowledge, we
may assume equal probability (p = 0.5) for the two possibili-
P1 1 - Po
ties (0,1). Therefore, from the standpoint of the receiver, the
log "1
P1 1 - Po
log -
probability of correctly identifying the bit being transmitted is
Po 1 - P1
Po 1 - P1
S = -
log
log
P1 1 - Po
Po 1 - P1
1 - Po
1 - p1
p = 0.5 because of chance alone. An operative remote-sensing
channel could then be expected to alter the probability of
correct identification to a value p = 0.5 + v, where the param-
eter \ satisfies 0<|v|<0.5. (The quantity may be positive
or negative depending on whether the paranormal channel
results in so-called psi-hitting or psi-missing.). Good psi func-
tioning on a repetitive task has been observed to result in v =
0.12, as reported by Ryzl [31]. Therefore, to indicate the
in which S is the slope, N is the number of trials, and d1 and
design procedApprovedfor Release 2000/08/08 : GIA RDR96-00789R003100120001-4 of receiver-
Vp = 0.1 and design a communication system on this basis.
generated responses to the target bit is compiled until either
351

352
CPYRGHT
PROCEEDINGS OF THE IEEE, MARCH 197
Approved For Release 2000/08/08 : CIA RDP96 00789R003100120001-4
1.0
RELIABILITY (probability of correctly
determining message being transmitted)
0.8
0.6
0.4
0.2
1
0
+0.4 +0.2
0 -0.2 -0.4
¿ (psi parameter)
Fig. 19. Reliability curve for sequential sampling procedure (Po = 0.4,
p: = 0.6, a = 0.01, 8 =0.01).
the upper or the lower limit line is reached, at which point a
decision is made to accept 0 or 1 as the bit being transmitted.
Channel reliability (probability of correctly determining
message being transmitted) as a function of operative psi
parameter w is plotted in Fig. 19. As observed, the sequential
sampling procedure can result in 90 percent or greater reliability
with psi parameters on the order of a few percent.
Implementation of the sequential sampling procedure re-
quires the transmission of a message coded in binary digits.
Therefore, the target space must consist of dichotomous ele-
ments such as the white and green cards used in the experi-
ments by Ryzl.
In operation, a sequence corresponding to the target bit (0
or 1) is sent and the cumulative entries are made (Fig. 18) until
a decision is reached to accept either a 1 or a 0 as the bit being
transmitted. At a prearranged time, the next sequence is
begun and continues as above until the entire message has been
received. A useful alternative, which relieves the percipient of
the burden of being aware of his self-contradiction from trial
to trial, consists of cycling through the entire message repeti-
tively and entering each response on its associated graph until
a decision has been reached on all message bits. The authors
have used this technique successfully in a pilot study, but a
discussion of this would take us beyond the intended scope of
this paper.
From the results obtained in such experiments, the channel
bit rate can be ascertained for the system configuration under
consideration. Furthermore, bit rates for other degrees of
reliability (i.e., for other Po, P1, @, and B) can be estimated by
construction of other decision curves over the same data base
and thus provide a measure of the bit rate per degree of
reliability.
In summary, the procedures described here can provide for a
specification of the characteristics of a remote-sensing channel
under well-defined conditions. These procedures also provide
for a determination of the feasibility of such a channel for
particular applications.
APPENDIX B
REMOTE-VIEWING TRANSCRIPT
Following is the unedited transcript of the first experiment
with an SRI volunteer (S6), a mathematician in the computer
science laboratory, w
viewing. The target, determined by random procedure, wa
White's Plaza, a plaza with fountain at Stanford University
(shown in Fig. 8).
As is our standard protocol, the experi
menter with the subject is kept ignorant of the specific targer
visited as well as the contents of the target pool. The experi
menter's statements and questions are italics.
Today is Monday, October 7th. It is 11:00 and this is a re.
mote viewing experiment with Russ Targ, Phyllis Cole, and Ha.
Puthoff. In this experiment Hal will drive to a remote site
chosen by a random process. Phyllis Cole will be the remote
viewer, and Russ Targ is the monitor. We expect this experi-
ment to start at twenty minutes after eleven and run for fif-
teen minutes.
It is just about twenty minutes after eleven and Hal should
be at his target location by now.
Why don't you tell me what kind of pictures you see and
what you think he might be doing or experiencing.
The first thing that came to mind was some sort of a large,
square kind of a shape. Like Hal was in front of it. It was a
... not a building or something, it was a square. I don't know
if it was a window, but something like that so that the bottom
line of it was not at the ground. About where his waist was, at
least. That's what it seemed to me. It seems outdoors some-
how. Tree.
Does Hal seem to be looking at that square?
I don't know. The first impression was that he wasn't, but I
have a sense that whatever it was was something one might
look at. I don't know if it would be a sign, but something that
one might look at.
Can you tell if it is on the ground or vertical?
It seemed vertical.
I don't have a sense that it was part of anything particular.
It might be on a building or part of a building, but I don't
know. There was a tree outside, but I also got the impression
of cement. I don't have the impression of very many people
or traffic either. I have the sense that he is sort of walking
back and forth. I don't have any more explicit picture than
that.
Can you move into where he is standing and try to see what
he is looking at?
I picked up he was touching something-something rough.
Maybe warm and rough. Something possibly like cement.
It is twenty-four minutes after eleven.
Can you change your point of view and move above the
scene so you can get a bigger picture of what's there?
I still see some trees and some sort of pavement or some-
thing like that. Might be a courtyard. The thing that came to
mind was it might be one of the plazas at Stanford campus or
something like that, cement.
Some kinds of landscaping.
I said Stanford campus when I started to see some things in
White Plaza, but I think that is misleading.
I have the sense that he's not moving around too much.
That it's in a small area.
I guess I'll go ahead and say it, but I'm afraid I'm just putting
on my impressions from Stanford campus. I had the impres-
sion of a fountain. There are two in the plaza, and it seemed
that Hal was possibly near the, what they call Mem Claw.
What is that?
It's a fountain that looks rather like a claw. It's a black
sculpture. And it has benches around it made of cement.
Are there any buildings at the place you are looking at? Are
Approved For Release 2000/08/08: CIA-RIP96200789R003100120001 kind of a courtyard.

