Discovery Lab Answers

Remote Viewing: Data, Experiments, and Commentary

Coverage note: The studies below are those currently held in the ESP-Nexus library on this question. The library’s share of the full published literature on remote viewing has not been measured, so what follows is a summary of what the library holds — not a settled account of the field as a whole.

The evidence base here spans roughly five decades of remote viewing research, from early SRI trials in the 1970s through recent meta-analyses and field studies. The picture is directionally mixed: most rows lean positive, but null results and one below-chance result are present, and the metrics used across studies are not comparable to one another — so no single pooled figure can honestly represent them all.

What the studies show

The 23 result rows in the evidence table draw on 21 studies and report outcomes in five different, non-interchangeable metrics: standardized effect sizes (Cohen’s d, z/√N), raw hit rates, p-values only, z-scores, and correlation-based measures. These cannot be averaged into a single bottom-line figure; the table below shows them as reported.

Settled signals

The Katz–Tressoldi meta-analysis (the Tressoldi 2023 row, covering 40 effect sizes from studies 1974–2022) is the most comprehensive quantitative synthesis in this evidence set. Its frequentist random-effects result — ES = 0.36 (z/√N), 95% CI [0.23, 0.48], p = 8.7×10⁻¹⁰ — represents an above-chance aggregate across that sample of studies [see the Remote Viewing Meta-Analysis page]. Dunne’s PEAR laboratory accumulated large-trial datasets over decades; the remote perception combined result (653 trials, p = 3×10⁻⁸) and the REG/RMC/Pendulum meta (z = 7.0, p = 6.5×10⁻¹¹) reflect that scale [see the Dunne remote viewing page]. The early SRI Pat Price series (Targ 1975; Targ 2019) contributed the foundational formal trials that launched government-sponsored programs.

Unsettled signals

Three results run counter to the positive trend and must be stated plainly:

• Wahbeh (2021): The Quick Remote Viewing Task across 534 participants returned an overall null result.
• Wahbeh (2022): A pre-post Extended Perception measure returned ES = −0.16 (Hedges’ g), a below-chance (psi-missing) direction, with a hit rate of 0.19.
• Wiseman (2010): Blind judging across 4 trials returned null.

Beyond direction, the Tressoldi (2023) meta-analysis carries high between-study heterogeneity (I² ≥ 50%), which means the 40 effect sizes in that synthesis are not consistent with one another — something about how, where, or with whom studies were conducted accounts for substantial variation that the mean effect size alone does not capture.

The Escola-Gascon (2023) results also show a large gap between believers (ES = 0.853, Cohen’s d) and the total sample (ES = 0.457, Cohen’s d), suggesting that participant attitudes or self-selection may be a substantial moderator.

Commentary: government programs and statistical evaluation

Jessica Utts’s 1995–1996 statistical evaluation of the Stargate program concluded that the experimental record met conventional scientific significance thresholds — a judgment made on the basis of effect-size comparison, not merely p-values [see the Utts remote viewing page]. Her commentary explicitly addressed what constitutes an operationally useful hit rate — a question she noted had not been defined in advance by the program’s customers — and she co-authored a prior critical review of an NRC committee that had not been shown the bulk of the SRI data.

The Star Gate Archives (reviewed in) document the SRI-era mass-screening work, which concluded that approximately 1% of the general population showed a natural remote viewing ability — a finding echoing Rhine’s earlier discovery of gifted subjects, and the subsequent difficulty replicating those results with unselected participants.

Skeptical critiques

What critics argue. Methodological critiques of early remote viewing experiments focused on sensory cues — the argument, associated with David Marks (1981, cited in), that successful judging could be explained by inadvertent cues in transcripts rather than anomalous cognition. A related line held that judging procedures in studies such as those at PEAR were ambiguous or inadequately documented. Utts, Hansen, and Markwick identified specific randomization and judging-procedure problems in the PEAR remote viewing experiments [see the Utts page]. The American Institutes for Research 1995 review applied what Utts characterized as ambiguous criteria to evaluate the operational utility of anomalous cognition, and its treatment of the NRC report — which she noted had not been shown the relevant data — troubled her.

What the experimental data show. Utts noted that by the mid-1990s, Ray Hyman — a persistent critic — acknowledged that the experiments being assessed were “free of the methodological weaknesses that plagued the early research”. The Katz–Tressoldi meta-analysis applied inclusion criteria that excluded studies not meeting methodological thresholds, and the resulting ES = 0.36 persisted even in the full 40-effect-size sample. The null results from Wahbeh (2021, 2022) come from the post-Stargate era with large samples, providing a counterweight to the positive aggregate.

Analysis. The sensory-cue critique drove protocol tightening across the field; later studies incorporated blind judging and double-blind designs specifically to close those gaps. The Tressoldi (2023) meta’s high heterogeneity leaves open which study features predict success or failure — moderator analyses exist but have not resolved this. The Wahbeh null and psi-missing results, using a standardized task across hundreds of participants, represent the largest recent negative evidence in this set and have not been directly reconciled with the positive meta-analytic aggregate.

For broader context on the phenomenon, protocols, and debates, see the Remote Viewing phenomena page.