Discovery Lab Answers

Remote Viewing: Overall Results from the ESP-Nexus Library

Coverage note (read first): The evidence below reflects the studies currently in the ESP-Nexus library. The library’s share of the published literature on remote viewing has not been measured, so this is a summary of what the library holds — not a settled account of the field as a whole.

The 23 result rows span studies from 1975 through 2026, covering laboratory experiments, meta-analyses, and applied operational contexts. The picture is predominantly positive in direction, but includes null and below-chance results that are a genuine part of the record.

Direction of results across the library

Of the 23 result rows:

• Most report positive (above-chance) outcomes — across a range of protocols, sample sizes, and participant types.
• Three rows report null or below-chance outcomes, which must be treated as real findings, not anomalies to explain away.

The three non-positive outcomes are:

• Wahbeh (2021) — the Quick Remote Viewing Task across 534 participants returned a null overall result.
• Wahbeh (2022) — a pre-post change measure on the same task returned a below-chance (psi-missing) Hedges’ g of −0.16.
• Wiseman (2010) — blind judging across 4 trials returned a null result.

These are not peripheral studies — they used defined protocols, and two of them come from the same research group running the same task at scale.

Key positive findings (selected)

The metrics across studies are not comparable — the rows include standardized effect sizes, raw hit rates, z-scores, p-values, and prevalence proportions. They cannot be averaged or pooled. The table below organizes what the rows actually report, grouped by metric type where possible.

Standardized effect sizes (Cohen’s d or z/√N)

z-scores and p-only results

Hit rates

Unsettled signals — you must factor these in

The evidence block flags these explicitly:

1. Directions disagree across studies. The null results from Wahbeh (2021) and Wiseman (2010), and the below-chance result from Wahbeh (2022), sit alongside a majority of positive findings. That disagreement is real and not resolved by the current library.

2. High between-study heterogeneity (I² ≥ 50%) appears in at least one pooled result. Tressoldi (2023) applied a random-effects model to 40 studies — appropriate precisely because the studies are not interchangeable. A consistent average effect does not mean consistent effects across conditions, labs, or participants.

3. Moderator effects are visible but not fully explained. Escola-Gascon (2023) found the believer subgroup effect size (0.853) was substantially larger than the total-sample effect (0.457), suggesting psi-attitude, target type, or both are functioning as moderators. What drives that difference is an open question.

4. Some results come from very small N. Targ (2019) reports 9 trials; Dunne (1979) reports 8 trials with 2 participants; Piao (2023) reports 100 trials with a single participant. High significance in small-N studies warrants caution about generalization.

5. Metric heterogeneity is substantial. The rows span standardized effect sizes, raw hit rates, z-scores, p-values, and prevalence proportions — these cannot be pooled into a single bottom-line number, and any attempt to do so would be misleading.

Skeptical critiques

What critics argue. The most sustained published critique of remote viewing research targets methodological controls: sensory leakage, inadequate blinding of judges, optional stopping, and the influence of experimenter effects on results. Ray Hyman, in his analysis of the SRI and SAIC remote viewing programs, argued that the evidence did not meet the threshold for establishing a genuine anomaly, emphasizing procedural irregularities and the difficulty of ruling out conventional explanations.

What the experimental data show. Jessica Utts, reviewing the same SAIC body of work, concluded in a 1996 Journal of Scientific Exploration assessment that the effect was real and replicable at a statistically significant level, with effect sizes in the small-to-medium range. The Tressoldi (2023) meta-analysis of 40 studies found a pooled effect with a CI of [0.23, 0.48] and applied trim-and-fill and other publication-bias corrections. The Escola-Gascon (2023) total-sample result across 634 participants also reached significance. Against this, Wahbeh (2021) — with 534 participants, the largest single sample in the library — returned a null result on the Quick Remote Viewing Task, and Wiseman (2010) returned null across blind judging in 4 trials.

Analysis. Utts and Hyman examined overlapping evidence and published divergent conclusions in 1996; neither has been superseded by a definitive independent adjudication. The Wahbeh null results represent an independent laboratory failing to find the effect with a large sample and a standardized instrument. Independent replication across diverse laboratories, with pre-registered protocols and blinded analysis, remains the unsettled methodological frontier — the Akin, Tressoldi, and Katz (2026) experts’ guidelines paper reflects the field’s own ongoing effort to standardize conditions that would make such replication meaningful.

For deeper treatment, the Remote Viewing phenomenon page synthesizes the broader literature, and the spoke pages for Debra Lynne Katz and Marilyn J. Schlitz cover the meta-analytic and replication dimensions in detail.