CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
Parapsychological Monographs
No. 1
A Review of Published Research
on the
Relationship of Some Personality
Variables to ESP Scoring Level
GORDON L. MANGAN
Victoria University, Wellington, New Zealand
Published in 1958
PARAPSYCHOLOGY FOUNDATION, INC.
29 West 57th Street, New York 19, N. Y.
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
Nash and Richards (26) in 1947 first investigated the relationship be-
tween a measure of intelligence and scores obtained in a series of PK tests.
The I.Q. scores of their 48 college subjects, obtained from the Higher
Examination of the Otis Self-Administering Tests of Mental Ability,
showed a very small correlation ( -. 12) with PK scores.
Summary on Intelligence and ESP
The nature of the relationship between intelligence and ESP scoring
level is still undefined. Valid objections, which preclude any clear-cut
conclusions being drawn, can be levelled at most of the studies that have
been made.
In the first place, they have often involved too few subjects, a fact which
makes generalization difficult, despite some high correlations. Again,
the same intelligence test was never used by two investigators, and since
different tests may be sampling different aspects of intellectual ability,
the results are not strictly comparable. In addition, not all the intelligence
scales or estimates used are of equal validity, and in two cases, the in-
vestigation of the relationship between intelligence and ESP scoring
level was a side-issue to the main experiment.
One tentative conclusion, however, may be drawn. There seems to
be one factor conducive to a correlation between ESP scoring level and
intelligence, namely, when the "best" estimate of scores is used as the
ESP criterion.
By the use of the "best" estimate of scores rather than averages for the.
ESP criterion, Humphrey found that the correlation between intelligence
ratings and ESP scoring increased. An estimate based on the best results
achieved should eliminate those fluctuations due to factors other than
intelligence, such as boredom and fatigue, which are known to affect
scoring level, and give a purer estimate of ESP to be correlated with in-
telligence. Obviously the overall average run scores need not be an ac-
curate reflection of the subject's real ESP ability.
Humphrey's findings particularly suggest cither that the more intelli-
gent subjects have better ESP, or that the obtained correlations between
intelligence and ESP scoring are merely indicative of the subjects' adapt-
ability to the test situation. No more definite judgment can be made at
this stage.
[ 10 ]
INTEREST
(. E. Stuart was on
personality factors an-
experiment involving :
their effect on ESP sco
role of "affectability"
mate of his success in a
score he has just prev
In 19446 Stuart (52
list of 60 items. Subjec
which varied from "li
lia, which included cv
college students, was
in the experiment. The
concealed stimulus pict
which were closely re'
gate whether the subj
pictures influenced the
These clairvoyance
matching method, der
tio i. The total EST
out only one non-sig.
scores made on the int
to the five attitudes ro
was no evidence, how
of the stimulus picture;
Swart then separat
thore which fell near (
Aextreme) from the ni
bis subjects to be "af
"level of aspiration" ?
at chance, and "unat
Paper, Stuart equatro
considered the extren
group as the "unaffcc
In the drawing ti
"altectable" group he
boas vatistically signif
da kward displaceme
sopht negative deviaty
taxe deviation (P ... 000
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
Summary of ESP and Interest Ratings
The successful discrimination between high and low scoring ESP sub-
jects on the basis of ratings on both the full Interest Inventory and on
the restricted 14-item scale, which was reported by Stuart and Humphrey
in carlier investigations, did not hold up as well in the later series. The
results of these later series, however, are not published in their entirety,
but are merely briefly mentioned by Humphrey in a review (19). Whether
this decreased efficiency reported was in fact due to the lack of a real
relationship between interest ratings and ESP scoring level, or whether
it was due mainly to widely differing psychological conditions, such as
number of runs per subject, or type of ESP test, which obtained during
the later series, cannot be determined from the information available.
Inspection of the items of the full scale indicate that they cover fairly
well the full range of student activity and interest. Stuart equated
"affectability" with range of interest; this fact, added to the pervasiveness
of the scale, seems to indicate that mid-range subjects may be those who
are moderate in their interests and who maintain a reasonably temperate
attitude towards their environment.
Inspection of the 14 items of the restricted scale, however, suggests
that they measure what could be loosely described as "social adjustment";
perhaps it would be more correct to say that the scale is heavily weighted
in favor of the more social or extravertive activities. The two scales appear
to be measuring somewhat different factors, and it would seem essential
to analyse the scales against established criteria in order to get at what
each scale basically is measuring. Without information so secured, we
can merely conclude that although both scales, to a different degree,
separate high and low ESP scorers, the personality traits concerned in
this differentiation remain in doubt.
[ 14 ]
INTROVERSION-EXTRAV
Humphrey first reported an ESP
Personality Inventory in 1945 (13).
the Earlham College Series I (GES
and the Humphrey-Pratt Precognit
jects; Bernreuter ratings on 6 perso
ulliciency, introversion, dominand
variousness-were correlated with
cognition series, the CR of the diffe
quarters of the record page was tak
the correlations between Bernreuter
cant; subjects who were stable, extra
however, tended to score positively
upposite characteristics tended to so
Humphrey (16) later utilized the
determine a cut-off point on the s
high and low scoring ESP subjects.
frets were judged to be extraverted o
thry scored above or below the 50t
" highly significant positive deviatie.
truverts scored at chance. The CR
1 .- P scores for the two groups was
Tab
ESP Scoring Levels of Extra
Subjects Scoring
above Chance
Letraverts
14
lotroverts
Totals
5
19
' In the basis of these results it was p:
'y 'atth percentile on the Bernreu
I'm on ESP card tests than those w
13
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96 05787R000400100014-4
sov ras:
The two series on which the prediction was tested were the Pratt-Humph-
rey Precognition and the unpublished Lawrence Clairvoyance Series.
In the Pratt-Humphrey series, the ten extraverts had a deviation of
+-56, and the nine introverts a deviation of -- 34. The CR of the difference
was significant (P == . 02). In the Lawrence series, the 9 extraverts made
a deviation of +48, the 12 introverts a deviation of -18. The CR of the
difference was non-significant (P == . 08). The total of 19 extraverts from
the two series made a deviation of +104, and the 21 introverts a devia-
tion of -52. The CR of this difference was significant (P == . 005).
As shown in Table 5, the consistency of this separation was significant
(P = . 005) with 74 per cent of the extraverts scoring above chance and
76 per cent of the introverts scoring at chance or below.
Attempts at Repetition
Caspar (5) administered the Bernreuter Inventory to 20 subjects and
obtained 2 GESP and 2 BT runs from cach. He classified his subjects
as extraverts or introverts on the basis of whether they scored above or
below the 50th percentile on the scale. The extraverts had a deviation
of 4-26, and the introverts a deviation of -18. The CR of the difference
was suggestive (P === . 03). Eight of the fourteen extraverts scored above
chance, but none of the six introverts did. When evaluated by the exact
method, the results are significant (P = . 02).
Although only two studies have been reported with the Bernreuter,
it appears, to be a very promising research tool. In both studies, high and
low scoring ESP subjects were separated with a high degree of consistency.
In the Nicol and Humphrey study (27) correlations were obtained
between ESP scores (Known and Unknown runs) and two measures
of introversion-extraversion. Factor T of Guilford's STDCR Inventory
is called Thinking Introversion-Extraversion. The thinking introvert is
given to reflective thinking and analyzing himself and others, while
the opposite holds true for the thinking extravert. The correlations be-
tween Factor T and the known ESP scores was 4 -. 10, with the Unknown
scores +.37,* and with total ESP scores -1 -. 33.
Factor S of this same test is called Social Extraversion; it correlated
+ -. 29 with Known ESP scores, +.21 with Unknown scores, and +.34
with total ESP scores. None of these correlations was significant, but a
significant correlation (+.54 ** ) was found between Social Extraversion.
and Self-Confidence (Factor I ) and a suggestive correlation (+.37*)
was found between Thinking Extraversion and Self-Confidence. The
latter correlations have value in this study. Self-confidence was found
to be the factor most highly correlated with total ESP score (r + .55 ** ).
