DECISION AUMENTATION THEORY: TOWARD A MODEL OF ANOMALOUS MENTAL PHENOMENA
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIARDP9600791R0002002800025
Release Decision:
RIFPUB
Original Classification:
U
Document Page Count:
43
Document Creation Date:
November 4, 2016
Document Release Date:
October 27, 1998
Sequence Number:
2
Case Number:
Publication Date:
May 16, 1995
Content Type:
RP
File:
Attachment  Size 

CIARDP9600791R0002002800025.pdf  2.4 MB 
Body:
it~pfa~dl~n~Rbtl~"fl~~Q$t9u0~r~~kf~~~4~P180002002800025 16 May 1995
Decision Augmentation Theory:
Toward a Model
of
Anomalous Mental Phenomena
ay
Edwin C. May, Ph.D
Science Applications International Corporation
Menlo Park, CA
Jessica M. Utts, Ph.D.
University of California, Davis
Division of Statistics
Davis, CA
and
S. James P. Spottiswoode
Science Applications International Corporation (Consultant)
Menlo Park, CA
Abstract
Decision Augmentation Theory (DAT) holds that humans integrate information obtained by anoma
lous cognition into the usual decision process. The result is that, to a statistical degree, such decisions
are biased toward volitional outcomes. We introduce our model and show that the domain over which it
is applicable is within a few standard deviations from chance. We contrast the theory's experimental
consequences with those of models that treat anomalous effects as due to a force. We derive mathemat
ical expressions for DAT and for forcelike models using two distributions, normal and binomial. DAT
is testable both retrospectively and prospectively, and we provide statistical power curves to assist in the
experimental design of such tests. We show that the experimental consequences of our theory are dif
ferentfrom those of forcelike models except for one special case.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 1
~1~L~~fR5~1~i~n~r~a~ar~~`~~f~1~180002002800025 16 May 1995
Introduction
We do not have positive definitions of the effects that generally fall under the heading of anomalous
mental phenomena. In the crassest of terms, anomalous mental phenomena are what happens when
nothing else should, at least as nature is currently understood. In the domain of information acquisi
tion, or anomalous cognition (AC), it is relatively straightforward to design an experimental protocol
(Honorton et al., 1990, Hyman and Honorton, 1986) to assure that no known sensory leakage of in
formationcan occur. In the domain of macroscopic anomalous perturbation (AP), however, it is often
very difficult.
We can divide anomalous perturbation into two categories based on the magnitude of the putative ef
fect. Macro AP include phenomena that generally do not require sophisticated statisticai analysis to
tease out weak effects from the data. Examples include inelastic deformations in strain gauge experi
ments, the obvious bending of metal samples, and a host of possible "field phenomena" such as teleki
nesis, poltergeist, teleportation, and materialization. Conversely, microAP covers experimental data
from noisy diodes, radioactive decay and other random sources. These data show small differences
from chance expectation and require statistical analysis.
One of the consequences of the negative definitions of anomalies is that experimenters must assure that
the observables are not due to "known" effects. 'IYaditionally, two techniques have been employed to
guard against such interactions:
(1) Complete physical isolation of the target system.
(2) Counterbalancedcontrol and effort periods.
Isolating physical systems from potential "environmental" effects is difficult, even for engineering spe
cialists. It becomes increasingly problematical the more sensitive the AP device. For example Hubbard,
Bentley, Pasturel, and Issacs (1987) monitored a lazge number of sensors of environmental variables
that could mimic perturbational effects in an extremely isolated piezoelectric strain gauge. Among
these sensors were threeaxis accelerometers, calibrated microphones, and electromagnetic and nu
clear radiation monitors. In addition, the strain gauges were mounted in agovernmentapproved en
closure to assure no leakage (in or out) of electromagnetic radiation above a given frequency, and the
enclosure itself was levitated on an air suspension table. Finally, the entire setup was locked in a wn
trolled access room which was monitored by motion detectors. The system was so sensitive, for exam
ple, that it was possible to identify the source of a perturbation of the strain gauge that was due to inno
cent,gentleknocking onthe door of the closed room. The financial and engineering resources to isolate
such systems rapidly become prohibitive.
The second method, which is commonly in use, is to isolate the target system within the constraints of
the available resources, and then construct protocols that include control and effort periods. Thus, we
trade complete isolation for a statistical analysis of the difference between the control and effort peri
ods. The assumption implicit in this approach is that environmental influences of the target device will
be random and uniformly distributed in both the control and effort conditions, while anomalous effects
The Cognitive Sciences Laboratoryhasadoptedthetermanomalousmentalphenomenainsteadofthemorewidelyknownpsi.
Likewise, we use the terms anomalous cognition and anomalous perturbation forESP andPK, respectively. We have done so
because we believe that these terms are more naturally descriptive of the observables and are neutral with regard to mecha
nisms. These new terms will be used throughout this paper.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 2
~~~q,~,~y~?~~~~8 f~~a~l~~~t(,~]j~180002002800025 16 May 1995
will tend to occur in the effort periods. Our arguments in favor of an anomaly, then, are based on statis
tical inference and we must consider, in detail, the consequences of such analyses.
Background
As the evidence for anomalous mental phenomena becomes more widely accepted (Bem and Honor
ton, 1994, Utts,1991, Radin and Nelson, 1989) it is imperative to determine their underlying mecha
nisms. Clearly, we are not the first to begin thinking of potential models. In the process of amassing
incontrovertible evidence of an anomaly, many theoretical approaches have been examined; in this sec
tion we outline a few of them. It is beyond the scope of this paper, however, to provide an exhaustive
review of the theoretical models; a good reference to an uptodate and detailed presentation is Stokes
(1987).
Brief Review of Models
T~vo fundamentally different types of models of anomalous mental phenomena have been developed:
those that attempt to order and structure the raw observations in experiments (i.e., phenomenological
models), and those that attempt to explain these phenomena in terms of modifications to existing physi
cal theories (i.e., fundamental models). In the history of the physical sciences, phenomenological mod
els, such as the Snell's law of refraction or Ampere's law for the magnetic field due to a current, have
neazly ahvays preceded fundamental models, such as quantum electrodynamics and Maxwell's theory.
In producing useful models of anomalies it may well be advantageous to start with phenomenological
models, of which DAT is an example.
Psychologists have contributed interesting phenomenological approaches. Stanford (1974a and 1974b)
proposed PSIMediated Instrumental Response (PMIR). PMIR states that an organism uses anoma
lous mental phenomena to optimize its environment. For example, in one of Stanford's classic experi
ments (Stanford, Zenhausern, Taylor, and Dwyer 1975) subjects were offered a covert opportunity to
stop a boring task prematurely if they exhibited unconscious anomalous perturbation by perturbing a
hidden random number generator. Overall, the experiment was significant in the unconscious tasks; it
was as if the pazticipants were unconsciously scanning the extended environment for any way to provide
amore optimal situation than participating in a boring psychological task!
As an example of a fundamental model, Walker (1984) proposed a literal interpretation of quantum
mechanics and posited that since superposition of eigenstates holds, even for macrosystems, anoma
lous mental phenomena might be due to macroscopic examples of quantum effects. These ideas
spawned a class of theories, the socalled observation theories, that were either based upon quantum
formalism conceptually or directly (Stokes, 1987). Jahn and Dunne (1986) have offered a "quantum
metaphor" which illustrates many parallels between these anomalies and known quantum effects. Un
fortunately, these models either have free parameters with unknown values, or are merely hand waving
metaphors. Some of these models propose questionable extensions to existing theories. For example,
even though Walker's interpretation of quantum mechanical formalism might suggest wavelike prop
erties of macrosystems,the physics data to date not only show no indication of such phenomena at roam
temperature but provide considerable evidence to suggest that macrosystems lose their quantum coher
ence above 0.5 Kelvins (Washburn and Webb, 1986) and no longer exhibit quantum wavelike behavior.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 3
~~~~~~~ ~~~B~~S,~~~a~~~~~~~~~1 80002002800025 16 May 1995
This is not to say that a comprehensive model of anomalous mental phenomena may not eventually
require quantum mechanics as part of its explanation, but it is currently premature to consider such
models as more than interesting speculation. The burden of proof is on the theorist to show why sys
tems, which aze normally considered classical (e.g., a human brain), are, indeed, quantum mechanical.
 That is, what are the experimental wnsequences of a quantum mechanical system over a classical one?
Our Decision Augmentation Theory is phenomenological and is a logical and formal extension of Stan
ford'selegant PMIR model. In the same manner as eazly models of the behavior of gases, acoustics, or
optics, DAT tries to subsume a large range of experimental measurements into a coherent lawful
scheme. Hopefully this process will lead the way to the uncovering of deeper mechanisms. In fact DAT
leads to the idea that there may be only one underlying mechanism of all anomalous mental phenome
na, namely a transfer of information from future to past.
Historical Evolution of Decision Augmentation
May, Humphrey, and Hubbazd (1980) conducted a careful random number generator (RNG) experi
mentwhich was distinguished by the extreme engineering and methodological care that was taken to
isolate any potentially known physical interactions with the source of randomness (D. Druckman and J.
A. Swets, page 189,1988). It is beyond the scope of this paper to describe this experiment completely;
however, those speck details which led to the idea of Decision Augmentation are important for the
sake of historical completeness. The authors were satisfied that they had observed a genuine statistical
anomaly and additionally, because they had developed an accurate mathematical model of the random
device, they were assured that the deviations were not due to any known physical interactions. They
concluded, in their report, that some form of anomalous data selection had occurred and named it Psy
choenergetic Data Selection.
Following a suggestion by Dr. David R. Saunders of MARS Measurement and Associates, we noticed in
1986 that the effect size in binary RNG studies varied on the average as one over the square root of the
number of bits in the sequence. This observation led to the development of the Intuitive Data Sorting
model that appeazed to describe the RNG data to that date (May, Radin, Hubbard, Humphrey, and
Utts,1985). The remainder of this paper describes the next step in the evolution of the theory which is
now named DecisivnAugmentation Theory.
Decision Augmentation TheoryA General Description
Since the case for ACmediated information transfer is now well established (Bem and Honorton,
1994) it would be exceptional ifwe did not integrate this form of information gathering into the decision
process. For example,, we routinely use realtime data gathering and historical information to assist in
the decision process. Why, then, should we not include AC in the decision process? DAT holds that AC
information is included along with the usual inputs that result in a final human decision that favours a
"desired" outcome. In statistical parlance, DAT says that a slight, systematic bias is introduced into the
dceision process by AC.
This philosophical concept has the advantage of being quite general. Tb illustrate the point, we describe
how the "cosmos" determines the outcome of awelldesigned, hypothetical experiment. 1b determine
the sequencing of conditions in an RNG experiment, suppose that the entry point into a table of ran
dom numbers will be chosen by the square root of the bazometric pressure as stated in the weather re
Approved For Release 2000/08/10 :CIARDP960079180002002800025 4
~~~~r~~~~~~~~~e~~a~>~'~ar~~l~g~9~(~t~~180002002800025 16 May 1995
port that will be published seven days hence in the New York mimes. Since humans are notoriously bad at
predicting or controlling the weather, this entry point might seem independent of a human decision; but
why did we "choose" seven days in advance? Why not six or eight? Why the New York Times and not the
London Times? DATwould suggest that the selection of seven days, the New York Times, the barometric
pressure, and squaze root function were better choices, either individually or collectively, and that other
decisions would not have led to as significant an outcome. Other nontechnical decisions may also be
biased by AC in accordance with DAT. When should we schedule a Ganzfeld session; who should be the
experimenter in a series; how should we determvne a specific order in a tripolar protocol? DAT ex
plainsanomalous mental phenomena as a process of judicious sampling from a world of events that are
unperturbed. In contrast, forcelike models, hold that some kind ofinentallymediated force perturbs
the world. As we will show below, these two types of models lead to quite different predictions.
It is important to understand the domain in which a model is applicable. For example, Newton's laws
are sufficient to describe the dynamics of mechanical objects in the domain where the velocities are very
much smaller than the speed of light, and where the quantum wavelength of the object is very small
compared to the physical extent of the object. If these conditions are violated, then different models
must be invoked (e.g., relativity and quantum mechanics, respectively). The domain in which DAT is
applicable is when experimental outcomes are in a statistical regime (i.e., a few standard deviations
from chance). In other words, could the measured effect occur under the null hypothesis? This is not a
sharpedged requirement but DAT becomes less apropos the more a single measurement deviates from
meanchanceexpectation (MCE). We would not invoke DAT, for example, as an explanation of levita
tion if one found the authors hovering near the ceiling! The source of the statistical variation is unre
stricted and maybe of classical or quantum origin, because a potential underlying mechanism for DAT
is precognition. lay this means, experiment participants become statistical opportunists.
