TESTING FOR THE CRYPTOLOGIC ASPECTS OF ESP (AN INITIAL REPORT)

• Document Type: CREST
• Collection: STARGATE
• Document Number (FOIA) /ESDN (CREST): NSA-RDP96X00790R000100030056-4
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 10
• Document Creation Date: November 4, 2016
• Document Release Date: December 1, 2011
• Sequence Number: 56
• Publication Date: January 12, 1969
• Content Type: REPORT

Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
Li
(b)(1)
J~;t:~r -~.?t ~ 2 JAN 196$ (b)(3)
SUBJECT: Testing for the Crrptologic Aspects of ESP (an initial report)
I. PURPOSE: ^
, P I>.
The purpose of the ESP experiments herein described iv threefold:
A. To determine that effect hunMnn extra-perceptive capabilities,
temporarily categorized as telepathy and clairvoyance, could
have on the CONSEC posture of the United States. The prime
target areas being considered are:
(1) Code and cipher keying materials and encrypted text
(2) Underlying plain text of.encrypted messages
(3) Knowledge of message originators or processors
B. To determine what factors enhance the phenomena known as
telepathy and clairvoyance and :?rhat factors detract from its
operation. Factors thought by researchers to contribute to
perfor^snce success are listed in appendix A. Those included in
preliminary tests and fully controlled tests are indicated by
asterisks. It is hoped that knowledge of such factors will
provide useful defensive security k-.no::lcdge, should a COIMS21;
threat be realized in this area.
C. To find talented subjects so that A and B above can be
deterr:iined.
a r=
It was decided at an early stage in the project that 'combinations of
code groups and th^.i r meanings ,?.ould be used as tar ;et material for 1':;,'.c ".'f
the tests. The reasons for their choice were:
A. It made the tests related to the Agency mission. ,A?-fr1 ~ti/ ~,~ u~ jo- e.. '
'J At -'t5~,.:.p g 1? i~~JS'T V l~it+an t!r/; :ys;~W:Rs r^! >wa --?Ar1.~" o~a~ta~7'j-us CdS,~i~
B. By using coribinations, it would be possible to obtain 100
guesses out of only 50 transmissions. It was considered that
the resulting abbreviation of each testing session :could inhibit
the decline factor *:-hich has so often detracted from the results
of previous parapsychological tests.
personality of the subject.
C. Subjects could be convinced that the test involved a real
Agency problem ;,hich L-.^redia:ely increased interest and attention.
D. The material permitted flexibility: Symbols or meanings
having idios~ cratic significance for he subject could be
isolated, fhich pe .fitted s--me tailoring of the test to the
-4U *40ta-014
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
iVC,M :{crag ,t+a~.,t e~ _ r e.~-aacl'W
E. Code group -moaning provided a constant context, a fixed
environmental condition or relationship.
F. It was not difficult to adapt the material to statistical procedures
already established in prior psychological, and parapsychological
research.
III. PRELLl4ThARY TESTS - (Purpose.Iiethod)
The purpose of the preliminary tests was to test the test, and
thereby, to ascertain if it were a useful tool. Those who acted as
percipients (receivers) in the preliminary tests were young college
graduates, about 40`; of them male and about boo female. The groups
tested were from virtually every state in the union. The attitude
toward the test was typically one of considerable interest combined
with an ambivalent hope-fear attitude o:? individuals toward exhibiting
the phenomena themselves.
A. Randormess - The order of the 50 transmissions was put into
a random matrix such as that in appendix B1. The random pattern
itself was chosen by the tester from available Agency produced
crypto-.material. (Since only 5 possible combinations of code
groups and meanings :'rere used and each future combination was
automatically designated A. B, C, D. or E. these letters rather
than the actual combinations themselves were randomized). The
randomized matrix was kept in a safe prior to the teat to preclude
access by anyone, e:.ceptr the tester and the transmitter themselves.
The combinations were randomized at the last minute as a further
'safeguard. (appendix B2).
B. Conditioning - The peecipients were given an ESP pep talk
before the test to heighten their enthuziasn, and &-i at-Los--pt t--as
made to have t h.-_ studento verbalize theii thoughts on the subject
in order to enhance th- conditions of nor-pal. cost unicatior o in
the hope that channels of unusual co: F runicati ons would thereby
also be enhanced. Students were encouraged to choose code group
meanings c.=hi--h had significance to them. Although in one session
the' transmitter was permitted to establish the timing of the
transmissions, this was found to make the test too lengthy. It
was finally decided that transmitting at a steady 10 second interval
was most comfortable for all concerned.
C. Physical conditions and controls - During the best controlled
preliminary test, the percipients were located in a room about 40
feet away from the transmitter who was in an outside hallway in
an alcove (see appendix C). There was a monitor posted with the
percipients and with the transmitter. Timing was passed by word
of mouth from a monitor who was stationed with the transmitter.