.A person with a high score on Factor S is characterized as being social,
as one who tends to seek social contacts and enjoys the company of others,
while low scores indicate shyness and seclusiveness.
Summary of Introversion-Extraversion and ESP Scoring Levels
In all the studies reviewed in this section, it was found that extraver-
sion was associated with higher ESP scores than introversion. This factor,
or more precisely, the scales on which this factor is measured, scparated
out high and low scorers with a high degree of consistency. Unfortunately,
however, it is not clear which aspects of behavior are included under the
term extraversion, and for evaluative purposes it would seem essential
to have more specific information on the factors underlying this broad
[16]
comprehensive category. 1'
muension scales, such as
the factor, and it is uncerta
tictors as, for example, so
ford questionnaire.
An alternative has been
ford-Martin or Cattell, wh
of highly correlated traits
better estimate of extraver
analysis, to correct for the
This method has been use.
and the direction is a prer
ents of extraversion is no:
these lines would have mit
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96-60787R000400100014-4
tested were the Pratt-Humph-
Lawrence Clairvoyance Series.
extraverts had a deviation of
-- 34. The CR of the difference
e series, the 9 extraverts made
iation of -18. The CR of the
The total of 19 extraverts from
and the 21 introverts a devia-
is significant (P == . 005).
this separation was significant
erts scoring above chance and
e or below.
Inventory to 20 subjects and
ich. He classificd his subjects
whether they scored above or
Le extraverts had a deviation
- 18. The CR of the difference
arteen extraverts scored above
When evaluated by the exact
.02).
cported with the Bernreuter,
Lool. In both studies, high and
h a high degree of consistency.
) correlations were obtained
wn runs) and two measures
Guilford's STDCR Inventory
m. The thinking introvert is
g himself and others, while
<travert. The correlations be-
was 4 -. 10, with the Unknown
.33.
I Extraversion; it correlated
Unknown scores, and ++.34
lations was significant, but a
between Social Extraversion
ggestive correlation (+.37*)
n and Self-Confidence. The
y. Self-confidence was found
total ESP score (r + .55 ** ).
characterized as being social,
-enjoys the company of others,
siveness.
Scoring Levels
., it was found that extraver-
han introversion. This factor,
factor is measured, separated
of consistency. Unfortunately,
avior are included under the
moses it would seem essential
actors underlying this broad
comprehensive category. Part of the difficulty lics in the fact that single
dimension scales, such as Bernreuter, may not give a pure measure of
the factor, and it is uncertain to what extent it can be identified with such
factors as, for example, social and thinking extraversion on the Guil-
ford questionnaire.
An alternative has been to use a multiple trait scale, such as the Guil-
ford-Martin or Cattell, where all the overlapping material of a number
of highly correlated traits, which together should give a progressively
better estimate of extraversion, is utilized, and by the use of regression
analysis, to correct for the degree of overlap between the various traits.
This method has been used with some success by Nicol and Humphrey,
and the direction is a promising one. Some clarification of the compon-
ents of extraversion is necessary, however, before further work along
these lines would have much value.
.
1
[17 ]
-
Approved For Release 2001/03/07 . CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96-09787R000400100014-4
EXPANSION-COMPRESSION RATINGS AND ESP SCORING
In 1942, while at Stanford University, Stuart (51) developed a tech-
nique for judging similarities between four concealed target pictures and
the drawings made by a subject attempting to reproduce the pictures.
This technique, called the preferential matching technique (PMT),
was used by Stuart to analyse the large collection of drawings he ob-
tained at Stanford; the latter provided the data to which the expansion-
compression ratings were applied.
Paula Elkisch (8) has devised a projective test which utilizes the
form qualities of children's drawings. Certain features of the drawings
are considered to indicate neurotic trends, and these features are measured
in terms of four criteria: rhythm-rule, complexity-simplexity, integra-
tion-disintegration, and expansion-compression (E-C). The E-C criter-
ion was the only one which subsequently proved successful in discrimina-
ting high and low scoring ESP subjects.
Elkisch defines expansion and compression as follows: "Expansion
stimulates the imagination dynamically. It conveys an atmosphere
of freedom, courage, adventure, and may be a symptom of vitality
and of healthily developed extraversion. Expansion stands for a direc-
tion toward the surrounding world; for the potential ability of making
contact ... Compression conveys a feeling of discomfort, of being shut
in, of pressure and compulsion. Compression may be, if connected with
other traits, a symptom of a neurotically developed introversion, even
of a compulsion-neurosis. Compression stands for isolation."
Certain aspects of expansion-compression can be fairly objectively
described. For example, in making drawings, the compressives use only
a small a. nount of the available space, their drawings are cramped and
badly proportioned, lines are light and feathery, they use too many con-
ventional forms-houses, boats, etc. By means of these characteristics
it is possible to make an overall assessment of expansion-compression.
Drawing Tests
In the first reported E-C research, Humphrey (20) in 1946 used the
data from four series of clairvoyance drawings obtained by Stuart. Of
the 96 subjects involved, 41 were rated expansive and 55 compressive.
The drawings from cach group were scored by the preferential matching
technique. With mean chance expectation at 40.0, the mean ESP score
for the expansive group turned out to be 41.88, for the compressive
group only 37.45. The difference in average scoring level between the
two groups has a significant value (P = . 003).
Although there was no significant overall deviation in his data, Stuart
had found significant backward displacement (P = . 003) which had
been the main ESP effect. Displacement data were not available for
one series, but a comparison was made between backward displace-
ment scores of the expansives and compressives on the remaining series.
[18]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
With mean chance expectat
of 29.54 and the 33 compre
between the scores of the tw
Following this successful
Humphrey applied the san
used the data of the Stuar!
Individual Tests. A total of
each. With mean chance cx
sive group was 38.23, for t
in scoring level between the
the E-C rating made a succ
with GESP drawings. There
with the expansives now scor
From these two reports it
compression, discriminates 1
voyance and GESP drawing
pressives are the positive so-
the nature of the ESP test,
A logical follow-up was th
test results. The first study
Humphrey (45). The 186
inked to make drawings in
then given 2 BT card runs. 1
non-significant. The drawin
a subject was expansive or co
expansive, 97 compressive.
The average run score of
pressives 4.79, but the diff
However, since there were
fried two sessions, a total
tent. The expansives had a
taking this consistency into
Isseen the expansives and co
In a large scale experim
Smith, and McMahan (53)
Lsp subjects on the basis c
wues of clairvoyance tests. 'T
ried 4 clairvoyance card 1
rah subject did 4 spontaneo
sadly for the purpose of com
wwwier different conditions.
In the group series, 63 st
Howed by 2 clairvoyance
": >wings, all in one session.
The overall results of the
*** group series were non-si
ave subjects scored above
- + hance, but the difference
: tt^ group series the 23 cx
wespressives below chance
twant (P = . 01). When
wearing level was highly
It the card tests the total
**** was no significant sep

Approved For Release 2001/03/07 : CIA-RDP96- 3787R000400100014-4
ratings were compared against PK scores, it was found that both groups
scored slightly above chance.
Summary of E-C ratings and ESP scoring
This review indicates that the E-C ratings were not always successful
in separating high and low scoring ESP subjects. The best evaluation
of the overall efficiency of E-C ratings is Humphrey's 1951 review article
(19). In this she states that in 10 of the 12 clairvoyance drawings series
evaluated up to that time, the expansive subjects, as a group, obtained a
higher average ESP score than did the compressive subjects. If the prob-
abilities associated with the difference in each series are combined by
Fisher's method, the E-C difference, considered in its entirety, was signifi-
cant (P == . 005), although the overall ESP results of the scries were in-
significant.