Development of a Formal Model
While DAT may have implications for anomalous mental phenomena in general, we develop the model
in the framework of understanding experimental results. In particular, we consider anomalous per
turbationversus anomalous cognition in the form of decision augmentation in those experiments whose
outcomes are in the fewsigma, statistical regime.
We define four possible mechanisms for the results in such experiments:
(1) Mean Chance Expectation. The results are at chance. That is, the deviation of the dependent vari
able meets accepted criteria for MCE. In statistical terms, we have measurements from an unper
turbed parent distribution with unbiased sampling.
(2) Anomalous Perturbation. Nature is modified by some anomalous interaction. That is, we expect
an interaction of a "force" type. In statistical parlance, we have measurements from a perturbed
parent distribution with unbiased sampling.
(3) Decision Augmentation. Nature is unchanged but the measurements are biased. That is, AC in
formationhas "distorted" the sampling. In statistical terms, we have measurements from an unper
turbed parent distribution with biased sampling.
(4) Combination. Nature is modified and the measurements are biased. That is, both anomalous ef
fects are present. In statistical parlance, we have conducted biased sampling from a perturbed paz
ent distribution.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 5
im~ne~0l~fgmnl'a?66036'tifhrdB~l~ANPP.1180002002800025 16 May 1995
General Considerations and Definitions
Since the formal discussion of DAT is statistical, we will describe the overall context for the develop
ment of the model from that perspective. Consider a random variable, X, that can take on continuous
values (e.g., the normal distribution) or discrete values (e.g., the binomial distribution). Examples of X
might be the hit rate in an RNG experiment, the swimming velocity of single cells, or the mutation rate
of bacteria. Let Ybe the average ofXcomputed over n values, where n is the number of items that are
collected as the result of a single decisionone trial. Often this may be equivalent to a single effort
period, but it also may include repeated efforts. The key point is that, regardless of the effort style, the
average value of the dependent variable is computed over the n values resulting from one decision
point. In the examples above, n is the sequence length of a single run in an RNG experiment, the num
ber ofswimming cells measured during the trial, or the number ofbacteriacontaining test tubes present
during the trial. As we will show below, forcelike effects require that the Zscore, which is computed
from the Ys, increase as the square root of n. In contrast, informational effects will be shown to be inde
pendent of n. '
Assumptions for DAT
We assume that the parent distribution of a physical system remains unpenurfied; however, the mea
surements of the physical system are systematically biased by some ACmediated informational pro
cess.
Since the deviations seen in experiments in the statistical regime tend to be small in magnitude, it is safe
to assume that the measurement biases will also be small; therefore, we assume small shifts of the mean
and variance of the sampling distribution. Figure 1 shows the distributions for biased and unbiased
measurements.
Figure 1. Sampling Distribution Under DAT.
The biased sampling distribution shown in Figure 1 is assumed to be normally distributed as:
where and ar are the mean and standard deviation of the sampling distribution.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 6
~ppNSf~rAGD~g~he'P~Qraddl~86#O~GI9180002002800025 16 May 1995
Assumptions for an Anomalous Perturbation_Model
DAT can be contrasted to forcelike effects. With a few exceptions reported in the literature of "field"
phenomena, anomalous perturbation appears to be relatively "small." Thus, we begin with the assump
tionthat aputative anomalous force would give rise to a perturbational interaction, by which we mean
that, given an ensemble of entities (e.g., binary bits, cells), an anomalous force would act equally on each
member of the ensemble, on the average. We call this type of interaction microAP.
Figure 2 shows a schematic representation of probability density functions for a parent distribution un
derthe microAP assumption and an unperturbed parent distribution. In the simplest microAP model,
the perturbation induces a change in the mean of the parent distribution but does not effects its vari
ance. We pazameterize the mean shift in terms of a multiplier of the initial standard deviation. Thus,
we define an APeffect size as:
~i  ?o~
,u~ = o~o
where ?l and ?p are the means of the perturbed and unperturbed distributions, respectively, and where
op is the standard deviation of the unperturbed distribution.
Figure 2. Parent Distribution for microAP.
For the moment, we consider Epp as a parameter which, in principle, could be a function of a variety of
variables (e.g., psychological, physical, environmental, methodological). As we develop DAT for specif
ic distributions and experiments, we will discuss this functionality of Epp
Calculation of E(Z~
We compute the expected value and variance of ZZ for mean chance expectation and under the force
likeand information assumptions. We do this for the normal and binomial distributions. The details of
the calculations can be found in the Appendix; however, we summarize the results in this section. 'Table
1 shows the results assuming that the parent distribution is normal.
7
Approved For Release 2000/08/10 :CIARDP960079180002002800025
~~~~t~~~i1~~t~eT~~a~O~'~r~Q~j~a~1R0002002800025 ig May 1995
tit
Q
Mechanism
uan
y
MCE
microAP
DAT
Var(Z~
2
2(1 ? 2s~n)
2(oZ + 2?=oz)
Table 2 shows the results assuming that the parent distribution is binomial. In this calculation, pp is the
binomial event probability and op = ~/pp ~ .
ti
 Mechanism
Quan
ty
MCE
microAP
DAT
E(ZZ)
1
1 + e~(n  1) + oo (1  2po)
?; + 0,
V~ZZ)
2 + nrr2 (1 600)
0
2(1 + 2e~n)'
2(0= +2?=o=)
* The variance shown assumespn = 0.5 and n> 1. See the Appendix for other cases.
We wish to emphasize at this point that in the development of the mathematical model, the parameter
Epp for micro AP; and the parameters ?Z, and al in DAT may all possibly depend upon n; however, for
the moment, we assume that they are all nindependent. We shall discuss the consequences of this as
sumptionbelow.
Figure 3 displays these theoretical calculations for the three mechanisms graphically.
E(ZZ)
large
microAP
small
DAT
Figure 3. Predictions of MCE, microAl; and DA'1:
Approved For Release 2000/08/10 :CIARDP960079180002002800025 $
~y~~~~~~~~@~Q(~T~1 80002002800025 16 May 1995
Within the constraints mentioned above, this formulation predicts grossly different outcomes for these
models and, therefore, is ultimately capable of separating them, even for very small perturbations.
Retrospective Tests
It is possible to apply DAT retrospectively to any body of data that meet certain constraints. It is critical
to keep in mind the meaning of nthe number of measures of the dependent variable over which to
compute an average during a single trial following a single decision In terms of their predictions for
experimental results, the crucial distraction between DAT and the microAP model is the dependence
of the results upon n; therefore, experiments which are used to test these theories must be those in
which n is manipulated and participants are held blind to its values. May, Spottiswoode, Utts and James
(1994) retrospectively apply DAT to as many data sets as possible, and examine the wnsequences of any
violations of these criteria.
Aside from these considerations, the application of DAT is straight forward. Having identified the unit
of analysis and n, simply create a scatter diagram of points (Z? n) and compute a least square fit to a
straight line. 'Fables 1 and 2 show that for the microAP model, the square of the effect size is the slope of
the resulting fit. A Student's ttest may be used to test the hypothesis that the effect size is zero, and thus
test for the validity of the microAP model If the slope is zero, these same tables show that the intercept
maybe interpreted as an AC strength parameter for DAT. A followon paper will describe these tech
niques indetail (May, Spottiswood, and Utts,1994).
Prospective Tests
A prospective test of DAT could not only test whether anomalous effects occurred, but would also dif
ferentiatebetween microAP and DAT. In such tests, n should certainly be adoubleblind parameter
and take on at ]east two values. If you wanted to check the prediction of a linear functional relationship
between n and the E(Z~) that issuggested bymicroAP model, the more values of n the better. It is not
possible to separate the microAP model from DAT at a single value of n.
In any prospective test, it is helpful to know the number of runs, N, that are necessary to determine with
95% confidence, which of the two models best fits the data. Figure 4 displays the problem graphically.
Figure 4. Model Predictions for the Power Calculation.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 9
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Decision Augmentation Theory: Toward a Model of AMP 16 May 1995
Under microAP, 95% of the values of ~Zwill be greater than the point indicated in Figure 4. Even if the
measured value of ~Zis at this point, we would like the lower limit of the 95% confidence interval for this
value to be greater than the predicted value under the DAT model. Or:
E(Z,Z,P)  1.645 o"P  1.960 0`"' >_ E,~c (ZZ).
Solving for N in the equality, we find:
I 3.605 vv.
N = lE,,~, (ZZ)  Esc (Zz)

(1)
Since opp > 6pC, this value of Nwill always be the larger estimate than that derived from beginning with
DAT and calculating the confidence intervals in the other direction.
Suppose, from an earlier experiment, one can estimate asingletrial effect size for a specific value of n,
say nY. Tb determine whether the microAP model or DAT is the proper description of the mechanism,
we must conduct another study at an additional value of n, say n2. We use Equation 1 to compute how
many runs we must conduct at n2 to assure a separation of mechanism with 95% confidence, and we use
the variances shown in'I~ables 1 and 2 to compute Upg Figure 5 shows the number of runs for anRNG
like experiment as a function of effect size for three values of n2.
We chose nl =100 bits because it is typical of the numbers found in the RNG database and the values of
n2 shown are within easy reach of today's computerbased RNG devices. For example, assuming az =
1.0 and assuming an effect size of 0.004, a value derived from a publication of PEAR data (Jahn, 1982),
then at nl =100,?l = 0.004 x ~/~ = 0.04 andEAC(ZZ) =1.0016. Suppose n2 =104, then Epp(ZZ) _
1.160 and oAp =1.625. Using Equation 1, we find N =1368 runs, which can be approximately obtained
from Figure S. That is in this example, 1368 runs are needed to resolve the microAP model from DAT
at nZ =104 at the 9S% confidence level. Since these runs are easily obtained in most RNG experiments,
an ideal prospective test of DAT, which is based on these calculations, would be to conduct 1500 runs
randomly counterbalanced between n = 102 and n = 104 bits/trial. If the effect size at n = 102 is near
0.004, than we would be able to distinguish between microAP and DAT with 95% confidence.
Figure S. Runs Required for RNG Effect Sizes
Approved For Release 2000/08/10 :CIARDP960079180002002800025 10
Appproved For Release 2000/08/10 :CIARDP960079180002002800025
DB~ision Augmentation Theory: Toward a Model of AMP 16 May 1995
Figure b shows similar relationships for effect sizes that are more typical of anomalous perturbation
experiments using biological target systems (May and Vilenskaya,1994).
In this case, we chose n1 = 2 because it is easy to use two targets simultaneously. If we assume an effect
size of 0.3 and az =1.0, at n2 =10 we compute Ep~(Z2) =1.180, Epp(Zz) =1.900, app = 2.366 and N =
140, which can be approximately obtained from Figure 6.
We have included n2 =100 in Figure 6, because this is within reach in cellular experiments although it is
probably not practical for most biological experiments.
Figure 6. Runs Required for Biological Effect Sizes
We chose n1 = 2 units for convenience. For example in a plant study, the physiological responses can
easily be averaged over two plants and n2 =10 is within reason for a second data point. A unit could be a
test tube containing cells or bacteria; the collection of all ten test tubes would simultaneously have to be
the target to meet the constraints of a valid test.
The prospective tests we have described so far are conditional; that is, given an effect size, we provide a
protocol to test if the mechanism for the anomalies is micro tAP or DAT. An unconditional test does not
assume any effect size; all that is necessary is to collect data at a large number of different values of n,
and fit a straight line through the resulting ZZs. The mechanism ismicroAP if the slope isnonzero and
maybe DAT if the slope is zero.
Stouffer's Z Tests
One consequence of DAT is that more decision points in an experiment lead to stronger results, because
an operator has more opportunity to exercise AC abilities. We derive a test criteria to determine wheth
er aforcelike interaction or an informational mechanism is a better description of the data.
Consider two experiments of M decisions at n1 and N decisions at n2, respectively. Regardless of the
mechanism, the Stouffer's Z for the first experiment is given by:
Approved For Release 2000/08/10 :CIARDP960079180002002800025 11
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Decision Augmentation Theory: Toward a Model of AMP 16 May 1995
M
~~~u
Z,'~ = 1=i = niMel,
where e1,~ is the effect size for one decision and where s1 is the average effect size over the M decisions.
