Meta-Analysis in Parapsychology

Meta-analysis combines effect-size estimates across multiple independent studies to produce a single pooled estimate and examine heterogeneity. Parapsychology has used meta-analytic techniques since the mid-1980s to summarize its most-studied paradigms, ganzfeld ESP, forced-choice precognition, micro-PK RNG experiments, and DMILS, and to test whether apparent effects survive under corrections for publication bias, quality weighting, and pre-registration.

Overview

Meta-analysis is a statistical method for pooling results across many independent studies. Rather than judging one experiment at a time, it combines effect-size estimates, standardized measures of how far a result departs from chance, into a single pooled value, and then tests whether the studies agree among themselves (a property called heterogeneity). In parapsychology the method has been used since the mid-1980s to summarize the field’s most-studied designs: ganzfeld telepathy, forced-choice extrasensory perception (ESP), micro-PK experiments with random number generators (RNGs), and distant mental interaction with living systems (DMILS).

A key point is that meta-analysis is a tool, not a verdict. Positive and negative meta-analyses of the very same paradigm coexist in the literature, and the differences usually trace to which studies were included, how study quality was weighted, and what assumptions the analysts brought. The pooled effects in psi research are also typically small. A small effect is not, by itself, evidence that nothing is there, but it does demand careful handling of publication bias and heterogeneity before any claim is made.

Protocol

A meta-analysis in this field generally follows a fixed sequence. Analysts first define a clear search and inclusion rule, for example, all ganzfeld telepathy trials meeting a stated standard between two dates. Each study is then converted to a common effect-size scale so that experiments of different sizes can be compared. Studies are weighted, often by sample size or by a quality rating, and a pooled estimate is produced together with a test of whether the spread of results is wider than chance would predict.

Recent work has refined this template by separating procedural moderators from the headline estimate. One meta-analysis of telepathy ganzfeld studies, for instance, modeled how features of the sender–receiver setup changed success rates, finding that letting senders hear the receiver’s running commentary raised hit rates while a post-session review period lowered them.[1] Such moderator analysis is now a routine part of the protocol, because a single pooled number can hide systematic differences between procedures.

History

The modern history begins with a 1985 ganzfeld meta-analysis that brought formal effect-size pooling into the field’s mainstream debate. It was answered the same year by a detailed critical appraisal that questioned the database’s quality and the handling of multiple analyses, an exchange that set the template for nearly every psi meta-analysis since: a quantitative summary followed by a methodological rebuttal. The following year the two analysts published a joint communiqué stating that “there is an overall significant effect in this data base that cannot reasonably be explained by selective reporting or multiple analysis,” while continuing to differ over the degree to which the effect constitutes evidence for psi and agreeing that the final verdict awaited experiments conducted by a broader range of investigators according to more stringent standards.[19] In 1991 Jessica Utts reviewed the field’s meta-analyses for a statistics audience in Statistical Science, arguing that most nonstatisticians do not appreciate the connection between statistical power and successful replication of experimental effects.[21]

The most cited episode came in 1994, when a pooled “autoganzfeld” database was published in a mainstream psychology journal and argued for a replicable information-transfer anomaly.[2] Five years later a second meta-analysis of thirty newer ganzfeld studies from seven laboratories reported no overall effect, concluding that the earlier replication had not held up.[3] These two papers, published in the same venue, became the standard illustration that inclusion criteria and study selection can drive opposite conclusions from the same paradigm. Parallel summaries appeared for other designs, including an early pooling of RNG experiments[4] and a meta-analysis of dice-throwing studies spanning half a century.[5]

The exchange that followed the 1999 null became the method’s most instructive case. Storm and Ertel combined the older and newer ganzfeld databases into a unified set of 79 studies with a mean effect size of 0.138 (Stouffer z = 5.66, p = 7.78 × 10−9) and calculated that 857 unpublished nonsignificant studies would be needed to bring that result down to chance.[22] Milton and Wiseman replied that problems in the early ganzfeld studies make it difficult to draw strong conclusions from meta-analyses that include them, and noted that a single strongly positive study published after their cutoff, by Dalton in 1997, was by itself enough to pull the otherwise null database of post-1986 studies into overall statistical significance.[23] Bem, Palmer, and Broughton then had three independent raters score all forty post-communiqué replications for adherence to the standard ganzfeld protocol: the forty studies combined yielded a 30.1% hit rate (ES = .051, Stouffer Z = 2.59, p = .0048, one-tailed), the standard replications a 31.2% hit rate (ES = .096, Z = 3.49), and the non-standard replications 24.0% (ES = −.10, nonsignificant), with effect size significantly correlated with the degree of adherence to the standard protocol.[24] The three papers describe one body of experiments; the conclusions diverge on which studies are pooled and how a replication is defined.