353
CPYRGHT
PUTHOFF AND TARG: PERCEPTUAL CHANNEL FOR INFORMATION
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Usually at some places there should be a building, large or
small that the courtyard is about. Look at the end or the sides
of the courtyard. Is there anything to be seen?
I have a sense that there are buildings. It's not solid build-
ings. I mean there are some around the periphery and I have a
sense that none of them are very tall. Maybe mostly one story,
maybe an occasional two story one.
Do you have any better idea of what your square was that
you saw at the outset?
No. I could hazard different kinds of guesses.
Does it seem part of this scene?
It .. . I think it could be. It could almost be a bulletin board
or something with notices on it maybe.
Or something that people are expected to look at. Maybe a
window with things in it that people were expected to look at.
What kind of trees do you see in this place?
I don't know what kind they are. The impression was that
they were shade trees and not terribly big. Maybe 12 feet of
trunk and then a certain amount of branches above that. So
that the branches have maybe a 12 foot diameter, or some-
thing. Not real big trees.
New trees rather than old trees?
Yeah, maybe 5 or 10 years old, but not real old ones.
Is there anything interesting about the pavement?
No. It seems to be not terribly new or terribly old. Not
very interesting. There seems to be some bits of landscaping
around. Little patches of grass around the edges and periph-
eries. Maybe some flowers. But, not lush.
You saw some benches. Do you want to tell me about them?
Well, that's my unsure feeling about this fountain. There
was some kind of benches of cement. Curved benches, it felt
like.
They were of rough cement.
What do you think Hal is doing while he is there?
I have a sense that he is looking at things trying to project
them. Looking at different things and sort of walking back
and forth not covering a whole lot of territory.
Sometimes standing still while he looks around.
I just had the impression of him talking, and I almost sense
that it was being recorded or something. I don't know if he
has a tape recorder, but if it's not that, then he is saying some-
thing because it needed to be remembered. It's 11:33. He's
just probably getting ready to come back.
ACKNOWLEDGMENT
The authors wish to thank the principal subjects, Mrs. Hella
Hammid, Pat Price, and Ingo Swann, who showed patience and
forbearance in addition to their enthusiasm and outstanding
perceptual abilities. We note with sadness the death of one of
our subjects, Mr. Price. We express our sincere thanks also to
Earle Jones, Bonnar Cox, and Dr. Arthur Hastings, of SRI, and
Mrs. Judith Skutch and Richard Bach, without whose en-
couragement and support this work could not have taken
place.
REFERENCES
[1] J. R. Smythies, Ed., Science and ESP. London, England: Rout-
ledge, 1967.
[2] C. Evans, "Parapsychology-What the questionnaire revealed,"
New Scientist, Jan. 25, 1973, p. 209.
[3] A. Gauld, The Founders of Psychical Research. New York:
Schocken Books, 1968. See also W. Crookes, Researches in the
Phenomena of Spiritualism. London, England: J. Burns, 1874.
18, 1974.
[5] D. D. Home, Lights and Shadows of Spiritualism. New York:
G. W. Carleton, 1877.
[6] J. Coover, Experiments in Psychical Research. Palo Alto, CA:
Stanford Univ. Press, 1917.
[7] G. Estabrooks, Bull. Boston Society for Psychical Research, 1927.
See also [12, pp. 18-19].
[8] L. T. Troland, Techniques for the Experimental Study of Telep-
athy and Other Alleged Clairvoyant Processes. Albany, NY, 1928.
[9] J. B. Rhine, New Frontiers of the Mind. New York: Farrar and
Rinehart, 1937.
[10] J. Pratt and J. B. Rhine et al., Extra-Sensory Perception after
Sixty Years. New York: Henry Holt, 1940.
[11] C. Scott, "G. Spencer Brown and probability: A critique,"
J. Soc. Psychical Res., vol. 39, pp. 217-234, 1958.
[12] G. R. Price, "Science and the supernatural," Science, vol. 122,
pp. 359-367, 1955.
[13]
-, "Apology to Rhine and Soal," Science, vol. 175, p. 359,
1972.
[14] J. B. Rhine, "A new case of experimenter unreliability," J.
Parapsychol., vol. 38, pp. 215-225, June 1974.
[15] S. G. Soal and F. Bateman, Modern Experiments in Telepathy.
London, England: Faber and Faber, 1953.
[16] C. Scott and P. Haskell, ""'Normal' explanation of the Soal-
Goldney experiments in extra-sensory perception," Nature, vol.
245, pp. 52-54, Sept. 7, 1973.
[17] C. E. M. Hansel, ESP-A Scientific Evaluation. New York:
Scribner, 1966.
[ 18] J. B. Rhine and J. G. Pratt, "A review of the Pearce-Pratt dis-
tance series of ESP tests," J. Parapsychol., vol. 18, pp. 165-177,
1954.
[19] J. G. Pratt and J. L. Woodruff, "Size of stimulus symbols in extra-
sensory perception," J. Parapsychol., vol. 3, pp. 121-158, 1939.
[20] S. G. Soal and H. T. Bowden, The Mind Readers: Recent Experi-
ments in Telepathy. New Haven, CT: Yale Univ. Press, 1954.
[21] C. Honorton, "Error some place!" J. Commun., vol. 25, no. 1
(Annenberg School of Commun.), Winter 1975.
[22] M. Ryzi, "Training the psi faculty by hypnosis," J. Amer. Soc.
Psychical Res., vol. 41, pp. 234-251, 1962.
[23] CIBA Foundation Symposium on Extra Sensory Perception.
Boston, MA: Little, Brown, 1956.
[24} M. Ryzl and J. Pratt, "A repeated-calling ESP test with sealed
cards," J. Parapsychol., vol. 27, pp. 161-174, 1963.
[25]
-, "A further confirmation of stabilized ESP performance in a
selected subject," J. Parapsychol., vol. 27, pp. 73-83, 1963.
[26] J. Pratt, "Preliminary experiments with a 'borrowed' ESP sub-
ject," J. Amer. Soc. Psychical Res., vol. 42, pp. 333-345, 1964.
[27] J. Pratt and J. Blom, "A confirmatory experiment with 'bor-
rowed" outstanding ESP subject," J. Amer. Soc. Psychical Res ..
vol. 42, pp. 381-388, 1964.
[28] W. G. Roll and J. G. Pratt, "An ESP test with aluminum targets,"
J. Amer. Soc. Psychical Res., vol. 62, pp. 381-387, 1968.
[29] J. Pratt, "A decade of research with a selected ESP subject: An
overview and reappraisal of the work with Pavel Stepanek,"
Proc. Amer. Soc. Psychical Res., vol. 30, 1973.
[30] C. Shannon and W. Weaver, The Mathematical Theory of Com-
munication. Urbana, IL: Univ. Illinois Press, 1949.
[31] M. Ryzl, "A model for parapsychological communication,"
J. Parapsychol., vol. 30, pp. 18-31, Mar. 1966.
[32] C. Tart, "Card guessing tests: Learning paradigm or extinction
paradigm," J. Amer. Soc. Psychical Res., vol. 60, p. 46, 1966.
(33] M. Ullman and S. Krippner, with A. Vaughan, Dream Telepathy.
New York: Macmillan, 1973.
[34] C. Honorton, "State of awareness factors in psi activation," J.
Amer. Soc. Psychical Res., vol. 68, pp. 246-257, 1974.
[35] Proc. 2nd Int. Congr. Psychotronic Research (Monte Carlo).
Cotati, CA: Int. Assoc. Psychotronic Res., 1975.
[36] L. L. Vasiliev, Experiments in Mental Suggestion. Hampshire,
England: ISMI Publ., 1963.
[37] I. M. Kogan, "Is telepathy possible?" Radio Eng., vol. 21, p. 75,
Jan. 1966.
[38]
-, "Telepathy, hypotheses and observations," Radio Eng., vol.
22, p. 141, Jan. 1967.
[39]
-, "Information theory analysis of telepathic communication
experiments," Radio Eng., vol. 23, p. 122, Mar. 1968.
[40]
-, "The information theory aspect of telepathy," RAND Publ.,
Santa Monica, CA, p. 4145, July 1969.
[41] A. S. Presman, Electromagnetic Fields and Life. New York:
Plenum, 1970.
[42] Y. A. Kholodov, Ed., Influence of Magnetic Fields on Biological
Objects, JPRS 63038, NTIS, Springfield, VA, Sept. 24, 1974.
[43] Y. A. Kholodov, "Investigation of the direct effect of magnetic
fields on the central nervous system," in Proc. 1st Conf. Psycho-
tronic Res., JPRS L/5022-1 and 2, Sept. 6, 1974.
[4] R.
Andloved ForRelease 2gbprogjoad CIA-RDE 96 00789R003100120001-4
tio
[44] D. Mennie, "Consumer electronics." IEEE Spectrum, vol. 12,
Comov, and A. R. Luria,

