NSA-RDP96X00790R000100040007-7

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
-
.
JOURNAL
of the
Society for Psychical Research
VOLUME 43 No. 727 March 1966
AN ATTEMPTED VALIDATION OF THE
'RÝZL TECHNIQUE' FOR TRAINING
ESP SUBJECTS
by JOHN BELOFF AND IAN MANDLEBERG
I. INTRODUCTION
NOTHING could do more at the present time to advance the cause
of parapsychology as a science than some method of developing a
dependable level of ESP ability in an ordinary volunteer subject.
When the senior author read in this Journal (Rýzl 1962) about the
work going on in Prague towards this end, based on the use of
an hypnotic training technique, he was anxious to try and follow
it up. Unfortunately, the article in question dealt almost exclusi-
vely with the attainments of one remarkable subject Miss J.K.
while Ryzl's many subsequent publications are all devoted to the
card-guessing exploits of another star subject: Mr P.S. Con-
sequently, although by now over 500 volunteers have undergone
some amount of training at Ryzl's hands, over the past 15 years or
so, the published literature is still conspicuously lacking in any
· systematic reports on attempts to validate the 'Rýzl technique'.1
Our first aim in undertaking the investigation we are about to
report was to make good this deficiency.
In the Summer of 1964 J.B., who had kept up a regular corres-
pondence with Dr Ryzl, went to Prague for a month to try and
learn at first hand about Dr Ryzl's methods. Unfortunately it
1 We are not overlooking the excellent 'Cambridge ESP-IIypnosis
Experiments 1958-64' (see this Journal, 43, June 1965, 77-92) but we
presume that the Cambridge group would agree that their investigation,
which was begun before the publication of Ryzl's findings, is somewhat
tangential to the specific claims of the 'Ryzl technique' as here described.
229
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
andy, and no trainee subjects
wue wien available and, in fact, much of the time was taken up
with a joint experiment with the ever-ready P.S. as subject (see
J. of Parapsychol. 29, 1965, 1.12). But P.S., though he has, by
dint of his extraordinarily long career as a high-scorer, become
the most celebrated of Ryzl's subjects, is by no means a typical
product of the Ryzl stable. Very early on in his carcer he began
to specialize in card-guessing to the exclusion of any other task
and soon, moreover, dispensed entirely with hypnosis. Yet; the
'Rýzl technique', as we are here using this expression, is con-
cerned primarily with clairvoyance of what might be called the
'eyeless sight' variety as distinguished from the familiar card-
guessing kind, that is to say the subject, when successful, is sup-
posed to experience a quasi visual percept of the target object, he
does not just make a verbal response that happens to be correct.
The Rýzl technique appears to involve two essential stages:
(i) Repeated hypnotic sessions during which the subject
becomes practised at experiencing to order vivid visual
hallucinations.1
(ii) Further hypnotic sessions during which the subject is
encouraged to try ascertaining selected target objects which
should now give rise to veridical hallucinations. Special
emphasis at this stage is laid on combining an attitude of
intense concentration with an emptying the mind of all
distracting thoughts and images to make it as receptive as
possible to incoming impressions.
It is important to note that it is not hypnosis as such that is the
crux of the 'Rýzl technique' but rather hypnosis as an avenue to
the attainment of veridical hallucination. In our investigation we
have accordingly deliberately avoided the use of the customary
ESP tasks, involving lengthy sequences of rote-guessing, in favour
of tasks involving only one trial at a sitting, in the course of which
the subject is urged to make an all out effort to visualize the target
in question.
During the Autumn Term '64 about 50 volunteer students
were each given at least one hypnotic session and from this
number 20 were eventually selected to take part in the critical
validation experiment reported below. Two criteria only were
considered in making the selection: (a) whether the subject was
1 We are using the term 'hallucination' in keeping with Ryzl's own usage
. but to avoid misunderstanding the reader should note that we are refer-
ring to the type of hallucination which occurs with the eyes closed as in
dreams. Some may prefer the expression 'vivid vimml imagery'.
230
sufficiently susceptible to hypnosis and (b) whether the subject
was keen enough to agree to come every week for the remainder
of the academic year if necessary, for a one-hour session. It may
be of interest to note that of our original sample of 50 not more
1
than 4 had to be rejected as being incapable of attaining a
trance state within a reasonable period of time. Before the students
went away for the Christmas Vacation each of the selected 20
subjects had undergone 5 preliminary training sessions involving
a wide variety of hypnotic exercises and some tentative assorted
ESP tests.
For the critical validation experiment we planned for the Spring
Term, we decided to use both some 'open' tests, i.e. tests in which
the target might be anything at all within a certain category, as
far as the subject is informed, as well as 'selective answer' tests
where the subject knows what the alternatives are. Although the
latter have become much more common in parapsychological
experimentation, since they lend themselves more readily to
precise quantitative evaluation, Rýzl was of the opinion that his
subjects performed more successfully on tests of the former sort,
presumably because they discourage mere guessing habits in
favour of careful attention to the subject's own impressions.