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
D. Experimental Controls - In order to separate those
have a more active ESP capability from those who might
simple test was designed with no stringent controls. 7
are the same as those of the Rhine cards, but in this i
5 love-hate syi bols (cross, swastika, heart, U.S. flag,
dove)- were substituted for the usual sy 'bols. It was ac
decided that any one who achieved 12 or more correct gue
9 or less correct guesses would be included in the exper.
group. It is interesting to note that the two symbols :?rL
received best, even by subjects who only guessed 5 or 6 c
.were the cross (22% of the correct guesses) and the ?U.S.
(2O of the correct guesses), followed by the swastika (1E.
the correct guesses).
E. Results - In this test, each combination of code group',
code meaning eras counted as two guesses. The nean chance e.
(A1CE) for 100 guesses where the possibilities are 1 in 5 is
correct (although it developed that the actual mean for 35 s~
and 3500 guesses was less than 181). The expected standard d
(SD) is plus-or-minus 4. All of this indicates that those sV
who scored more than 24 correct guesses show, a deviation in f=
of the ESP hypothesis. Correct co-binations ware also studied
statistically. The same laws apply; however, the expectance is
this case is for the IM ."A to be 2 correct with an SD of 1 in eit
direction.
(1) The 12 members of the ex-periaental group showed a tot?
positive deviation of 26 hits above the expected SD of 60 for 12
people.
(2) The 10 members of 'the control group showed a total posit
deviation of 5 hits above the expected SD of 50 for 10 people. b
control group also showed doable the total negative deviation of
the positive group below the SD--a minus 8.
(3) Within the experimental group there were 4 individuals *1.,nc
scored double or more the number of correct hits expected. One of
these individuals eras who scored 29 correct guesses
on the code groups or meanings more than two times the standard
deviation: This subject also scored an unusually high number of
hits on gue33ing combinations -- 6 instead of the expected 2.
(!t) Within the control group no e:itraordinary positive deviation
-occurred except where one subject guessed 5 combinations correctly;
however, 5 fairly significant nogative deviations from ME were
scored: two of ;onus 7 and three of minus 5.
rsl% ?r~.e.r?sa
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96X00790R000100030056-4
(3) Timing was provided by using synchronized electric clocks
activated by a single short phone call from the tester who stayed
with the transmitter (under :monitoring) . The ten second -interval
was maintained by the monitor at each end verbalizing each nu.?nbered
interval.
(!t) The tests were marked by
at the receiver site.
(5) 5 or 10 warm-up questions were used at the insistence
.of the receiver.
(6) In test two the receiver was kno,m to be having a
personal disagreement with the transmitter but the test was
to continua in order to observe the effect this might
(7) In test three another trans. was added to the first
transmitting in hopes that this offset any negative attitudes
that might have carried ., from the conditions existing under
test Ni'. 2. Not this transmitter was retained because
miss '.ad such good rapport with her during the prelLminary
tests. The receivers room was subjected to jackhamI?let-like noises.
The transmitter site was also subject to temporary disruptive noise
(door knocking).
(8) Test three was conducted under extremely adverse conditions.
The receiver was under constant scrutiny by 2 guest obserTers.
CONTFOLLED TEST :'ESULTS :
A. Under prelI:dna test r' s 7 ns where test conditions were not
strictly controlled, hiss scored a total of 80 correct
guesses of rode group3 and code :awnings (individually tabulated)
out of a total o+ 00 au~:sses. :.ith a standard deviation i'r i !C
being 60, .Miss scored hors than Zt tines the expected deviation.
'Similarly the subject scored correctly on 19 co-bi.Zation -where 'N
is 6 and the SD 2 or :acre than 6 tLmas the expected rate.
B. Under fully cc:itrolled conditions, :is!; 0 still continued
to score si,-nificantly but lover than under the preliminary tests.
' ? ! i t h the s :.us~:bBr of total guesses involved as in the prelir..inarf
gists i:iss ~: ade 77 correct guesses for individual code group
id meaning on about 2.7a tines the expected rate. The subject also
teased 10 ao.ibinations instead of the expected 6 or twice the
iected rats.
4s.I_ lrss,r t F ''n .mss. ~~
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96X00790R000100030056-4
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96X00790R000100030056-4
r.fr..7.~ ~ea
U NU i.%740
C. It should be noted that Miss ~ scored highest in both
categories on the last test (th3 most difficult in terms of conditions).
On this occasion she was asked if she thought she had gotten any one
combination correct nere often than any other. She correctly replied
SCO-attack which constituted 4 of the 6 correct combinations vhich
she pss:;d on the third test. (In two preliminary tests
miss had duplicated this feat. In both instances the code
meaning had been the turd retreat. In the first such instance
Miss guessed 6 combinations correctly, 5 of which contained
the meaning "Re-treat". The total combinations involved were the
usual 50. Under sLmilar conditions 1? iss 0 again guessed 5
-combinations correctly, 4 of which contained the meaning "Retreat").
D. It should be noted that on the occasion of her final test,
-Miss 0 was asked if. anything disturbed her during the test.
(The transmitters had estimated that around question 2I there had been
a disturbing series of knocks on the door of their room. This
question was checked off on the answer sheet). Miss 0 replied
that around question 20 she felt like giving up -- an intriguing
coincidence, the exact same thoughts.were going through the trans-
mitter's mind at that time.