354
CPYRGHT
PROCEEDINGS OF THE IEEE, MARCH 1976
Approved For Release 2000/08/08 . CIA-RDP96-00789R003100120001-4
"parapsychology: Fiction or reality?" Questions of Philosophy,
vol. 9, pp. 128-136, 1973.
[46] R. Cavanna, Ed., Proc. Int. Conf. Methodology in PSI Research.
New York: Parapsychology Foundation, 1970.
[47] E. D. Dean, "Plethysmograph recordings as ESP responses," Int.
J. Neuropsychiatry, vol. 2, Sept. 1966.
[48] C. Tart, "Physiological correlates of psi cognition," Int. J. Para-
psychol., no. 4, 1963.
[49] D. H. Lloyd, "Objective events in the brain correlating with
psychic phenomena," New Horizons, vol. 1, no. 2, Summer 1973.
[50] J. Silverman and M. S. Buchsbaum, "Perceptual correlates of con-
sciousness; A conceptual model and its technical implications for
psi research," in Psi Favorable States of Consciousness, R. Ca-
vanna, Ed. New York: Parapsychology Foundation, pp. 143-
169, 1970.
[51] J. Kamiya, "Comment to Silverman and Buchsbaum," ibid., pp.
158-159.
[52] D. Hill and G. Parr, Electroencephalography. A Symposium on
Its Various Aspects. New York: Macmillan, 1963.
[53] T. D. Duane and T. Behrendt, "Extrasensory electroencephalo-
graphic induction between identical twins," Science, vol. 150,
p. 367, 1965.
[54] K. Osis, ASPR Newsletter, no. 14, 1972.
[55] R. L. Morris, "An exact method for evaluating preferentially
matched free-response material," J. Amer. Soc. Psychical Res.,
vol. 66, p. 401, Oct. 1972.
[56] G. R. Schmeidler, "PK effects upon continuously recorded tem-
peratures," J. Amer. Soc. Psychical Res., vol. 67, no. 4, Oct.
1973.
[57] W. Scherer, "Spontaneity as a factor in ESP," J. Amer. Soc.
Psychical Res., vol. 12, pp. 126-147, 1948.
[58] R. Targ, P. Cole, and H. Puthoff, "Techniques to enhance man/
machine communication," SRI, Menlo Park, CA, Final Rep.,
NASA Contract NAS7-100, June 1974.
[59] R. Ornstein, The Nature of Human Consciousness. San Fran-
cisco, CA : Freeman, 1973, ch. 7 and 8.
[60] R. W. Sperry, "Cerebral organization and behavior," Science,
vol. 133, pp. 1749-1757, 1961.
[61] O. Bilaniuk and E. C. G. Sudarshan, "Particles beyond the light
Barrier," Phys. Today, vol. 22, May 5, 1969.
[62] W. Pauli and C. G. Jung, Eds., The Interpretation of Nature and
the Psyche (Bollingen Ser. LI). Princeton, NJ: Princeton Univ.
Press, 1955.
[63] M. A. Persinger, "ELF waves and ESP," New Horizons Trans.
Toronto Society for Psychical Research, vol. 1, no. 5, Jan: 1975.
[64|
-, "The paranormal-P. II: Mechanisms and models," M.S.S.
Information Corp., New York, 1974.
[65] B. Julesz, Foundations of Cyclopean Perception. Chicago, IL:
Univ. Chicago Press, 1971.
[66] H. Puthoff and R. Targ, in Psychic Exploration-A Challenge for
Science, J. White, Ed. New York: Putnam, 1974, pp. 522-542.
[67] G. Feinberg, "Precognition-A memory of things future?" in
Proc. Conf. Quantum Physics and Parapsychology (Geneva,
Switzerland). New York: Parapsychology Foundation, 1975.
[68] E. P. Wigner, "The problem of measurement," Amer. J. Phys.,
vol. 31, no. 1, p. 6, 1963.
[69] J. J. Freedman and J. F. Clauser, "Experimental test of local
hidden variable theories," Phys. Rev. Lett., vol. 28, no. 14, p.
938, Apr. 3, 1972.
[70] J. F. Clauser and M. A. Horne, "Experimental consequences of
objective local theories," Phys. Rev. D, vol. 10, no. 2, p. 526,
July 15, 1974.
[71] D. Bohm and B. Hiley, "On the intuitive understanding of non-
locality as implied by quantum theory" (Birkbeck College,
London, England), Feb. 1974, Preprint.
[72] J. S. Bell, "On the problem of hidden variables in quantum
theory," Rev. Mod. Phys., vol. 38, no. 3, p. 447, July 1966.
[73] H. Stapp, "Theory of reality," Lawrence-Berkeley Lab. Rep.
LBL-3837, Univ. California, Berkeley, Apr. 1975.
[74] A. Einstein, B. Podolsky, and N. Rosen, "Can quantum-mechani-
cal description of physical reality be considered complete?" Phys.
Rev., vol. 47, p. 777, May 15, 1935.
[75] R. H. Dicke and J. P. Wittke, Introduction to Quantum Mechan-
ics. Reading, MA: Addison-Wesley, 1960, ch. 7.
[76] E. H. Walker, "Foundations of paraphysical and parapsychologi-
cal phenomena,' in Proc. Conf. Quantum Physics and Para-
psychology (Geneva, Switzerland). New York: Parapsychology
Foundation, 1975.
[77] O. Costa de Beauregard, "Time symmetry and interpretation of
quantum mechanics," Lecture delivered at Boston Colloquim for
Philosophy of Science (Feb. 1974), Foundations of Physics (in
press).
[78 | J. A. Stratton, Electromagnetic Theory. New York: McGraw-
Hill, 1941.
[79] A. Sinkov, Elementary Cryptanalysis-A Mathematical Approach.
New York: Random House, 1968.
[80] P. Hoel, Introduction to Mathematical Statistics, 2nd ed. New
York: Wiley, 1954, p. 27.
[81] R. Taetzsch, "Design of a psi communications system," Int. J.
Parapsychol., vol. 4, no. 1, p. 35, Winter 1962.
.. ..
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
ADVANCES IN REMOTE-VIEWING
ANALYSIS.
BY EDWIN C. MAY, JESSICA M. UTTS, BEVERLY S. HUMPHREY,
WANDA L. W. LUKE, THANE J. FRIVOLD, AND VIRGINIA V. TRASK
ABSTRACT: Fuzzy set technology is applied to the ongoing research question of
how to automate the analysis of remote-viewing data. Fuzzy sets were invented to
describe, in a formal way, the subjectivity inherent in human reasoning. Applied
to remote-viewing analysis, the technique involves a quantitative encoding of target
and response material and provides a formal comparison. In this progress report,
the accuracy of a response is defined as the percent of the intended target material
that is described correctly. The reliability is defined as the percent of the response
that was correct. The assessment of the remote-viewing quality is defined as the
product of accuracy and reliability, called the figure of merit. The procedure is
applied to a test set of six remote-viewing trials. A comparison of the figures of
merit with the subjective assessments of 37 independent analysts shows good
agreement. The fuzzy set technology is also used to provide a quantitative defini-
tion of target orthogonality.
Human analysts are commonly used to evaluate free-response
data. Although there are many variations, the basic idea is that an
analyst, who is blind to the actual result, is presented with a re-
sponse and a number of target possibilities, one of which is the in-
tended target. The analyst's task is to decide what is the best re-
sponse/target match, and frequently includes rank-ordering the
targets from best to worst correspondence with the response. It is
beyond the scope of this report to provide a critical review of the
extensive literature on this topic.
One aspect, however, of this type of evaluation is that analysts
are required to make global judgments about the overall match be-
tween a complex target (e.g., a photograph of a natural scene) and
an equally complex response (e.g., written words and drawings). In
a recent book, Dawes (1988) has discussed various decision algo-
rithms in general and the difficulty with global techniques, such as
those used in rank-order evaluation, in particular.' According to
Dawes, the research results suggest that global decisions of this type
are not as good as those based on smaller subelements that are later
CPYRGHT
:
' We are indebted to Professor D. Bem, Cornell University, for directing us to
this valuable source of information.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
.
1

194
The Journal of Parapsychology
Advances in Remote-Viewing Analysis
195
combined. (See Dawes, 1988, chap. 10, for references to the re-
search.) Humans appear to be capable of deciding what the appro-
priate variables should be in complex decision processes, but they
have proved to be unreliable at combining these variables to arrive
atoi single decision. Linear algorithms are consistently better at this
lager task. Therefore, it seems prudent to develop evaluation tech-
nQues that are less sensitive to global decision processes and rely on
combinations of more restrictive decisions.
Honorton (1975) has pointed out an additional difficulty inher-
ero in a global rank-order approach. Asking an analyst to rank-
order a small set of target possibilities converts the free-response
experiment into a forced-choice one, at least on the part of the an-
al@st. It is obvious that in doing so, much quantitative information
istlost. For example, a near perfect correspondence between re-
Ponse and target will receive only as much "credit" as one that just
boely allowed an analyst to discriminate among the possibilities.
ir
If multiple analysts are used, addition problems arise concerning
beranalyst reliability. If an individual analyst judges a number of
r
sponses in a series, within-analyst consistency becomes an individ-
problem.
To address these difficulties, various computer-automated pro-
Aures have been suggested in an attempt to reduce the inter-
amalyst reliability while increasing within-analyst consistency. For ex-
ples, see Honorton (1975), Humphrey, May, Trask, and
Tuomson (1986), Humphrey, May, and Utts (1988), Jahn, Dunne,
ad Jahn (1980), May (1983), May, Humphrey, and Mathews
(1985), and Targ, Puthoff, and May (1977).
RDP9690789R003190120001-
In this paper we present the current status of an ongoing re-
search topic. We are not yet ready to propose that the techniques
described here be used for free-response analysis; however, we hope
Pinspire the community to develop a proper set of subvariables so
that the problems inherent in global decision processes can be
avoided.
Finally, we present a successful application of the mathematical
techniques for quantifying target orthogonality for a complex target
Bol.
Background
Substantial progress has been made in methods for evaluating
remote-viewing experiments since the publication of the initial re-
mote-viewing (RV) effort at SRI International (Puthoff & Targ,
1976). This paper outlines some of the progress and presents the
details for one particular method."
Two basic questions are inherent in the analysis of any remote-
viewing data, namely, how is the target defined, and how is the re-)
sponse defined.
In a typical outbound RV experiment, definitions of target and
response are particularly difficult to achieve. The protocol for such
an experiment dictates that an experimenter travel to some ran-
domly chosen location at a prearranged time; a viewer's task is to
describe that location. One method of trying to assess the quality of
the RV descriptions in a series of trials is to require that an analyst 7
visit each of the sites and attempt to match responses to them. While
standing at a site, the analyst has to determine not only the bounds
of the site, but also the site details that are to be included in the
analysis. For example, if the target location was the Golden Gate
Bridge, the analyst would have to determine whether the buildings
of downtown San Francisco, which are clearly and prominently vis-
ible from the bridge, were to be considered part of the target. The
RV response to the Golden Gate Bridge target could be equally
troublesome, because responses of this sort are typically 15 pages of
dream-like free associations. A reasonable description of the bridge
might be contained in the response; it might be obfuscated, how-
ever, by a large amount of unrelated material. How is an analyst to2
approach this problem of response definition?
The first attempt at SRI at quantitatively defining an RV re-
sponse involved reducing the raw transcript to a series of declarative
statements called concepts (Targ et al., 1977). Initially, it was de-
cided that a coherent concept should not be reduced to its compo-
nent parts. For example, a small red VW car would be considered a
single concept rather than four separate concepts, small, red, VW
and car. Once a transcript had been "conceptualized," the list of
concepts constituted, by definition, the RV response. The analyst
rated the concept lists against the sites. Although the response was
well defined by this method, no attempt was made to define the tar-
get site.
In 1982, a procedure was developed to define both the target
and response material (May, 1983). It became evident that before a
site can be qualified, the overall remote-viewing goal must be clearly
defined. If the goal is simply to demonstrate the existence of the
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
" Although the term remote viewing is used throughout this paper, the analysis
techniques can easily be applied to any free-response data.
.