Ordinarily, a validation experiment requires a control group
who are matched against the experimental group but who do not
receive the special treatment accorded the latter. In the present
instance, however, it would obviously be a prodigious waste of
time to run such a control group until we were in a position to
demonstrate that the experimental group (a) performed at a con-
sistently above chance level and (b) improved as a function of
practice. Accordingly, all 20 subjects were assigned to the one
experimental group. Before the experiment commenced a full
description of our proposals was sent to Dr Rýzl for comment and
he, we are glad to say, graciously gave it his blessing. The
experiment now to be described occupied all ro weeks of the Spring
Term '65.
II. THE EXPERIMENT
I. Subjects
Twenty student volunteers, 10 male, 10 female, all of whom had
previously undergone five preliminary hypnotic training sessions.
2. Experimenters
Fourteen of the above subjects were allocated to I.M. who
became their regular hypnotist-experimenter, the remaining six
231
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
case the non-hypnotist took over the duties of note-taker.
3. Procedure
Each subject came once a week for 10 weeks for a one-hour
session. When the subject was comfortably settled in his easy
chair our first job was to get him securely hypnotised. Standard
methods of induction were used, ordinarily the subject would be
told to fixate a certain spot on the centre lamp-fixing while the
hypnotist would count slowly to ten interspersing his counting
with suggestions of drowsiness. After a while we found that
most subjects could be put into a trance if the hypnotist merely
snapped his fingers while they were gazing up at the light. Once
hypnosis had supervened much further effort was expended in
trying to deepen and stregthen the trance state. Sometimes the
subject would be asked to estimate on a six-point scale the level
of trance he had attained.
When induction was complete, and this might casily occupy
the first 10 minutes of the session, the subject would be blind-
folded. Although Rýzl preferred to rely on screening we had
decided that blindfolding would permit greater flexibility. This,
as we soon discovered, demanded a certain amount of improvisa-
tion. Thus, we began by using a pair of underwater goggles with
the eye-pieces blacked out and the rubber surrounds stuck down
to the subject's face at the critical region of check and nose with
strips of 1-inch oxide plaster. Then, during the third session, for
the sake of greater comfort for the subject, we introduced a blind-
fold made from a double thickness of felt fastened with elastic
and stuck to the subject's face as before. Finally, during the
seventh session, we substituted for this a specially designed face-
mask which we believed offered the greatest security with the least
discomfort. It was made of wire-mesh covered over in a double
thickness of felt and came down under the subject's chin and
was fastened by elastic round his head and neck. It also pro-
truded at the forchead to allow a free circulation of air.
A tray was then set before the subject, resting on the arms of
his easy chair, and the informal exercises for that session would
commence. These involved the subject in trying to obtain an
impression of some target (a picture, an inscription, an object,
etc.) that had been placed on his tray or in trying to receive a
colour impression merely by touching some coloured material.
For these informal exercises we improvised continually, always
with a view to stimulating as much as possible the subject's
232
with the subject, tell him how he was doing, exhort him or caution
him as the situation demanded:
The second half, roughly, of each hourly session, was devoted
to the critical tests (see below). During a critical test there was,
of course, no interchange between subject and experimenter.
Ordinarily, the test would be announced and then nothing further
would be said until the subject had finished giving his responses.
Occasionally, however, the experimenter found it necessary in
view of the semi-stuporous condition of some subjects to inter-
ject some remark such as 'have you finished yet?' or 'take your
time', etc. Originally we had decided not to give knowledge of
results on any of the critical tests. For the sake of morale, how-
ever, we did later make a practice of telling the subject every
time he scored a direct hit on any of the 'selective answer' tests.
A verbatim record, as near as possible, was kept of everything
the subject said throughout an entire session and, in addition,
his responses on the 'object' and 'picture' tests (see below) were
tape-recorded.
As an incentive, half-a-crown was awarded for each direct hit
plus a lump sum of ten shillings at the end for goodwill. Despite
the frustrating nature of the work subjects were most co-operative
and we were pleasantly surprised to find that there was little
difficulty in maintaining good motivation and morale.
4. The Tests
In all, 6 critical tests were used: 3 of the 'selective answer'
kind, 3 of the 'open' kind. Dice were thrown at each session
to determine in which order the tests should be administered (at
first we gave the tests in what we assumed would be an order of
difficulty but we discovered that the use of a fixed order resulted
in subjects anticipating their responses before the next test had
been announced). Dice were also thrown to determine which
particular target would be used for each test. Each subject was
given one trial only on each of the 6 tests described below and,
since there were 20 subjects each doing 10 sessions the number of
responses recorded per test was: 20 x 10 =: 200.
(1) The Clock Test
The clock used was an ordinary domestic clock with a circular
dial of 12.5 cm diam. set into a square frame. The hours were
each marked with an arabic numeral. The following extract
233
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
Jour witt of the Society for Psychical Research
[VOL. 43, No. 727
from the Instruction Sheet circulated beforehand to all subjects
explains the exact nature of the task:
'A clock will be set before you on your tray with its big hand always
pointing to the 12 o'clock position and its little hand pointing to one
of the hour positions as randomly determined anew on each occasion.
YOUR TASK is to ascertain which of the 12 positions the little hand is
pointing to. Your initial ascertainment must be done without touching
the clock but you may then run your fingers over the glass cover and
revise your answer if you wish. Your revised response will then be
accepted as final for the purpose of scoring.'
