CIA-RDP96-00792R000400030001-0

Approved For Release 2000/00/1
Experimental Dream Telepathy-Clairvoyance and
Geomagnetic Activity
Michael Persinger, PhD, Laurentian University, Sudbury,
Ontario, Canada, and Stanley Krippner, PhD, Saybrook Institute,
San Francisco, California, U.S.A.
Introduction
Several
studies
involving putative
spontaneous
telepathic-clairvoyant
(T-C) experiences concerning
death or crisis to family members or to friends
have suggested a geomagnetic contribution
to
these phenomena.
There is some
evidence (e. g., Schaut & Persinger. 1985)
that
spontaneous
experiences
have occurred when the geomagnetic activity was
lower (i.e., calmer) than the
days before or afterwards
and
lower
than means of the monthly values.
decided to
study whether
this
pattern
WAS
evident
in
experimental
T-C
experiences as well.
The experiments
in
T-C dreams
grew
out of Montaque Ullman's
(1969)
communication
appearing
in the context
We
observations,
in
his
psychiatric
practice,
of
ostensible anomalous
some
dreams
reported
by
his
clients.
Ullman decided
to
inquire whether this phenomenon would appear in
dreams that were experimentally monitored
under
controlled
conditions
that
would seem to exclude alternative interpretations.
project by Stanley Krippner,
NOW
professor
of
psychology
at
Saybrook
Ullman was
joined in this
Institute,
San Francisco,
and Charles Honorton,
now director
of
the
Psychophysical Research Laboratories, Princeton,
New Jersey.
The experiments
were
carried
out
in the Dream Laboratory
of
Maimordes
Medical
Center.
'459
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0

Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0
Brooklyn, where Ullman served
director of
the
Community Mental Health
Center.
The typical procedure followed at
Maimonides
was for the percipient (or
subject) to arrive at the laboratory in time to meet the agent -- a person who
would soend much of the night focusing upon the contents of an art print.
The
percibient's
task
was to dream about this art print even though it would
be
selected once the percipient was isolated
from the agent.
The percipient
would also meet the two experimenters who would explain the procedures. (On a
few occasions, the art print was selected randomly, was not removed
from
the
sealed envelope, and no agent was used.
The percipient was simply instructed
to attempt dreaming about the art print. )
After electrodes
were
to
attached
the
percipient's
head
for
the
monitoring
of brain waves and eye movements, the percipient
a random number table, provided a number
that corresponded to one
on
a
sealed envelope containing an art print.
The
envelope was opened once the
agent
reached
his or her private room in a
distant part of the building.
This art print became the target on which the
agent focused during the course of the night (Ullman & Krippner, 1978).
The experimenters took turns monitoring the percipient's sleep.
Toward
the end of each period
of
rapid eye
movement
(REM) ,
the percipient
was
awakened by an experimenter via intercom and described any dream content that
could
be recalled.
These comments were
tape recorded
was a
morning
interview in which the percipient associated
to his or her dream recall.
The
interview
was
conducted double blind;
neither the
percipient
nor
the
460
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0
would
have
no
further contact with the agent
until
the following morning.
An experimenter
threw dice that, in combination with

Approved For Release 2000/08/11 : CIA-RDP96-0079ZKOU
experimenters knew the identity of the target nor the pool of art prints from
which the target had been randomly selected.
The target for a given night and the dreams for the night often contained
a number of striking similarities, suggesting that an anomaly (so-called
"telepathy" or "clairvoyance") occurred. For example,
on 23 May 1966
the
target was a print of a zebra painted by an unknown Indian artist.
The
percipient dreamed about a "horse show, " a "horse race, " and a "striped tie. "
But it could have been the case that any transcript of a night's dreams might
have contained passages of striking similarity to any picture
to which they
might have been compared (Child, 1985).
To evaluate the chance hypothesis, the Maimonides team obtained judgments
of similarity between the dream
content and each of the other potential
targets in the pool from which the actual target had been randomly
selected.
Typically, three judges were used who worked blind and incepencently from each
other with materials that
had
been
mailed to them.
They had no information
about which picture had been randomly selected as the target. Any extrachance
difference between targets and non-targets
in their
similarity to dream
content was considered an apparent anomaly.
Typically, the target pools used
by the judges were duplicates that had never been handled by the agents.
Although percipients sometimes evaluated their
own
creams
against
the
target pool (before they discovered the identity of
the
actual
target), and
although some experiments required
the
Judges
to
rate
target /cream
Similarities on a 100-point scale, the only form in which data
are available
for all sessions is a count of judges' hits and misses.
If the actual target
had been ranked in the upper half of the target oool (e.g.) #1, #8, #3 in a
461
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0

Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0
4
of
six)
for
similarity
to the dreams and post-sleep
interview, the
outcome was considered a hit.
If the actual target had been ranked in the
lower half of the oool
(e. g. ,
#4, #5, #6 in a pool of six), the outcome was
considered a miss.
The median score of
the
three judges was selected
to
determine hits and misses.
For
the
purposes
of
this
study,
the
ranks
were divided into four
categories. A "high hit" would
be
a rank in the top Quartile (e.g, #1 or #2
in a cool of eight; #1 in a pool of six), a "low hit" would be a rank in the
second
quartile (e. g. ,
#3
or
#4 in a pool of eight; #2 or #3 in a pool of
six). A "high miss" would be a rank in the third quartile (e. g., #5 or #6 in
a pool of eight; #4 or #5 in a pool of six); a "low miss" would be a rank in
the fourth quartile (#7 or #8 in a poll of eight; #6 in a pool of six).
The first night' each
subject spent
at
the
Maimonides Laboratory was
utilized and
the other nights were discarded.
This yielded 62 experimental
T-C
nights available for analysis -- 18 "high hits, " 29 "low hits, " 7 "high
misses, " and 8
"low misses. " The 62 cases represent
an
almost total
collection of subjects seen between 1962 and 1969 at Maimonides.
Procedure
It was decided to use geomagnetic activity as measured by the AA index.
Northern hemisphere AA values
were
collected for the 3 days before, the day
of,
and the
3
days
after the day that each subject began
the dream
experiment.
These AA values were determined from Mayaud's (1973)
data and
consequent monthly uodates.
Mean monthly values were also
listed.
Subjects
(cases) were coded according to gender and to the closeness of their dreams'
462
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0

Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0
correspondence to the target (i.e., "high hit" or HH; "low hit" or LH; "high
miss" or HM; "low miss" or LM). In other words, these four groups represented
judges' ranks
of
successive order from strongest
"hits"
£
to
strongest
"misses. "
The major design involved MANOVA (multiple analyses of variance).
The
repeated measure was the AA values for the three days before, the day of, and
the three days after the day the experiment was held (a total of 7 days). The
range of the key day plus/minus 3 days was selected
because geomagnetic
activity within this period tends to be correlated, particularly with a given
day plus/minus
1
or 2 days.
Exceot
for
specific
periodicities,
the
intercorrelations between
geomagnetic
activity
on a day and more than three
days before or afterwards
1
are not statistically significant. "The two major
main factors (non-repeated) were gender (male vs.
female) and grous. Because
the numbers of subjects within the HM and
LM
groups
were small
(N=7,
8
respectively),
additional
analyses were
completed
with
these two groups
combined.
The major analyses involved loo base 10 transformations of the daily AA
values. This was completed in order to reduce the contributions from single
outlier cases (days)
and to
increase
the
homogeneity of variance between
groups and
between
repeated
measures.
However ,
all
analyses
were also
completed using the original AA values as square root transformations because
the
latter were not as
extreme
the
1ος
as
modifications.
9
posteriori
contrasts were
completed
using
correlatec
t-tests
for each
OrOUD
and
Independent t-tests (between groups on a given day) to determine the source of
Any factor by repeated measure interactions.
463
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0
was
ntal
high
otal
Idex.
: day
Jrean
and
Jects
weamsy
ar
#2
he
of
in
in