nuvances in nemore-viewing Analysis
197
V phenomenon, then anything that is perceived at the site is im-
ortant. But if the goal is to gain specific information about the RV
process, then possibly specific items at the site are important
'hereas others remain insignificant.
In 1984, work began on a computerized evaluation procedure
May gs al., 1985), which underwent significant expansion and re-
nemEt during 1986 (Humphrey et al., 1986). The mathematical
ormamm underlying this procedure is known as the "figure of
hierit"<FM) analysis. This method is predicated on descriptor list
technology, which represented a significant improvement over ear-
er "conceptual analysis" techniques, both in terms of "objectifying"
he andysis of RV data and in increasing the speed and efficiency
ith which evaluation can be accomplished. Humphrey's technique,
hich Was based on the pioneering work of Honorton (1975) and
s expansion by Jahn, Dunne, and Jahn (1980), was to encode tar-
et ang response material in accordance with the presence or ab-
ence of specific elements.
It IScame increasingly evident, however, that this particular ap-
lication of descriptor lists was inadequate in providing discrimina-
ors th& were "fine" enough to describe a complex target accurately,
nd urgble to exploit fully the more subtle or abstract information
bonten?of the RV response. To decrease the granularity of the RV
valuation system, therefore, a new technology would have to allow
he analyst a gradation of judgment about target and response fea-
ires rather than the hard-edged (and rather imprecise) all-or-noth-
ng binary determinations. Requiring an analyst to restrict subjective
idgment to single elements rather than to complete responses is
onsistgot with the research reported by Dawes (1988).
A pseliminary survey of various disciplines and their evaluation
hetho& (spanning such diverse fields as artificial intelligence, lin-
uistic@and environmental psychology) revealed a branch of math-
maticsuknown as "fuzzy set theory."
uzzy
Concepts
Fuzzy set theory was chosen as the focal point of the RV analyt-
al tedyniques because it provides a mathematical framework for
odeling situations that are inherently imprecise. Because it is such
h important component in the analysis, a brief tutorial will be pre-
ented to highlight its major concepts.
' We wish to thank S. James P. Spotiswoode and D. Graff, CE, for directing us
the fuzzy set literature and for many helpful discussions.
TEST CITY
Membership Value
1.0
0.8
0.5 0.3 0.2
0.1
0.01
0.0
1
2
3
4
5
6 7 8 9 10
..
15
30
Population in 100 Thousands
Figure 1. The fuzzy set "kind-of-small" cities.
In traditional set theory (i.e., crisp sets), an element either is or
is not a member of a set. For example, the crisp set of cities with
population equal to or greater than 1,000,000 includes New York
City, but not San Francisco. This set would also not include a city
with a population of 999,999. The problem is obvious. There is no
real difference between cities with populations of 1,000,000 and
999,999, yet one is in the set and the other is not. Humans do not
reason this way; therefore, something other than crisp sets is re-
quired to capture the subjectivity inherent in RV analysis.
Fuzzy set theory introduces the concept of degree of membership.
Herein lies the essence of its applicability to the modeling of impre-
cise concepts. For example, if we consider the size of a city, we
might define certain fuzzy sets, such as very small cities or kind-of-small
cities. Using kind-of-small cities as a fuzzy set example, we might sub-
jectively assert that a city with a population of 100,000 is definitely
such a city, but a city with a population of 400,000 is only a little bit
like a kind-of-small city. As depicted in Figure 1, fuzzy set theory al-
lows us to assign a membership value between 0. and 1 that repre-
sents our best subjective estimate as to how much each of the pos-
sible city populations embodies the concept kind-of-small. In this
example, a population of 700,000 assigned a membership value of
0.3.
Clearly, a different set of membership values would be assigned
to the populations for the fuzzy sets very small cities, medium cities,
large cities, and so forth; a population of 100,000 might receive a
value of 0.2 for very small cities, but a value for 1.0 for kind-of-
small cities, depending on context, consensus, and the particular
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

198
The Journal of Parapsychology
Advances in Remote-Viewing Analysis
199
application. These membership values can be obtained through con-
sensus opinion, a mathematical formula, or by several other means.
Crisp sets are special cases of fuzzy sets, in which all membership
values are either zero or one. By using membership values, we are
abu to provide manipulatable numerical values for imprecise natu-
raglanguage expressions; in addition, we are no longer forced into
moking inaccurate binary decisions such as, "Is the city of San Fran-
ciso large-yes or no?"
In this example, the crisp set of all cities defines the universal set
ofglements (USE). The crisp set of cities with populations of one
miljon or more is a subset of USE. The fuzzy sets very small, kind-
of amall, medium, and large cities are fuzzy subsets of USE.
d For Release 2000/08/08
Umversal Set of Elements
Since targets and the responses will be defined as fuzzy sets, we
mast specify a USE. The universal set of elements can be quite gen-
a and include all aspects of a given target pool, or it can be tai-
loged to a specific experiment to test a given concept (e.g., include
onby geometric shapes). Since the method of fuzzy set analysis crit-
ically depends on the choice of USE, we provide one example that
we derived from a target pool used in earlier experiments. What
follows is only an example of how one might construct a USE. The
we use is not generally applicable to other target pools or other
experiments.
We constructed our USE by including a list of features present
in photographs from the National Geographic magazine with ele-
ments obtained from the RV responses in earlier experiments. This
USE is presented in Appendix A as the actual coding forms. For
the target features, we focused on direct visual elements. (In the
general case, other perceptual dimensions can be considered.) In
the case of the RV response-derived elements, an effort was made
to preserve the vocabulary used by the viewers. Some of the ele-
ments, therefore, are either response-dependent or target-depen-
dept or both, whereas others, particularly at the more abstract lev-
elo appear to be more universal across possible USEs.
This universal set of elements is structured in levels, ranging
from the relatively abstract, information poor (such as vertical lines),
to the relatively complex, information rich (such as churches). The
current system is structured into seven primary and three secondary
levels of elements; the main intent of this structure is to serve as a
heuristic device for guiding the analyst into making judicious con-
crete element assignments based on rather abstract commentary.
The use of levels is advantageous in that each element level can be
weighted separately and used or not, as the case may be. This ena-
. bles various combinations of levels to be deployed to identify the
optimal mix of concrete versus abstract elements. Of course, any
such weighting scheme must be determined in advance of any ex-
periment.
The determination as to which elements belonged on which leveld
was made after consideration of two primary factors: (1) the appar-
ent ability of the viewers to be able to resolve certain features, cou-
pled with (2) the amount of pure information thought to be con -;
tained in any given element. Some of these "factor one'
determinations were based on the combined anecdotal experience
of analysts and monitors in the course of either analyzing or con-
ducting numerous RV experiments; some were determined empir-
ically from post hoc analyses of viewers' abilities to perceive variousa
elements in previous experiments.
The "factor two" determinations were made primarily by arrang-
ing the elements such that an element at any given level representsd
the sum of its constituent elements at lower levels. For example, aj
port element (Level 7) could be considered to include canal (Level 6)
and partially bounded expanse of water (Level 5). The world is not af
very crisp place and not all its elements are amenable to hierarchical
structuring. Certain violations of the "factor two" rule appear,
therefore, throughout the USE example. It should be noted, how-
ever, that some of the more glaring violations were largely driven
by the "factor one" determinations (i.e., the viewers' abilities to dis-
cern certain elements) enumerated above.
To emphasize once again, it is very important to realize that this
universal set of elements was constructed to match our particular
special targets, viewers, and requirements. They are shown here to
illustrate the procedure. Any particular application of fuzzy set tech-
nology to the analysis of free-response material requires an a priori
construction of an individualized, and improved, USE specific to the
target pool and the goals of the experiment.
Target Fuzzy Sets
Each target is defined as a fuzzy set constructed by assigning a
membership value to each of the elements in the USE (see Appen-
dix A). In general, membership values can vary continuously on the
interval [0,1]. In this application they represent human judgment
CPYRGHT
Approved For Release 2000/08/08 : CIA RDP96-00789R003100120001-4