Under the microAP assumption that the effect size, EI, is constant regardless of n, Stouffer's Z in the
second experiment is given by:
Z~2> = V n2N Ztl>.
s VV n1M s
Under the DAT assumption that the effect size is proportional to 1/~/n, the Stouffer's Z in the second
experiment becomes:
Z~2> _ ~ Zt'~.
s M s
As in the other tests of DAT, if data are collected at two values of n, then a test between these Stouffer's
Z values may yield a difference between the competing mechanisms.
Discussion
We now address the possible ndependence of the model parameters. A degenerate case arises if epp is
proportional to 1/~/n; if that were the case, we could not distinguish between the microAP model and
DAT by means of tests on the n dependence of results. If it were the case that in the analysis of the data
from a variety of experiments, participants, and laboratories, the slope of a ZZ VS n linear leastsquares
fit were zero, then either Epp = 0.0 or Epp is proportional to 1/~/n, the accuracy depending upon the
precision of the fit (i.e., errors on the zero slope). An attempt might be made to rescue the microAP
hypothesis by explaining the 1/~/n dependence of epp in the degenerate case as a fatigue or some other
time dependence effect. That is, it might be hypothesized that anomalous perturbation abilities would
decline as a function of n; however, it seems improbable that ahumanbased phenomenon would be so
widely distributed and constant and give the 1/~/n dependency in differing protocols needed to imitate
DAT. We prefer to resolve the degeneracy by wielding Occam's razor: if the only type of anomalous
perturbation which fits the data is indistinguishable from AC, and given that we have ample demonstra
tions of AC by independent means in the laboratory, then we do not need to invent an additional phe
nomenon called anomalous perturbation. Except for this degeneracy, a zero slope for the fit allows us
to reject all microAP models, regazdless of their ndependencies.
DAT is not limited to experiments that capture data from a dynamic system. DAT may also be the mech
anism in protocols which utilize quasistatic target systems. In aquasistatic target system, a random
process occurs onlywhen a run is initiated; a mechanical dice thrower is an example. Yet, in a series of
unattended runs of such a device there is always a statistical variation in the mean of the dependent
variable that maybe due to a variety of factors, such as Brownian motion, temperature, humidity, and
possibly the quantum mechanical uncertainty principle (Walker, 1974). Thus, the results obtained will
ultimately depend upon when the run is initiated. It is also possible that asecondorder DAT mecha
nism arises because of protocol selection; how and who determines the order in tripolar protocols. In
second order DAT there maybe individuals, other than the formal subject, whose decisions effect the
Approved For Release 2000/08/10 :CIARDP960079180002002800025 12
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Decision Augmentation Theory: Toward a Model of AMP 16 May 1995
experimental outcome and are modified by AC. Given the limited possibilities in this case, we might
expect less of an impact from DAT.
In surveying the range of anomalous mental phenomena, we reject the evidence for experimental mac
ro AP because of poor artifact control and accept the evidence for precognition and microAP because
of the large number of studies and the positive results of the metaanalyses. We believe that DAT, there
fore, might be a general model for anomalous mental phenomena in that it reduces mechanisms for
laboratoryphenomena toonly onethe anomalous transtemporal acquisition of information.
Acknowledgements
Since 1979, there have been many individuals who have wntributed to the development of DAT. We
would first like to thank David Saunders without whose remark this work would not have been. Beverly
Humphrey kept the philosophical integrity intact at times under extreme duress. We are greatly appre
ciative of Zoltan Vassy, to whom we owe the Zscore formalism, to George Hansen, Donald McCarthy,
and Scott Hubbard for their constructive criticisms and support.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 13
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Decision Augmentation Theory: Toward a Model of AMP 16 May 1995
References _
Bem, D. J. and Honorton, C. (1994). Does psi exist? Replicable evidence for an anomalous process of
information transfer. Psychological Bulletin. 115, No. 1, 41$.
_ Druckman, D and Swets, J. A. (Eds.) (1988). Enhancing Human Performance. Issues, Theories, and
Techniques. Washington D.C., Nation Academy Press.
Honorton, C., Berger, R. E., Varvoglis, M. P., Quant, M., Derr, P., Schechter, E. I., and Ferrari, D. C.
(1990) Psi Communication in the Ganzfeld. Journal of Parapsychology, 54, 99139.
Hubbard, G. S., Bentley, P. P., Pasturel, P. K., and Isaacs, J. (1987). A remote action experiment with a
piezoelectric transducer. Final Report Objective II, Task 3 and 3a. SRI International Project
1291, Menlo Park, CA.
Hyman, R. and Honorton, C. (1986). A joint communiquc: The psi ganzfeld wntroversy. Journal of
Parapsychology. 50, 351364.
Jahn, R. G. (1982). The persistent paradox of psychic phenomena: an engineering perspecitve.
Proceedings of the IEEE. 70, No. 2,136170.
Jahn R. G. and Dunne, B. J. (1986). On the quantum mechanics of consciousness, with application to
anomalous phenomena. Foundations of Physics. 16, No 8, 721772.,
May, E. C., Humphrey, B. S., Hubbard, G. S. (1980). Electronic System Perturbation'Ibchniques. Final
Report. SRI International Menlo Park, CA.
May, E. C., Rodin, D. I., Hubbard, G. S.,' Humphrey, B. S., and Utts, J. (1985) Psi experiments with
random number generators: an informational model. Proceedings of Presented Papers Yol I. The
Parapsychological Association 28th Annual Convention, 'Rifts University, Medford, MA, 237266.
May, E. C. and Vilenskaya, L. (1994). Overview of Current Parapsychology Research in the Former
Soviet Union. Subtle Energies 3, No 3.4567.
Radin, D. I. and Nelson, R. D. (1989). Evidence for consciousnessrelated anomalies in random
physical systems. Foundations of Physics. 19, No. 12,14991514.
Stanford, R. G. (1974a). An experimentally testable model for spontaneous psi events I. Extrasensory
events. Journal of theAmerican Society for Physical Research, 68, 3457.
Stanford, R. G. (1974b). An experimentally testable model for spontaneous psi events II. Psychokinetic
events. Journal of the American Society for Physical Research, 68, 321356.
Stanford, R. G., Zenhausern R., 'Paylor, A., and Dwyer, M. A. (1975). Psychokinesis as psimediated
instrumental response. Journal of the American Society for Physical Research, 69,127133.
Stokes, D. M. (1987). Theoretical parapsychology. In Advances in Parapsychological Research 5.
McFarland & Company, Inc. Jefferson NC, 77189.
Utts, J. (1991). Replication and metaanalysis in parapsychology. Statistical Science. 6, No. 4, 363403.
Walker, E. H. (1974). Foundations of Paraphysical and Parapsychological phenomena. Proceedings of
an International Conference: Quantum Physics and Parapsychology. Oteri, E. Ed. Parapsychology
Foundation, Inc. New York, NY,153.
Walker, E. H. (1984). A review of criticisms of the quantum mechanical theory of psi phenomena.
Journal of Parapsychology. 48, 277332.
Washburn S. and Webb, R. A. (1986). Effects of dissipation and temperature on macroscopic quantum
tunneling in Josephson junctions. In New Techniques and Ideas in Quantum Measurement Theory.
Greenburger, D. M. Ed. New York Academy of Sciences, New York, NY, 6677.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 14
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Decision Augmentation Theory: Toward a Model of AMP
Appendix
Mathematical Derivations for the Decision Augmentation Theory
In this appendixwe develop the formalism for the Decision Augmentation Theory (DA7~. We consider
cases for mean chance expectation, forcelike interactions, and informational processes under two as
sumptions~normality and Bernoulli sampling. For each of these three models, we compute the ex
pectedvalues of Z and ZZ, and the variance of Z~'
Mean Chance Expectation 
Normal Distribution
We begin by considering a random variable,X, whose probability density function is normal, (i.e., N(Fcp,
ap~)t). After many unbiased measures from this distribution, it is possible to obtain reasonable ap
proximations to pp and aa2 in the usual way. Suppose n unbiased measures are used to compute a new
variable, Y, given by:
Yk = n ~ XJk
Jet
Then Yis distributed as N(Np, a? 2), where ant = apZ/n. If Z is defined as
Z=Y~F~o
a
then Z is distributed as N(t7, 1) and E(Z) is given by:
Since Yar(Z) =1 = E(ZZ)  E2(Z), then
The Yar(Z2) = E(Z4)  EZ(Z2) = E(Z4) 1. But
? We wish to thank Zoltan Vassy for originally suggesting the ZZ formalism.
f Throughout this appendix, this notation means: x
N~~~) = t eo.s~x~ ~ .
o~
(1)
(2)
Approved For Release 2000/08/10 :CIARDP960079180002002800025 Appendix: 1
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Decision Augmentation Theory: Toward a Model of AMP
EMCEIZ?) = 1 I z4eos:2dz = 3.
YarMCE(Zz) = 2. (3 )
Bernoulli Sampling
Let the probability of observing a one under Bernoulli sampling be given by pp. After n samples, the
discrete Zscore is given by:
ZknPo
o'o
Qo = Po(1  Po)
and k is the number of observed ones (0 Ck C n}. The expected value of Z is given by:
n
EMCE\Z) _ ~ ~ ~(k  nPo)Bk(n~Po)~ (4 )
n
B~(n~Po) = k Po(1  Po)"k,
The first term in Equation 4 is the E(k) which, for the binomial distribution, is npd. Thus
E~cE(Z) _ ~ ~(k  nPo)B~(n~Po) = 0.
6o n x_o
The expected value of ZZ is given by:
EMCElZ2) = Yar(Z) + EZ(Z),
Yar(k nPo
_ + 0,
nao
z
EMCElZ2) = nQq = 1.
0
As in the normal case, the Yar(Z2) = E(Z4)  E2(Z2) = E(Z4) 1. But*
* Johnson, N. L, and S. Kotz, DucrieteDistributiong John Wiley & Sons, New York, p. 51, (1969).
(S)
(6)
Approved For Release 2000/08/10 :CIARDP960079180002002800025 Appendix? 2
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Decision Augmentation Theory: Toward a Model of AMP
EMCE\Z4) _
So,
n
nza'4 ~(k  nPo)4B~(n~Po)
Ok0
= 3 + nd2 (1 60'0).
0
YaroMCE(ZZ) = 2 + no'2 (1 600) = 2  n ~ (Po = 0.5). (7 )
0
ForceLike Interactions
Normal Distribution
Under the perturbation assumption described in the text, we let the mean of the perturbed distribution
be given by/cp+ Eopop, where Fop is an anomalousperturbation strength parameter, and in the general
case may be a function of n and time. The parent distribution for the random variable, X, becomes
N(pp+ eap0a, 002). As in the meanchanceexpectation case, the average of n independent values ofX,
is Y ~ N(Np+ eapop, 0? 2). Let
Y=fro+e~a'o+dy,
For a mean of n samples, the Zscore is given by
Z=Y,ao=e~0o+dy_s n+
where i; is distributed as N(0, 1) and is given by dy / a,,. Then the expected value of Z is given by
and the expected value of Z2 is given by
E~(ZZ) = E,u.([E,~ ~ + ~]Z) = new + E(~2) + 2e,~ ~E(~)
= 1 + ern,
(8)
(9)
since E(~ = 0 and E(~Z) =1.
Ingeneral, Z2 isdistributed as anoncentral X2with 1 degree of freedom and noncentrality parameter
nip 2, X2(1, nip 2). Thus, the variance of Z2 is given by`
Yar",~(ZZ) = 2(1 + 2ne~). (1Q )
Bernoulli Sampling
As before, let the probability of observing a one under mean chance expectation be given by pp, and the
discrete Zscare be given by:
Approved For Release 2000/08/10 :CIARDP960079180002002800025 Appendix: 3
Approved For Release 2000/08/10 :.CIARDP960079180002002800025
Decislon Augmentation Theory: Toward a Model of AMP
o'o
where k is the number of observed ones (0 < k C n). Under the perturbation assumption, we let the
mean of the distribution of the singlebit probability begiven by p1 = pp + sapc(p, where eQp is an anoma
lousperturbation strength parameter. The expected value of Z is given by:
h
E~(Z) = 1 ~(k  nPo)Bk(n~Pi)~
SOY"ks0
n
Bk(n~Pi) = k Pi(1  Pi)"k,
The expected value of Z becomes
n
E~(Z) = 1 ~ ~k(n~Pi) 'nPo
QOY" k~0
Since eap = E(Z)/~, so Eap is also the binomial effect size. The expected value of ZZ is given by:
E~(ZZ) = Yar(Z) + EZ(Z),
_ Yar(k  npo s
nag + r~rt,
0
0
Expanding in terms of pl = pp + eapoh,
E~(ZZ) = 1 + E p(n  1) + 60 (1  2Po)?