A record was taken after each trial of the actual position of the
hour-hand and the subject's call.
(2) The Block Test
FIG. 1. Plan of Matrix to scale showing black target-block at Cz.
16 cm
-8cm-
>
Notation of Matrix:
The blocks used in this test were plastic toy building blocks,
2.4 cm x 1.6 cm x 1 cm high, manufactured by Lego Ltd who
kindly supplied us free of charge with sets of coloured blocks that
had all been extruded from the identical mould to make sure that
any surface irregularities should be equalized as between the
different colours. The following extract from the Instructions
Sheet explains the exact nature of the task:
234
MARCH 1966]
The 'Ryzl Technique' for Training ESP Subjects
"The target material in this case consists of a 4X 4 matrix of "Lego"
blocks affixed to a green Lego base. All the blocks except one are white,
the one odd block is black and its position in the matrix will be randomly
determined anew on each occasion. YOUR TASK is to pick off with your
fingers the black target block, though you may finger all the blocks as
much as you please until you have located it. Note that the matrix is
so placed as to fall readily into four 2 X 2 sub-matrices and some credit
will be given for selecting the correct quadrant even when the target
unit is incorrect.'
The position of the target-block and the position identified by
the subject was recorded at each trial.
(3) The Colour-Card Test
The target materials were made up from Windsor & Newton
gummed colour papers stuck on to 2" x 2" squares of card. Five
different colours were used as targets but to discourage guessing
the subjects were not told which five colours would be used. For
purposes of scoring an identification was counted as correct if
the appropriate colour-name Blue, Red, Orange, Yellow or Green
was ascribed to the appropriate target. Once again the exact
nature of the task is best explained by quoting directly from the
Instructions Sheet:
'Five squares of coloured card will be set out in front of you in a row,
mounted on black paper and covered over with a transparent perspex
sheet. One of these squares is different from the other four. YOUR
TASK is to say whether the odd square is Ist, 2nd, 3rd, 4th or 5th from
the left, to identify its colour and to identify the different colour of the
four other squares. Both the position of the odd square and the par-
ticular combination of colours used will be randomly selected anew on
each occasion. You are free, if you wish, to run your fingers over the
perspex sheet.'.
At the termination of each trial a record was made of the 3
critical independent variables of the target layout and of the
subject's response on each.
(4) The Word Test
The five target words used were:
A. BANANA B. FREEZE C. KISS D. GOOD E. THUMP
These words had been chosen arbitrarily but with the special
proviso that no two words should have any letters in common.
We had hoped this would facilitate matching of the subject's
responses when it came to analysing the results. Words were
made up from 1}" high, thick solid black, capital, sans-serif
235
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
-----------

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
Sheet:
1
'A single English word in large black capitals of standard dimensions
on a white ground will be presented to you on the tray where it will be
covered over by a transparent perspex sheet. YOUR TASK is to read the
target word or, failing that, to mention as many of the letters belonging
to it as may come through to you. Note that only 5 different words
will be used as targets, the particular target word on each occasion
being randomly selected anew on each occasion. You may, if you wish,
run your fingers over the perspex cover.'
At the end of each trial a record was taken of the target word and
of the sequence of letters given in the subject's response.
(5) Picture Test
The 5 target pictures used were as follows:
A. Cramond Harbour: Large colour-photo of waterside
scene on a calm sunny day but with no human figures.
Three small yachts are visible without sails, there are trees,
grass, rocks, a swan and a stone house in the background on
extreme left.
B. Charlie Chaplin: Black-and-white full-face still of Chaplin
from 'City Lights'. Figure shown from above the waist.
C. Little girl: Coloured magazine advertisement showing
small girl in red dress eating soup, spoon in mouth.
D. Music Room: Coloured photo of an interior. Foreground:
'cello resting on chair, music open on music stand, books
lying on desk at extreme right. Background: grand piano
to left with metronome on top, picture on rear wall of old
city with canals, potted indoor tree in right corner. Carpet:
deep pink ground with elaborate scrollwork pattern in buff.
E. Lady with Ginger Hair: A full-face colour portrait-photo
from Vogue of Jane Asher. Long loose ginger hair falling
about shoulders, broad grin, arms akimbo, wearing mauve
jumper and green tartan skirt.
Instructions Sheet:
"The targets for this test are various photos such as one finds in
illustrated magazines. The target picture will be laid in front of you
on the tray and YOUR TASK will be to describe it as fully as you can.
Touch will not be used. Note that only five different pictures will be
used, the particular target being randomly selected anew on each
occasion.'
236
A. Brandy glass.
B. Flower-pot.
C. Black doll: a realistic jointed doll in sitting position
wearing off-white coat and hood with red trim.
D. Brass fish: an articulated Indian brass fish with 'ruby'
· eyes.
E. Egg: a hard-boiled egg in shell.
Instructions Sheet:
'Here the target is some object (which might be anything within
reason providing that it has no characteristic odour and emits no sounds)
which will be placed on the small table to the right of your chair on a
felt pad. YOUR TASK will be to describe the object as fully as you can
and to try whenever possibly to identify the object by name. Note that
only 5 different objects will be used, the particular target being ran-
domly selected anew on cach occasion. We regard this as the most
important of the 6 tests.'