Key findings

Across the free-response designs, later pooled summaries continued to report effects modestly above chance, with ganzfeld procedures producing the strongest results and no sign of a decline over roughly four decades.[6] A combined free-response meta-analysis reached similar conclusions, and a review tying several ganzfeld meta-analyses together reported a statistically significant overall ESP effect across more than a hundred publications.[7]

The forced-choice design tells a quieter story. A meta-analysis of 141 forced-choice ESP studies from 1987 to 2022 found a reliably positive but very weak effect, near the edge of statistical noise yet consistent across experimenters and target types.[8] A broader classical-and-Bayesian review argued that the cumulative evidence from more than two hundred non-local perception studies meets standards comparable to those used in clinical medicine.[9]

A meta-analysis of 141 forced-choice ESP studies from 1987 to 2022 found a reliably positive but very weak effect, consistent across experimenters and target types.[8]

For mind–matter interaction, an early RNG pooling reported a small but persistent influence on random output across many investigators,[4] and later analysts maintained that publication bias and quality concerns do not plausibly erase the cumulative result.[10] Related meta-analytic work covered the sense of being stared at[11] and the anomalous anticipation of future events.[12]

Landmark meta-analyses and framing agreements in the reference library

The entries below are the landmark meta-analyses of parapsychological research paradigms currently held in full text in the ESP-Nexus reference library, together with the methodological agreements that frame them, listed chronologically. Two founding documents of the 1985 exchange, Honorton’s meta-analysis of the original ganzfeld database and Hyman’s critical appraisal of the same studies, are not yet held; that episode is represented by the joint communiqué the two authors published the following year. Each entry reports the figures the paper itself states, and the entries report different kinds of number, Stouffer z values, standardized effect sizes, correlations, Cohen’s d, and raw hit rates, which cannot be combined into one bottom-line figure. For one of these databases the library’s own coverage has been measured: ESP-Nexus holds 57 of the 78 studies in the ganzfeld database analyzed by Tressoldi and Storm (2024), and pooling the held 57 under the same random-effects specification gives 0.085, against 0.077 computed for the full 78 studies; the paper itself prints .074, and the small gap between 0.077 and .074 is the site’s recomputation under the same specification versus the published rounding. That comparison is the site’s own holdings check, not a published finding; it means any summary built from the library’s ganzfeld holdings runs slightly more positive than the full published database.

