Discovery Lab Answers

Dean I. Radin — Data Summary

Coverage note: The studies below reflect what is currently held in the ESP-Nexus library. The library’s share of the total published literature on Radin’s work has not been measured, so this is a summary of what the library holds, not a settled account of his full research output. The picture may skew toward studies that reached publication, and the balance of positive, null, and below-chance results in the complete literature is unknown.

Radin’s profile on ESP-Nexus is at Dean I. Radin, PhD.

The evidence table (rendered below your answer) draws from 66 studies and 81 result rows. The metrics across those rows are not comparable — they include standardized effect sizes, raw hit rates, z-scores, and correlation-family statistics — so no single pooled “overall effect size” can be stated honestly. What follows is a study-by-study and thematic account.

Research Themes and Headline Results

Global Consciousness Project — New Year’s Eve

Radin (2025) examined how Global Consciousness Project random-number generator data behaved during New Year’s Eve midnight transitions across multiple years, applying six analytical methods (mean shift, chaotic attractor, sample independence, fractal dimensions, nonlinear dependencies, and population-weighted analyses). The principal-components composite at midnight reached a combined result reported in the evidence table. A subsequent corrigendum identified variance inflation in the original PCA composite (inflation factor approximately 1.67 rather than 1.00), which reduced the midnight composite p-value; the corrected figure is an OCR-extracted value pending verification against the published corrigendum text. Radin notes that midnight remained the most statistically significant of all 100 time points tested even after correction, and that the pattern of coherent deviation across multiple analytical methods was unchanged in direction.

Excess Correlation in Water pH

Radin (2025) conducted a preliminary independent replication of a claimed “excess correlation” effect — correlated pH changes between isolated beakers of water stimulated by a magnetic field halo. Across 50 experimental and control runs, a linear mixed-effects analysis followed by bias-corrected and accelerated (BCa) bootstrap found a statistically significant but small-magnitude pH shift in the remote beaker during the stimulation phase. A separate subset of 26 sessions showed a significant main effect for Condition and a significant Phase × Condition interaction before autocorrelation correction; after the Durbin-Watson-motivated BCa bootstrap was applied, the Phase × Condition interaction became non-significant (p = 0.111) in that subset. Radin flags multiple uncontrolled variables — field strength, magnetic sequence, water quantity, halo orientation relative to magnetic north, geomagnetic field strength — as factors that could affect replicability.

Double-Slit / Optical Interference

The Radin et al. (2012) study audit page covers six experiments (137 participants, 250 sessions) testing whether focused human attention was associated with measurable reduction in a double-slit interference fringe pattern; meditators showed larger effects than non-meditators, and control sessions showed no effect. The Radin et al. (2016) study audit page covers a distant variant of the same paradigm. Kauffman (2023) in the evidence table reports a cumulative result across 30 double-slit experiments at five laboratories.

Presentiment / EEG / Physiological Anticipation

Several rows in the evidence table concern Radin’s electroencephalographic and physiological anticipation work — the paradigm he termed “presentiment” — including trend analyses comparing observed versus resting slopes in optical-interference sessions (Radin 2024, Radin 2025 observer-quantum-interference page).

Intention Effects on Water / Plants

Radin (2021) in the evidence table reports an indirect test of practitioner intention on distilled water.

Survey and Experience Research

Wahbeh (2022, 2023) rows include a quick remote-viewing task, near-death experience persuasiveness ratings, and workshop-based transformative experience reports. The overall remote-viewing performance row for Wahbeh (2021) is flagged as null in the evidence table.

Card-Guessing / Precognition

Alibalaei (2025) — a 40-day yoga course study measuring pre- vs. post-course precognition via paired t-test on 4,160 trials — is null in the evidence table (d = −0.027, p = .783). Wahbeh (2023) completers on a telepathic/pre-selected task show a positive hit-rate result.

Quantitative Overview (selected headline rows)

The table below lists the headline result row for each study in the evidence table, with direction and metric. Because the metrics differ, the columns are not summable — scan them column by column only within the same metric type.

Unsettled Signals

Two signals the evidence table explicitly flags must be stated directly:

Direction disagreement across studies. Not all results point the same way. Radin (2023) Experiment 1 (concentrate vs. relax sessions) is below chance (psi-missing, z = −3.0). Wahbeh (2022) remote-viewing pre-post change is below chance (g = −0.16). Wahbeh (2021) overall remote-viewing performance is null. Alibalaei (2025) is null (d = −0.027). Nine result rows in total record null or below-chance outcomes.

Metric heterogeneity. The 81 result rows use five different metric families — standardized effect sizes, hit rates, p-values alone, z-scores alone, and correlation-family statistics. These cannot be averaged or pooled into a single bottom-line number; the table above preserves the distinction.

Skeptical Critiques

What critics argue. A persistent methodological objection to Radin’s double-slit and physiological-anticipation work concerns the use of within-subjects designs without adequate blind controls, and the possibility that subtle cues or demand characteristics in the experimental setting — rather than any anomalous effect — could drive the observed deviations. Critics have also raised concerns about flexibility in analysis pipelines (choice of transformation, smoothing window, and permutation method) that may inflate apparent effect sizes even when the pre-registered primary outcome does not survive correction.

What the experimental data show. Radin (2025) included a preregistration for his primary test and applied BCa bootstrap correction for autocorrelation. Radin (2025) applied the same BCa procedure after the Durbin-Watson statistic indicated autocorrelation, and the Phase × Condition interaction did not survive that correction in the 26-session subset (p = 0.111), while a separate 50-session combined result remained significant. The corrigendum is itself the result of Radin responding to a statistical critique about variance inflation in the original PCA — the correction reduced but did not eliminate the midnight signal.

Analysis. The pattern across the evidence table shows that corrections for autocorrelation and variance inflation consistently reduce effect magnitudes, and in at least one instance (the 26-session subset in the water pH study) move an interaction below conventional significance. Independent replication outside Radin’s own laboratory is reported in some lines of work (Kauffman 2023 cumulates five laboratories; Radin 2025 is itself framed as an independent replication attempt) but the degree to which those replications were conducted fully independently of Radin’s involvement varies and is not fully resolved in the retrieved sources. The below-chance and null results in the Wahbeh series, conducted at the same institution (IONS), show that not all paradigms run at that laboratory produce positive outcomes.