Humphrey reported that 54% of the 140 expansive subjects scored
above chance, while only 42% of the 345 compressive subjects scored
above chance. If these percentage figures are evaluated for consistency
of group scoring, a significant chi square of 6.03 (1 d. f.) is found (P =. 01).
In analyzing the GESP drawings, it was discovered that the compressive
subjects had a higher average ESP score than did the expansives in cight
of the nine series evaluated. Humphrey states that the difference between
the two types of subjects for all series pooled is statistically significant,
but the method of evaluation is not specified.
There were 29 experiments completed in which clairvoyance card
tests and drawings were given cach subject. In 17 of these, the expansives
made a positive deviation on card tests while the compressives had a
negative deviation; in nine experiments this direction was reversed,
and in three no difference between the two groups was found. There
was a deviation of -1-62 for the 955 runs of the expansive subjects and a
deviation of -51 for the 1949 runs of the compressive subjects; the
difference between them was insignificant.
The difference in average run score for the 26 Duke series was of
borderline significane (P
=== approx. . 02), while the three non-Duke
series showed a non-significant reversal of effect. Another interesting
point reported by Humphrey was that the four series in which subjects
were tested individually gave a much larger difference than that found
in the group-test series.
In these series where the E-C rating was applied to clairvoyance ran
scores, the psychological conditions varied widely from series to series.
The E-C rating was based on one drawing in some series, on two draw-
ings in others; sometimes four drawings were used. The ratings were ad-
ministered before the card runs in some of the series, in others after the
runs.
On the basis of her experience, Humphrey suggests that the E-C rating
is not dividing subjects according to whether they will score positively
or negatively, but rather according to the type of hit distribution they
will give. Although compressive subjects as a group gave negative ESI'
scores, closer analysis of the results revealed that this score was due to
the bad beginning and that compressives are quite capable of making
high positive ESP scores after they are "warmed up".
It was also observed that an individual's drawings may change from
expansive to compressive within a single session, or between sessions
with the ESP scores tending to reflect these changes. The E-C rating:
therefore appear to be indicative of the subject's temporary mood.
[ 22 ]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
ADJUSTMENT K
QUESTIONNAI
Although projective techniq
ally used to assess an individ
personality inventories and qu
an overall adjustment index o
lated to adjustment. An exar
Insecurity Questionnaire.
A short form of the Mas
used by Smith and Humphre
The secure subjects averaged
4.71 in 166 runs; the different
not significant.
A later article by Stuart, of
Test when applied to drawing
of this experiment, the secure
viation on the drawings than
Series the insecure subjects sco
Irlow chance. None of these
significant. The differences on
msignificant, with the secure
(4 5.07 and the insecure subjec
The Heston Personal Adjus
wo obtain an overall index of
tev ile scores for each of the six
that the majority of subjects w
Laned positive deviations and
"JAained negative deviations.
waived was consistent at a sug
Rivers (30) administered the
talents and a college group o
& hool group the subjects who
"saturity", "feelings of inadeg
? cher scores on the clairvoya
> very slight relationship to th
' ar of the mental health ratis
earl of the college subjects.
Though in the Nicol and
Lesst rating was derived from
wersal correlations having a
fryvitality factors which could
r .justment. Among those
4.44. ESP scoring were freedo

Approved For Release 2001/03/07 : CIA-RDP96-30787R000400100014-4
.. . ....... .. . . .. . .. .. .. .......
lucky disposition), freedom from nervous tension, emotional stability,
calm trustfulness, and low irritability level.
Summary on Adjustment Ratings from Questionnaires and ESP Scoring Levels
With the exception of Rivers' study, the results of the research in-
cluded in this section all point toward the conclusion that higher ESP
· scores are obtained by subjects possessing the personality characteristics
generally included under the label of "good personal adjustment."
Whether well-adjusted subjects score higher because of greater co-
operation, quicker adaptation to the experimental situation, better ability
to establish rapport with the experimenter, freedom from personal in-
hibitions, some combination of these factors, or other unsupected factors
is a matter for further research.
COMBINATIONS OF I
ESP :
Some of the researches rev
the relationship between ESI
three personality measureme
cussed previously, however, a
the various measurements sin
of any combination of th
This section will be devote
the relationship of these cor
scoring level.
In a review article Humph
in average score was obtainc
treme-low ratings on the In
rither rating was considered s
level between these combined
and represents the pooled ro
Expansion-compression rat
scries and the difference in
ESP Scoring Iv
Subjects Scori
above Chan
Expansive
Malrange
High
26
( onpressive
36
Totals
62
"pre-high group and the :
+ pronounced, but becat
# difference was not as s.
; combination of person
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
[ 24]
:
:
1

Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
found to be significam
Cable 7.
of personality measure.
3). The highest average
bjects (43.93) and the
subjects (36.81). The
= . 0007). The differ-
nsecure and the com-
5).
tality combination was
ults of the two studies
shows an average score
subjects, and 4.61 for
.c CR of the difference
5). In the carlier (45)
· all six classes, which
-C-secure subjects had
re subjects. If the re-
ix gave results accord-
(23) where a "pure
there was an average
re subjects and 5.85
CR of the difference
with the E-C results,
& clairvoyance to the
recalled that a bat-
tered to 36 subjects
1 conditions (know-
-s runs under known
5 ** and emotional
nie inter-correlation
-ining them, there-
ich eliminates the
correlation between
: (R == +.65) was
For each factor. By
Ierive an equation
:If-confidence and
two measures to-
han either measure
= predicted group
'idually, however,
deviated sharply
at the magnitude
number of runs
-ence scores were
s in the session,
*2, after 12 runs
> found in Hum-
phrey's study on introversion-extraversion (16), and raises the question
of the optimal number of runs to be used in studies utilizing personality
measurements.
In a later study, based upon data gathered in the 1952 study and a
Liter 1953 series, Nicol and Humphrey (28) attempted to discover whether
subjects could correctly identify successful ESP calls. Subjects were re-
quested to place a check mark beside each call which they felt was a hit.
This, of course, was done before the subject was informed of his success.
The method used to evaluate whether an awareness of ESP had been dem-
onstrated was to compare the proportion of checked hits against
checked misses.
The authors reported that the 34 subjects represented in the pooled
Unknown runs were successful in identifying correct calls to a very
significant degree (P == . 0003). This effect did not hold up for the 22
subjects represented in the Known runs. Since only the Unknown runs
gave significant results, these alone were considered when the attempt
was made to discover if "conviction of success" was related to personality
factors.
Only those subjects who gave an average of five to ten checks per run
were included in any of the statistical evaluations. The checking success
of the confident and unconfident subjects were compared, and it was
found that the 17 unconfident subjects had a significant (P =-. 0006)
excess of checked hits over misses; the checking success of the 12 cmo-
tionally unstable subjects was also highly significant (P = . 002).
On the surface, these findings appear to have considerable theoretical
importance. If, on the basis of personality tests, certain groups of subjects
could be selected who "sometimes know when they're right," the pro-
gress of ESP research would be considerably advanced. However, there
are certain criticisms which can fairly be leveled at the experimental
procedure. For instance, it seems questionable to include only subjects
having an average of 5-10 checks per run in the overall evaluation. Because
of the well-known bias resulting from atypical scores in computing an
average, it would appear that a more appropriate measure of central
tendency, such as the mode, might have been employed to select sub-
jects. An interesting comparison would have been to present the overall
cvaluation in terms of all runs having 5-10 checks, rather than making
the subject the basic unit.
Another point deserving attention is that there appears to be some
grounds for assuming that checking behavior per se is a function of self-
confidence. Since the authors mention that quite persistent urging and
coaxing v as resorted to in an effort to obtain the desired 5-10 checks,
it seems reasonable to assume that subjects who were unresponsive to
such prodding could be considered as lacking in confidence. Yet it was
these same "unconfident" subjects who were excluded from consideration
when the role of confidence upon checking success was investigated.