and, thus, were constrained to vary in steps of 0.1. In addition, they
must represent the perceptual dimension used to construct the USE.
In our example, membership values were assigned to each element
for each of the targets, according to a consensus (on an clement-by-
element basis) reached by three analysts. This approach was used to
mitigate the potential influence of any single coder's biases and idio-
syncukies. A numerical assignment, p (0 ≤ p < 1, in steps of 0.1),
was roade for each element in response to the following question:
How @isually important is this element to this photograph?
Eroded by this method, the fuzzy sets served as a formal defi-
nitiorPof the targets for the analysis. It should be noted that our
USE Defined targets in terms of visual importance.1 If other dimen-
sions are of interest (e.g., conceptual, functional, allegorical), the
JSE Would have to be revised to incorporate them.
Ingin actual experimental series, it is critical that the target fuzzy
ets be defined by analysts before the series begins. Because of the
potendal information leakage owing to bias on the part of the ana-
yst, it's an obvious mistake to attempt to define the target fuzzy set
on a Orget-by-target basis in real time or post hoc.
Respond Fuzzy Sets
To.define RV response fuzzy sets, membership values p are as-
igned for each element in the USE by asking: To what degree am
(the)analyst) convinced that this element is represented in this re-
ponsao For example, if a response explicitly states "water," then the
membership value for the water-element should be 1. If, however,
he response is a rough sketch of what might be waves, then the
membership value for the water-element might be only 0.3, depend-
ng or the specificity of the drawing. This definition of membership
value & quite general and can be used in most applications.
In Our example, responses were coded according to this defini-
ion (enit still using the USE in Appendix A). The assigned p's for
he taggets and responses were one-digit fuzzy numbers on the in-
erval -¿ 0,1] (e.g., 0.1, 0.2, 0.3, etc.). In some rare cases, two-digit
ssignments (e.g., 0.05, 0.15, 0.25, 0.35, etc.) were made; any finer
ssignagents, however, were deemed to be meaningless. Thus, the
espo Be was defined as its fuzzy subset of the USE.
' Implied visual importance was ignored. For example, in a photograph of the
Grand Canyon that did not show the Colorado River, water, river, and so on would
e scored as zero. By definition the target was only what was visible in the photo-
raph.
In an actual experimental series, each response fuzzy set is cre-
ated by analysts who are blind to the intended target.
Fuzzy Set Definition of Figure of Merit
Once the fuzzy sets that define the target and the response
ive
been specified, the comparison between them to provide a figure of
merit (FM) is straightforward. In previous work (Humphrey et al.,
1986), we have defined accuracy as the percent of the target matovial
that was described correctly by a response. Likewise, we have de-
fined reliability (of the viewer) as the percent of the response Bat
was correct. The FM is the product of the two; to obtain a high EM,
a response must be a comprehensive description of the target and
be devoid of inaccuracies. The mathematical definitions for accuracy
and reliability for the jth target/response pair are as follows. met
1}(R;) and ((T;) be the membership values for the kth element'in
USE for the ith response and the jth target, respectively. Then ce
accuracy and reliability for the ith response applied to the jth target
are given by:
EW.min{p.(R.),p.(T))}
accuracyij = @j =
*
Ew.min{p.(R.), p.(T;)}
reliability;j = Tij =
where the sum over k is called the sigma count in fuzzy set termingl-
ogy, and is defined as the sum of the membership values. We hoe
allowed for the possibility of weighting the membership values woch
weights W, in order to examine various level/element contributicas
to the FM. The index, k, ranges over the entire USE.
For the above calculation to be meaningful, the p's for the ter-
gets must be similar in meaning to the p's for the responses. As we
noted above, in our definition of the membership values, this is It
the case. The target u's represent the visual importance of the
ment relative to the scene, and the response p's represent the da
gree to which an analyst is convinced that the element is repr-
sented in the response regardless of its relevance to that response
With advanced viewers it might be possible to change the defi-
nition of the response u's to match the definition of the target u's.
In that case, the viewer must not only recognize that an element is
CPYRGHT
Approved For Release 2009/08/08 : CIA-RDP96-0G789R00319021 20004

202
The Journal of Parapsychology
Advances in Remote-Viewing Analysis
203
present in the target, but must also provide information as to how
visually important it is. This ability is currently beyond the skill of
most novice viewers. Alternatively, we have opted to modify the tar-
get n definition by using the fuzzy set technique of a-cuts. In our
examrole, an a-cut is a way to set a threshold for visual importance.
All target elements possessing that threshold value or higher are
considered to be full members of the target set. In fuzzy set par-
lanceBan a-cut converts a fuzzy set to a crisp one. The result is that
the target set is now devoid of detailed visual information: a poten-
ial (@get element is either present or absent in the target set, re-
gardlegs of its actual visual importance. Even with this conceptual
change in the target definition, the FM formalism described above
remagis applicable, because a crisp set can be considered as a fuzzy
set with all membership values equal to 0 or 1. It is important to
recognize that the a-cut is only applied to the target set; the re-
spon
set remains fuzzy.
Assessment of Quality of the Remote Viewing
If is difficult to arrive at a general assessment of how well a given
response matches a specified target. The ideal situation is to obtain
some>absolute measure of goodness of match. Although the FM is
an approximation to this measure, it is impossible to assess the like-
lihode of a particular FM value because it requires knowledge of the
viewes's specific response bias for the session. It is possible to deter-
mine general response biases (May et al., 1985), but that knowledge
is ong useful on the average. For example, a viewer may love rock
climbing and may spend most of his free time involved in that ac-
tivity@Thus, the general response bias would probably entail aspects
of mountains, rocks, ropes, and so forth. Suppose, however, that the
view, spent the evening previous to a given RV session on a ro-
mantis moonlight sail on San Francisco Bay. For this specific RV
sessim, the response bias might include romantic images of the
moonlit water, lights of the city, and bridges.
0789R003.106120001
ge current solution to the problem is to provide a relative as-
sessiRent of FM likelihood. A relative assessment addresses the fol-
lowing question: "How good is the response matched against its in-
tended target, when compared to all possible targets that could have
been chosen for the session?" This is not ideal, since the answer de-
pends on the nature of the remaining targets in the pool. An ex-
ample of the worst-case scenario illustrates the problem. Suppose
that the target pool consisted of 100 photographs of waterfalls, and
the viewer gave a near-perfect description of a waterfall. (We as-
sume that this description is not fortuitous.) An absolute assessment
of the resulting FM should be good, whereas a relative assessment
will be low. The worst-case scenario can be avoided, to a large dg
gree, by carefully selecting the target pool. (See the later section
Quantitative Definition of Target Orthogonality.")
To provide a relative assessment of the likelihood of a given FM
we define the score for one session to be the number of targets, ny
out of a total, N, that have an FM equal to or higher than the FIT
achieved by the correct match." The answer to the question: "Give?
this response, what is the probability of selecting a target that would
match it as well as or better than the target selected?" is n/N.
Consecutive RV responses by the same viewer are not statisticallo
independent, nor can the responses be considered to be random in
any sense. The statistically independent random element in the seg
sion is the target. Since targets are selected with replacement, undes
the null hypothesis of no psi, the collection of scores derived over 9
series of m trials constitutes a set of independent random variable
each with a discrete uniform distribution. Under the null hypothe-
sis, the mean chance expectation for the score in each session
given by (N + 1)/2 and the variance is given by (N2 - 1)/12. If
is the sum of scores from a series of remote viewings, then the prot
ability of K, under the null hypothesis, can be obtained from the
exact distribution for the sum of ranks given by Solfvin, Kelly, and
Burdick (1978):
K
N
p(K or less) = Σ Σ (-1)"(m) (a - bN - 1)
m - 1
a == 6=0
If m is large, then the sum-of-ranks distribution is approximately
normal and K/m has a mean of (N + 1)/2 and a variance of
(N2 - 1)/12m. Thus, a z score can be computed from:
0.5(N + 1) - :
K
m .
z(K or less) =
N2 - 1
12m
' N must be the size of the target pool from which each target was randomly
selected, and for this theoretical discussion, we assume no ties.
CPYRGHT
Approved For Release 2000/08/08 ; GIA-RDP96-00789R003100120001-4

nuowners in nemoir-viewing analysis
205
Ground Truth
To determine whether the new analytical approach was effective,
a standard had to be developed against which.it could be measured.
It was determined that this standard-known as "ground truth"-
should consist of a "real-world" normalized consensus about the de-
gree &f correspondence between RV responses and their intended
targeg.
TOachieve this objective, we presented analysts (chosen from the
general SRI staff) with the same test case of six remote-viewing re-
spons@s and their associated targets. The test case was the data from
a single viewer (177) taken from an experimental series in a 1986
photomultiplier tube experiment (Hubbard, May, & Frivold, 1987).
The @sponses (i.e., two to five pages of rudimentary drawings with
some "associated descriptive words) were fairly typical of novice
viewer,output and represented a broad range of response quality.
The @@gets consisted of six photographs of outdoor scenes selected
from @ National Geographic magazine target pool of 200. Thus, this
data @ was ideally suited for an analysis testbed. Appendix B con-
ains the "best" and "worst" trials (Sessions 9005 and 9004, respec-
ivelyPfrom this series in the form of their responses, their intended
targets and their fuzzy set encodings (see the next section).
Each analyst was asked individually for his subjective judgment
about the degree of correspondence between the remote-viewing re-
sponse and their respective intended targets. The "degree of cor-
responulence" was purposely undefined; the analysts had to formu-
late their own criteria. The only information provided was that
responses typically begin with small bits of information and even-
tuallySulminate in a composite drawing at the end. Appendix C
contages the coding form that was used to obtain "ground truth."
Each analyst was instructed to examine all of the responses and
heir Btended targets. Then, on a session-by-session basis, he was
asked@?(1) to assess the degree of correspondence between the
remoteviewing response and its intended target, and (2) to register
this correspondence assessment by making a vertical hash mark
across 10-cm scale ranging from "none" to "complete."
TOperform the ground truth analysis, distance measurements
were taken from the left end point of each scale to the vertical slash
mark For each assessment. Let the distance obtained for the kth ses-
sion from the jth analyst be given by x ;... To account for analysts'
biases, the x,, were normalized by a z transformation,
3.4 = * 14 - Hj,
where p, and o, are the mean and standard deviation of the jth an
alyst's distance scores, x) ... The effect of this transformation is toc
convert an analyst's absolute subjective opinion to a relative one. For
the jth analyst, the largest zj, indicates that the degree of corre-
spondence for response/target k is higher than any other pair in the
series. It does not indicate overall quality. This type of transforma-
tion was necessary since we wished to combine the assessments from
a number of different analysts.
To combine the assessments across analysts, we computed the
mean z score for each response/target pair, k, as:
3) = N. E30
Na j=1
where Na is the number of analysts. The number of analysts was
determined by the data. For the best response/target pair (i.e., ses-
sion 9005, k = 5) we computed the percent change of z, for every
additional analyst. When the addition of two new analysts produced
consecutive changes of less than 2%, the process was considered
complete. For this data set, 37 analysts were required before this
condition was met. Figure 2 shows the normalized mean for each
target/response pair, and represents a relative assessment of remote-
viewing quality. These means constitute the basis for the ground
truth against which the fuzzy set technique was measured. We re-
cognize that this definition of ground truth is based on global deci-
sions and may not be most optimal (Dawes, 1988).
Results of the Fuzzy Set Analysis
To effect a meaningful comparison between ground truth and
the figure of merit analysis, we also analyzed the same RV series
that served as the ground truth set by the fuzzy set figure of merit
method. The fuzzy set membership values (u's) for the six targets
and six responses were consensus coded by five analysts ranging
from expert. to novice. A typical spread of u assignments was ± 0.1
with an occasional outlier. Some of the elements were vigorously de-
bated until a consensus was reached. Accuracies, reliabilities, and
CPYRGHT
Approved For Release 2000/08/08 . CIA-RDP96-007891
120001-4
:

