Discovery Lab Answers

Telepathy Data in the ESP-Nexus Library

Coverage note (required): The evidence table below reflects the studies currently held in the ESP-Nexus library. The library’s share of the published literature on telecommunication and general telepathy has not been measured, so treat this as a summary of what the library holds — not a settled account of the field.

The library currently holds 25 studies generating 26 result rows on the `gesp_telepathy` phenomenon. The metrics across these rows are not all comparable — 14 rows report raw hit rates and 11 report standardized effect sizes — so they cannot be pooled into a single “overall” figure. The pattern is described metric by metric below.

What the Evidence Shows

Hit-rate studies

The largest body of work comes from Rupert Sheldrake’s telecommunication telepathy program, in which participants guess which of several known contacts is about to call, text, or email them. Across multiple experiments spanning 1999–2024, Sheldrake’s studies consistently report hit rates above the relevant chance baseline, typically 25% in four-caller designs.

The 2025 meta-analysis by Sheldrake, Stedall, and Tressoldi brings together 26 telecommunication telepathy experiments from 15 papers (2003–2024) using a random-effects model. The pooled result across those experiments is reported as very significantly above chance. Importantly, the same meta-analysis reports that when tests were run under precognitive conditions (where the caller was selected after the guess), hit rates fell to chance — a result the authors treat as evidence that the positive results in the standard telepathy condition reflect something other than precognition or chance coincidence.

Two studies in the library return null results on hit-rate metrics:

• Stedall (2025) — a preregistered test with 117 participants over 604 trials — reports an overall hit rate effectively at chance.
• Rey (2010) — a Zener-card telepathy design with 96 participants across 2,400 trials — also reports a null result.

Schmidt (2009) shows a split picture: Study 1 alone was null, but across three combined studies the pooled result reached significance.

Standardized effect-size studies

Silberstein (2024) reports the largest standardized effect in the library for this phenomenon, though it involves only 10 participants. Vernon (2020) reports a small positive Cohen’s d but with a confidence interval spanning zero — a null by conventional criteria. Sheldrake (2025)’s meta-analytic standardized effect size is modest in magnitude.

A non-comparable row

Johnstad (2020) reports a prevalence proportion — the share of psychedelic-experience interviewees who reported telepathic impressions — rather than a hit rate or effect size from a controlled experiment. This figure describes self-reported phenomenology in a qualitative sample and is not comparable to the experimental rows.

Summary Table

Unsettled Signals (Required Disclosure)

The evidence base carries genuine disagreement that must be stated plainly:

• Six result rows report null or below-chance outcomes. Stedall (2025)’s preregistered test and Rey (2010)’s large Zener-card study both return hit rates indistinguishable from chance. Vernon (2020)’s confidence interval crosses zero. Schmidt (2009) Study 1 is null independently.
• Effect sizes vary widely — from Cohen’s d near zero (Sheldrake 2009, 6,000-trial pool) to 0.85 (Silberstein 2024, 10 participants). This heterogeneity is not explained in the rows the library holds.
• The corpus is heavily concentrated in one research group. Nearly all positive hit-rate rows come from Rupert Sheldrake and collaborators. Schmidt (2009) is the main independent replication attempt in this library; it produced a marginal combined-studies result but a null in the primary Study 1 analysis.
• Selected vs. unselected participants differ. Sheldrake, Stedall & Tressoldi (2025) note that selected participants showed significantly higher hit rates than unselected participants — a subgroup pattern the library does not fully resolve.

Skeptical Critiques and Discussion

What critics argue. Philosophical and methodological objections to telepathy research have a long history. Charles Sanders Peirce, engaging with the early Gurney et al. corpus, argued that statistical arguments used to rule out chance were deeply confused, and that the apparent impossibility of conducting controlled experiments that could rule out clairvoyance as an alternative to telepathy was a profound methodological problem — one the field had not resolved. The experimental critique centers on two issues still active today: (1) whether positive results can be explained by sensory leakage, subtle cuing, or procedural artifacts rather than information transfer, and (2) whether the body of positive results reflects publication bias rather than a genuine effect.

What the experimental data show. The 2025 meta-analysis by Sheldrake, Stedall & Tressoldi addresses the publication bias question by reporting that tests run under precognitive conditions — structurally identical to the telepathy condition but with caller selection after the guess — fell to chance, which the authors argue is inconsistent with a simple artifact explanation. The videotaped series (Sheldrake 2003) was designed specifically to minimize cheating; hit rates remained significantly above chance in 271 filmed trials. The Stedall (2025) preregistered, independent test, however, returned a null result across 604 trials with 117 participants — the most methodologically prospective study in the library.

Analysis. Peirce’s core objection — that controlling for clairvoyance and ruling out procedural confounds in telepathy research is structurally difficult — has been noted by researchers as still open. Independent replication outside Sheldrake’s own program is sparse in the library’s holdings; Schmidt (2009) provides the clearest example and produced mixed results across three studies. The preregistered Stedall (2025) null adds a data point that the meta-analytic positive picture does not absorb without qualification. Whether the positive hit-rate pattern reflects a genuine anomaly, a methodological artifact, or a selection effect in who participates and who is studied has not been settled by the data the library holds.

For a fuller synthesis, see the ESP-Nexus Telepathy phenomenon page, and for the automated telephone and SMS testing program specifically, the Automated Tests for Telephone Telepathy Using Mobile Phones study page. Dean Radin’s contributions to ganzfeld telepathy meta-analysis are covered on his scientist page.