Precaution: The target material for all the tests other than the
clock and the Lego Block Matrix were kept in a locked suitcase
throughout the whole period of the experiment except when
actually in use. Coded symbols only were used in recording
results in the notebook to prevent leakage of information.
III. RESULTS
I. The Selective Answer Tests
TABLE I. CLOCK TEST
(a) Overall totals:
11 == 200
(i) Direct Hits (D.HI.)
₱ =1/12
Total =25
M.C.E. = 200/12 = 16-6'
Dev. = +7.83 (corrected for continuity)
S.D. = 3.9
C.R. = +2.01 P<0.05
(ii) Divergence Scores (S)
Total =534
M.C.E. = 3 x 10 x 20 - 600
Dev. = +66
S.D. = 192/6 =25.16
C.R. = + 2.62 P<0.005
237
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7

(V)
(i) Distribution of all 200 calls:
Divergence from Target Position 40
41
4,
4.
As
fo: 25
40
31
35
30
Σ
f .: (16.6) (33-3) (33-3) (33'3) (33'3) (33'3) (16-6') (200)
27
12
200
where f. = observed frequency of calls
f. = expected frequency of calls
Chiª (d.f. = 5) =7'46 N.S.
(ii) Distribution of Extreme Calls:
40, 41 43, 4,
Σ
fo: 65
39
104
f .: (52)
(52) (104)
Chi" (d.f. = 1) =6.50 P<0.or
(c) Individual Scores (SA & D.H.)
n = 10 trials each
Rank Order of Sa Score
1 2 3 4 5 6 7 8 0 10 11 12 13 14 15 16 17 18 19 20
Σ
SA:
18ª 22
22 22 22 23
25 25 27 27
27 28 28 31 31 34 35 36 36
53+
D.H .:
3
1 4
2 1
O
3
2
I
I
I
1
1
0 0 0 1 0
25
SA : M.C.E. = 30, S.D. = 3.63
D.H .: M.C.E. = 0.83, S.D. = 0.874
(d) Session by Session Totals:
n = 20Ss
Serial No. of Session
Ist 2nd 3rd 4th 5th 6th 7th 8th 9th 10th E
S4: 57 46 66 41 58 60 57 49 42" 58
534
D.H .? !. 3 2
3 3
4
+ 3
3 3
O
25
S:
M.C.E. = 60, S.D. = 7.96
D.H .: M.C.E. = 1-6', S.D. = 1'24
*= Poor
(e) Position Preference in Calling:
Call ('o'clock')
1 2 3 4
5
6 7 8 9 10 11 12
Σ
fo: 9 21 19 19 24 11 22 21 19 12 16 7 200
f. for all positions assuming no preference = 16-6'
Chiª (d.f. = 11) =20.56 P<0.05
(1) The Clock Test
Scores on this test can be reckoned either in the usual way, in
terms of the number of direct hits, or, more sensitively, by means
238
wancnt 1900]
the 'Kyzl Technique' for Training ESP Subjects
of a divergence score (Fisk & Mitchell 1953). Here we present
both analyses. Table I (a) (i) gives the overall total for Direct
Hits which yields an above chance deviation of Critical Ratio 2.01.
This, on a one-tail test, is significant beyond the 5% level of
confidence. The overall Divergence Score is given in Table I
(a) (ii). Note that for a clock-target the mean divergence will be
3 hrs per trial, hence on 200 trials the Mean Chance Expectation
will be 600 hrs. Scores less than 600 represent above chance
scoring. Here we have an overall above chance deviation with
C.R. = +2.62 and, on a one-tail test, this is significant beyond
the 0.5% level of confidence.
The Distribution of calls over the 6 possible divergence
categories is shown in Table I (b) (i). On the hypothesis that
the subject is tending in the direction of the target position we
would expect to find a pile-up of responses in the neighbouring
categories and, conversely, a tailing-off of responses towards the
opposite end. Table I (b) (i), however, shows that the full dis-
tribution of responses does not differ from chance expectation,
presumably because the middle categories contribute virtually
nothing to the overall chi2 index. In Table I (b) (ii), therefore,
we have presented the results of a chi? analysis of the extreme calls
only which gives a value of 6.50 for one degree of freedom. Since,
on a one-tail test, a chi? index of 5.41 may be taken to correspond
to a P of o.or, for one degree of freedom, such a result is easily
significant at the 1% level.
Individual scores have been set out for easy inspection in
Table I (c). Here it can be seen that 2 subjects attain diver-
gence scores that are individually significant at or beyond the 1%
level. A calculation using the Poisson formula for expected
frequency of rare events reveal that the odds against finding even
2 subjects in a sample of 20 with scores of this magnitude is
about 50-1. We may also note that a third subject attained a
score of 4 out of 10 direct hits. An exact probability calculation,
based on the binomial formula, indicates that a score of not less
than this magnitude cannot be expected to occur by chance more
than once in a hundred times.
Table I (d) presents the session by session totals in terms both
of divergence scores and direct hits. It can be seen that in 2 out
of the 10 sessions the overall total exceeded 2 S.D.s above
chance expectation but, alas, no progressive trend in scoring
emerges. This table is, of course, critical for the training hypo-
thesis.
