Article
A report of knowing something unexpectedly may matter deeply to the person who experienced it. A research claim asks a narrower question: can a defined procedure show information transfer or influence after ordinary routes have been controlled? Three commonly discussed paradigms test different outcomes, so their findings should not be collapsed into one verdict.
Ganzfeld studies ask about anomalous information transfer
Ganzfeld experiments use a relaxed, low-stimulation setting and ask a participant to identify a target selected or held elsewhere. A 2010 review by Storm, Tressoldi, and Di Risio covered free-response research more broadly; its 29-study Ganzfeld subset produced a positive aggregate. That subset did not exceed the review’s non-Ganzfeld noise-reduction subset, and the paper was not exclusively a Ganzfeld meta-analysis.[1]
An earlier review by Julie Milton and Richard Wiseman combined 30 Ganzfeld studies from seven laboratories and reported a null result. The reviews used different databases and interpretive choices, so they are not a simple matched rerun. The positive aggregate and the null aggregate belong in view together; neither by itself establishes exact zero or verified telepathy.[2]
Retroactive tasks test a specific timing claim
Bem’s 2011 paper reported positive results on tasks designed to test whether a later event could influence an earlier response, including a retroactive-facilitation task. The outcome was performance on those experimental measures. The paper made the claim testable, but a reported effect is not proof that information traveled backward in time; independent replication and controls for ordinary explanations matter.[3]
The 2025 Transparent Psi Project ran three studies with 26,483 participants and 420,472 critical trials. It failed to reproduce the targeted above-chance effect from Bem’s Experiment 1. Some exploratory results pointed in the opposite direction in two studies, while the third was approximately at chance; those exploratory results do not establish a different psi effect. This constrains the tested task, not every imaginable claim.[4]
Random-number tests ask about a different claim
Psychokinesis studies using random-number generators ask whether intention is associated with a deviation in a machine’s output. Bösch and colleagues reviewed 380 studies and reported a very small, heterogeneous aggregate effect. Their analysis noted that publication bias could explain the pattern. That possibility was raised as an explanation, not proven as the cause of every result.[5]
Maier and colleagues later tested a specified micro-psychokinesis alternative in a large protocol involving 12,571 participants. Their Bayesian analysis reported evidence against that tested alternative under the model and prior they specified. It is evidence about that protocol, not a universal impossibility result. The word quantum in a random-number study does not show that a mind controls physical measurement.[6]
Keep each result attached to its test
These findings are not equally strong, and each uses a distinct outcome: target selection, performance after a retroactive cue, or machine output. A positive aggregate can motivate better tests without proving paranormal information transfer. A failed replication limits the targeted effect without settling every form of unusual experience. The reviewed sources do not establish telepathy, precognition, or psychokinesis as a verified human ability.[1][2][4][5][6]
What would make you more confident in one of these claims: a clearer target, stronger blinding, a larger study, or a successful independent replication?