200
The Journal of Parapsychology
Advances in Remote-Viewing Analysis
201
TABLE 1
FUZZY SET QUANTITIES FOR "GROUND TRUTH" SERIES
Session
Accuracy
Reliability
Figure of
merit
Rank
Fractional
rank
9001
.317
.484
.153
80
.403
9002
.273
.477
.130
103
.515
9003
.358
.571
.205
31
.155
9004
.212
.379
.080
142
.713
9005
.573
.594
.340
3
.015
9006
.298
.555
.165
13
.068
-1.5
9001
WX0X12
9003
9004
9005
Session Number
Figue 2. Normalized mean for each target/response pair.
figures of merit were calculated for each targetresponse pair (Table
1). Igoshould be noted that the encoding was a post hoc exercise, but
becase the assignment for each element in the USE had to be de-
fended before a consensus was reached, the FMs shown in Table 1
con@Mute reasonable estimates of their "blind" equivalents. Appen-
dix >> shows the target and response elements that were scored from
the Whiversal set (see Appendix A) for Sessions 9004 and 9005. As
an
Kample of the fuzzy calculation, Appendix B also shows the re-
+P96-00789R003100120001-4
0.500
Fuzzy Sets
Ground Truth
0.375
Figure of Merit
0.250
0.125
0
9001
9002
9003
9004
9005
9006
Session Number
Figure 3. Comparison with ground truth.
CPYRGHT
-Approved For Release 2009/08/08 :CARDP 96
sults of the target a-cut, the fuzzy intersection, and the accuracy
reliability, and figure of merit for Session 9005. Table 1 also shows
the absolute and relative ranks from a target pool of 200. To dete
mine the absolute rank for each session, we calculated figures
merit for all 200 targets in the pool and placed them in numeric
order from the largest to the smallest. The absolute rank is just the
position (from the top) of the FM corresponding to the intende
target. Ties were resolved by choosing the next larger integer rard
number to the centroid of the ties. The fractional rank number can
be considered a p value for an individual session and is equal to t12
absolute rank/200. Using Equation 1, the overall p value for the
combined six trials is .052 (N = 200, K = 372, m = 6). Using
approximation (Equation 2), we compute z = 1.633, p < . 05, 1
demonstrate that for six trials, the approximation in reasonable. F
completeness, we compute the effect size (r = 0.67).
To compare the results of the fuzzy set analysis with those of the
ground truth, we linearly renormalized the ground truth figures too
be within the interval [0,1] and to possess the same maximum an&
minimum. As can be seen from Figure 3, the results from the fuzzo
set analysis system parallel those obtained by a consensus of the 37,
analysts each making a subjective assessment of the matches,
-00789R003100120001-4
These results imply that the combination of (1) the structure of
the USE (i.e., the linguistic hierarchical structure), (2) the fuzzy se
mathematics, and (3) a consensus approach to assessing the fuzzs
sets themselves provided a reasonable representation of the subject
tive scoring of the same data by a large number of individuals.
A Quantitative Definition of Target Orthogonality
It is often of interest to define how similar or dissimilar targets
are to each other. For example, free-response experiments like the
Approved For Release 2000/08/08
1.5
Normalized Mean
0.5
-0.5

ganzfeld often use target packets, with the unselected targets in a
packet serving as decoys for judging. Assigning potential targets to
packets would be easier with some measure of target orthogonality.
Target definition for the purposes of this mode of analysis is ex-
actly the same as the one described (i.e., a given target is defined by
its fuzzy subset of the USE, which has been coded to reflect the vis-
ual importance of each target element). The average number of ele-
ment of the total of 131, that was assigned a nonzero value for the
targes in our pool of 200 was approximately 37, indicating that the
fuzzyDset representation of the target pool is rich in visual infor-
mation. We used this information to determine the degree to which
the torget set contains visually similar targets.
Itis beyond the scope of this paper to describe the extensive
workpin the literature seeking to find algorithmic techniques that
mimi human assessments of visual similarity. One recent article de-
scribes techniques similar to the one we used (Zick, Carlstein, & Bu-
desc ?. 1987).
Ver begin by defining the similarity between target i and target j
to be normalized fuzzy set intersection between the two target sets:
2
*
" the index k ranges over the entire USE. We have allowed for
the possibility of weighting the membership values with weights W,
to egamine various level/element contributions to the target similar-
itiess
Pr N targets, there are N(N- 1)/2 unique values (19,900 for N
= 2:00) of Sy. The values i and j that correspond to the largest value
of Sp represent the two targets that "look" most similar. Suppose
anoger target m is chosen and S_, and S_, are computed. If both of
these values are larger than S ... (for all n not equal to i or j), then
target m is assessed to be most similar to the pair ij. The process of
grouping targets based on these similarities is called cluster analysis.
Using this process, 200 targets were grouped into 19 clusters,
suckthat the targets are similar within a cluster, and dissimilar be-
tween clusters. Table 2 provides an overview of the 19 clusters
found from the total analysis of the 200 targets. Some of the names
appear to be quite similar, but, in fact, these sets are visually quite
distinctive. Figure 4 shows the graphic output of a single cluster in
TABLE 2
NAMES OF THE 19 CLUSTERS
No.
Name
No.
Name
1
Flat towns
11
Cities with prominent geometries
2
Waterfalls
12
Snowy mountains
3
Mountain towns
13
Valleys with rivers
4
Cities with prominent structure
14
Meandering rivers
5
Cities on water
15
Alpine scenes
6
Desert/water interfaces
16
Outposts in snowy mountains
7
Deserts
17
Islands
8
Dry ruins
18
Verdant ruins
9
Towns on water
19
Agricultural scenes
10 Outposts on water
detail. A much more complex-and visually difficult to und
stand-graph is generated for the full cluster analysis and is no
included here; this smaller subset, therefore, has been chosen to
illustrative of the whole analysis. All targets in this particular samp
cluster are islands; the island in each photograph is visible in its ed
tirety. Except for one outlier (i.e., a hexagonal building covering
island), the islands fall into two main groups (i.e., with and without
Linear Geometries
(e.g., Runways)
1138
1128
1186
Many Structures
1179
(c.g., Town)
1177
Ruins
-> 1083
1185
Flat And Verdant
1139.
1140
Sand Bars
1088
1161
Mountains
1049
1038
Hexagonal Building Covering Island
-- >1003
0.0
0.2
0.4
0.6
0.8
1.0
1 - Si,j
Figure 4. Cluster analysis of island targets.
CPYRGHT
Approved For Release 2000/08/98 CIA-RDP96-00789R003100120001-4
Release-2000/08/08 : CIA-RDP96-
Sy =
Ew,min{H,(T),μ.(T;)}
-
whe
1198
Target 1xxx
Number
1
8