Summary on Combined Personality Measures and ESP
In all the reports reviewed in this section, a higher degree of separa-
tion was obtained between subjects when combined rather than single
personality measures were used. This suggests that the expression of
ESP may be dependent upon a number of personality factors working in
combination and that the most profitable method of selection for ob-
taining high and low scoring ESP subjects would be to use a battery of
personality tests rather than single measures.
[27]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96-09787R000400100014-4
Some of these reports also give indications that through the use of
combined personality measures and more refined methods of statistical
analysis, it may be possible to show a relationship between personality
characteristics and amount of ESP, rather than merely sign of deviation, as
has been found in studies employing a single personality measure and a
simple statistical evaluation.
ATTITUDES OF BI
The most active worker i
ESP scoring level has been
(37) on an investigation in
of an individual interview.
possibility of ESP (sheep) (
The sheep classification was
undecided about the possil"
After the subjects had 1
testing environments were
The goats were placed in a
as many as 50 runs per te
periodic knowledge of result
working conditions, were g
of the ten runs in the testing
Three series of clairvoya
shecp and goats were tested
The next article (38) repo
of the sheep for the 3 poole
the 574 runs of the goats .
was significant (P == . 001)
not significant.
Since the sheep and the g
experimental conditions, !
lationship between the atti
since the possible effects o
results, attitude of the expe
the differences in scoring le
of these three series, therefe
tical evaluations made in th
It seems more appropriate
data from which the hype
score of the sheep would
than the average run score:
idered as providing an i
hypothesis.
Later Series
In all experiments inclur
were tested under identica
college students, and all w:
number of runs per subject
trader can find a full :
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
[ 28.]

Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
or a very negative response to the questions. 'Do you believe in the exist-
ence of ESP?' and 'Do you believe you possess ESP abilities?' If we split
our group into high-low categories, neither category including many
subjects with very positive attitudes, one way or the other, we find that
our results do not substantiate those obtained by Schmeidler and Bevan."
The results of these others workers' published data bearing on the
sheep-goat classification are collated in Table 10.
Inspection of Table 10 reveals that in 5 out of 6 cases, the sheep had
higher ESP scores than the goats; that the sheep, with one exception,
had positive deviations, while all the goats, with 2 exceptions, had
negative deviations. The non-sheep non-goat subjects who, for conven-
ience and for comparative purposes have been lumped together as in-
decisives, had deviations which showed considerable variation.
The consistency of the group scoring levels, which was reported in
one study only (55), is shown in Table 11.
Table 11
ESP Scoring Levels of Sheep and Goat Groups (Petrof)
Group
Subjects Scoring
above Chance
Subjects Scoring
below Chance
Totals
Sheep
11
7
18
Goats
1
9
10
12
16
28
P
18!
10!
12!
16!
+ 18!
· 10!
12!
16!
28!
.011
11! 7! I! 9!
28!
12!
6! 0! 10!
Table 11 shows that the majority of sheep scored above chance and
the majority of goats below chance. Since this pattern of scoring was
predicted from Schmeidler's results, only a one-tailed probability is
reported; this has a statistically significant value (1P == . 01).
Having reviewed all the researchos which can be considered as at-
tempts to repeat Schmeidler's findings, the question which needs to be
answered is "Can these studies be interpreted as confirmation of Sch-
meidler's findings?"
The crucial problem is obviously that of the criterion on which the
sheep-goat differentiation is to be made. Schmeidler herself changed the
criterion as her experiments progressed. In the series reported in 1943
(38), subjects were merely questioned as to their attitude to psychic
phenomena in general, telepathy and clairvoyance in particular; the
sheep were those who wondered if such phenomena would occur, or who
believed in their reality, the goats those who rejected the possibility.
In the tables presented in the report, however, the two categories are
labelled "open-minded" and "expect to score at chance". There seems to
be something of a contradiction here. The goats, who rejected the
possibility of ESP, would certainly expect to score at chance; on the
other hand, it is possible to imagine a sheep who accepts the reality
of ESP phenomena and who nevertheless expects to score at chance in
[ 34 ]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
the test situation. This could
In her later series, Scha
that paranormal success in
goats as those who denied ;
success under the condition
Schmeidler used essentially
items in the sentence com!
ject's attitude to the test sit
tion on his attitude of belici
Bcvan's criteron was soi
subjects whether they acces
not they were goats; if they
were demonstrated, they w
measured by the technique
"no" or "don't know", the
themselves on a continuum
a category of indecisives.
Schmeidler's work, the ind
In series A of his experin
believed in ESP (sheep),
or whether they disbelieve
subjects were asked three (
means?", "Do you believe
you believe that you you!
points out, question three
ject's belief in his own ESP
He reports that, in the limi
be compared with her resu
4.89 hits per run, and the
presented.
Kahn's criterion was wl
cally possible (1) in this |
stances. He found that one
here only", that is, in th
Table 10 as indecisives,
criterion, they should be
the "impossible anywhere
jects on whether they ex
below chance. This overl
however, treats this as a
the subject in the exper
Eilbert considered bot
in ESP and thinks he wi
LSP but doubts that he
who were doubtful abor
pletely or who gave con
is similar to Schmeidler':
Woodruff and Dale a
believe in the existence
abilities?", "I think my
chance', 'at chance', '
made no overall shecp-g
tionnaire. The subjects
terms of classification on
..

.
Approved For Release 2001/03/07 : CIA-RDP96 00787R000400100014-4
. 'Do you believe in the exist-
-ess ESP abilities?' If we split
er category including many
ty or the other, we find that
1 by Schmeidler and Bevan."
dished data bcaring on the
e 10.
Jut of 6 cases, the sheep had
sheep, with one exception,
s, with 2 exceptions, had
t subjects who, for conven-
een lumped together as in-
-iderable variation.
-els, which was reported in
»at Groups (Petrof)
Ets Scoring
- Chance
Totals
7
18
9
10
6
28
-
18!
10!
12!
16!
28!
12!
6! 0! 10!
scored above chance and
is pattern of scoring was
one-tailed probability is
c (IP == . 01).
ran be considered as at-
cstion which needs to be
as confirmation of Sch-
e criterion on which the
idler herself changed the
: series reported in 1943
eir attitude to psychic
rance in particular; the
"na would occur, or who
rejected the possibility.
, the two categories are
chance". There seems to
pats, who rejected the
core at chance; on the
who accepts the reality
is to score at chance in
the test situation. This could be a matter of confidence rather than belief.
In her later series, Schmeidler defined sheep as those who thought
hat paranormal success in the experiment was at least a possibility,
zaats as those who denied that there was any possibility of paranormal
"access under the conditions of the experiment. In her 1954 P-F study,
chicidler used essentially the same criterion, although some of the
items in the sentence completion questionnaire, used to rate the sub-
."ct's attitude to the test situation as such, furnished additional informa-
son on his attitude of belief.
Bevan's critcron was somewhat different. He first of all asked his
subjects whether they accepted ESP as an established fact. If they did
not they were goats; if they did, after laboratory methods of testing ESP
were demonstrated, they were asked, "Do you think that ESP can be
measured by the techniques just explained to you?" If the answer was
"no" or "don't know", the subject was disqualified. All subjects placed
themselves on a continuum from belief to disbelief; Bevan thus obtained
a category of indecisives. For the purpose of comparing Bevan's and
Schmeidler's work, the indecisives should be combined with the sheep.
