Discovery Lab Answers

Harald Walach, PhD is a psychologist and complementary medicine researcher whose work spans parapsychology theory, distant-healing trials, micropsychokinesis experiments, and mindfulness research. His central contribution to parapsychology is applying Generalized Quantum Theory (GQT) to reframe psi phenomena — not as causal signals, but as instances of generalized nonlocal correlations. The ESP-Nexus profile for Walach is at Harald Walach, PhD.

Experiments

Walach’s experimental record, as reflected in the retrieved sources, covers several distinct lines of work:

Matrix experiment and its replications. The most extensively documented work in the sources is the “Matrix experiment,” originally reported as Walach et al. (2020), which examined correlations between behavioral and physical variables using a nonparametric matrix method applied to random event generator (REG) data. The original study (503 participants) produced a positive result. Walach then commissioned and co-authored two preregistered direct replications of that experiment — one conducted in Dresden by Karolina Kirmse, one larger — reported jointly in Walach et al. (2021) ^[1]. Both replications failed to reproduce the original finding.

Transferred neural potentials / EEG correlations. The sources mention a line of work connected to Wackermann et al. (2003), examining whether connected participants show transferred neural potentials. Walach is cited in ^[1] as having commissioned two large-scale replications in the same laboratory, which were reportedly positive but never published.

Micropsychokinesis / REG studies. Walach examined whether human intention correlates with REG output, reanalyzed through the lens of nonlocal correlations rather than direct causation. This is described across ^[1] and the ESP-Nexus spoke page on nonlocal correlations.

Distant-healing randomized controlled trials. The ESP-Nexus profile notes Walach conducted distant-healing RCTs applying pharmaceutical-grade clinical methodology to test nonlocal intention effects.

Feldenkrais Functional Integration RCT. Walach co-authored a randomized controlled study (Brummer, Walach, and Schmidt, 2018) ^[3] examining whether a single Feldenkrais Functional Integration session increases body contact surface area in the supine position — a complementary medicine study rather than a psi study, demonstrating his methodological range in clinical trial design.

Theoretical and position papers. Walach co-authored the theoretical framework paper applying GQT to parapsychology (2014) ^[4] with Walter von Lucadou and Hartmann Römer, and has published position papers on the philosophy of science as it relates to anomalistics ^[2][5].

Methodology

Matrix method. The core design of the Matrix experiment uses a correlation matrix of 2,025 cells, filled by Spearman rank correlations between variables across participants. The predetermined criterion counts correlations significant at p ≤ .1 (one-sided), comparing the observed count to the chance expectation (approximately 202–203 significant correlations in a 2,025-cell matrix at that threshold). Sensitivity analyses at lower p-value thresholds were also reported ^[1].

Preregistration. Both replications in Walach et al. (2021) were preregistered, making them comparatively strict direct replications by the authors’ own description ^[1]. The protocol followed the predecessor experiment’s design closely, including the use of TrueRNG hardware for random number generation. A programming error was discovered in the RNG normalization, corrected before subsequent use ^[1].

Randomization and blinding. The Feldenkrais RCT ^[3] used randomization with a crossover design, adjusted alpha levels (α = 0.025) for two parallel assessments, and objective pressure-mat measurements alongside self-report scales. The paper reports group comparability at baseline, confirming the randomization produced equivalent groups.

Nonlocal correlation framework. Methodologically, Walach argues that psi phenomena should not be modeled as signals (causal influences) but as nonlocal correlations — a distinction that affects both how experiments are designed and how power analyses apply. He explicitly questions whether classical power calculations are appropriate for effects theorized to be non-signal in nature ^[1][4].

Sample sizes in the replications. The two replication experiments together enrolled 104 participants, compared to 503 in the predecessor. Walach et al. (2021) calculate that, using the predecessor’s effect size (approximately r = .38), the smaller replication had 69% power and the larger had 88% power to detect the effect — though they note that classical power reasoning may not apply to non-signal effects ^[1].

Data

The structured evidence block for this question is empty, meaning no audited quantitative rows have been loaded for this query. The retrieved excerpts from Walach et al. (2021) ^[1] do contain some figures; where the surrounding text is intact and attribution is unambiguous, they are reported below. Where text is fragmented, figures are not reported.

Matrix experiment replications (Walach et al., 2021 ^[1]):

The combined sample across both replications was 104 participants ^[1]. Effect size from the predecessor was approximately r = .38 ^[1] — this figure appears in intact context in the source text. Both replications yielded results conforming to chance expectation.

Feldenkrais RCT (Brummer, Walach, and Schmidt, 2018 ^[3]):

The study found a significant change in body contact surface area after treatment (factor time: p < 0.0001, η²p = 0.90). A significant time × group interaction was found on the left side for participants beginning with the left side (p = 0.016; η²p = 0.62), with a weaker effect on the right side ^[3]. These are complementary medicine outcomes, not psi outcomes.

Overall pattern. Across the psi-relevant experimental record in the retrieved sources, the picture is: one positive original Matrix experiment result, followed by two preregistered direct replications by the same research group that both failed to reproduce it. Walach et al. (2021) frame this pattern as evidence that the replication problem in parapsychology may be “unsurmountable” when using direct experimental approaches — consistent with his theoretical position that psi effects, if real, are nonlocal correlations that destabilize under direct scrutiny ^[1].

Skeptical critiques

What critics argue. The most substantive critical engagement present in the sources is internal to Walach et al. (2021) itself ^[1]: the paper confronts the failure of two preregistered replications of a previously positive result and engages directly with the possibility that the original finding was a false positive or an artifact of the analytic approach. The paper also cites a commissioned replication of a different experiment — one in which trained meditators were supposed to affect an optical interference pattern — that also failed to replicate ^[1]. Walach and colleagues in ^[1] acknowledge that the predecessor Matrix experiment (503 participants) was never independently replicated before the two failed direct replications were attempted.

What the experimental data show. Both preregistered replications reported in Walach et al. (2021) produced results at chance expectation ^[1]. The programming error discovered in the RNG normalization was corrected before analysis, so it does not account for the null results. Walach et al. (2021) ^[1] note that together the two replications enrolled 104 participants — substantially fewer than the 503 in the predecessor — and calculate adequate power under classical assumptions, though they argue classical power may not be the right framework for non-signal effects.

Analysis. The critique most directly supported by the sources is the one Walach’s own team raises: that positive results in this line of research have not survived preregistered direct replication, even when conducted by the original investigators. The paper’s title — “Nailing Jelly” — reflects the authors’ own assessment that direct replication may be structurally unable to confirm psi effects, given the theoretical model. Whether that theoretical rescue of the null results is persuasive or constitutes post-hoc rationalization is a question the sources raise but do not settle ^[1][4]. Independent replication of the Matrix experiment outside the original group has not been reported in the retrieved sources.