210
The Journal of Parapsychology
manmade elements). The natural islands include three similar
mountain islands, two sandbars, and two flat verdant islands.
Using cluster analysis in conjunction with fuzzy set analysis pro-
videsgor a quantitative definition of sets of targets that are similar
to each other within a cluster, but visually different across clusters.
Orthogonal clusters can be used to provide visual decoy targets for
traditonal rank-order judging.
Recommendations and Conclusions
To apply the analysis in its present form to a long RV series is
quitedabor intensive and, from the results shown in Figure 3, is
most dikely not justified since this fuzzy set technique approximates
human assessment. As we stated in the introduction, however, we
are providing only a progress report of ongoing research. Because,
of thedecision concepts described in Dawes (1988) and the obvious
benelas of an automated evaluation system, the effort to improve
whatWas described in this paper is certainly justified. The proce-
dureosan be used "as is" to improve and quantify target orthogo-
nality:
SDeral future research areas are suggested to improve the tech-
niques described in this paper. The use of both inter- and intra-level
weighing factors needs to be examined systematically. In the analy-
sis described above, all levels and elements were accorded equal
weigle. The ideal goal would be to determine the optimal weighted
mix of abstract versus concrete elements, as a means to achieving
the following objectives:
IRefinement of the cluster analysis for targets, in an effort to
Usimulate, as closely as possible, what is meant by "visual simi-
larities" between targets.
3789R003100120001-4.
2
Refinement of the analysis of responses, in an effort to
achieve even greater correlations between the fuzzy set figure
Sof merit analysis and various forms of ground truth.
A@other area that requires examination in some detail is the
USEgand the hierarchical nature of its structure. It is probable that
some elements are more appropriate than others; furthermore, they
might be more effectively structured in a semantic network as op-
posed to a true hierarchy. If a hierarchical structure is retained,
then some attention must be paid to the formulation of logical con-
sistency rules that govern element use. This would include numeri-
Advances in Remote-Viewing Analysis
211
cal relationships governing the membership values (p's) of higher-
order elements (e.g., port) vis-à-vis the combined value of their con-
stituent parts (e.g., city, river, boats, jetties, commercial).
One inadequacy of the system is that it atomizes conceptual!
"units." For example, if the response element is red box, information
is lost in separating red from box. Current research in fuzzy set the-
ory indicates that fuzzy aggregates of fuzzy elements-"fuzzy sets of
fuzzy sets"-are mathematically complex but possible. Some effort
should be made to determine whether this technology could be im!
plemented as a means to capturing the information content of theu
RV response with greater accuracy.
For the visual analysis, research into visual similarities between
pictures of natural scenes may serve as a potential refinement tool"
The aim here would be to enhance the visual orthogonality of ranky
order analysis decoy targets as much as possible. Experiments ing
normal perception of similarities would assist in determining
whether scenes are perceived as similar because of their low-level
geometries, concrete elements, or some combination of factors. The
ultimate aim would be to refine the target cluster analysis such that
it closely simulates ground truth representations of orthogonality.
.
APPENDIX A
CODING FORMS FOR THE UNIVERSAL SET OF ELEMENTS
The following coding forms illustrate the use of a universal set of elements
(USE) that matched our particular special targets, viewers, and requirements,
We constructed our USE by including a list of features present in photographs
from the National Geographic with elements obtained from the remote-
viewing responses in earlier experiments.
1
.
CPYRGHT
Approved For Release 2000/
10&. CIA
6-00789R00
120001-4
Ippraved For Releas
... . ... . .... .... .. ..
..
.........

C17
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CONCRETE DESCRIPTOR LEVELS I
Experiment:
That:
Resp./Targ:
Coder:
Viewer:
LEVEL
SINGLE STRUCTURES
SUDSTRUCTURES
font
,
carte
palace
10
church (other religious buildings, monastery)
mosque
.
pagoda
-
cotesum (stadium,
amphitheater, arena)
10
boats (berges)
·
bridge
·
[dem flock, spillway)]
12
(cars, trucks, trans;)
13 1
column
.
16
spiro (mineret tower)
tounten
"
arch
wel (eg , the Great Wain)
monument
-
&
roads
CONCRETE DESCRIPTOR LEVELS II
Experiment:
THal:
Resp./Targ:
Coder:
Viewer:
LEVEL
SETTLEMENT
ELEVATION
LAND/WATER
INTERFACE
NO
WATER OR
VEGETATION
VEGETATION
AMBIENCE!
FUNCTION
port (harbor)
{orchards)
recreational
religious
7 .
.
agricultural
commerole
34
ruine
(incomplets
buildings)
mesa
(platecu)
watertal
desert
L
forest
31 /
wilderTisse
37
glacier
41
Jungle
urban
-
30
canal (channel,
manmade
waterway)
swamp
(manıh)
(pastorı)
131
Autorical
43
eingie peak
unbounded large
expense of water
(ocean, 504)
8
vegetation
loun (village)
bumpe, humpe.
completely bounded
mounds)
expanse of water
(laks, pool, pond)
45
chy
44
mountains
partially bounded
5
cliff(s)
expense of water (bay)
[plein, delta]
Island
veley (clent,
river (stream, creek)
guy. treg.
depression)
costine
canyon
.
[crater .. bowl-
2
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Figure Al. Coding form.
Figure A2. Coding form.
agricultural
N
Industrial
[oasia)
mechanical
technical
hils (slopee.
=
[motorized vehicles

215
Approved For Release 2000/08/08 -CIA-BDP96-00789R003100120001-4
Experiment:
Trial:
Resp./Targ:
Coder:
Viewer:
QUALITIES
LEVEL
COLOR
OTHER
VISUAL
IMPLIED
TEXTURE
IMPLIED
TEMPERATURE
IMPLIED
MOVEMENT
AMBIENCE
yshow 71
shiny (reflective)
8
smooth
hot
Nowing
congested (cluttered.
dense, busy)
orange 72
[gold]
=
fuzzy
cold (snow, ice) 00
other
serene (peaceful.
red
(silver)
grany tsandy.
crumbly )
humid
imphed
movement 07
unhurned.
unfrenetc)
blue
74
[chrome]
rocky (ragged.
rugged. jagged.
8
[copper]
rucolad. rough)
8
purple
(pink)
obscured (fuzzy.
dim, smoky}
H
&
ordered (aligned)
2
black
70
old
white
weathered
foroded.
grey
incomplete)
ARCHETYPES
STRUCTURE
ELEVATION
INTERFACE
UNIQUENESS
AMBIENCE
building(s)
(structure(s))
nse (vertical nse
as well as slope!
100
light'dark areas
(bg swains)
104
single (or cantz!)
predominant leafure
108
manmade
(or uttered)
8
nat
IC
boundaries
odd (or surprising)
105
juxtaposition of
elements
107
natural
1C2
tand water
interface
103
interface (horizon)
ABSTRACT DESCRIPTOR LEVELS II
Experiment:
Trial:
.Resp./Targ:
Coder:
Viewer:
2-D & 3-D GEOMETRIES
LEVEL
RECTILINEAR
FORMS
CURVILINEAR
FORMS
MIXED
FORMS
IRREGULAR
FORMS
REPEAT
MOTIF
10
rectangle
(square,
box)
112
circle (oval.
sphere)
cyinder
irregular
forma
(HTegutN
features)
110
repeat motif
.
triangle
109
(trapezoid,
pyramid)
119
{hemisphere,
2
other polygonal
110
(> 4 sides: hexagon,
octagon. etc.)
cross-heich
(and)
1-D GEOMETRY
stepped
127
ac (curve)
130
meandenng curve
120
parallel lines
120
wave form (ripples)
121
vertical lines
120
spiral
122|
honzontal lines
1
123
diagonal lines
1124
V-shape
125
inverted V-shape
120
other angies
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Figure A3. Coding form.
Figure A4. Coding form.
The Journal of Parapsychology
214
CPYRGHT
Advances in Remote-Viewing Analysis
4
open (spacious.
vast, expansive)
brown 11
cloudy (foggy.
misty)
(beige)
&
disordered tjumbled
unalignedl
.
113
(torus)
cone
SOTNCircle
dome)
land sky
dry (and)
closed In
(claustrophobic)

nuvances in remote-Viewing Analysis
217
APPENDIX B
FUZZY SET ANALYSIS TESTBED
The following pages show the targets, responses, and analysis for two
remo@-viewing trials.
Approved For Release 2000/0
Figur 81. Target for Session 9004.
Task is to identify slide
in PMT holder
Target
tall-vertical feature
.
break
Figure B2. Page one of the response (Session 9004, Target 1094)
-
CPYRGHT
6-00789R003100120001-4
i
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4

218
The Journal of Parapsychology
Advances in Remote-Viewing Analysis
219
Target
Approved For Release 2000/08/08 : GIA-RDP96-00789R003100120001-4
Target
rising in background
+ boundary
1
A
0
structur
D
smooth- fine graininess
It- openings
flat
Freak
rectangular opening
1
vertical
feature
smooth
texture
-
flat
playa- like
Figut@ B3. Page two of the response (Session 9004, Target 1094).
FLAT
END of Session
Figure B4. Page three of the response (Session 9004, Target 1094).
CPYRGHT
Approved For Release 2000/08/08 : CTA-RDP96-00789R003100120001-4

TABLE BT
TARGET-RESPONSE 9004
Element
Name
Target
Response
20
Roads
0.30
0.00
23
Agricultural fields
0.05
0.00
32
Urban
0.00
0.50
33
Rural, pastoral
0.60
0.50
44
Town, village
0.00
0.50
45
City
0.00
0.40
46
Single peak
0.70
0.00
47
Hills, slopes, bumps, mounds
0.10
0.40
48
Mountains
0.00
0.60
49
Cliffs
0.00
0.10
60
Vegetation, trees
0.30
0.00
64
Blue
0.50
0.00
65
Green
0.30
0.00
69
White
0.10
0.00
70
Grey
0.20
0.00
76
Obscured, fuzzy, dim, smoky
0.20
0.00
77
Cloudy, foggy, misty
0.20
0.00
79
Weathered, eroded, incomplete
0.00
0.10
80
Smooth
0.00
1.00
81
Fuzzy
0.20
0.00
82
Grainy, sandy, crumbly
0.20
1.00
90
Other implied movement
0.20
0.00
91
Congested, cluttered, busy
0.10
0.30
92
Serene, peaceful, unhurried
0.40
0.00
93
Closed in, claustrophobic
0.00
0.10
94
Open, spacious, vast
0.60
0.00
95
Ordered, aligned
0.00
0.40
97
Buildings, structures
0.00
1.00
98
Rise, vertical rise, slope
0.60
1.00
99
Flat
0.30
1.00
100
Light/dark areas
0.10
0.00
101
Boundaries
0.30
1.00
103
Land/sky interface
0.50
0.00
104
Single predominant feature
0.60
0.00
105
Odd juxtaposition, surprising
0.30
0.00
106
Manmade, altered
0.20
0.80
107
Natural
0.70
0.20
108
Rectangle, square, box
0.00
1.00
109
Triangle, pyramid, trapezoid
0.60
0.00
115
Cone
0.60
0.00
117
Irregular forms
0.00
0.20
118
Repeat motif
0.10
0.60
119
Stepped
0.10
0.70
120
Parallel lines
0.10
0.00
121
Vertical lines
0.10
1.00
122
Horizontal lines
0.10
0.00
123
Diagonal lines
0.40
0.00
125
Inverted V-shape
0.70
0.00
126
Other angles
0.00
0.10
Figure B5. Target for Session 9005.
Task is to describe slide
Target in PS 345
Target
interface
curing
·
Land
water
Great
Figure B6. Page one of response (Session 9005, Target 1005).
CPYRGHT Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
L

222
The Journal of Parapsychology
Advances in Remote-Viewing Analysis
223
Tsu get
Approved For Release 2000/08/08 : GIA-RDP96-00789R003100120001-4.
strep restical
rise
1 mound-like
- curved
Break
Target
steep
-
very sharp
Jop 08
break
Figure B7. Page two of response (Session 9005, Target 1005).
Target
- sleep rete
squasfeatures
-ridge
- stepped
CPYRGHT
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
break
1
Figure B8. Page three of response (Session 9005, Target 1005).