(12)
If pp = 0.5 (i.e., a binary case) and n> 1, then Equation 12 reduces to the E(ZZ) in the normal case,
Equation 9.
We begin the calculation of Yar(ZZ) by using the equation for the jth moment of a binomial distribution
m! _ ~~~(4 + per)"~ I ~_o.
Since Yar(ZZ) = E(Z4)  EZ(Z2), we must evaluate E(Z4). Or,
E~(Z4) = nZO~ ~(k  nPo)'Bk(n~Pi)?
Ok.p
Expanding n'Zap'4(k  npp)4, using the appropriate moments, and subtracting EZ(ZZ), yields
Yar(ZZ) = Co + C, n + C_, n '. (13 )
Approved For Release 2000/08/10 :CIARDP960079180002002800025 Appendix: 4
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Decision Augmentation Theory: Toward a Model of AMP
x
Co=236s~+lOsp+8QO(12,po)(12s p)+6s~?P,
0
3
Cl = 4s~(1  sue) + 4 ao (1  2,po), and
C_, = 48  6(s~  3]Z + 12 00 (1  2Po) + (1 ~7s p) + 63a (1  2po)(12po  12Po + 1).
0 0
Under the condition that sap 1, E < 1, and pp = 0.5, the variance reduces to that derived under the normal
distribution assumption. Or,
Yare,~(ZZ) ~ 2(1 + 2ns~). (14 )
Information Process
Normal Distribution
The primary assumption in this case is that the parent distribution remains unchanged, (i.e., N(?p, op2).
It further assumes that because of ananomalouscognitionmediated bias the sampling distribution is
distorted leading to a Zdistribution as N(?ac o'ac2) ~ In the most general case, ?a~ and oar may be func
tions of nand time.
The expected value of Z is given by (by definition)
Eac(Z) = ?~
(15)
The expected value of Z2 is given by definition as
The Yar(ZZ) can be calculated by noticing that
(16)
a z
Q  XZ,~ 1, a~`~.
So the Yar(Z~) is given by
Approved For Release 2000/08/10 :CIARDP960079180002002800025 Appendix: 5
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Decision Augmentation Theory: Toward a Model of AMP
/ z
Yarl Z2~ = 2 1 + 2~?`
Luc 1 mac
YarAC(ZZ) = 2((74ac + ~ac~ac)?
(17)
Bernoulli Sampling
As in the normal case, the primary assumption is that the parent distribution remains unchanged, and
that because of apsimediated bias the sampling distribution is distorted leading to a discrete Zdis
tributioncharacterized by?~ (n) and Qac 2(n). Thus, by definition, the expected values of Z and Z2 are
given by
EACIZ) = fuac
EB ((ZZ 2 + QZ
ACl )  N'ac ac?
For any value of n, estimates of these parameters are calculated from N data points as
N
?,~ = N~z~,and
/1
2
a,~
__ N N zi 2
(N  1) ~ N  ~ac.
i1
The Yar(ZZ) for the discrete case is identical to the continuous case.
Yarac(ZZ) = 2(v;~ + 2?;i aa~).
(19)
Approved For Release 2000/08/10 :CIARDP960079180002002800025 Appendix? 6
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Applications of Decision Augmentation Theory 14 May 1995
Applications
of
Decision Augmentation Theory
by
Edwin C. May, Ph.D
S. James Spottiswoode (Consultant)
Science Applications International Corporation
Menlo Park, CA
Jessica M. Utts, Ph.D.
University of California, Davis
Division of Statistics
Davis, CA
Christine L. James
Science Applications International Corporation
Abstract
Decision Augmentation Theory (DAT) provides an informational mechanism for a class of anomalous
mental phenomena which have hitherto been viewed as being caused by aforcelike mechanism. Under
specifiable conditions, DAT's predictions for statistical anomalous perturbation databases are differ
entfrom those of all forcelike mechanisms. For large random number generator databases, DAT pre
dicts azero slope for a least squares fit to the (Z? n) scatter diagram, where n is the number of bits result
ingfrom asingle run and Z is the resulting Zscore. We find a slope of {1.73f3.19) X 10 _6 (t = O.S43, df
=12! p = 0.295) for the historical binary random number generator database which strongly suggests
that some informational mechanism is responsible for the anomaly. Ina 2sequence length analysis of a
limited set of RNG data from the Princeton Engineering Anomalies Research laboratory, we find that a
forcelike explanation misses the observed data by 8.6a; however, the observed data are within 1.10 of
the DAT prediction. We also apply DAT to one pseudorandom number generator study and find that its
predicted slope is not significantly different from the expected value for an informational mechanism.
We review and comment on six published articles that discussed DAT's earlier formalism (i.e., Intuitive
Data Sorting). We found two studies that support aforcelike mechanism. Our analysis of Braud's 1990
hemolysis study confirms his finding in favor of an influence model over a selection one (p = 0.023), and
Braud and Schlitz (1989) demonstrated aforcelike interaction in their remote staring experiment (p =
0.020). We provide six circumstantial arguments against an influence hypothesis. Our anomalous
cognition research suggests that the quality of the data is proportional to the total change of Shannon
entropy. We demonstrate that the change of Shannon entropy of a binary sequence from chance is in
dependent ofsequence length; thus, we suggest that a fundamental argument supports DAT over influ
ence models. In our conclusion, we suggest that, except for one special case, the physical random num
ber generator database cannot be explained by any influence model, and that contradicting evidence
from two experiments on biological systems should inspire more investigations in a way that would al
low valid DAT analyses.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 1
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Applications of Decision Augmentation Theory 14 May 1995
Introduction
May, Utts, and Spottiswoode (1994) proposed Decision Augmentation Theory as a general model of
anomalous mental phenomena.` DAT holds that anomalous cognition information is included along
with the usual inputs that result in a final human decision that favours a "desired" outcome. In statisti
calparlance, DAT says that a slight, systematic bias is introduced into the decision process by anoma
louscognition.
This concept has the advantage of being quite general. We know of no experiment that is devoid of at
least one human decision; thus, DAT might be the underlying basis for anomalous mental phenomena.
May et al. (1994) mathematically developed this concept and constructed a retrospective test algorithm
than can be applied to existing databases. In this paper, we summarize the theoretical predictions of
DAT, review the criteria for valid retrospective tests, and analyze the historical random number genera
tor (RNG) database. In addition, we summarize the findings from one prospective test of DAT and
comment on the published criticisms of an earlier formulation, which was then called Intuitive Data
Sorting. We conclude with a discussion of the RNG results that provide a strong circumstantial argu
ment against aforcelike explanation. As part of this review, we show that one biologicalAP experi
ment is better described by an influence model.
Review of Decision Augmentation Theory
Since the formal discussion of DAT is statistical, we will describe the overall context for the develop
ment of the model from that perspective. Consider a random variable, X, that can take on continuous
values (e.g., the normal distribution) or discrete values (e.g., the binomial distribution). Examples of X
might be the hit rate in an RNG experiment, the swimming velocity of single cells, or the mutation rate
of bacteria. Let Ybe the average of X computed over n values, where n is the number of items that are
collected as the result of a single decisionone trial. Often this may be equivalent to a single effort
period, but it also may include repeated efforts. The key point is that, regardless of the effort style, the
average value of the dependent variable is computed over the n values resulting from one decision
point. In the examples above, n is the sequence length of a single run in an RNG experiment, the num
ber ofswimming cells measured during the trial, or the number ofbacteriacontaining test tubes present
during the trial. As we will show below, forcelike effects require that the Zscore, which is computed
from the Ys, increase as the square root of n. In contrast, informational effects will be shown to be inde
pendent of n.
Under DAB we assume that the underlying parent distribution of a physical system remains unper
turbed; however, the measurements of the physical system are systematically biased by an ACmediated
informational process. Since the deviations seen in actual experiments tend to be small in magnitude, it
is safe to assume that the measurement biases are small and that the sampling distribution will remain
normal; therefore, we assume the bias appears as small shifts of the mean and variance of the sampling
distribution as:
? The Cognitive Sciences Laboratory has adopted the termanomolousmentaiphenomena insteadof the more widelyknownpsi.
Likewise, we use the terms anomclous cognition and ananalousperturbation forESP and PK, respectively. We have done so
because we believe that these terms are more naturally descriptive of the observabies and are neutral in that they do not imply
mechanisms. These new terms will be used throughout this paper.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 2
Appprlications of Deicision Augmentation hepDy 960079180002002800025 14 May 1995
where ?z and aZ are the mean and standard deviation of the sampling distribution. Under the null hy
pothesis, ?z = 0.0 and az = 1.0.
Review of an Influence Model
For comparison's sake, we summarize a class of influence models. We begin with the assumption that a
putative anomalous force would give rise to a perturbational interaction bywhich we mean that given an
ensemble of entities (e.g., random binary bits), an anomalous force would act equally on each member
of the ensemble, on the average. We call this type of interaction microAP.
In the simplest microAP model, the perturbation induces a change in the mean of the parent distribu
tion but does not effect its variance. We parameterize the mean shift in terms of a multiplier of the
initial standard deviation. Thus:
?i = ?o + Eer o'o~
where ?t and ?a are the means of the perturbed and unperturbed distributions, respectively, and where
aq is the standard deviation of the unperturbed distribution. epp can be considered the AP effect size.
Under the null hypothesis for binary RNG experiments, ?1 = ?p = 0.5, ap = 0.5, and epp = 0.
The expected value and the variance of Z2 for mean chance expectation and under the forcelike and
information assumptions for the normal distribution are shown in Table 1. The details of the calcula
tionscan be found in May, Utts, and Spottiswoode (1994).
Quantit
Mechanisms
y
MCE
MicroAP
DAT
E(ZZ)
1
1 + e,;Pn
?2 + v,
Var(Z~)
2
2(1 + 28~n)
2(vZ + 2?za=)
Figure 1 graphically displays these theoretical calculations for the three mechanisms.
E(ZZ)
large
microAP
s all
DAT
Figure 1. Predictions of MCE, microAl; and DAT
Approved For Release 2000/08/10 :CIARDP960079180002002800025 3
~p rloved For Release 2000/08/10 :CIARDP960079180002002800025
pp ications of Decision Augmentation Theory 14 May 1995
This formulation predicts grossly different outcomes for these models and, therefore, is ultimately ca
pable of separating them, even for very small effects. The important differences are in the slope and
intercept values. MCE gives a slope of zero and an intercept of one. Dat predicts a slope of zero, but an
intercept greater than one, and MicroAP predicts an intercept of one, but a slope greater than zero.
Monte Carlo Verification
The expressions shown in Table 1 are representations which arise from simple algebraic manipulations
of the basic mathematical assumptions of the models. Tb verify that these expressions give the expected
results, we used a published pseudo random number generator (Lewis, 1975) with wellunderstood
properties to produce data that mimicked the results under three models (i.e., MCE, microAP and
DAT). Our standard implementation of the pseudoRNG allows the integers in the range (0,2151) as
potential seeds. For the sequence lengths 100, 500, 1000, and 5000, we computed Zscores for all pos
sibleseeds with an effect size of 0.0 to simulate MCE and an effect size of 0.03 to simulate microAP. 'Ib
simulate DAT, we used the fact that in the special case where the effect size varies as 1 /~/n, microAP
and DAT are equivalent. For this case we used effect sizes of 0.030, 0.0134, 0.0095, and 0.0042 for the
above sequence lengths, respectively. Figures 2ac show the results of 100 trials, which were chosen
randomly from the appropriate Zscore data sets, at each of the sequence lengths for each of the mod
els. In each Figure, MCE is indicated by a horizontal solid line at Z2 = 1.
The slope of a least squares fit computed under the MCE simulation was (2.81 f2.49) x 106, which
corresponded to apvalue of 0.812 when tested against zero, and the intercept was 1.007~O.OOS, which
corresponds to apvalue of 0.131 when tested against one. Under the microAP model, an estimate of
the effect size using the expression in Table 1 was eAp = 0.028$0.002, which is in good agreement with
0.03, the value that was used to create the data. Similarly, under DAT the slope was (2.4457.10) X
10$ which corresponded to a pvalue of 0.515 when tested against zero, and the intercept was
1.OSOf0.001, which corresponds to apvalue of 2.4 X 104when tested against one.
Thus, we are able to say that the Monte Carlo simulations confirm the simple formulation shown in
Table 1.