We have thought it worth including, in 'Table I (e), the dis-
tribution of calls relative to the clock face. A chi? analysis shows
239
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
is due mainly to the unpopularity of the 12 and I o'clock positions
plus the popularity of the 5 o'clock position. This bias is an
unfortunate weakness in what otherwise makes an excellent test
of ESP. A similar chi? analysis was carried out, though not
shown here, on the distribution of target positions and this
demonstrated that there was no significant deviation from the
random distribution that was intended as a result of tossing dice.
We also examined the matrix of target/call plottings for the
possibility that subjects might be making confusions of symmetry
in guessing at clock positions. No such tendency, however, could
be discerned.
(2) The Block Test
The probability of a direct hit, on this test, is 1/16. The test,
however, is specially designed to permit an analysis in terms of
whether the responses fall within a specified 'target-area'. An
analogy would be that of taking a target used in archery and
attempting to find out whether there was any tendency for the
arrows to converge towards the centre by describing progressively
larger circles about the centre and then estimating the expected
frequency of shots inside as against outside the given circle. From
Table II (a) we see that, while the number of direct hits fails to
reach an acceptable level of significance, an analysis in terms of the
octant (1/8) and quadrant (1/4) divisions of the target-matrix con
be considered significant at the 5% level, if we take chi2 of 2.71
(for d.f. = 1) as corresponding to a P of 0.05 on a one-tail test.
Table II (b) presents the distribution of individual scores in
terms of how many subjects scored totals of either 1, 2, 3 . . . or 10
direct hits in their 10 trials. It can be seen by inspection that the
observed frequencies do not differ significantly from the series of
expected frequencies listed in brackets beneath as estimated on a
TABLE II. BLOCK TEST
(a) Overall Totals:
11 = 200
fo:
33
61
108
fe: (12'5)
(25)
(50)
(100)
Chi? (d.f. == 1): 2.58
2.93*
3.23*
1.28
where
D.II. = Direct Hit
Oct. = within same octant as target-block
Quad. = within same quadrant as target-block
Half =within same left-right bisection as target-block
* = P <0.05
240
o
I
2
3
4
5-10
8
I
I
H
o
20
where f. = observed number of subjects with specified D.H. score
fe=exact probability estimate of number expected by bi-
nomial distribution
* = P <0.005
(c) Session by Session Totals:
n =20 Ss.
Serial No. of Session
1 2 3 4 5 6 7 8 9 10
Σ
D.II .: 0 3 2 3 1 2 2 4 1 0
Quad .: 5 8 4 9 5
6
7
8
4 5
18
19
D.H .: M.C.E. = 20/16 = 1.25, S.D. = 1.08
Quad .: M.C.E. = 20/4 =5, S.D. = 1.94
* = P<0.05
binomial distribution where p = 1/16, q == 15/16. The one surprise
is that even one subject should score as many as 4 direct hits
out of 10. The exact probability calculation given by the bi-
nomial formula indicates that a score of not less than this magni-
tude cannot be expected to occur by chance more than once in
about 400 replications. Morcover, a further calculation based on
the Poisson distribution, indicates that odds against finding even
one such case in a sample of 20 subjects is more than 20-1.
Finally, 'Table II (c) gives the session by session progress chart
in terms of the incidence both of direct hits and quadrant hits.
As shown, in only one session was the number of direct hits in
excess of 2 S.D.s above chance expectation and once again we
look in vain for any signs of improvement over the 10 sessions.
(3) Colour-Card Test
As previously explained, this test was only partially of the
selective answer type, inasmuch as subjects were not told which
colours would be used as targets. Optimistically, we had been
hoping that results would be significant even if analysed as if the
test had been selective i.e. as if p = 1/5. When this proved not to
be the case we decided to make an empirical estimate of p based
on the observed proportion of guesses falling within the target
range. Now, since each subject makes two colour calls (one for the
odd card, one for the others) this gives a total of 400 colour calls.
But only 336 of these correspond to our target colours. Hence we
may say that the probability of being correct on any given colour-
241
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
D.H.
18
Oct.
Quad.
Half
fo
9
f .: (10'49)
(6.99)
(2.10)
(0.37)
(0.04)
(0.01)
(20:00)

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
colour combination (1.e. correct on 'odd' and on 'others ) was
P1 .Ps=(1/5 x 336/400) (1/4 x 336/400) =0.035. And since the odd
card could be in any of 5 different positions the chances of being
correct both on colour combination and position of odd card was:
0.035 x 0.2 =0.007. From Table III (i), it can be seen that 4
direct hits were recorded as against a chance expectation of 1-4.
By the Poisson analysis for rare events this is only just significant
at the 5% level. Separate analyses in terms of position only and
colour combination only, set out in Table III (ii) and (iii), respec-
tively, gave non-significant results.
TABLE III. COLOUR-CARD TEST
Overall Totals
(i) Direct Hits on Position and Colour Combination
n = 200
Total: = 4
M.C.E. (where p =0.007 see text) =1'4
P (=prob. of getting at least 4 hits in 200 trials on Poisson dis-
tribution for rare events) =0.05 approx.
(ii) Direct Hits on Position only.