VII. CONTROLLED TEST CONCLUSION : -
A. The test methods used, if used carefully, are probably effective
for the isolation and further testing of some persons who have a
measurable ESP capability.
B. Since Hiss I scored more significantly under conditions
which were more competitive and encouraged greater intragroup come .vni-
cation, these conditions should probably be given more emphasis in
future tests.
C. n three tests (t::o preliminary and one fully controlled)
Miss seeped to display eft affinity for a particular code
meaning. If such an affinity should prove to have any durability,
it could be of measurable benefit to c yptanalysis .
D. Since Miss scored more significantly when there were t:o
transmitters, the possibility of enhancing percipient performance by 0 of
using groups of transmitters should be given greater emphasis. In
this connection, it might also be beneficial to score groups of
receivers in a collective manner, i.e.', if 3 out of 5 people guess
combination A. it is considered the group answer.
R. It is considered that other test methods involving more realistic
situations might measurably enhance test results. It appears that
personality clashes between transmitter and receiver tend to cause
a lack of interest on the part of the persons concerned, which may
result in a near average performance by even talented subjects.
?6
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96X00790R000100030056-4
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96X00790R000100030056-4
F. Miss ability to produce near random guesses of
target. ma eria appeared to be considerably above average.
G. The above test results indicate that the hunan ESP capability
could be a significant threat to U.S. COMSEC in all of the
categories indicated in paragraph IA. For this reason the National
Security Agency cannot afford to remain ignorant of the scientific
advances in utilizing or developing this field.
VIII. RECOI?ZIE DAT ION : `
A. It is strongly recommended that the Agency seek to acquire all
available information on the subject of ESP research and that our
own modest testing effort be permitted to continue and even to
expand slightly in the direction of. greater subject detection and
testing.
B. It is recor. ended that the Agency's psychological testing
program be amended to include a test such as that reported above
which would serve two purposes.
(1) To isolate ?super guessers" or, personswc;ho have a knack
for cryptenalysis.
(2) To isolate persons who have a high ESP potential.'
C. It is recommended that future ESP tests concentrate on the
collective testing and scoring of groups of receivers and
transmitters.
D. It is recoirmnded that some real material be included for
ESP targeting in future tests.
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96X00790R000100030056-4
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
APPENDIX A
Factors Considered to Benefit the Operation of ESP Tests:
xcc 1. Creating a realistic situation.
?xxx 2. Getting transmitter and receiver to concentrate on sanding and
receiving with each person considering the other as the intended target.
xxx Conditioning the subjects to react to the material properly
to consider the test -important.
xx !t. Testing subjects who have demonstrated an ESP capability by some
degree.
ooc 5. Give feeling of success to sender and receiver by praising positive
and negative deviations.
xxx;. 6. Transmitting of targets in some sort of context or natural
environment.
xx ?7. Limiting each testing session to a short period.
xxx 8. Building on success by eliminating unsuccessful materials or
conditions and conserving successful ones.
acx 9. Trying to open as many channels of normal hu*+.en comminications
as possible - and opening them wide!
aoc 10. Synchronization of sender and receiver.
: 11. Using groups of transmitters.
12. Testing groups of receivers in song collective-way so as to
3cover even slight trends in res7onses which defy chance.
Included in proli*minary tests x
Included in controlled tests ,cc
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
rte. ry ? -%c-ra.ata
. I. Initial Random I;atrix of code group-code meaning combination
positions derived f ro,ti Agency Random sequences.
Position of Transmitted Group Agency Random Sequence
1 A = LJTA - Women and
Children
2 E.
3. B
C
5 A
6 A
7 . A
50 B
II. ? Combinations receive their A. B, C, D, and E designations by means
two more random sequences used to mate the meitbers of the combination,
below.
.II. After an arbitrary A, B, C, D. or 4 designator is appended to each
group and meaning, a random sequence is used to determine how they are'
mated. A separate rardon sequence is used to rearrange the order of
ode groups and code neaain s. If for example, E ?s Iould. be the first of
ive letters to occur in the randora sequence assi, ted to code ;groups,
'elovr) would be tha first partner of Combination A. Similarly, if B
o be the first letter to occur in the random sequence assip.ed to the
!aninss, then. TUor:en and. Children would be the code meaning assigned to
.lion A. The same nathod. is used to determine the remainder of
the
tic; s to be transmitted.
SCO Attack A Code.'Combination A =
WA Women and Children
14RV Target
BXY ' Killed C
PId: Women and Children D
W'TA Intelligence. E
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
L=:r'1 two a ti+:a ti:Jra
Participants of the AD/HOC testing group
include:
Mr.
Dr.
Mis
Mrs
Aar -'^.I~ .~'~ ~:/Tn7M?/
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01 : NSA-RDP96XO079OR000100030056-4
 Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4
,
PRELPI.I3IAPY TEST
Sr ARP ,.
APP C
21
Declassified in Part - Sanitized Copy Approved for Release 2011/12/01: NSA-RDP96XO079OR000100030056-4