Retrospective Tests
It is possible to apply DAT retrospectively to any body of data that meet certain constraints. It is critical
to keep in mind the meaning of nthe number of measures of the dependent variable over which to
compute an average durvng a single trial following a single decision In terms of their predictions far
experimental results, the crucial distinction between DAT and the microAP model is the dependence
of the results upon n; therefore, experiments which are used to test these theories ideally should be
those in which experiment participants are blind ton, and where the distribution of n does not contain
extreme outliers.
Aside from these considerations, the application of DAT is straight forward. Having identified the unit
of analysis and n, simply create a scatter diagram of points (Z? n) and compute a weighted least square
fit to a straight line. Table 1 shows that for the microAP model, the slope of the resulting fit is the square
of the AP effect size. Astudent's ttest may be used to test the hypothesis that the APeffect size is zero,
and thus test for the validity of the microAP model If the slope is zero, these same tables show that the
Approved For Release 2000/08/10 :CIARDP960079180002002800025 4
A roved For Release 2000/08/10 :CIARDP960079180002002800025
~~pllcations of Decision Augmentation Theory 14 May 1995
and thus test for the validity of the microAP model If the slope is zero, these same tables show that the
intercept may be interpreted as a strength parameter for DAT. In other words, an intercept lazger than
one would support the DAT model, while a slope greater than zero would support the microAP model.
If the DAT strength is presumed to be constant (i.e., ?l and al are constant) then an additional test is
possible. That is, in two experiments involvingN at n1 and M at n2 decisions, respectively, DAT predicts
that Stouffer's Z's of these experiments should be in the ratio of ~/NT1G1 and ~/1V71G7' x ~/nT for AP.
o.o ~
0 1400 2400 3800
(a)
/~
,/
/'
/'
~~
..
2400// 3600
IC~
Figure 2. Z2 vs n for Monte Carlo Simulations of MCE, microA1? and DAT.
Historical Binary RNG Database
Radin and Nelson (1989) analyzed the complete literature (i.e., over $00 individual studies) of con
sciousnessrelated anomalies in random physical systems. They demonstrated that a robust statistical
anomaly exists in that database. Although they analyzed this data from a number of perspectives, they
report an average Z / ~/n effect size of approximately 3 x 104, regardless of the analysis type. Radin
and Nelson did not report pvalues, but they quote a mean Z of 0.645 and a standard deviation of 1.601
for 597 studies. We compute asinglemean tscore of 9.844, df = 596 (p = 3.7 x 10Z~.
We returned to the original publications of all the binary RNG studies from those listed by Radin and
Nelson and identified 128 studies in which we could compute, or were given, the average Zscore, the
Approved For Release 2000/08/10 :CIARDP960079180002002800025 5
A~~rcaiions of Decis3~on2Augmentation TheoDryP960079180002002800025 14 May 1995
number of runs, N, and the sequence length, n, which ranged fr_ om 16 to 10,000. For each of these stud
 ies we computed:
(1)
 Since we were unable to determine the standard deviations of the Zscores from the literature, we as
sumed that sz = 1.0 for each study. We see from 'Fable 1 that under mean chance expectation the ex
pectedvariance of each Z2 is 2.0 so that the estimated standard deviation for the ~ for a given study is
~/~67A.
Figure 3 shows a portion of the 128 data points (ZZ,n). MCE is shown as a solid line (i.e., ZZ =1), and
the expected bestfit lines for two assumed AP effect sizes of 0.01 and 0.003, respectively, are shown as
short dashed lines. We calculated a weighted (i.e., using NI2.0 as the weights) least squares fit to an a +
b*n straight line for the 128 data points and display it as alongdashed line. For clarity, we have offset
and limited the ZZ axis and have not shown the error bars for the individual points, but the weights and
all the data were used in the least squares fit. We found an intercept of a = 1.0360.004. The 1stan
dard error for the intercept is small and is shown in Figure 3 in the center of the sequence range. The
tscore for the intercept being different from 1.0 (i.e., t = 9.1, df = 126, p = 4.8 x 102~ is in good
agreement with that derived from Radin and Nelson's analysis. Since we set standard deviations for all
the Z's equal to one; and since Radin and Nelson showed that the overall standard deviation was 1.6, we
would expect that our analysis would be more conservative than theirs because a larger standard devi
ationwould increase our computed value for the intercept.
The important result, however, was that the slope of the bestfit line was b = (1.73f3.19) x 10 _6
(t = 0.543, df = 126y p = 0.295), which is not significantly different from zero. Adding and subtracting
one standard error to the slope estimate produces and interval that encompasses zero. Even though a
very small AP effect size might fit the data at large sequence lengths, it is clear in Figure 3 what happens
at small sequence lengths; an eAp = 0.003, suggests a linear fit that is significantly below the actual fit.
2000 4000 6000
Sequence Length (n)
Figure 3. Binary RNG Database: Slope and Intercept for Best Fit Line
Approved For Release 2000/08/10 :CIARDP960079180002002800025 6
A~pl~ca~fions?of Dec~s~on AugmentationlTheory960079180002002800025 y4 May 1995
The sequence lengths from this database are not symmetric nor are they uniformly distributed; they
 contain outliers (i.e., median = 64, average = 566). Figure 4 shows that the lower half of the data, how
ever, is symmetric and nearly uniformly distributed (i.e., median = 35, average = 34). Since the criteri
onfor avalid retrospective test is that n should be uniform, or at least not contain outliers, we analyzed
  the two median halves independently. The intercept for the weighted bestfit line for the uniform lower
half is a =1.0220.006 (t = 3.63, df = 62, p = 2.9 x 10~, and the slope is b = (0.0343.70) x 10_4
(t = 0.010, df = 62, p = O.SO4). The fits for the upper half yield a =1.0640.005 (t =13.47, df = 62, p
=1.2 X 104j) and b = (4.52f2.38) x 10_d (t = 1.903, df = 62, p = 0.969), for the intercept and
slope, respectively.
Since the best retrospective test for DAT is one in which the distribution of n contains no outliers, the
statistically zero slope for the fit to the lower half of the data is inconsistent with a simple AP model.
Although the same conclusion wind be reached from the fits to the database in its entirety (i.e., Figure
3), we suggest caution in that this fit could possibly be distorted by the distribution of the sequence
lengths. That is, a few points at large sequence lengths can easily influence the slope. Since the slope for
the upper half of the data is statistically slightly negative, it is problematical to assign an imaginary AP
effect size to these data. Mare likely, the results are distorted by a few outliers in the upper half of the
data.
o.oo ~ ~ ~ ~ ~ ~ ~ ~
O 20 40 60 80 100
Sequence Length (n)
Figure 4. Historical Database: Distribution of Sequence Lengths < 64.
From these analyses, it appears that Z2 does not linearly depend upon the sequence length; however,
since the scatter is so large, even a linear model is not a good fit (i.e., XZ =171.2, df = 125, p = 0.0038),
whereX~ is a goodnessoffit measure in general given by:
where the of are the errors associated with data point yj, fj is the value of the fitted function at point j, and
v is the number of data points.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 7
ov d For elea a 000/08/10 :CIARDP960079180002002800025
~~~ica~ions of ecisT~on ugmentation Theory 14 May 1995
A "good" fit to a set of data should lead to anonsignificant X2. The fit is not improved by using higher
order polynomials (i.e.,X2 =170.8, df =124; XZ =174.1, df =123; for quadratics and cubits, respective
ly). If, however, the AP effect size was any monotonic function of n other than the degenerate case
where the AP effect size is exactly proportional to 1 /~/n, it would manifest as a nonzero slope in the
regression analysis.
Within the limits of this retrospective analysis, we conclude for RNG experiments that we must reject all
influence models which propose a shift of the mean of the parent distribution.
Princeton Engineering Anomalies Research Laboratory RNG Data
The historical database that we analyzed does not include the extensive RNG data from the Princeton
Engineering Anomalies Research (PEAR) laboratory since the total number of bits in their experi
ments exceeds the total amount in the entire historical database. For example, in a recent report Nel
son, Dobyns, Dunne, and Jahn (1991) analyze 5.6 X 106 trials all at n = 200 bits. In this section, we
apply DAT retrospectively to their published work where they have examined other sequence lengths;
however, even in these cases, they report over five times as much data as in the historical database.
Jahn (1982) reported an initial RNG data set with a single operator at n = 200 and 2,000. Data were
collected both in the automatic mode (i.e., a single button press produced 50 trials at n) and the manual
mode (i.e., a single button press produced one trial at n). From a DAT perspective, data were actually
collected at four values of n (i.e., 200, 2000, 200 X SO =10,000, and 2000 X SO =100, 000). Unfortunately
data from these two modes were grouped together and reported only at 200 and 2, 000 bit/trial. It would
seem, therefore, we would be unable to apply DAT to these data. Jahn, however, reports that the differ
entmodes "...give little indication of importance of such factors in the overall performance." This qual
itativestatement suggests that the microAP model is indeed not a good description for these data, be
cause, under microAP, we would expect stronger effects (i.e.,higher Zscores) at the longer sequence
lengths.
Nelson, Jahn, and Dunne (1986) describe an extensive RNG and pseudoRNG database in the manual
mode only (i.e., over 7 X 106 trials); however, whereas Jahn provide the mean and standazd deviations
for the hits, Nelson et al. report only the means. We are unable to apply DAT to these data, because any
assumption about the standard deviations would be highly amplified by the massive data set.
As part of a cooperative agreement in 1987 between PEAR and the Cognitive Sciences Program at SRI
International, we analyzed a set of RNG data from a single operator.* Since they supplied the raw data
for each button press, we were able to analyze this data at two extreme values of n. We combined the
individual trial Zscores for the high and low aims, because our analysis is twotailed, in that we examine
Z2.
Given that the data sets at n = 200 and 100, D00 were independently significant (Stouffer's Z of 3.37 and
2.45, respectively), and given the wide separation between the sequence lengths, we used DAT as a ret
rospective test on these two data points.
? We thank R. Jahn, B. Dunne, and R. Nelson for providing this raw data for our analysis in 1987.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 8
Approved For Release 2000/08/10 :CIARDP960079180002002800025 14 Ma 1995
Applications of Decision Augmentation Theory Y
Because we are examining only two values of n, we do not compute abestfit slope. Instead, as outlined
in May, Utts, and Spottiswoode (1994), we compare the microAP prediction to the actual data at a
single value of n.
At n = 200, 5918 trials yielded ~ = 0.044 ~ 1.030 and ~ =1.063 f 0.019. We compute a proposed AP
effect size ~ 1 ~/~ = 310 X 103. With this effect size, we computed what would be expected under
the microAP model at n =100,000. Using the theoretical expressions in 'Fable 1, we computed ~ _
1.961 f 0.099. The 1sigma error is derived from the theoretical variance divided by the actual number
of trials (S97) at n =100, 000. The observed values were ~ = 0.100 f 0.997 and ~ =1.002 f O.OSO. A
ttest between the observed and expect values of ~ gives t = 8.643, df = 1192. Considering this t as
equivalent to a Z, the data at n = 100,000 fails to meet what would be expected under the influence
model by 3 t5a. Suppose, however, that the effect size observed at n =100, 000 (3.18 X 10 ~ better
represents the AP effect size. We computed the predicted value of ~ =1.00002 f 0.018 for n = 200.
Using a ttest for the difference between the observed value and this predicted one gives t = 2.398,
df=11,834. The microAP model fails in this direction by more than 2.3a. DAT predicts that ~ would
be statistically equivalent at the two sequence lengths, and we find that to be the case (t = 1.14, df =
6513, p = 0.127).
Jahn (19$2) indicates in their RNG data that "TYaced back to the elemental binary samples, these values
imply directed inversion from chance behavior of about one or one and a half bits in every one thou
sand...:' Assuming 1.S excess bits/1000, we calculate an AP effect size of 0.003, which is consistent with
the observed value intheir n = 200 data set. Since this was the value we used in our DAT analysis, we are
forced to conclude that this data set from PEAR is inconsistent with the simple microAP model, and
that Jahn's statement is not a good description of the anomaly.
We urge caution in interpreting these calculations. As is often the case in a retrospective analysis, some
of the required criteria for a meaningful test are violated. These data were not collected when the oper
atorswere blind to the sequence length. Secondly, these data represent only a fraction of PEAR's data
base.
A Prospective Test of DAT
In developing a methodology for future tests, Radin and May (1986) worked with two operators who
had previously demonstrated strong ability in RNG studies. They used apseudoRNG, which was
based on ashiftregister algorithm by Kendell and has been shown to meet the general criteria for "ran
domness" (Lewis,197S), to create the binary sequences.