0
4LJ
Target
open
ragged boundary
fine-grained
textwe
water
1
5,3)
upclose to
stepped feature
1
Figufg B9. Page four of response (Session 9005, Target 1005).
Target
openin distance
Land
- Flat
& steep
une
water boundary
Land
stepped
1 features
teep
drap
feels
stepped feature
curve
around
·
Figure B10. Page five of response (Session 9005, Target 1005).
satz
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
CPYRGHT
Approved For Release 2000/08/08 : CIA RDP96 00789R003100120001 4

TABLE B2
TARGET-RESPONSE 9005
Advances in Remote-Viewing Analysis
227
Approved For-Release 2006/08:08
Analysts' Instructions for Remote-Viewing Series 900X
Thank you for helping us perform a post hoc assessment of a seriescof
remote viewings. The targets were actually 35-mm slides that were attached
to a photomultiplier, a device to measure small amounts of light. We were
searching for possible physical correlates to remote viewing.
You will find in your packet 6 remote viewing responses labeled 900%
9006 respectively. Also shown is the target number of the intended photo-
graph. We have supplied the original, rather than the 35-mm slide.
We would like you to make a subjective judgment as to the degree f
correspondence between the remote viewing response and its associated to-
get. Familiarize yourself with the task by first looking at all the responses
and their intended targets. Then, on a session-by-session basis, rate your
assessments. You are completely free to define what is meant by "Degreecof
Correspondence." Indicate your judgment by marking one line. across the
appropriate continuous scale shown below. A vertical line near the "None"
end of the scale will indicate that you feel there is very little correspondedee
between that response-target pair. Likewise a vertical line near the "Como-
plete" end of the scale will indicate that you feel that there is a significant
degree of correspondence.
Many of the responses begin with a little information and build toward
a composite drawing at the end. Please assess the response in its entirety
best you can. Thank you again.
SESSION
DEGREE OF CORRESPONDENCE
TARGET
Nonc
Complete
9001
1034-
9002
1042
9003
1065
122
Horizontal lines
0.10
0.00
0
0.00
123
Diagonal lines
0.10
0.20
0
0.00
125
Inverted V-shape
0.00
0.20
0
0.00
127
Arc, curve
0.30
1.00
1
1.00
128
Wave form
0.00
0.10
0
0.00
Totals
21.20
22.00
12.60
14
Spire, minaret, tower
0.00
0.20
0
0.00
20
Roads
0.10
0.10
0
0.00
32
Urban
0.80
0.70
1
0.70
38
Canal, manmade waterway
0.00
0.10
0
0.00
44
'Town, village
0.00
0.30
0.00
45
City
0.90
0.70
0.70
46
Single peak
0.00
0.20
0
0.00
-17
Itills, slopes, bumps, mounds
0.00
0.10
0.00
54
Unbounded large expanse water
0.00
0.40
0
0.00
56
Partially bounded water
0.30
0.30
1
0.30
58
River, stream, creck
0.00
0 .- 10
0.00
59
Coastline
0.00
0.20
0)
0.00
60
Vegetation, trees
0.20
0.20
1
0.20
64
65
Green
0.20
0.00
1
0.00
Brown, beige
0.50
0.00
0.00
69
70
Grey
0.10
0.00
U
0.00
80
Smooth
0.10
0.00
0
0.00
81
Fuzzy
0.00
1.00
0
0.00
82
Grainy, sandy, crumbly
0.00
1.00
0
0.00
83
Rocky, ragged, rubbled, rough
0.00
1.00
0
0.00
91
Congested, cluttered, busy
0.10
1.00
0
0.00
95
Ordered, aligned
0.00
0.30
0
0.00
96
Disordered, jumbled, unaligned
0.30
0.00
1
0.00
97
Buiklings, structures
0.80
0.90
1
0.90
Rise, vertical rise, slope
0.00
1.00
1
1.00
100
Light/dark areas
0.10
0.00
0
0.00
101
Boundaries
0.20
1.00
1
1.00
102
Land/water interface
0.30
1.00
1
1.00
103
Land/sky interface
0.10
0.10
0
0.00
104
Single predominant feature
0.10
0.40
0
0.00
106
Manmade, altered
0.80
0.80
I
0.80
107
Natural
0.20
0.20
1
0.20
108
Rectangle, square, box
0.70
1.00
1
1.00
111
Cross-hatch, grid
0.30
0.00
1
0.00
112
Circle, oval, sphere
0.10
0.00
0
0.00
116
Sernicircle, dome, hemisphere
0.10
0.30
0
0.00
118
Repeat motif
0.40
0.80
0.80
119
Stepped
0.20
1.00
3
1.00
120
Parallel lines
0.30
0.30
1
0.30
121
Vertical lines
0.50
1.00
1
1.00
9004
1094
9005
1005
9006
1024
1
0.00
Blue
0.25
0.00
67
0.10
0.00
0
0.00
White
0.70
0.70
1
0.70
94
Open, spacious, vast
99
Flat
0.50
1.00
0
0.00
98
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4
Element Name
Target Response T. - oz
TOR
.
APPENDIX C
"GROUND TRUTH" INSTRUCTION AND CODING FORM
CPYRGHT
CIA-REP96-00789R003100120001-4
Accuracy = 0.573
Reliability = 0.594
Figure of merit = 0.340

REFERENCES
DAWES, R. M. (1988). Rational choice in an uncertain world. New York: Har-
court, Brace, Jovanovich.
HONORTON, C. (1975). Objective determination of information rate in psi
tasks with pictorial stimuli. Journal of the American Society for Psychical
Research, 69, 353-359.
HUBBARD, G. S., MAY, E. C., & FRIVOLD, T. J. (1987). Possible photon pro-
Oduction during a remote viewing task: A replication experiment. Final
Report, SRI Project 1291, SRI International, Menlo Park, California.
HUMPHREY, B. S., MAY, E. C., TRASK, V. V., & THOMSON, M. J. (1986).
TRemote viewing evaluation techniques. Final Report, SRI Project 1291,
SRI International, Menlo Park, California.
HUTOHIREY, B. S., MAY, E. G., & Uris, J. M. (1988). Fuzzy set technology
On the analysis of remote viewing. Proceedings of the 31st Annual Conven-
Bion of the Parapsychological Association (pp. 378-394).
JAHA, R. G., DUNNE, B. J., & JAHN, E. G. (1980). Analytical judging pro-
"cedure for remote perception experiments. Journal of Parapsychology,
84, 207-231.
MAP E. C. (1983). A remote viewing evaluation protocol. Final Report (re-
yised), SRI Project 4028, SRI International, Menlo Park, California.
MAGO E. C., HUMPHREY, B. S., & MATHEWS, C. (1985). A figure of merit
Analysis for free-response material. Proceedings of the 28th Annual Con-
. vention of the Parapsychological Association, (pp. 343-354).
PUTNJOFF, H. E., & TARG, R. (1976). A perceptual channel for information
Transfer over kilometer distances: Historical perspective and recent re-
-search. Proceedings of the IEEE, 64(3), 329-354.
SOGVIN, G. F., KELLY, E. F., & BURDICK, D. S. (1978). Some new methods
Tof analysis for preferential-ranking data. Journal of the American Society
For Psychical Research, 72(2), 93-110.
TALG, R., PUTHOFF, H. E., & MAY, E. C. (1977). State of the art in remote
Cviewing studies at SRI. 1977 Proceedings of the International Conference of
Cybernetics and Society (pp. 519-529).
ZICH R., CARLSTEIN, E., & BUDESCU, D. V. (1987). Measures of similarity
among fuzzy concepts: A comparative analysis. International Journal of
Approximate Reasoning, 1(2), 221-242.
SRPInternational
333 Ravenswood Av.
Medo Park, CA 94025
Division of Statistics
University of California, Davis
Davis, CA 95616
CPYRGHT
.
Approved For Release 2000/08/08 : CIA-RDP96-00789R003100120001-4