Total =50
M.C.E. = 200/5 =40
Dev. = +9.5 (corrected for continuity)
S.D. - 5.66
71 =200
C.R. = + 1.68 N.S.
(iii) Direct Hits on Colour Combination only.
Total =12
M.C.E. (where p =0.035 sce text) =7
Dev. = +4.5 (corrected for continuity)
S.D. = 2.6
C.R. = + 1.73 N.S.
72 =200
Comments on the Results of the Selective Answer Test
Results on the Clock Test and on the Block Test definitely
suggest the operation of x non-random factors. It is perhaps also
worth mentioning that an n of 200 is low in comparison with that
customary in most parapsychological experiments of a quantitative
kind so that a relatively high rate of scoring was necessary to
reach the significance levels here attained. On the Colour-Card
Test, however, it is doubtful whether anything more than mere
guesswork was involved. This was a big disappointment to us,
the more so as this test promised well in the preliminary informal
try-out. No sex differences emerged from any of these 3 tests.
After all the 200 response words had been tabulated under their
respective target words various criteria of similarity were applied
(prevalence of certain key letters, etc.) to see whether the target
words had had any influence on the responses given. In general,
the result was completely negative except in a few special cases
listed below where it certainly looks as if something more than
mere coincidence was responsible, although in the nature of the
case no estimate of probability can be adduced. These special
cases were as follows:
Target Word
Response Words
(where each response represents a
separate occasion)
FREEZE
FREEZE
FRIENE
GOOD
GOOD
WOOD
THUMP
HUMP
(5) The Picture Test
As previously stated, the subjects' spoken impressions were
recorded on tape. The tapes were then handed to a departmental
secretary (who was ignorant about what targets we had been using)
with instructions simply to make a typewritten transcript of all
the protocols. The final set of these protocols, duly coded, were
then submitted, together with the actual targets, to a colleague on
the staff of this department, whose duty it was to act as an inde-
pendent judge in matching protocols against targets. We instruc-
ted this judge to rate each protocol on a six-point scale for resem-
blance where o =a perfect description of one of the target pictures
and 5=impossible to identify with any of the target pictures.
Results were as follows: In all 199 separate protocols were
collected, but of this number in only 91 cases was any matching
attempted. Of the matches only 16 corresponded with the target
for that occasion as against a chance expectation of 91/5 = 18-2;
clearly, a non-significant result. We may note that of the 91
matches more than half (49) involved descriptions of outdoor
scenes, yet less than 1/5th (9) were given in response to Picture A.
'Cramond Harbour'.
The 3 protocols that were given the highest rating for re-
semblance were each matched against a picture other than the
actual target. For the sake of interest here they are in full:
'Picture of man's face. Dark . . . moustache . . . looks old-fashioned'
Match: B
Target: E
243
242
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
metrical and very light coloured . .. something, a small bit of red in
the middle, may be a person .. . piano on left-hand side ... I see
piano again and room . . . window open at back . .. a light-coloured
carpet, blue and buff claborate pattern'
Match: D
Target: C
'Picture of the sea . . . coast . .. cliff coming down .. . rocks .. . seals?
. .. sailing ships with no sails.'
Match A
Target: E
(6) Object Test
The matching procedure for this test was exactly the same as in
the case of the previous test. Results were as follows: For only
57 protocols was a matching attempted. Of this number only 9
corresponded with the actual target for that trial as against a
chance expectation of 57/5 = 10.4, again a clearly non-significant
result.
The following target objects were mentioned once but not in
conjunction with the appropriate target: Flower-Pot, Brandy-
Glass, Egg. On one occasion an egg-cup was mentioned when the
target happened to be an egg.
Comments on the Results of the Open Tests
Apart from the few word-responses listed there was no satis-
factory evidence of anything beyond mere guesswork on any of
the three open tests. The protocols on the picture test which we
have quoted in full look at first as if they might be displaced hits
but none of them are sufficiently circumstantial to warrant such a
conclusion. The utter failure of the object test which, in our
Instructions to the subjects, we had stressed as being the most
important of the battery was an acute disappointment.
3. Individual Differences
.
In the course of our investigation it definitely seemed to us
that some subjects showed more aptitude for the task than others
but none proved in the end to be uniformly superior on all tests
or at all times so there is little objective evidence to support this
impression. Two subjects, however, are worth special comment
in this connection. Mr R.C. was responsible for two of the five
exceptional word responses: GOOD and HUMP and, in addition,
made the score of 4 out of 10 direct hits on the clock test. Oddly
enough he was never a good hypnotic subject, it was hard to
bring him into any depth of trance, conceivably, but for this
handicap, he might have done much better. The other note-
no special signs of any unusual gifts. She then produced one
outstanding performance but although thereafter we devoted par-
ticular attention to her, always arranging it so that both experi-
menters were present at each of her subsequent sessions, we never
succeeded in eliciting again anything at all comparable. She was,
however, responsible for the score of 4 direct hits out of 10
trials on the Block Test, which, on an exact probability calculation,
corresponds to a P of 0.0025.
IV. DISCUSSION
Under this heading we shall consider the following questions:
Q.I. From an examination of the results we have presented above,
are we really justified in rejecting the null assumption that
they were due to chance alone?
