AMSAA RV EXPERIMENT

CIA-STARGATE

PDF Scan: PDF

Open AI Summary

Text

Body:  Approved For Release 2003/09/10 : CIA-RDP96-00788R001200230033-9
 AMSAA RV Experiment
 (1)  Independent replication of SRI experiments in RV with targeting
 by coordinates, is in advanced planning stages at AMSAA.  Initial assessment
 of the AMSAA experimental plan is reported here.  In general, the integrity
 of the experiment appears good.  Also an improved means of providing assurance
 of unequivocal experimental results is being attempted.  In spite of such an
 attempt, however, the initial assessment is that maintaining the SRI protocol
 in evaluating the raw trial results will present a severe limit to the
 credibility of the ultimate results.
 (2)  The contribution to confidence in the results of these tests should
 be high as regards the integrity of the experiment.  Different viewers and
 experimenters and a new target pool,,, together with double blind control
 by an independent group (AMSAA management) should provide certitude that the
 tests represent an unbiased independent replication of RV testing.
 (3)  Two aspects of the experiment cause some concern, and are discussed
 below:  the target characteristics, and the continued use of the basic
 SRI trial judging technique (or "protocol").
 (4)  Target Characteristics:
 The utilization of a large area (100 mile radius) for selection of some
 forty targets provides an opportunity, being exploited by AMSAA, for a target
 pool of fairly distinct and different target characteristics.
 However, the uniqueness of each of the targets is an order of magnitude
 less than that obtainable by utilizing simpler artificial targets, such as
 alphanumerics, geometric shapes, etc.  The impediment to the use of such
 artificially created targets is the often cited deleterious effect on the
 viewer of simplistic "non-interesting" target objects.  The success rate of
 remote viewer "hits" is averred to be greatly degraded when boredom is induced
 by attempting to view such simple targets.  This effect is apparently repeatedly
 observed in parapsychology experiments and presents the scientific experi-
 mentalist with a psychological analogue to Heisenberg's Principle:  The more
 precise is the measurement of the quality of an RV performance, the less
 likely is the probability of an RV occurrence.  Unless this effect is, indeed,
 a natural principle, other means of measurement of RV performance quality
 should be created, to permit unequivocal RV evaluation.   (One suggested set
 of "interesting targets" which are well defined, are moon topographical
 features.  These are well defined photographically, and unexamined in the
 experience of most of the population).
 It is to be observed that the AMSAA target selection criteria are specifically
 aimed at choosing each individual target with features, or combinations of
 features (geometry, color, action, materials, etc.) which are unique:  i.e.-
 any given "state vector" is "orthogonal" to all others.  Unfortunately, the
 features that are selected as the uniqueness-defining elements are subjectively
 Approved For Release 2003/09/10 : CIA-RDP96-00788R001200230033-9
 Approved For Release 2003/09/10 : CIA-RDP96-00788R001200230033-9
 arrived at.  The viewer ("subject") may well define a target ("state vector")
 without using many of the uniqueness-defining elements selected a priori.  Thus
 complete orthogonality of targets is lost, requiring a posteriori subjective
 judgment to be employed to evaluate the RV performance, (and deciding from
 amongst all targets that one which most closely matches the viewer's des-
 cription).
 Comments concerning the judging procedure follow inthe next section.
 The observation to be emphasized here concerns the doubt expressed by the
 committee about the attempt by the AMSAA target selection/description group
 to orthogonalize the targets in the face of a measuring device (the viewer,
 or "subject") that doesn't use the same target parameters to render a description
 (measurement) as the "target describers".  Putting numerical weighting factors
 on each of the target features, as is being planned by AMSAA, is regarded as
 a doubtful means to orthogona.lize what is probably a non-orthogonal set of
 targets as discerned by the remote viewer.  We just don't know the transfer
 function that characterizes the viewer well enough to be sure that heavily
 weighted features are not regarded by the viewer as unimportant--and
 conversely.
 Thus, in spite of a thoughtful and creative attempt to provide a set of
 unambigious, and unambiguously describable, unique targets, the committee
 presently feels that the authentication of Remote Viewing will depend on
 post trial subjective judgment.
 (5)  Judging Procedure:
 There is little planned departure from the SRI post trial judging procedure.
 The principal difference is the employment of several target descriptors to be
 utilized in comparing the remote viewer's "state vector" description with a
 priori descriptors ("state vector coordinates").  The appropriateness of heavy
 dependence on descriptor matching (and weighting) is questioned,  as elucidated
 in section (4), above.  In a pretrial exercise this problem was illustrated
 by failure of the judge to match any of the targets (in a limited trial pool)
 to the remote viewer's description.  By dissecting the viewer's description
 and subjectively arriving at specific descriptors, the judge was able to
 arrive at a match based on a sum of very low correlations of each individual
 descriptor with characteristics of the true target.   (The correlation of these
 descriptors with the other targets were lower).  This matching success was
 regarded by the judge as an achievement.  The committee, on the other hand,
 felt that this means of pulling the signal out of the noise may well have
 demonstrated that an apparent high numerical correlation was produced that
 did not in fact exist.
 Thus we are concerned that the praiseworthy effort to devise an objective
 and quantitative measure of RV performance quality may well generate apparent
 high correlations, reported with a precision which the subjective raw input
 does not justify.
 Approved For Release 2003/09/10 : CIA-RDP96-00788R001200230033-9
 Approved For Release 2003/09/10 : CIA-RDP96-00788R001200230033-9
 In any case, if the experiment proceeds as presently planned, the committee
 suggests that the (subjective) raw as well as derived "!quantitative" data and
 analyses be reported for each trial and its subsequent evaluation.
 (6)  Quantitative Analysis Techniques
 The quantization, weighting, and statistical treatment of the target
 features and remote viewer descriptions contain some techniques which
 should be further scrutinized.  One is the target feature weighting approach,
 which although structured to provide consistency, is highly subjective.  The
 other is the excessive resolution expected of human judges in deciding the
 degree of agreement between target features and viewer descriptions of
 target features.  A scale of seven is a clinically demonstrated limit of
 human resolution.  The planned resolution has a scale of ten.  As subjective
 as the evaluation procedure is, a scale of 5 levels or less is probably
 as large as is justified.
 Approved For Release 2003/09/10 : CIA-RDP96-00788R001200230033-9