In series A of his experiment, Caspar asked his subjects whether they
believed in ESP (sheep), whether they were undecided (indecisives),
or whether they disbelieved (goats). In the second series, however, his
subjects were asked three questions; "Do you know what the term ESP
means?", "Do you believe that ESP is a theoretical possibility?", "Do
you believe that you yourself have ESP ability?" As Caspar himself
points out, question three of the questionnaire, concerning the sub-
ject's belief in his own ESP ability, resembles most Schmeidler's criterion.
He reports that, in the limited part (Series B) of his experiment that can
be compared with her results, the sheep (sheep and indecisives) averaged
4.89 hits per run, and the goats 4.97; a more detailed analysis is not
presented.
Kahn's criterion was whether subjects thought that ESP is theoreti-
cally possible (1) in this particular experiment, (2) under other circum-
stances. He found that one group of subjects considered ESP "impossible
here only", that is, in the test situation. These have been entered in
Table 10 as indecisives, but, in accordance with Schmeidler's final
criterion, they should be included in the goat category, together with
the "impossible anywhere" group. Kahn further questioned his sub-
jects on whether they expected to score above chance, at chance, or
below chance. This overlaps with Schmeidler's initial criterion; Kahn,
however, treats this as a separate analysis, bearing on the confidence of
the subject in the experimental situation.
Filbert considered both those subjects who were rated as "believes
in ESP and thinks he will do well in the experiment" and "believes in
ESP but doubts that he will do well in the experiment" as sheep; those
who were doubtful about the whole thing, who rejected ESP com-
pletely or who gave contradictory responses, were goats. His criterion
is similar to Schmeidler's; his results may be fairly compared with hers.
Woodruff and Dale asked their subjects three questions; "Do you
believe in the existence of ESP?", "Do you believe you possess ESP
abilities?", "I think my results in the above experiment are 'above
chance', 'at chance', 'below chance'." Unfortunately, however, they
made no overall sheep-goat assessment on all three items of their ques-
tionnaire. The subjects' scoring averages can merely be presented in
terms of classification on each item singly.
[ 35 ]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96-85787R000400100014-4
In considering these various analyses, it appears that no strict answer
can be given to the question of whether Schmcidler's results have been
repeated. In the first place, her criterion was initially a shifting one,
and the criteria others workers used differed from hers, in some cases
considerably. In addition, there were differences existing in subjects
(high school, volunteers and college), differences in targets (ESP sym-
bols, IBM sheets), differences in number of runs per subject (4,5,6,8,12),
differences in ESP situation (clairvoyance and GESP), and differences
in the experimenters (seven different experimenters).
The question is an extremely important one, however, and some sort
of comparison, however crude, seems necessary. This is attempted in
Table 12 by fitting the various criteria to Schmeidler's as closely as
possible. Thus, since Schmeidler combined indecisive and sheep, in
Table 12 Bevan's, Petrof's and Eilbert's indecisives are combined with
their sheep. In Kahn's experiment, the indecisives were those who con-
sidered that ESP was "impossible here only," i.c. in the test situation.
These are included in the goat category in accordance with Schmeidler's
final criterion. Only that section of Caspar's results which he himself
claimed to be comparable with Schmeidler's results is included in Table
12. In the Woodruff and Dale experiment, no break-down is given for
the whole serics. Differentiation in terms of three items, each of which
partly includes the sheep-goat criterion, is presented here.
Table 12
Sheep-Goat Data of Other Workers
Sheep
Goats
Experimenter
Type ESP
Av.
Av.
Sub. Runs Dev. Score Sub. Runs Dev. Score
Bevan
GESP
Cl
20
232 +110
5.47
10
120
+2
5.02
Caspar
GESP
CI
4.89
4.97
Eilbert
Cl
37
185
+-39
5.21
4
20
-- 2
4.(2)
Kahn
CI
62
733
+-42
5.06
12
143
+13
5.000
Petrof
CI
29
232
+-1
5.00
10
80
-18
4.75
Dale and
Woodruff
(a)
CI
460
+20
5.04
1500
+-35
50;
(b)
CI
1040
-3
4.997
920
+-58
(c)
CI
1500
-9
4.99
460
-1-64
.
Inspection of the Table shows that in three cases the sheep (sher."
and indecisives) scored higher than the goats, in three cases the go
higher than the sheep. Although the various experimenters in !!
cases obtained successful discrimination of high and low ESP scor"
in terms of the sheep-goat criterion as each one defined it, these nr
not be regarded as repetitions of Schmcidler's results.
COMBINATIONS OF ROR
WITH ATTITUDES OF B
The Rorschach is a widely
ard cards, administered in a :
sponds by reporting what he
The underlying principle is t
such ambiguous material, the
self into the material. This s.
patterning of the subject's unt
some indications about many
he is rigid or flexible in his ap
sive, creative, anxious, intellec
A quantitive index of the si
through use of a check list de
inore check marks are given
to in an atypical manner, am
a single score representing the
In the ESP series, an intro
subjects then classified themse
completed 3 clairvoyance ru
their results as the target or
Proceeded until a total of 9
The group Rorschach test w
ink blots on a large screen. T
!rits.
The Rorschach records w
and subjects having 10 check
subjects with 11 or more ch
'a eliminate any possibility (
funcidler was kept ignora
'ren checked by an assistant
In preliminary work with
kaneidler noticed that w
with the sheep-goat rating, i
.. P scoring levels.
The poorly adjusted subje
that the difference between
the well adjusted subje.
"i'trin of well adjusted she.
('# well adjusted goats scor
" found in future series, an
: str Fall of 1945.
"When Rorschach data fre:
#wassents (41) were analyze
[36 ]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96 50787R000400100014-4
t appears that no strict answer
Schmeidler's results have been
n was initially a shifting onc,
ered from hers, in some cases
differences existing in subjects
Iferences in targets (ESP sym-
of runs per subject (4,5,6,8,12),
=: and GESP), and differences
erimenters).
t. onc, however, and some sort
-cessary. This is attempted in
to Schmeidler's as closely as
led indecisive and sheep, in
indecisives are combined with
idecisives were those who con-
ily," i.e. in the test situation.
accordance with Schmeidler's
war's results which he himself
's results is included in Table
«, no break-down is given for
of three items, each of which
presented here.
er Workers
Goats
Av.
Score Sub.
Runs Dev.
Av.
Scorc
5.47
10
120
++-2
5.02
4.89
4.97
5.21
4
20
-2
4.90
5.06
12
143
+-13
5.09
5.00
10
80
-18
4.78
5.04
1500
+-35
5.02
1.997
920
+58
5.06
1.99
460
-+-64
5.14
ree cases the sheep (sheep
, in three cases the goats
us experimenters in most
high and low ESP scorers
one defined it, these need
results.
COMBINATIONS OF RORSCHACH ADJUSTMENT RATINGS
WITH ATTITUDES OF BELIEF AND ESP SCORING LEVEL
The Rorschach is a widely used projective test consisting of 10 stand-
ard cards, administered in a set order; to these cards, the subject re-
sponds by reporting what he sees or what the blots represent to him.
The underlying principle is that in order to structure anything from
such ambiguous material, the subject must project something of him-
self into the material. This structuring is interpreted as reflecting the
patterning of the subject's unconscious needs and drives, thereby giving
some indications about many facts of his personality, such as whether
he is rigid or flexible in his approach to situations, whether he is impul-
sive, creative, anxious, intellectually ambitious, socially withdrawn.
A quantitive index of the subject's overall adjustment can be made
through use of a check list devised by Dr. Ruth Munroe (24). One or
more check marks are given for each Rorschach category responded
to in an atypical manner, and these check marks are added to obtain
a single score representing the subject's degree of adjustment.
In the ESP series, an introduction was given by Schmeidler and the
subjects then classified themselves as sheep or goats. The subjects next
completed 3 clairvoyance runs (a unit of 75 trials), and then checked
their results as the target order was read aloud to them. The testing
proceeded until a total of 9 runs had been completed in this fashion.