The operators were blind to which of nine different sequences (i.e., n =101, 201, 401, 701, 1001, 2001,
4001, 7001, 10001 bits)` were used in any given trial, and the program was such that the trials lasted for a
fixed time period and feedback was presented only after the trial was complete. Thus, the criteria for a
valid test of DAT had been met, except that the source of the binary bits was apseudoRNG.
We reanalyzed the combined data from this experiment with the current Zscore formalism of DAT.
For the 200 individual runs (i.e.10 at each of the sequence lengths for each of the two participants) we
found the best fit line to yield aslope = 4.3 X 10'8 f 1.6 X 10'6 (t = 0.028, df = ~ p = 0.489) and an
intercept =1.16 ~ 0.06 (t = 2.89, df = $, p = 0.01). The slope interval easily encompasses zero and is
? The original IDS analysis required the sequence lengths to be odd because of the logarithmic formalism.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 9
~~~r~~~tons ofDeep"iasl~on2Augmenfa~lon TfieoDy 960079180002002800025 14 May 1995
not significantly different from zero, the intercept significance level (p = 0.01) is consistent with what
Radin and May reported earlier.
Since the pseudoRNG seeds and bit streams were saved for each trial, it was possible to determine if
the experiment sequences exactly matched the ones produced by the shift register algorithm; they did.
Since their UNIXbased Sun Microsystems workstations were synchronized to the system clock, any
momentary interruption of the clock would "crash" the machine, but no such crashes occurred. There
fore, we believe no forcelike interaction occurred.
Tb explore the timing aspects of the experiment Radin and May reran each run with pseudoRNG seeds
ranging from 5 to +5 clock ticks (i.e., 20 ms/tick) from the actual seed used in the run. We plot the
resulting run effect sizes, which we computed from the experimental Fratios (Rosenthal, 1991), for
operator 531 in Figure 5. The estimated standard errors are the same for each seed shift and equal
0.057.
6 4 2 O 2 4 6
Figure 5. Seed Timing for Operator 531(298 Runs).
Radin and May erroneously concluded that the significant differences between zero and adjacent seed
positions was meaningful, and that the DAT ability was effective within 20 milliseconds. In fact, the
situation shown in Figure 5 is expected. Differing from true random number generators in which slight
changes in timing produce essentially the same sequence, pseudoRNGs produced totally different se
quences as afunction of single digit seed changes. Thus, it would be surprising if the seedshift display
produced anything but a spike at seed shift zero. We will return to this point in our analysis of some of
the published remarks on our theory.
From this prospective test of DAT, we conclude that for pseudoRNGs it is possible to select a proper
entry point into a bit stream to produce significant deviations from mean chance expectation that are
independent of sequence length.
The Literature: Review and Comment
We have identified six published articles that have commented upon the Intuitive Data Sorting theory,
the earlier name for DAT. In this section, we chronologically summarize and comment on each report.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 10
A~~icavt~Ons a~ Deces~on Augmeniaion Theory 960079180002002800025 14 May 1995
Walker September 1987
In his first of two criticisms of Intuitive Data Sorting (IDS), Walker (1987) suggested that his Monte
Carlo simulations did not fit the predictions of the model. He generated a single deviant set of 100 bits
(i.e., Z = 2.33, p = 0.01), and he inserted this same sequence as the first 100 bits of 400 otherwise ran
domlygenerated sequences ranging from 100 to 106 bits in length. Walker's analysis of these sequences
did not yield a least square's slope of 0.5 as predicted under the IDS formalism. Walker concluded that
the model was incorrect. Walker's sequences, however, aze not the type that are generated in AP ex
periments orthe type for which the IDS model is valid.
May et al. (1985) were explicit about the character of the sequences that fit the IDS model. Specifically,
Walker quotes May et al. "Using psiacquired information, individuals are able to select locally deviant
subsequences from a large random sequence." (Italics aze used in the original May paper.) The very next
sentence on page 249 of the reference says, "Such an ability, if mediated by precognition, would allow
individuals (subjects or experimenters) to initiate a collection unit of continuous samples (this has been
reported as a trial, a block, a run, etc.) in such a way as to optimize the final resuk. (Italics added here for
emphasis.) Walker continued, "Indeed, the only way the subject can produce results that agree with the
data is to wait for an extrachance run that matches the experimental run length." In the final analysis,
Walker actually supported our contention that individuals select deviant subsequences. Both from our
text and the formalism in our 1985 paper, it is clear that what we meant by a "large random sequence,"
was large compazed to the trial length, n.
In his second criticism of IDS in the same paper, Walker proposed that individuals would have to exhibit
a physiologically impossible control over timing (e.g., when to press a button). As evidence apparently
in favor of such an exquisite timing. ability, he referred to the data presented by Radin and May (1986)
that we have discussed above. (Please see Figure 5.) Walker suggested that Radin and May's result,
therefore, supported his quantum mechanical observer theory. It is beyond the swpe of this paper to
critique Walker's quantum mechanical models, but we would hope they do not depend upon his analysis
of Radin and May's results. The enhanced hitting at zero seed and the suppressed values ~ one 20 ms
clock tick that we show in Figure S is the expected result based upon the wellunderstood properties of
pseudoRNG's and does not represent the precision of the operator's reaction time.
We must consider how it is possible with normal human reactions to obtain significant scores, which can
only happen in 20 ms windows. In typical visual reaction time measurements, Woodworth and Sclilos
berg (1960) found a standard deviation of 30 ms. If we assume these human reactions are typical of
those for AC performance and are normally distributed, we compute the maximum probability of being
within a 20 ms window (i.e., centered about the mean) of 23.5%. For the worst case, the operators must
"hit" significant seeds less often than 23.5% of the time. Radin and May do not report the number of
significant runs, so we provide aworstcase estimate. Given that they quote apvalue of 0.005 for 800
trials, we find that 39 trials must be independently significant. That is, the accumulated binomial proba
bility is ~ OOS for 39 hits in 800 trials with an event probability of 0.05. This corresponds to a hitting rate
(i.e., 39/500) of only 7.8%, a value well within the capability of human reaction times. We recognize that
it is not a requirement to hit only on significant seeds; however, all other seeds leading to positive Z
scoresare less restrictive than the case we have presented.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 11
A~~~i?ca~ions ofDecels~on~lugmen/ta~ian TheoDryP960079180002002800025 14 May 1995
The zerocenter "spike" in Figure 5 misled Walker and others into thinking that exceptional timing was
 required to produce the observed deviations. As we have shown this is not the case, and, therefore,
Walker's second criticism of the IDS theory is not valid.
 Bierman 1988
Bierman (1988) attempted to test the IDS model with a gifted subject. His experimental design ap
peared tomeet most of the criteria for a valid test of the model; however, Bierman found no evidence of
an anomaly and stated that no conclusions could be drawn from his work. We encourage Bierman to
continue with this design and to be speck with what he would expect to see if DAT were the wrrect
mechanism compared to if it were not.
Braud and Schlitr 1989
Braud and Schlitz (1989) conducted an electrodermal PK experiment specii"ically to test the IDS model.
They argued that if the mechanism of the effect were informational, then. allowing participants more
opportunities to select locally deviant values of the dependent variable should yield stronger effects. In
their experiment, 12 electrodermal sampling epochs were either initiated individually by a press of a
button, or a1112 were determined as a result of the first button press. Braud and Schlitz hypothesized
that under IDS, they would expect to see a larger overall effect in the former condition. They found that
the single button press data yielded a significant result; whereas the multiple press data scored at chance
(tom&x[31 J = 214, p = 0.02, tm,,u;[31 J = 0.53). They correctly concluded, therefore, that their data
were more consistent with an influence mechanism than with an informational one.
One implication of their result, which is supported by Braud's 1990 study (see below), is that perhaps
there is something unique about biological systems that allow forcelike interactions, whereas physical
systems such as RNGs do not.
Vassy 1990
Vassy (1990) used a similar timing argument to refute the IDS model as did Walker (1987). Vassy gener
atedpseudoRNG single bits at a rate of one each 8.7 ms. He argued that if IDS were operating, that a
subject would be more likely to identify bursts of ones rather than single ones given the time between
consecutive bits. While he found significant evidence for the primary task of "selecting" individual bits,
he found no evidence that these hits were imbedded in excess clusters of ones.
We compute that the maximum probability of a hit within an 8.7 ms window centered on the mean of the
normal reaction curve with a standard deviation of 30 ms (Woodworth and Schlosberg,1960) is I1.5%.
Vassy quotes an overall Zscore for 100 runs of 239. From this, we compute a mean Z of 0.239 for each
run of 36 bits. 'Ib obtain this result requires an excess hitting of 0.717 bits, which corresponds to an ex
cesshitting rate of 2%. Given that 11.5% is the maximum one can expect with normal human reaction
times, Vassy's results easily allow for individual bit selection, and, thus, cannot be used to reject the DAT
model on the basis of timing.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 12
Approved For Release 2000/08/10 :CIARDP960079180002002800025
Applications of Decision Augmentation Theory 14 May 1995
Braud 1990
In a cooperative effort with SRI International, Braud (1990) conducted a biological AP study with hu
manred blood cells as the target system. The studywas designed, in part, as a prospective test of DAT, so
all conditions for a valid test were satisfied. Braud found that a significant number of individuals were
independently able to "slow" the rate of hemolysis (i.e., the destruction of red blood cells in saline solu
tion) in what he called the "protect" mode. Using data from the nine significant participants, Braud
found support in favor of microAP over DAT. Figure 6 shows the results of our reanalysis of all of
Braud's raw data using our more modern formalism of DAT.
O Effort Data
^ Control Data
x Predicted AP
? Predicted DAT
O 2 4 6 8 1D
Number of'Ibst'Iiibes
Figure 6. DAT Analysis of Hemolysis Data.
The solid line indicates the theoretical mean chance expectation. The squares are the mean values of
Z~ for the control data, and the error bazs indicate the 1standard error for the 32 trials in the study. We
notice that the control data with eight test tubes is significantly below chance (t = 2.79, df = 6~ p =
0.996). Compared to the chance line, the effort data is significant (t = 4.04, df = 31, p = 7.6 x YO5) for
eight test tubes and nearly so for n = 2 (t = 2.06, df = 3l, p = 0.051). The x at n = 8 indicates the
calculated value of the mean of Z~ assuming that the effect size at n = 2 was entirely because of AP;
similarly, the x at n = 2 indicates the calculated value assuming that the effect size, which was observed
at n = 8, was totally due to AP. These AP predictions are not significantly different from the observed
data (t = 0.156, p = 0.431, df = 62 and t = 0.906, p = 0.184, df = 62, at n = 2 and 8, respectively).
Whereas DAT predicts no differences between the data at the end points for n, we find a significant
difference (t =2.033, p = 0.023, df = 62). That is, to a statistical degree the data at n = 8, cannot be
explained by selection alone. Thus, we concur with Braud's original conclusion; these results indicate a
possible forcelike relationship between mental intent and biological consequences.
It is difficult to conclude from our analysis of a single study with only 32 trials that AP is part of nature;
nonetheless, this result is very important. Taken with the results of Braud and Schlitz (1989) the evi
dence ofpossible AP on biological systems is growing. May and Vilenskaya (1993) and Vilenskaya and
May (1995) report that the preponderance of the research on anomalies in the Former Soviet Union is
Approved For Release 2000/08/10 :CIARDP960079180002002800025 13
A proved For Release 2000/08/10 :CIARDP960079180002002800025
A~jilications of Decision Augmentation Theory 14 May 1995
the study of AP on biological systems. Their operators, as do ours, report their internal experiences
suggestive of a forcelike connection between them and their biological targets.
Dobyns 1993
Dobyns (1993) presents a method for comparing what he calls the "influence" and "selection" models,
corresponding to what we have been calling DAT and microAP. He uses data from 490 "tripolar sets" of
experimental runs at PEAR. For each set, there was a high aim, a baseline and a low aim condition.
The three values produced were then sorted into which one was actually highest, in the middle, and
lowest for each set. The data were then summarized into a 3 X 3 matrix, where the rows represented the
three intentions, and the columns represented the actual ordering. If every attempt had been success
ful, the diagonal of the matrix would consist of the number of tripolar sets, namely 490. We present the
data portion of Dobyns' Table from page 264 of the reference as our 'Table 2:
Scoring Data From Dobyns (1993)
Actual
Intention
High
Middle
Low
Total
High
180
167
143
490
Baseline
1S9
156
175
490
Low
151
167
172
490
Total
490
490
490
Dobyns computes an aggregate likelihood ratio of his predictions for the DAT and microAP models
and concludes in favor the the influence model with a ratio of 2$.9 to one.