Q.2. If the answer to Q. I is 'yes' are we really justified in invoking
ESP or is there some 'normal' explanation that could con-
ceivably account for the findings?
Q.3. If the answer to Q.2 is 'yes' does our investigation lend any
support to the claims that have been put forward for the
Rýzl technique?
A.I In experimental psychology a significance level of 5".5 is
usually considered sufficiently stringent to justify rejecting the
null assumption. But the experimental psychologist is usually
working within a fairly well-established framework of fact and
theory. This, unfortunately, is not the case in parapsychology
where accepting a paranormal hypothesis means, ipso facto,
challenging one or more well-entrenched assumptions regarding
the limits of what is physically possible. Accordingly it is proper
that a much more stringent level of confidence should be demanded
if the alternative to accepting the null hypothesis is to embrace
a paranormal explanation. On this reckoning the case for rejecting
coincidence as an explanation for our own findings is far from
overwhelming. Only two out of the three quantitative tests
attained even a modest level of significance; the strongest single
piece of evidence against the theory of coincidence was the overall
divergence score on the Clock 'T'est which gave odds of 200-1
against chance. Apart from this, the main case against a chance
hypothesis rests on the five exceptional word responses and the
details omitted here for reasons of space of Miss A.D.'s 4th
session.
R
2.45
2.44
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
au coulgation of this kind there are two main spurious
effects against which one has to be continually on one's guard:
(a) inadequate blindfolding of subjects or screening of objects and
(b) inadvertent cues imparted by the experimenter. Since the
prime purpose of our investigation was to train ESP rather than
to demonstrate, ab initio so to speak, the genuineness of ESP, we
did not bother to insure that our conditions were the most rigorous
possible believing that there would be ample opportunity for fool-
proof demonstration later on once we started making progress.
Now, of course, endowed with hindsight, we realize that we ought
to have given more thought to just how much could be salvaged
from the wreckage if the training aspect of the investigation were
to come to naught. For example at least one of the critical tests
should have been run under double-blind conditions which, at
one stroke, would have taken care both of (a) and (b). The double-
blind condition also eliminates a telepathic as opposed to a clair-
voyant interpretation of any positive results. Very late in the day
we did, in fact, introduce a supplementary test under double-
blind conditions. This consisted of one complete pack of Zener
cards, each card separately enclosed in its own envelope and duly
coded after shuffling. Each of our subjects was required to guess
his way through the pack, lifting each envelope up one at a time
and proceeding at his own rate. This was done twice over, on the
9th and roth sessions, before the final decoding. Alas, scoring
was found to be purely at a chance level.
As regards (a) the problem of blindfolding, here again our
inexperience betrayed us. Had we used our face-mask right from
the start we believe we could defend ourselves against any criticism
on this score, but it seems to be almost an axiom of conjuring that
there is no completely foolproof method of blindfolding the eyes
alone.1 Alternatively we could have worked instead with screens
although even then one has the problem of transporting the target
1 We are indebted to Mrs K. M. Goldney for drawing our attention to
a passage in T. H. Hall's The Strange Case of Edmund Gurney (Duckworth
1964), pp. 115-16 concerning the exploits of a professional conjuror, one
Kuda Bux, an Indian, who, in 1935, gave a demonstration of eyeless sight
after submitting to the most elaborate blindfolding precautions in the
presence of such notabilities as William McDougall, C. E. M. Joad,
S. G. Soal, Harry Price, and Mrs Goldney herself. Yet when later he was
asked by Mrs Goldney to let his head be put into a loose shoe-bag the
magician was forced to decline! We are further indebted to Dr Alan Gauld
for pointing out to us an early article by Eleanor Sidgwick 'On Vision with
Sealed and Bandaged Eyes' (Jour. S.P.R., Vol. I, 1884, pp. 84-6) des-
cribing Richard Hodgson's experiments on himself when he successfully
duplicated the performance of a young boy who, it was claimed, attained
eyeless sight when under hypnosis.
246
MARCH 1966]
The 'Ryzl Technique' for Training ESP Subjects
material into position with complete security. Moreover at the
parapsychical level Rýzl is of the opinion that screening is least
prejudicial to success when it is either directly in contact with the
target, as in the case of an envelope, or directly in contact with the
subject's eyes, as in the case of a blindfold, while the intermediate
position represents a maximal impediment. (Private communi-
cation).
We would like to mention that so far as we could tell blind-
folding made no actual difference to our subjects' performances.
With adults at least there would seem to be no residual vision once
the eyelids are closed under hypnotic suggestion.1 If Miss A.D.
had been peeping from under her blindfold in her extraordinary
session 4 it is hard to understand why she did not do so much more
effectively on the many occasions since the termination of the
experiment when we deliberately refrained from using any blind-
fold.
As regards (b) the question of minimal sensory information
unwittingly signalled by the experimenter, what psychologists call
the 'Clever Hans' effect with allusion to the exploits of a certain
so-called 'calculating' horse, it is impossible to exclude such a
possibility completely with any set-up short of the double-blind
condition. Naturally, being fully alive to this possibility, we took
all reasonable precautions against it but, as explained under
Procedure, we found that we had to relax occasionally our rule
never to say anything during a critical test. Here again the advan-
tages of a double-blind test is that it allows the experimenter to
say anything he thinks necessary by way of encouragement or
reassurance at no risk.