The group Rorschach test was administered by projecting slides of the
ink blots on a large screen. This was given cither before or after the ESP
tests.
" The Rorschach records were scored by Munro's check list method,
and subjects having 10 checks or fewer were rated as well adjusted, while
subjects with 11 or more checks were rated poorly adjusted. In order
to eliminate any possibility of bias when scoring the Rorschach records,
Schmeidler was kept ignorant of the subject's ESP score, which had
been checked by an assistant and then later double checked.
In preliminary work with 85 subjects from two earlier series (39),
Schmeidler noticed that when an adjustment rating was combined
with the sheep-goat rating, it was possible to obtain greater separation of
ESP scoring levels.
The poorly adjusted subjects scored at approximately the chance level,
but the difference between the sheep and goats became more marked
for the well adjusted subjects. She advanced the hypothesis that this
pattern of well adjusted sheep scoring higher than poorly adjusted sheep
and well adjusted goats scoring lower than poorly adjusted goats would
be found in future series, and large scale testing of this hypothesis began
in the Fall of 1945.
When Rorschach data from 250 subjects tested in 11 classroom cx-
periments (41) were analyzed, the difference in average run score found
[ 37]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
significant (P= . 0002) but
hance, thus confirming the
-'n in later experiments re-
cw article (32) presented a
experiments utilizing the
"ctober 1945 and December
3 in Table 13.
> Adjustment Ratings
No. Runs
Av. Score
3000
5.10
1879
5.17
1121
4.97
2205
4.95
856
5.10
1349
4.85
zen the average scores of
significant (P ===. 000003).
means of the poorly ad-
1.4).
and Goat Groups
jects Scoring
below Chance
Totals
85
209
91
150
176
359
7 (1 d.f.)
)001
tcd subjects arranged in
own indicates that when
were positive scorers,
hance scorers. The chi-
nly a one-tailed test of
ns were predicted from
COMBINATIONS OF RORSCHACH SEVEN SIGNS WITH
ATTITUDES OF BELIEF AND ESP SCORING
In an attempt to explore further the relationships between Rorschach
variables and ESP scoring, Schmeidler decided to analyze the 250 Ror-
schach protocols from her first work (41) for particular categories that
scemed to appear more frequently in the records of high and low scoring
subjects. She isolated 7 factors or signs whose presence in a subject's
record seemed to act as deterrents to ESP scoring.
If these seven signs are analyzed in terms of their interpretative signifi-
cance, three patterns of "response tendencies" seem to cmerge. A cold,
withdrawn, restricted attitude can be inferred from the presence of
F+%, Mr., and no shock; extreme impulsiveness or lack of emotional
control from the presence of CF+ and C+; and excessive, near-com-
pulsive mental activity or "quantity ambition" from the presence of
R+ and total movement-++. Thus, subjects who have even one of
these seven signs present in their record could be considered to have a
specific maladjustment which might prevent them from demonstrating
ÉSP.
After having empirically determined these seven signs from this
collection of 250 records, Schmeidler went on to gather new data from
other subjects to see if the seven signs would continue to show the same
relationship to ESP scoring. The two review articles (33, 34), which
report further testing with the Rorschach, indicate that absence of seven
signs continued to be associated with higher scoring, i.c., her data
show that sheep in whose records these signs do not appear score higher
than sheep in general, and goats from whose records the signs are ab-
Table 15
ESP Data of 250 Subjects from whom 7 Signs were Empirically Derived
Classification
7 Signs No. Subjects
No. Runs
Average Score
Sheep
Present
66
590
4.84
Absent
51
459
5.44
Goats
Present
62
559
5.09
Absent
71
638
4.73
sent score lower than goats in general. Table 15 shows the scoring levels
of the original 250 subjects from whose records the data were derived;
Table 16 shows the scoring level of 329 additional subjects whose rec-
ords were subjected to a similar analysis.
[39 ]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
.....

Approved For Release 2001/03/07 : CIA-RDP96-80787R000400100014-4
REACTIONS TO FRUSTRATION AND ESP SCORING
The Rosenzweig Picture-Frustration Study (P-F) is a projective tech-
nique used to obtain a measure of a person's reaction to frustration. It
consists of a booklet of 24 cartoons, cach depicting an unpleasant or
frustrating circumstance, such as missing a train, in which one person
makes a remark of frustrating significance, depriving or blaming the
other. The subject responds on behalf of the frustrated person. The
drawings are deliberately crude, having only indistinct facial features
and a minimum of background provided.
The test can be. scored for several different categories but so far only
three have been used for research in parapsychology. These three are
defined as follows:
Extrapunitiveness-refers to aggression overtly directed toward the
environment in the form of blaming some outside force for the frustra-
tion or of placing someone else under an obligation to solve the difficulty.
Intropunitiveness-aggression is expressed overtly by the subject against
himself in a martyrlike fashion with an acknowledgment of guilt or
shame, or by assuming the responsibility to clear up the situation.
Impunitiveness-aggression is evaded or avoided in any overt form,
and the situation is interpreted as being insignificant or no one's fault
or as likely to solve itself if the subject simply waits or conforms.
The first indication that the P-F might be a useful test in parapsychology
grew from a thesis study by L. Eilbert at CCNY. An article by Eilbert
and Schmeidler (7) reported that when the P-F scores of Eilbert's sub-
jects were divided into four quartiles, the differences between ESP scores
obtained by subjects in the first and fourth quartiles were suggestive
(P around .05). The correlation of -. 32 between extrapunitiveness and
ESP score was significant (P == . 01) but the correlation of +.28 for in-
tropunitiveness and +.22 for impunitiveness were only suggestive
(P = . 04 and .07 respectively).
Schmeidler (43) then attempted to see if similar results could be ob-
tained from analysis of P-F scores which she had obtained during several
years of testing. She had P-F scores for 446 subjects and obtained a
correlation of -. 09 between ESP scores and extrapunitiveness (P == ,03)
and a correlation of -+ -. 10 with impunitiveness (P = . 02). When her re-
sults were combined with Eilbert's, the correlation of -. 12 between
ESP scores and extrapunitiveness was significant (P -. 005), and the
correlation of +.12 with impunitiveness was also significant (P ==. 003).
These combined data were also analyzed by comparing the difference
in mean ESP score between the subjects scoring in the lowest 10% and
highest 10% of the Rosenzweig categories. The mean score of the Icast
extrapunitive (lowest decile) subjects was 5.20, while the mean score
of the most extrapunitive (highest decile) subjects was 4.86. This difference
in mean score was significant (P == ,01
impunitive subjects was 4.94, of the h
5.27; this difference in mean scores was
scoring directions were in all cases mi
than for the goats. In fact, most of the
tioned were independently significant
for the goats.
Despite the fact that significant co!
tween the P-F and ESP scores, the cor
the relationships measured would secin
might be expected since the P-F score
the subjects would respond to a mildly
life. This does not necessarily mea:
be expressed in an ESP situation. It
ideas as to how the subject interpreter
joyable experience, the aggressive ter
annoying situation would have little
ESP situation.
To test this assumption, Schmeidler
in a group setting with the P-F and
annoying the subjects found the ESP si
ance was based upon a combined scor
naire, a variation of the incomplete se
of a paragraph written on the subject'
sentence method contributed most he
Ratings were made along a 7 point
rating, the greater was the degree of
jects. Since the P-F scores were der
projection into a moderately frustrati
hand that only the P-F scores of sul
moderately frustrating would be co
annoyance ratings of 5 or 6 were sele
annoyed group.
Although the correlations between
266 subjects were in the expected
insignificant. However, when the
moderately frustrated subjects were
statistically significant for extrapuni
impunitiveness (r == +.21, P == . 0]
ently significant for the sheep, but
Schmeidler's interpretation of the
habitual response to mild frustration
and hostile while making ESP respe
mildly frustrating, and would there!
subjects who characteristically rea
punitive fashion would emphasize t
ately frustrating experiment and co-
force, make higher ESP scores. The it
correlation for the intropunitive go
virtue of being a goat, was probably
toward the experiment, he neverth
in a frustrating situation, and be
would take upon himself the respe.