However, there are serious problems with the methods used in Dobyns' paper. In this paper we outline
only two of the difficulties. Tb fully explain them would require a level of technical discussion not suit
able for a short summary such as this.
One problem is in the calculation of the likelihood ratio function using his Equation 6, which we repro
duce from page 265 of the reference:
B~P/4~ = P~PZ P~  rP~l~` PZ ~ P3 ~
4~q2 q3  Lg1J [4s] [q3,
wherep and q are the predicted rank frequencies for each aim under the influence and selection models,
respectively, and the n are the observed frequencies for each aim. We agree that this relationship cor
rectlygives the likelihood ratio for comparing the two models for one row of Table 2. However, immedi
atelyfollowingthat equation, Dobyns writes, "The aggregate likelihood of the hypothesis over all three
intentions may be calculated by repeating the individual likelihood calculation for each intention, and
the total likelihood will simply be the product of factors such as (6) above for each of the three inten
tions."
Approved For Release 2000/08/10 :CIARDP960079180002002800025 14
A ov d For ~ele~alse 000/08/10, : C~4RDP960079180002002800025
A~~li~cat~ons of ec s on ugmentation haory 14 May 1995
That statement is incorrect. A combined likelihood is found by multiplying the individual likelihoods
only if the random variables are independent of each other (DeGroot, 1986, p. 145). Clearly, the rows
of the table are not independent. In fact, if you know any two of the rows, the third is determined exact
ly. The correct likelihood ratio needs to build that dependence into the formula.`
A second technical problem with the conclusion that the data support the influence model is that the
method itself strongly supports the influence model. As noted by Dobyns, "In fact, applying the test to
data sets that, by construction, contain no effect, yields strong odds (ranging, in a modest Monte Carlo
database, from 8.5 to over 100) in favor of the influence model (page 268)." The actual data in his paper
yielded odds of 28.9 to one in favor of the influence model; however, this value is well within the re
portedlimitsfrom his "influenceless" Monte Carlo data.
Under DAT it is possible that ACmediated selection might occur at the protocol level, but the primary
way is through timinginitiating a run to capitalize upon a locally deviant subsequence. How this might
work in dynamic RNG devices is clear; wait until such a deviant sequence is in your immediate future
and initiate the run in time to capture it. With "static" devices, such as PEAR's random mechanical
cascade device, how timing enters in is less obvious. Under closer inspection, however, even with this
device there is a statistical variation among unattended control runs. That is, there is never a series of
control runs that give exactly the same mean. Physical effects, such as Browian motion, temperature
gradients, etc., can account for the observed variance in the absence of human operators. Thus, when a
run is initiated to capture favorable local "environmental" factors, even for "static" devices, remains
the operative issue with regard to DAT. Dobyns does not consider this case at all in his analysis. If DAT
enters in at the protocol selection, as it probably does, it is likely to be a secondorder contribution be
cause of the limited possibilities from which to select (i.e., six in the tripolar case).
Finally, a major problem with Dobyns' conclusion, which was pointed out when he Hirst presented this
paper at a conference (May, 1990), is that the likelihood ratio supports the influence model even for
their pseudoRNG data. Dobyns dismisses this finding (page 268) all too easily given. the preponder
ance of evidence that suggest that no influence occurs during pseudoRNG studies (Radin and May,
1986).
Aside from the technical flaws in Dobyns' likelihood ratio arguments, and even ignoring the problem
with the pseudoRNG analysis, we reject his wnclusions simply because they hold in favor of influence
even in Monte Carloconstructed unperturbed data.
Circumstantial Evidence Against an AP Model for RNG Data
Experiments with hardware RNG devices are not new. In fact, the title of Schmidt's very first paper on
the topic (1969) portended our conclusion, "Precognition of a Quantum Process." Schmidt lists PK as a
third option after two possible sources for precognition, and remarks, "The experiments done so far do
not permit a distinction (if such a distinction is at all meaningful) between the three possibilities." From
1969 onward, the RNG research has been strongly oriented toward a PK model. The term microPK,
itself, embeds the force concept further into the lexicon of RNG descriptions.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 15
~~~ii'~~~~r~~f ~~~~Tt ~~t~4ior~1~960079180002002800025 14 May 1995
In this section, we examine a number of RNG experimental results that provide circumstantial evidence
against the AP hypothesis. Any single piece of evidence could be easily dismissed; however, taken to
gether, they demonstrate a substantial case against AP.
Internal Complexity of RNG Devices and Source Independence
Schmidt (1974) conducted the first experiment to explore potential dependencies upon the internal
workings of his generators. Since by definition AP implies a force or influence, it seemed reasonable to
expect that an influence should depend upon the details of the target system. In this study, one genera
tor produced individual binary bits, which were derived from the Rdecay of 90Sr, while the other
"binary" output was a majority vote from 100 bits, each of which were derived from a fast electronic
diode. Schmidt reports individually significant effects with both generators, yet does not observe a sig
nificantdifference between the generators.
This particular study is interesting, quite aside from the timing and majority vote issues; the binary
streams were derived from fundamentally different physical sources. Radioactive Rdecay is governed
by the weak nuclear force, and electronic devices (e.g., noise diodes) are governed by the electromag
neticforce. Schematically speaking, the electromagnetic force is approximately 1,000 times as strong as
the weak nuclear force, and modern highenergy physics has shown them to be fundamentally different
after about 1010 seconds after the big bang (Raby, 1985). Thus, a putative APforce would have to
interact equally with these two forces; and since there is no mechanism known that will cause the elec
tromagneticand weak forces to interact with each other, it is unlikely that AP will turn out to be the first
coupling mechanism. The lack of difference between ~decay and noise diode generators was con
firmedyears later by May et al. (19$0).
We have already commented upon one aspect of the timing issue with regard to Radin and May's (1986)
experiment and the papers by Walker (1987) and Vassy (1990). May (1975) introduced a scheme to
remove any firstorder biases in binary generators that also is relevant to the timing issue. The output of
his generator was a match or antimatch between the random bit stream and a target bit. One mode of
the operation of the device, which May describes, included an oscillating target bitone oscillation per
bit at approximately 1 MHz rate.` May and Honorton (1975) and Honorton and May (1975) reported
significant effects with the RNG operating in this mode. Thus, significant effects can be seen even with
devices that operate in the microsecond time domain, which is three orders of magnitude faster than
any known physiological process.
Effects with Pseudorandom Number Generators
Pseudorandom number generators are, by definition, those that depend upon an algorithm, which is
usually implemented on a computer. Radin (1985) analyzed all the pseudoRNGs commonly in use and
found that they require a starting value (i.e., a seed), which is often derived from the computer's system
clock. As we noted above, Radin and May (1986) showed that the bit stream, which proved to be "suc
cessful" in a pseudoRNG study, was bitforbit identical with the stream, which was generated later,
but with the same seed. With that generator, at least, there was no change from the expected bit stream.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 16
Approved For Release 000/0841 Q : CII~RDP960079180002002800025 14 Ma 1995
Applications of Decision ugmen anon eory Y
Perhaps it is possible that the seed generator (i.e., system clock) was subjected to some AP interaction.
We propose two azguments against this hypothesis:
(1) Even one cycle interruption of a computers' system clock will usually invoke a system crash; an
event not often reported in pseudoRNG experiments.
(2) Computers use crystal oscillators as the basis for their internal clocks. Crystal manufacturers usual
 ly quote errors in the stated oscillation frequency of the order of 0.001 percent. That translates to
500 cycles fora 50 MHz crystal, or to 10 ?s in time. Assuming that the quoted error is a 10 estimate,
and that a putative AP interaction acts at within the f 20 domain, then shifting the clock by this
amount might account for only one seed shift in Radin and May's experiment. By Monte Carlo
methods, we determined that, given a random entry into seedspace, the average number of ticks to
reach a "significant" seed is 10; therefore, even if AP could shift the oscillators by 20, it cannot
account far the observed data.
Since computers inpseudoRNG experiments aze not reported as "crashing" often, it is safe to assume
that pseudoRNG results are only due to AC. In addition, since the results ofpseudoRNG studies are
statistically insepazable from those reported with true RNGs, it is also reasonable to assume that the
mechanisms aze similarly ACbased.
Precognitive AC
Using the tools of modern metaanalysis, Honorton reviewed the precognition cardguessing database
(Honorton and Ferarri,1989). This analysis included 309 separate studies reported by 62 investigators.
Nearly two million individual trials were contributed by more than 50,000 subjects. The combined ef
fect size was E = 0.0200.002, which corresponds to an overall combined effect of 11.40. 'Iivo impor
tant results emerge from Honorton's analysis. First, it is often stated by critics that the best results are
from studies with the least methodological controls. Tb check this hypothesis, Honorton devised an
eightpoint quality measure (e.g., automated recording of data, proper randomization techniques) and
scored each study with regard to these measures. There was no significant correlation between study
quality and study score. Second, if researchers improved their experiments over time, one would expect
a significant correlation of study quality with date of publication. Honorton found r = 0 246, df = 307, p
_ 2 X 10~. In brief, Honorton concludes that a statistical anomaly exists in this data that cannot be
explained by poor study quality or a large variety of other hypotheses including the file drawer; there
fore, apotential mechanism underlying DAT has been verified.
SRI International's RNG Experiment
May, Humphrey, and Hubbard (19$0) conducted an extensive RNG study at SRI International in 1979.
They applied stateoftheart engineering and methodology to construct two true RNGs, one based on
the ~decay of 137Pm and the other based on an MD20 noise diode from Texas Instruments. It is be
yond the scope of this paper to describe, in detail, the intricacies of this experiment; however, we will
discuss those aspects that are pertinent to this discussion.
Technical Details
Each of the two sources were battery operated and optically coupled to a Digital Equipment Corpora
tion I.SI 11/23 computer. Failsafe circuitry would disable the sources if critical physical parameters
(e.g., batteryvoltages and currents, temperature) exceed preset ranges. Both sources were subjected to
Approved For Release 2000/08/10 :CIARDP960079180002002800025 1~
Applications of Decision Augmentation TheoDy 960079180002002800025 14 May 1995
environmental testing which included extreme temperature cycles, vibration tests, E&M and nuclear
gamma and neutron radiation tests. Both sources behaved as expected, and the critical parameters,
such as temperature, were monitored and their data stored along with the experimental data.
A source was sampled at 1 KHz rate. After eight milliseconds, the resulting byte was sent tothe comput
erwhilethe next byte was being obtained. In this way, a continuous stream of 1 ms data was presented to
the computer. May et al. had specified, in advance, that bit number 4 was the designated target bit.
Thus each byte provided 3 ms of bits prior to the target and 4 ms of bits after the target bit.
A trial was defined as a definitive outcome from a sequential analysis of bit four from each byte. In
exchange for not specifying the number of samples in advance, sequential analysis requires that the
Type I and Type II errors, and the chance and extrachance hitting rate be specified in advance. In May
et al.'s twotailed analysis, a = (3 = 0.05 and the chance and extrachance hitting rate was O.SO and 0.52,
respectively. The expected number of samples to reach a definitive decision was approximately 3,000.
The outcome from a single trial could be in favor of a hitting rate of O.S2, 0.48, or at chance of 0.50, with
the usual risk of error in accordance with the specified Type I and Type II errors.
Each of seven operators participated in 100 trials of this type. For an operator's data to reach indepen
dentlystatistical significance, the operator had to produce 16 successes in 100 trials, where a success was
defined asextrachance hitting (i.e., the exact binomial probability of 16 successes for 100 trials with an
event probability of 0.10 is 0.04 where one less success is not significant). 'Iivo operators produced 16
and 17 successful trials, respectively.
Temporal Analysis
We analyzed the 33 trials from the two independently significant operators from May et al.'s experi
ment. Each of the 33 trials consisted of approximately 3,000 bits of data with 3 bits and +4 bits of 1
ms/bit temporal history surrounding the target bit. We argue that if the significance observed in the
target bits was because of AP, we would expect a large correlation with the target bit's immediate neigh
bors,whichare only ~ 1 ms away. As far as we know, there is no known physiological process that can be
cognitively, or in any other way, manipulated within a millisecond. We might even expect a 100% cor
relationunder the complete AP model.
We computed the linear correlation coefficients between bits 3 and 4, 4 and 5, and 3 and 5. For binary
data:
where ~ is the linear correlation coefficient and N is the number of samples. Since we examined three
different correlations for 33 trials, we computed 99 different values of N?Z. Four of them producedXZs
that were significant well within chance expectation. The complete distribution is shown in Figure 7.
We see that there is excellent agreement of the 99 correlations with theX2 distribution for one degree of
freedom, which is shown as a smooth curve.
