The question the review answered
In 1995, the American Institutes for Research reviewed the United States government-sponsored anomalous-cognition research program. Two principal reviewers, statistician Jessica Utts and psychologist Ray Hyman, examined the experimental record and reached different overall judgments. That disagreement matters, but it is often described too loosely. Utts concluded that psychic functioning had been established under the standards normally used in science and that the effect was replicable. Hyman did not accept that the experiments proved a paranormal mechanism, yet he agreed that the laboratory results contained statistically significant departures from chance.
That makes the honest headline stronger than a vague mystery pitch. The reported experimental effect was not merely an anecdote, and the central reviewers did not dispute that the statistics exceeded chance. Their dispute concerned what caused the effect, whether every alternative had been excluded, and how far the laboratory result could be carried into operational intelligence work.
Statistical evidence is not the same as perfect performance
Above-chance performance does not mean that every session succeeds, that every detail is accurate, or that a viewer can identify any target on demand. A small repeatable effect can coexist with many noisy individual trials. This is why one spectacular transcript and one empty transcript are both weak evidence on their own. The useful unit is a body of pre-recorded sessions compared with a stated chance baseline.
Utts described the observed effects as consistent across laboratories and conditions and argued that continued work should focus on how the effect operates. Hyman asked for additional independent replication and tighter exclusion of possible methodological artifacts. The final program evaluation also separated laboratory evidence from operational utility: reviewers did not establish that the program had produced actionable intelligence value consistently enough for continued intelligence use.
What this platform does next
RV Assistant takes the above-chance record as the reason to build better measurement, not as permission to inflate individual results. Targets remain hidden until the record is locked. Feedback is element by element. Silent matching is tested against blank, noise, mismatched and broad-description controls. A personal signal estimate appears only across enough eligible sessions and always carries uncertainty.
The project is therefore aimed at the next practical question: under which conditions can a person distinguish signal from imagination more reliably? A mass-scale platform can compare protocols, rest patterns, confidence calls and target types while publishing null findings as well as positive ones. The promise is not instant perfection. It is a systematic path from an established statistical effect toward repeatability—and, eventually, greater command.