He would, therefore, tend to minimiz
[42]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Pelease 2001/03/07 : CIA-RDP96-00787R000400100014-4
iter, and through securing positive
r in such a situation. Such findings
subject's scoring level in an ESP
know is how the ESP situation is
VALUE-RATINGS AND ESP
There is one article by Schmeidler reporting on the use of the Allport-
Vernon Study of Values (AVSV) in an ESP experiment (35). This test
indicates in which of six different value areas (theoretical, religious,
social, economic, political, or aesthetic) a subject seems to identify him-
self most. Scores are obtained in terms of percentile ranks and subjects
scoring high in one or two areas must necessarily score low in the remain-
ing oncs.
Although it had been found that sheep made higher ESP scores than
goats, it is apparent that the subjects' answers to the theoretical question
of whether ESP exists or not did not separate them into clearly distinct
groups with favorable or unfavorable attitudes toward the experiment.
Some of the sheep might find the experiment boring or irritating and
some of the goats might like competitive tasks and enjoy playing "guessing
games". Schmeidler had earlier suggested (44) that the sheep-goat
dichotomy would be most meaningful for subjects to whom theoretical
problems are important (that is, subjects with high theoretical scores on
the AVSV).
Table 19
ESP Data Arranged According to Percentile Rank on Theoretical
Scale of AVSV
Sheep
Goats
Diff. in
Percentile
No. Runs
Avc. Score No. Runs Ave. Score Ave. Score
P
All Subjects
504
5.30
455
4.93
.37
.002
Below 90
384
5.18
367
4.95
.23
.06
90 or Above
120
5.68
'88
4.85
.83
.002
95 or Above
40
5.95
24
4.38
1.57
.001
100
24
6.54
8
4.50
2.04
.006
The hypothesis stated before these data were gathered therefore was
that the difference in scoring level between the sheep and goats would
be greater for those subjects who had a strong theoretical orientation.
The problem of whether ESP could be demonstrated in the test situation
should then be one that takes on personal significance for these subjects,
since it is closely related to their systems of values or expectancies. Such
[ 45 ]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

1
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
subjects would presumably identify more closely with the purpose of the
experiment, that is, to show the presence or absence of ESP.
A total of 63 subjects from four different psychology classes were
tested in a classroom setting. Each subject was supposed to classify him-
self as a sheep or goat, make 8 ESP runs, and complete the AVSV. The
theoretical scale of the AVSV was then scored and subjects receiving a
percentile rank of 90 or above were considered to be theoretical subjects.
Table 19 shows the results of the various breakdowns which were made
to compare theoretical and non-theoretical subjects.
In Table 19 it is shown that the difference between the mean scores
of the non-theoretical sheep and goats was not significant (P == . 06),
but when the theoretical sheep and goats are considered, the difference
between their average scores is over three times as great as the difference
of the non-theoretical subjects (P = . 002). From the table, it appears
that the differences in scoring level continue to become larger as the
degree of theoretical orientation becomes more marked; the P values
associated with these differences are significant or highly suggestive. The
interpretation advanced is that subjects who place increasing emphasis
on theoretical values are able to exhibit a corresponding increase or
decrease in their ESP score.
Generally, the number of cases in cach category is too small for such
generalization. In addition, however, when the three categories (90 or
above, 95 or above, and 100) in Table 19 are considered as discrete
rather than continuous categories (ie., 90-94, 95-99, 100), as they should
be in any valid comparison of scoring levels, the differences in scoring
Table 20
ESP Data Arranged According to Percentile Rank on Theoretical
Scale of AVSV (Amended Figures)
Shecp
Goats
Diff. in
Percentile
No. Runs Ave. Score No. Runs Ave. Score Ave. Score
P
All Subjects
504
5.30
455
4.93
.37
.002
Below 90
384
5.18
367
4.95
.23
.06
90 or Above
120
5.68
88
4.85
.83
.002
90 to 94
80
5.55
64
5.03
.52
.06
95 to 99
16
5.07
16
4.32
76
.14
100
24
6.54
8
4.50
2.04
.006
.. .
level between the sheep and goats at each level of theoretical orientation
cease to be significant except in the case of the 3 subjects on the 100th
percentile. These amended figures are shown in Table 20. It is apparent
that although there are significant differences in scoring level between
theoretical and non-theoretical sheep and goats as groups, the im-
pressive progression of theoretical level with ESP scores does not stand up
under strict evaluation.
[ 46 ]
CONCL
From this review of the pe
studies, it seems that some pr
the personality characteristic:
subjects. As a generalization
somewhat extraverted, secure
ably disposed towards ESP, al
tend to score high, while su
tend to score low.
It was stated at the begin
propriate to review the ESP
two basic approaches of Hu
spects; on the one hand, in
other in the consistency of
In general, Humphrey mn
of questionnaires, or from a
qualities exhibited in drawir
either by herself or by other
although she did have some
rived from the ESP materi:
reuter and the Stuart Inter
It is generally recognized
limitations. Regardless of il
be remembered that Hump>
"surface" traits like expr
measuring instrument itself
tend to give rise to spurious
the well-known "halo" effec
Bernreuter and Guilford-X
factor of the attitude of th
condition his responses to a
A second factor is the te
shown to affect responses o1
a similar influence on secu
apply particularly to the
the fact that some subject
from expansive to compre
would, presumably, chany
unreliability lies in the fx
"A drawings displayed 201
Mi probably the explanatio
with such scales as the M1:
general explanation appca
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
.
......

Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
y with the purpose of the
nce of ESP.
psychology classes were
supposed to classify him-
omplete the AVSV. The
and subjects receiving a
Lo be theoretical subjects.
lowns which were made
bjects.
between the mean scores
t significant (P = . 06),
considered, the difference
as great as the difference
om the table, it appears
10 become larger as the
e marked; the P values
r highly suggestive. The
Lace increasing emphasis
rresponding increase or
Dry is too small for such
three categories (90 or
e considered as discrete
-99, 100), as they should
e differences in scoring
Rank on Theoretical
gures)
Diff. in
Score Ave. Score
P
3
.37
.
.002
15
.23
.06
15
.83
.002
3
.52
.06
2
76
.14
0
2.04
.006
f theoretical orientation
3 subjects on the 100th
"able 20. It is apparent
scoring level between
ts as groups, the im-
cores does not stand up
CONCLUDING REMARKS
From this review of the pertinent data of most of the ESP-Personality
studies, it seems that some progress has been made towards determining
the personality characteristics of groups of high- and low-scoring ESP
subjects. As a generalization, we might judge that subjects who are
somewhat extraverted, secure, temperate, well-adjusted, who are favour-
ably disposed towards ESP, and who have a high theoretical value system
tend to score high, while subjects who possess opposite characteristics
tend to score low.
It was stated at the beginning of this monograph that it scemed ap-
propriate to review the ESP-Personality research in two sections. The
two basic approaches of Humphrey and Schmeidler differ in two re-
spects; on the one hand, in type of measuring instrument used, on the
other in the consistency of the results achieved.
In general, Humphrey made her personality assessments by means
of questionnaires, or from a more or less objective estimate of certain
qualities exhibited in drawings. Her results were usually not repeatable
cither by herself or by other experimenters working along similar lines,
although she did have some repeated success with the E-C rating de-
rived from the ESP material itself, and partial success with the Bern-
reuter and the Stuart Interest Inventory.
