American Institutes for Research

The American Institutes for Research (AIR) is a nonprofit behavioral and social science research organization founded in 1946, headquartered in the Washington, D.C. area. It conducts applied research in education, health, workforce development, and related fields under government and private contracts. AIR's primary relevance to intelligence history is its 1995 evaluation of the STAR GATE program, the classified remote viewing research effort that had operated under various designations since 1978 under Defense Intelligence Agency and, from 1991, CIA sponsorship.¹

The 1995 STAR GATE Evaluation

When the CIA took over management of the STAR GATE program in 1991 and began a review of its scientific and operational value, it contracted AIR to conduct an independent assessment. The evaluation was led by two researchers with opposing predispositions toward the subject: Ray Hyman, a University of Oregon cognitive psychologist and longtime skeptic of parapsychology who had evaluated SRI remote viewing research since 1973, and Jessica Utts, a University of California Davis statistician with expertise in experimental design and more open views on the statistical evidence for psi phenomena.¹

The AIR report, formally titled "An Evaluation of Remote Viewing: Research and Applications," was delivered to the CIA on September 29, 1995. Jessica Utts concluded in her portion of the report that the statistical evidence across the accumulated STAR GATE experiments was sufficiently robust to warrant the conclusion that some form of anomalous cognition was occurring - she found the effect size comparable to that considered sufficient for publication in mainstream behavioral science. Ray Hyman's portion of the report argued that the experimental protocols contained sufficient methodological weaknesses, including inadequate controls against sensory leakage and experimenter bias, that no firm conclusions could be drawn.¹

The overall AIR assessment, reconciling these divided conclusions, recommended against continuing the program. The CIA accepted this recommendation. STAR GATE was declassified and formally terminated in 1995, with the program's archives transferred to the American Institute of Research and later made available through declassification requests. Approximately 80,000 pages of documents were eventually released by the CIA.¹

Significance of the Split Finding

The divided Utts-Hyman finding reflected a genuine methodological disagreement that had characterized the entire history of parapsychology research: whether statistically significant but small effect sizes in carefully controlled experiments constitute evidence for anomalous phenomena, or whether undetected methodological artifacts explain such results. Utts's positive assessment was unusual in that she was a mainstream academic statistician, not a parapsychology proponent; her endorsement of the statistical case attracted attention from researchers on both sides of the debate.

The AIR evaluation was criticized from multiple directions. Proponents of the research, including former program participants and researchers at Stanford Research Institute, argued that Hyman's portion of the review applied higher methodological standards than would be applied to equivalent evidence in other fields, and that the operational record - instances of remote viewing that had produced actionable intelligence - was not adequately weighted. Skeptics argued that any positive finding was the product of accumulated methodological weaknesses that neither evaluator had fully identified.¹

Role of Ray Hyman

Ray Hyman had been involved in government-sponsored evaluations of parapsychology research since 1973, when the DARPA contracted him to evaluate Uri Geller's sessions at SRI. His consistent position was that methodological controls in remote viewing research were insufficient to rule out conventional explanations. The 1995 AIR evaluation represented the culmination of more than two decades of his critical engagement with the field. His portion of the report documented specific methodological concerns about the protocols used in STAR GATE's operational and research phases and argued that the database did not support conclusions about anomalous information transfer.¹