A.3: As we have already seen, no progressive trends of any kind
were anywhere discernible in our results and not one subject has
yet emerged with even a rudimentary skill in the use of ESP.
From this point of view it cannot be said that our investigation
offers any corroboration of the claims which Dr Rýzl has made,
and still persists in making, for his training technique, on the basis
of his own experiences. The interesting question is what accounts
for this striking discrepancy? Lack of reproducibility has become
such a notorious bugbear of parapsychological research generally
that we feel no effort should be spared to throw whatever light
one can on this vexatious question.
1 The same cannot however be said to hold in the case of young children
as J.D. discovered to his dismay when he began experimenting on blind
colour discrimination using his 7-year-old daughter. But children are
notoriously difficult to hypnotize at all ; ecurely.
247
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
.-

Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7
.it un pitwit Hwwwnet tin et pusSUl svurtes of ulc ulstitpality
must be considered: (1) In the subjects, (2) in the experimenters,
and (3) in the methods. So far as (I) is concerned, the subjects
in both cases were drawn mainly from a student population and
it seems unlikely that British students are any less gifted para-
psychically than their Czech counterparts. Nor have we any
reason to suspect that our subjects might be more sceptical
initially of the possibility of ESP, in other words that we had more
'goats' among our group, for it seems intrinsically improbable that
our subjects would have sacrificed so much time for no better
purpose than to vindicate their disbelief. We feel therefore that
unless we were very unlucky in our sample, we can probably
discard source (1).
As regards (2), however, it has often been remarked that where
parapsychical phenomena are concerned a particular experimenter
may, in some subtle way, be just as indispensable to a particular
sort of manifestation as the ostensible subject. The only satis-
factory way of settling the query in the present case would be for
Rýzl to come to Edinburgh and tackle our subjects. Unfortunately,
until the Czech authorities adopt a more liberal policy towards
the right to travel this course of action is not open to us.
This brings us finally to (3). Here the most plausible sugges-
tion is that, owing to our lack of skill and experience in hypnotism,
we were unable to make competent use of the training procedure.
Thus, as we have already explained, Rýzl lays great emphasis on
the capacity for vivid visual hallucination as the essential pre-
requisite for success in ESP. We must admit that none of our
subjects were capable of hallucinating with a degree of realism
that would make the experience comparable, subjectively, to an
actual perception. On the other hand, the two best visualisers in
our group were both most disappointing when it came to the ESP
tests while, on the contrary, our two most hopeful ESP subjects
(R.C. & A.D.) were neither particularly accomplished visualisers.
Yet Rýzl insists that his subjects did learn to tell the difference
between the merely imaginary vision and the true veridical
hallucination. Thus he writes with reference to his object tests:
'real "guessing" occurred rarely, if ever. Usually the subject
either asserted that he did not see anything, or he described his
visions and then they were-as a rule-veridical.' (Private com-
munication.) Our subjects, on the contrary, as we have already
seen, hardly ever declined to describe their impressions in this
situation but on no single occasion were these veridical. Yet, in
attempting to arrive at some conclusion on this point, it must not
be forgotten that, at the beginning of this academic year both
248
authuis wlic yune fitn ttime
Rýzl training procedure. It is scarcely surprising if Rýzl with
years of experience behind him can approach the task with very
much greater confidence and authority and this in turn may well
inspire his subjects to greater achievements.
V. SUMMARY AND CONCLUSIONS
The attempt we have here described to validate the Rýzl
technique for the training of ESP under hypnosis was unsuccess-
ful: there were no indications that our subjects were in fact
developing any ESP ability as a result of the treatment to which
we were exposing them.
Some evidence, however, was presented and discussed which
strongly suggests to us that some ESP was operative in the course
of the investigation, notably in the Clock Test, the Block Test, the
Word Test and in the performance of one particular subject
Miss A.D.
The authors are of the opinion that, all things considered,
further work in this direction would be warranted. It would be
especially valuable if, during the next academic year, parallel
experiments could be run simultaneously in Prague and Edinburgh
with a view to discovering just when and where and why any
discrepancies occur. Even if it may be very premature to talk of
training ESP, in the sense of a method of eliciting it from any
normal subject, it would be a considerable advance to have a
better method of cultivating whatever potential talent may be
around.
ACKNOWLEDGMENTS
We are much indebted to C.H.K. whose generous donation to
Edinburgh University for parapsychological research enabled
I.M. to be taken on in the Department of Psychology as a full-time
research assistant. We are also grateful to Prof. J. Drever for
providing us with accommodation and facilities in his department.
REFERENCES
G. W. Fisk and A. M. J. Mitchell (1953). 'ESP experiments with
Clock Cards: A New 'Technique with Differential Scoring.' Jour.
Soc. Psych. Res., 37, 1953, 1-14.
M. Rýzl (1962). "Training the P'si-Faculty by Hypnosis.' Jour. Soc.
Psych. Res., 41, 1962, 234-52.
M. Rýzl (1965). 'ESP experiments with P.S.' Part II. (Unpublished
Monograph.)
249
Approved For Release 2008/04/01 : NSA-RDP96X00790R000100040007-7