We wnclude, therefore, that there was no evidence beyond chance to suggest that the target bit neigh
bors were affected even when the target bit analysis produced significant evidence for an anomaly. So,
knowing the physiological limitations of the human systems, we further concluded that the observed
effects could not have arisen due to ahumanmediated force (i.e., AP).
Approved For Release 2000/08/10 :CIARDP960079180002002800025 113
~~~r~~~ons of~~ce~ion~ g~ie~tla~ibn Tr~~~y 960079180002002800025 14 May 1995
~
XZ (df = 1)
Figure 7. Observed and Theoretical Correlation Distributions.
Mathematical Model of the Noise Diode
Because of the unique construction parameters of Texas Instrument's MD20 noise diode, May et al.
were able to construct a quantum mechanical model of the detailed workings of the device. This model
contained aU known properties of the material and it's construction parameters. For example, the band
gap energy in Si, the effective mass of an electron or hole in the semiconductor, and the impurity con
centration were among the parameters for the model. The model was successful at calculating the
diode's known and measured behavior as a function of temperature. May et al. were able to simulate
their RNG experiment down to the quantum mechanical details of the noise source. They hoped that
by adjusting the model's parameters so that the computed output agreed with the experimental one,
that they could gain insight as to where the influence "entered" the device.
May et al. were not able to find a set of model parameters that mimicked their RNG data. For example,
changing the band gap energy for short periods of time; increasing or reducing the electron's effect
mass; or redistributing or changing the impurity content produced no unexpected changes in the device
output. The only device behavior that could be effected was its known function of temperature.
Because of the construction details of the physical RNG, this result could have been anticipated. The
changes that could be simulated in the model were all in the microsecond domain because of the details
of the device. Both with the RNG and in its model, the diode's multiMHz output was filtered by a
100KHz wide bandwidth filter. Thus, any microsecond changes would not pass through the filter. In
short, because of this filtering, the RNG was particularly insensitive to potential changes of the physical
parameters of the diode.
Yet solid statistical evidence for an anomaly was seen by May et al. Since the diode device was shown
mathematically and empirically to be insensitive to environmental and physical changes, these results
must have been as a result of AC rather than AP. In fact, this observation coupled with the bit timing
argument, which we have described above, led May et al. to question forcelike models in RNG studies
in general.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 19
Appproved For Release 2000/08/10 :CIARDP960079180002002800025
A{Sjsllcations of Decision Augmentation Theory 14 May 1995
Summary of Circumstantial Evidence Against AP
We have identified six circumstantial arguments that, when taken together, provide increasingly diffi
cult requirements that must be met by a putative AP force. In summary, the RNG database demon
strates that:
(1) Data are independent of internal complexity of the hardware RNG device.
(2) Data are independent of the physical mechanism producing the randomness (i.e., weak nuclear or
electromagnetic).
(3) Effects with pseudorandom generators are statistically equivalent to those observed with true
hardware generators.
(4) Reasonable AP models of mechanism do not fit the data.
(5) In one study, bits which are ~ 1 ms from a "perturbed" target bit are themselves unperturbed.
(6) A detailed model of a diode noise source, which includes all known physics of the device, could not
simulate the observed data streams.
In addition, AC, which is a mechanism to describe the data, has been confirmed in nonRNG experi
ments. We conclude, therefore, an AP force that is consistent with the database must
? Be equally coupled to the electromagnetic and weak nuclear forces.
? Be mentally mediated in times shorter than one millisecond.
? Follow a 1 /~/ n law.
For these to be true, an AP force would be at odds with an extensive amount of verified physics and
common behavioral observables. We are not saying, therefore, that it cannot exist; rather, we are sug
gestingthat instead of havrrig to force ourselves to invent a whole new science, we should look for ways
in which AP might fit into the present world view. In addition we should invent informationbased and
testable alternate mechanisms for the experimental observables.
Discussion and Conclusions
Our recent results in the study of anomalous cognition (May, Spottiswoode, and James, 1994) suggest
the the quality of AC is proportional to the change in Shannon entropy. Following Vassy (1990), we
compute the change in Shannon entropy for anextrachance, binary sequence of length n. The total
change of entropy is given by:
dS=SpS,
where for an unbiased binary sequence of length n, So = n, and S is given by:
S =  nPiloBzPi  n(1  Pi) /oSx(1  Pi)?
I.etpl = 0.5 (1 + e) and assume that e, the effect size, is small compared to one (i.e., typical RNG effect
sizes are of the order of 3 X 10~. Using the approximation:
we find that S is given by:
S=nn21n2
Approved For Release 2000/08/10 :CIARDP960079180002002800025 20
~p~i~ica~ions of~ecls$on2i4ugmenta~ion Theory 960079180002002800025 14 May 1995
or that the total change of entropy for a biased binary sequence is given by;
z
dS= SoS=n e
21n2
Since our analysis of the historical RNG database shows that the effect size is proportional to 1 / ~/n,
the total change of Shannon entropy becomes a constant that is independent of the sequence length:
We have seen in our other AC experiments (May, Spottiswoode, and James, 1994) that the quality of
the data is proportional to the change of the target entropy. In RNG experiments the quality of the data
is equivalent to the excess hitting, which according to DAT is mediated by AC and should be indepen
dent of the sequence length. We have shown above that the quality of RNG data depends upon the
change of target entropy and is independent of the sequence length. Therefore we suggest that the
change of target entropy may account for successful AC and RNG experiments.
Braud's study of AP on red blood cells and Braud and Schlitz's study on electrodermal effects imply that
there is something unique about living systems. Before we would be willing to declare that AP is a valid
mechanism for biological experiments, more than two, albeit well designed and executed, studies are
needed.
When DAT is applied to the RNG database, a simple forcelike perturbational model fails, by many
orders of magnitude, as a viable candidate for the mechanism. In addition, when viewed along with the
collective circumstantial arguments against aforcelike explanation, it is clear that another model is
required. Any new model must explain why quadrupling the number of bits in the sequence length fails
to produce a Zscore twice as large.
Given that one possible information mechanism (i.e., precognitive AC) can, and has been, indepen
dently confirmed in the laboratory, and given the weight of the empirical, yet circumstantial, arguments
taken together against AP, we conclude that the anomalous results from the RNG studies arise not be
cause of amentally mediated force, but rather because of a human ability to be a mental opportunist by
making ACmediated decisions to capitalize on the locally deviant circumstances.
Generally, we suggest that future studies be designed in such a way that the criteria, as outlined in this
paper and in May, Utts, Spottiswoode (1994), conform to a valid DAT analysis. Our discipline has
evolved to the point where we can no longer be satisfied with yet one more piece of evidence of a statisti
cal anomaly. We must identify the sources of variance as suggested by May, Spottiswoode, and James
(1994); limit them as much as possible; and apply models, such as DAT, which can begin to shed light on
the physical, physiological, and psychological mechanisms of anomalous mental phenomena.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 21
a~pl~catlons of Dec~s~on Augmentation 1,~eO~ 960079180002002800025 14 May 1995
References _
Bierman, D.J. (198$). Testing the IDS model with a gifted subject., Theoretical Parapsychology, 6, 3136.
Braud, W. G. and Schlitz, M. J. (1989). Possible role of Intuituve Data Sorting in electrodermal
biological psychokinesis (BioPK). The Joumal of the American Society for Psychical Research, 83,
No. 4, 289302.
Braud, W. G. (1990). Distant mental influence of rate of hemolysis of human red blood cells. 7h.eJoumal
of theAmerican Society for Psychical Researah, 84, No. 1,124.
DeGroot, Morris H. (1985). Probability and Statistics, 2nd Edition. Reading, MA: AddisonWesley
Publishing Co.
Dobyns, Y. H. (1993). Selection versus influence in remote REG anomalies. Joumal of Scientific
Exploration. 7, No. 3, 259270.
Honorton, C. and May, E. C. (1975). Volitional control in a psychokinetic task with auditory and visual
feedback. Research in Parapsychalogy,1975, 9091.
Honorton, C. and Ferrari, D. C. (1989) "Future Ti;lling:" Ametaanalysis offorcedchoice precognition
experiments, 19351987. Joumal of Parapsychology, S3, 281308.
Jahn, R. G. (19$2). The persistent pazadox of psychic phenomena: an engineering perspecitve.
Proceedings of the IEEE. 70, No. 2,136170.
Lewis, T. G. (1975). Distribution Sampling for Computer Simulation. Lexington, MA: Lexington
Books.
May, E. C. (1975). PSIFI: Aphysiologycoupled, noisedriven random generator to extend PK studies.
Research in Parapsychology, 1975, 2022.
May, E. C. and Honorton, C. (1975). A dynamic PK experiment with Ingo Swann. Research in
Parapsychology,1975, 8889.
May, E. C., Humphrey, B. S., Hubbard, G. S. (19$0). Electronic System Perturbation Techniques. Final
Reporx SRI International Menlo Park, CA.
May, E. C., Radin, D. I., Hubbazd, G. S., and Humphrey, B. S. (1985). Psi experiments with random
number generators: an informational model. Proceedings of Presented Papers Yol I. The
Parapsychological Association 2$th Annual Convention, Tufts University, Medford, MA, 237266.
May, E. C. (1990). As chair for the session at the annual meeting of the Society for Scientific
Exploration in which this original work was presented, I pointed out the problem of the likelihood
ratio for the pseudorandomnumbergenerator data from the floor of the convention.
May, E. C., Spottiswoode, S. James P., and James, C. L. (1994). Shannon entropy as an Intrinsic Target
property: Toward a reductionist model of anomalous cognition. Submitted to The Joumal of
Parapsychology.
May, E. C., Utts, J. M., Spottiswoode, S. J. (1994). Decision augmentation theory: Tbwazd a model of
anomalous mental phenomena. Submitted to The Joumal of Parapsychology.
May, E. C. and Vilenskaya, L. (1994). Overview of current pazapsychology reseazch in the former Soviet
union. Subtle Energies. 3, No 3.4567.
Nelson, R. D., Jahn, R. G., and Dunne, B.J. (1986). Operatorrelated anomalies in physical systems and
information processes. Jonsnal of the Society for Psychical Research, 53, No. 803, 261285.
Nelson, R. D., Dobyns, Y. H., Dunne, B. J., and Jahn, R. G., and (1991). Analysis of Variance of REG
Experiments: Operator Intention, Secondary Parameters, Database Structures. PEAR
Laboratory Tbchnical Report 91004, School of Engineering, Princeton University.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 22
A~j~,ca~lons o?~Decls~on Augm/~r~~~~idn~heo DP960079180002002800025 14 May 1995
Raby, S. (1985). Supersymmetry and cosmology. in Supersymmetry, Supergravity, and Related Topics.
Proceedings of the XVth GIFT International Seminar on Theoretical Physics, Sant Feliu de
Guixols, Girona, Spain. World Scientific Publishing Co. Pte. Ltd. Singapore, 226270.
Radin, D. I. (1985). Pseudorandom Number Generators in Psi Reseazch. Journal of Parapsychology. 49,
No 4, 303328.
Radin, D. I. and May, E. C. (1986). 'I~sting the Intuitive Data Sorting mode with pseudorandom
number generators: A proposed method. The Proceedings of Presented Papers of the 29th Annual
Convention of the Parapsychological Association, Sonoma State University, Rohnert Park, CA,
539554.
Radin, D. I. and Nelson, R. D. (1989). Evidence for consciousnessrelated anomalies in random
physical systems. Foundations of Physics. 19, No. 12, 14991514.
Rosenthal, R. (1991). Metaanalysis procedures for social research. Applied Social Research Methods
Series, Vol. 6, Sage Publications, Newbury Pazk, CA.
Schmidt. H. (1969). Precognition of a quantum process. Journal of Parapsychology. 33, No. 2, 99108.
Schmidt. H. (1974). Comparison of PK action on two different random number generators. Journal of
Parapsychology. 3$, No. 1, 4755.
Vassy, Z. (1990). Experimental study of precognitive timing: Indications of a radically noncausal
operation. Joumal of Parapsychology. 54, 299320.
Vilenskaya, L. and May, E. C. (1994). Anomalous mental phenomena research in Russia and the
Former Soviet Union: A follow up. Submitted to the 1994 Annual Meeting of the
Parapsychological Association.
Walker, E. H. (1987). A comparison of the intuitive data sorting and quantum mechanical observer
theories. Jouma! of Parapsychology, 51, 217227.
Woodworth, R.S. and Schlosberg H. (1960). Experimental Psychology. Rev ed. New York Holt. New
York.
Approved For Release 2000/08/10 :CIARDP960079180002002800025 23