It is generally recognized that the questionnaire method has severe
limitations. Regardless of the stability of the factor itself, and it must
be remembered that Humphrey was largely concerned with transitory,
"surface" traits like expansion-compression, security-insecurity, the
measuring instrument itself is subject to irrelevant influences which
tend to give rise to spurious measurements. In self-rating scales, there is
the well-known "halo" effect, and the amount of "halo" in such scales as
Bernreuter and Guilford-Martin is considerable. The strong general
factor of the attitude of the subject to the experimental situation may
condition his responses to a considerable degree.
A second factor is the temporary mood of the subject. This has been
shown to affect responses on the Bernreuter scale, and it probably exerts
a similar influence on security-insecurity assessments. It would seem to
apply particularly to the expansion-compression ratings, judging from
the fact that some subjects rated by one judge were found to change
from expansive to compressive in the one experimental session, and
would, presumably, change from day to day. An additional source of
unreliability lies in the fact that ratings by two judges on the same set
of drawings displayed not a great deal of consistency. The second factor
is probably the explanation of the non-repeatability of the E-C studies;
with such scales as the Maslow and Bernreuter, however, the first, more
general explanation appears more pertinent.
[ 47 ]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4

Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
Schmeidler generally used attitude classifications and projective
techniques. She obtained consistent results, and her experiments were
generally repeatable. Insofar as the sheep-goat classification is con-
cerned, however, the question remains of precisely what factors are in-
volved in this differentiation. In the first place, is it possible for a subject
to give an unequivocal answer to the question of his attitude towards
parapsychology, which is a multi-dimensional subject? Ile may accept
one aspect of psi (telepathy, for example), and reject another (clair-
voyance, for example); in such a case, differentiation must obviously
be made along these lines. Further, it is possible that in addition to the
theoretical acceptance of ESP other factors such as confidence, interest
in the experiment, and willingness to co-operate might be concerned
in the sheep-goat differentiation. If these additional factors are involved.
the subject's answer might merely reflect much deeper multiphasic
motivational factors.
Concerning the personality measurements obtained from projective
tests, it is generally agreed that the factors measured on Rorschach and
the P-F Scale are basic fundamental aspects of personality structure.
Because of the endurance of this structure, one would expect to get
repeatability of differentiation in terms of Rorschach and P-F criteria
providing the tests themselves arc reliable. When we describe separa-
tion in terms of Rorschach or P-F variables, we are describing a somewhat
gross estimate in cach case, and it scems reasonable enough to assume
that the Rorschach estimate of adjustment and the P-F estimates of
extrapunitiveness and intropunitiveness, in their gross evaluation, arc
reliable enough measures. Since there has been repeated success in dis-
criminating high and low scorers on the basis of these criteria, we imply
that there is a relationship between these deeper factors and E SP.
It must be remembered that in all ESP experiments, the role of the
experimenter is a vital one. A factor which might contribute to
consistency or lack of it in any series of ESP experiments is the delicate
experimenter-subject relationship. The effect of such a factor is very
difficult to estimate, as it involves the personalities of the experimenter
and the subject, and their interaction. In considering this problem of
consistency of results, however, cognizance should be taken of the possible
effects of such a factor.
It must be emphasized that at this stage of ESP-personality research,
. more successful predictions of ESP scoring levels have been made on a
group than on an individual basis. Certainly the greatest amount of re-
scarch effort has been directed towards differentiation of scoring levels
on the basis of single personality measurements. This is a separation in
terms of direction rather than amount of deviation, and as such, is
generally not discriminating enough for the purposes of individual pre-
diction. For example, though Schmeidler's poorly adjusted group, as
a group, scored around chance, the variation in range of individual
scores, from very high to very low, was statistically significant.
Better prediction of direction of group deviation has resulted from the
use of combinations of personality measurements, rather than single
dimensions. Evidence for the efficiency of such combinations is offered
by Humphrey with combinations of E-C and Interest ratings, and E-C
and Security-Insecurity ratings, by Schmeidler with combinations of
sheep-goat and adjustment criteria, sheep-goat and "absence of seven
signs" criteria and sheep-goat and value ratings and by Nicol and Hum-
phrey with a combination of confidence and emotional stability factors.
[ 48 ]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
These combinations permitted
measures used in isolation.
Schmcidler's AVSV study is
the sheep-goat attitudinal class
linear relationship between ESH
orientation. Although no stric
predictions were made for grou
(numbering 1-5). Onc must poi
with theoretical orientation is
criticisms notwithstanding, this
this area.
Of major importance is the
some success in predicting ind
personality ratings, using inu
level of success reported is not
the approach most promising
In the final evaluation, it aĮ
unique factors in a subject's
possesses marked tendencies
stimulated to competition, it
predicting the direction, and,
ESP deviation. The question
characteristics possessed by th
similar in kind to those posses:
above chance, and whether
therefore, might reasonably
the characteristics possessed
to be one of the major proble
The answer may well come
intensive study of the person
jects, and direct comparison
displayed by groups of subje
on the other, from develop
techniques for selecting indiv
levels, solely on the basis of
tests and assessments.

Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4
cations and projective
d her experiments were
at classification is con-
sely what factors are in-
is it possible for a subject
of his attitude towards
subject? He may accept
d reject another (clair-
ntiation must obviously
> that in addition to the
h as confidence, interest
itc might be concerned
mal factors are involved,
ich deeper multiphasic
btained from projective
surcd on Rorschach and
of personality structure.
ic would expect to get
schach and P-F criteria
ien we describe separa-
re describing a somewhat
nable cnough to assume
id the P-F estimates of
ir gross evaluation, are
repeated success in dis-
" these criteria, we imply
factors and E SP.
eriments, the role of the
ch might contribute to
periments is the delicate
f such a factor is very
ties of the experimenter
idering this problem of
I be taken of the possible
SP-personality research,
have been made on a
greatest amount of re-
iation of scoring levels
This is a separation in
tion, and as such, is
oses of in'dividual pre-
ly adjusted group, as
n range of individual
ly significant.
has resulted from the
s, rather than single
combinations is offered
rest ratings, and E-C
with combinations of
d "absence of seven
d by Nicol and Hum-
ional stability factors.
These combinations permitted greater differentiation than any of the
measures used in isolation.
Schmeidler's AVSV study is a further step in this direction. Once
the sheep-goat attitudinal classification was known, there appeared a
linear relationship between ESP scoring level and degrec of theoretical
orientation. Although no strictly individual predictions were made,
predictions were made for groups which in some cases were very small
(numbering 1-5). Onc must point out that the progression of ESP scores
with theoretical orientation is not as impressive as it appears; these
criticisms notwithstanding, this study is an important contribution in
this arca.
Of major importance is the study by Humphrey and Nicol reporting
some success in predicting individual ESP scores from a knowledge of
personality ratings, using multiple regression analysis. Although the
level of success reported is not high, the method is a valuable one, and
the approach most promising.
In the final evaluation, it appears clear that if something is known of
unique factors in a subject's personality make-up, if, for example, he
possesses marked tendencies towards social participation, or is easily
stimulated to competition, it is possible to utilize this information in
predicting the direction, and, to a much lesser degree, the amount of
ESP deviation. The question still remains of whether the personality
characteristics possessed by the rare individual high-scoring subject are
similar in kind to those possessed by groups of subjects who score slightly
above chance, and whether the relative difference in scoring level,
therefore, might reasonably be attributed to differences in amount of
the characteristics possessed or to motivational factors. This appears
to be one of the major problems in this arca of ESP personality research.
The answer may well come from two sources-on the one hand, from
intensive study of the personality makeup of the few high-scoring sub-
jects, and direct comparison with what is known of the characteristics
displayed by groups of subjects who score positively, as a group, and,
on the other, from development of better experimental and statistical
techniques for selecting individuals and predicting their probable scoring
levels, solely on the basis of measurements on a number of personality
tests and assessments.
- -- 41
[49]
Approved For Release 2001/03/07 : CIA-RDP96-00787R000400100014-4