Approved For Release 2000/00/1 . VIA KOT JO 007 JAN
Correlated t-tests with p values set at p (.01 (to reduce the effect of
multiple t probabilities) were completed for each group for the AA values
of
each of the 7 days during the experiment and the AA average for the month in
which the experience occurred.
The latter analyses were completed
to
determine the absolute activity of the days of the experience compared to the
typical monthly values rather
than restricting the analyses to the relative
differences between the key day and the days before and afterwards. All
analyses were completed using SPSSX software on a DEC2020 computer.
Results
MANOVA for all 62 subjects according to the three major groups (HH, LH,
LM plus HM), gender,
and the seven repeated measures (key days plus/minus 3
days) of geomagnetic activity (AA values) demonstrated no significant group or
pencer interactions. However, there was a significant (F=2.53, df=12, p =. 003)
interaction between groups and the geomagnetic activity over days for the log
base transformed AA values (Figure 1). There was also a nearly significant
daily difference (F=2.54, df=6, 336, p =. 03).
There were neither gender by day
or gender by group by day interactions.
The group by day interaction was
significant (p =. 01) for the absolute AA values as well.
Six of the experimental days involved the testing of two subjects rather
than one.
To determine if this slight modification in procedure may have
altered the geomagnetic contribution to the psi effect,
additional
analyses
were completed on those days (cases) where only one subject was tested.
The
basic results are shown in Figure 2. There was no appreciable change in the
geomagnetic temporal pattern and the significant day by group interactions
were not affected.
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0
46

Approved For Release 2000/08/11 : CIA-RDP96-00792R000400
When all four groups
were
analyzed in a similar design, the same basic
results were noted. There were no significant main effect differences between
the four groups or between genders. Again, there was a significant
group
by
days interactions (F=2.02, df=18, 324, p=, 009) and a between days difference.
Removing the second main (non-repeated) factor of gender and simply using the
four groups did not change the effect.
There were still significant repeated
day effects (F=3.76,
df=6,348, p =. 001)
and a day by group (F=22. 31, df=18,
p =. 003) interaction.
The basic results are presented in Figure 3.
A
posteriori
tests
showed
that the only
near-significant
difference
between days for the four groups occurred on 3 days before the key day (-3).
On this day, the geomagnetic activity was higher (F=3.40, df=58, o=, 02).
This
finding was confirmed by Duncan's analysis (o=, 05).
Correlated
t-tests
for
within group comparisons (set at o =. 01) demonstrated that
the nights
of the
experiments for Group I (HH) were quieter than day
-3
( == 2.99, df=17) and. - 2
(t=3.04, df=17).
The
of
night
the
experience
(plus 1) was significantly
quieter than the average of the month (t =4.55, df=17, o =. 001).
Both
the
Rey
day
(t=2.82)
and the days -1 (t=2.68) and plus 2 (t=3.84) were quieter tham
the month.
For the second group (LH), correlated t-tests
indicated that only day -3
(t=2.80) was significantly quieter than the monthly averages.
There were no
Significant
differences between the
key days
and the
days before
afterwards.
The third group
(HM) demonstrated
an unusual
Dattern.
The
Geomagnetic activity on day -3 was higher (t=4.81)
than the
monthly average
wille the activity on days olus 1 (t=7.78) and plus 2 (t=3.71) were lower than
monthly averages.
In addition, day plus 1
was
significantly
lower
the
465
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0

Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0
(t=9.80) than day -3. Finally, for group four (LM), there were no significant
differences between geomagnetic activity between any of the days.
Combining
groups three and four (df=14), no statistically significance differences were
found between the geomagnetic activity on any of the days. However, activity
on days plus 1
{t=3.71,
df=14),
plus 2 (t=22.87), and plus 3 (t=2.75) were
lower than for monthly averages.
Discussion
The
results
of the HH group in this study are basically similar to the
V-shaped pattern
geomagnetic activity
that
has been observed
during
spontaneous T-C experiences (e. g .. Persinger & Schaut, 1986).
However, the HM
group in the Maimonides data did show a temporal pattern that was similar to
that of the group
that
demonstrated the strongest psi effect.
Because the
third day before the key
day
demonstrated such elevated geomagnetic activity
in the HM group, we suspect that the apparent
V-shape is misleading and that
magnetic storms had been in progress. Our .hypothesis was confirmed as can be
seen
Figure 4. Analyses of the geomagnetic activity for the HM
group
indicated that day -3 (three days before the key day) displayed significantly
greater geomagnetic activity. than the monthly average.
This
pattern
is
strongly reminiscent of geomagnetic storms.
On
the
other hand, the HH group, the one that showed the greatest psi,
demonstrated
a
pattern
where there
was
no
pre-experience
elevation
in
geomagnetic activity compared to monthly values.
Instead, there was a sudden
decrease in geomagnetic activity; this activity was significantly
lower
than
for the month of for the days before or afterwards.
This latter pattern and
not the pattern displayed by the HM g
group is more typical of profiles than
466
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0

Approved For Release &
have been found in the cases of spontaneous T-C.
The use of AA values must be viewed with some caution.
Data collected at
several stations is
averaged
to
produce global
and
hemispheric
results.
Various
types
of
monitoring
devices are used,
ranging
from
crude
sophisticated.
It will come as no
surprise to learn that frequently there is
little or no correlation between daily
geomagnetic activity monitored at the
various stations,
especially those 1000 kilometers or more distant .. The AA
values are derived from a logarithmic process that omits much
of the data in
order to derive averages.
Nevertheless, this approach does serve the ourooss
of determining the existence of electromagnetic storms.
The best index for the study of anomalous behavior of laboratory subjects
would
be
one
based
on readings from a magnetometer in or adjacent
to
the
laboratory
itself.
Optimally,
these readings should
be
considered
relationship to solar
and
lunar effects, competing electromagnetic
fieic
effects (e. g., radiation),
and
biologicai
cycles of
the
subiects
be: no
studied.
If significant data are still obtained
soitals and magnetic
storms
243-270).
Therefore,
robust
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0
under
these
circumstances,
potential mechanisms involved can be studied.
Would the absence of electrical
Storms on nights of telepathy-clairvoyance
"hits'
mean
that
the
calm
environment
affects the
subject's receptivity?
Or
do electrical
storms
Interfere with the "transmission"
of the "signal" in some way?
There is no
conclusive evidence that
geomagnetic
activity exerts detectable effects upor
Maman behavior, although
preliminary
data
collected
Becker
(Becker
&
Wider. 1985) suggests a
relationship
between
admission
Yates
.
merta:
. 467

correlation between
AA
values
and anomalous laboratory behavior would be of
interest not only to parapsychologists but to many other researchers as well.
References
Becker. R.O .. & Seloen, G. (1985).
Electromagnetism and the
foundation
of life.
New York: william Morrow.
Child, I. L. (1985).
Psychology and anomalous observations: The question
of ESP in creams.
American Psychologist, 40, 1219-1230.
Mayaud,
P. N. (1973).
A
hundred
year
series of
geomagnetic data
1868-1967. IAGA Bulletin, NO.
33.
Persinger, M. A., & Schaut, B. B. (1986).
Manuscript
submitted
to the
Journal of the American Society for Psychical Research.
Schaut,
B.
B. ,
&
Persinger.
M.
A.
(1985).
Subjective telepathic
experiences, geomagnetic activity and the
ELF hypothesis.
Part
I. Data
analyses,
Psi Research. 4 (1). 4-20.
Ullman, M.,
&
Kriponer,
S. (1978).
Experimental dream studies.
In M.
Ebon (Ed. ).
The
Sionet
handbook of parapsychology (pp.
409-422). New York:
New American Library.
Ulimanı,
M.
(1969).
Telepathy and dreams.
Experimental
Medicine
and
Surgery. 27, 19-38.
Approved For Release 2000/08/11 : CIA-RDP96-00792R000400030001-0
468