Hyman and Honorton (1986)[19]
Design and scaleJoint statement in the Journal of Parapsychology closing the authors’ 1985 exchange over the original ganzfeld database, Honorton’s quantitative meta-analysis and Hyman’s critical appraisal of the same studies; rather than another round of debate, the two collaborated on a communiqué emphasizing points of agreement, with recommendations on randomization, judging and feedback procedures, multiple analysis and statistics, and documentation
Reported resultThe authors state that “there is an overall significant effect in this data base that cannot reasonably be explained by selective reporting or multiple analysis”; they continue to differ over the degree to which the effect constitutes evidence for psi, and agree that the final verdict awaits future experiments conducted by a broader range of investigators according to more stringent standards
Radin and Nelson (1989)[4]
Design and scaleMeta-analysis of experiments testing whether human intention correlates with the output of electronic random number generators: 152 references from 1959 to 1987 describing 832 studies by 68 investigators, 597 experimental and 235 control studies, with studies reported only as nonsignificant conservatively assigned z = 0
Reported resultControl studies conformed to chance expectation while experimental studies showed what the authors describe as unequivocal non-chance effects; the relation between assessed study quality and effect size was nonsignificant, which the authors read as evidence that the database is not compromised by poor experimental methodology
Honorton and Ferrari (1989)[20]
Design and scaleMeta-analysis of forced-choice precognition experiments published in the English-language parapsychological literature between 1935 and 1987: 309 studies in 113 publications by 62 investigators, with nearly two million individual trials contributed by more than 50,000 subjects
Reported resultA small but reliable overall effect: combined Stouffer z = 11.41; cumulating by investigator rather than by study gave a combined z of 12.13 with a mean investigator effect size of 0.033, and 23 of the 62 investigators (37%) had overall outcomes significant at the 5% level; the authors calculate a “fail-safe N” of 14,268 unreported studies
Milton and Wiseman (1999)[3]
Design and scaleMeta-analysis of thirty ganzfeld ESP studies conducted after the 1986 joint communiqué and published 1987 to 1997, retrieved from 14 papers by 10 principal authors at 7 laboratories, totaling 1,198 individual trials; the cumulation method (Stouffer) was specified in advance
Reported resultA nonsignificant overall result: Stouffer z = 0.70, p = .24, one-tailed, with a mean effect size of 0.013 (SD = 0.23); of three internal effects retested from the Bem and Honorton database, only one was confirmed
Storm and Ertel (2001)[22]
Design and scaleComment on Milton and Wiseman (1999) recombining the ganzfeld databases: a unified old-plus-new set of 79 ganzfeld and autoganzfeld studies built from Honorton’s 28 studies, Milton and Wiseman’s 30, the authors’ own 11, and Bem and Honorton’s 10
Reported resultMean effect size 0.138 (SD = 0.30), Stouffer Z = 5.66, p = 7.78 × 10−9; the authors calculate that 857 unpublished nonsignificant studies would be needed to bring the result down to chance, and report that a combined new database of 40 studies gives ES = .050 with Z = 1.88 (p = .03)
Bem, Palmer, and Broughton (2001)[24]
Design and scaleReanalysis of the 40 post-communiqué ganzfeld replications (Milton and Wiseman’s 30 studies plus 10 published after their cutoff date), with three independent raters scoring each study’s adherence to the standard ganzfeld protocol
Reported resultAll 40 studies combined yield a 30.1% hit rate, ES = .051, Stouffer Z = 2.59, p = .0048, one-tailed; standard replications obtain a 31.2% hit rate (ES = .096, Z = 3.49, p = .0002) while non-standard replications obtain 24.0% (ES = −.10, Z = −1.30, nonsignificant), and effect size correlates significantly with adherence to the standard protocol
Sherwood and Roe (2003)[25]
Design and scaleReview and effect-size comparison of dream-ESP studies conducted since the end of the Maimonides dream-ESP program, pooling the Maimonides series and the post-Maimonides studies separately
Reported resultCombined effect size r = 0.33 (95% CI 0.24 to 0.43) for the Maimonides studies and r = 0.14 (95% CI 0.06 to 0.22) for the post-Maimonides studies; the Maimonides studies were significantly more successful (p < 0.05), which the authors suggest may reflect procedural differences, post-Maimonides receivers tending to sleep at home without deliberate awakening from REM sleep
Schmidt, Schneider, Utts, and Walach (2004)[26]
Design and scaleTwo meta-analyses of distant-intentionality experiments measuring electrodermal activity: 36 studies of direct mental interaction with living systems, and 15 remote-staring studies
Reported resultDirect mental interaction: d = .11, p = .001, with a best-evidence synthesis of 7 studies yielding d = .05 (p = .50); remote staring: mean d = 0.13, p = .01; the authors conclude there are hints of an effect but also a shortage of independent replications and theoretical concepts
Bösch, Steinkamp, and Boller (2006)[27]
Design and scaleMeta-analysis of experiments testing whether random number generator output correlates with human intention: 380 experimental studies and 137 control studies drawn from 117 reports, analyzed under both fixed-effects and random-effects models
Reported resultA statistically significant but very small overall effect, with study effect sizes strongly and inversely related to sample size and extremely heterogeneous; the three largest studies showed results opposite to intention, and without them both models showed a highly significant effect in the intended direction; the 137 control studies were nonsignificant; a Monte Carlo simulation showed that the small effect size, the sample-size relation, and the extreme heterogeneity could in principle be a result of publication bias
Storm, Tressoldi, and Di Risio (2010)[6]
Design and scaleMeta-analysis of free-response studies in three categories, ganzfeld, nonganzfeld noise reduction, and standard free response; the period of analysis runs from March 1997 to 2008 for the ganzfeld and noise-reduction categories and from 1992 to 2008 for standard free response
Reported resultA homogeneous set of 29 ganzfeld studies yielded a mean effect size of 0.142 (Stouffer Z = 5.48, p = 2.13 × 10−8), with 483 hits in 1,498 trials, a 32.2% hit rate against 25% chance expectation; a homogeneous nonganzfeld noise-reduction set of 16 studies yielded a mean effect size of 0.110 (Stouffer Z = 3.35)
Storm, Tressoldi, and Di Risio (2012)[28]
Design and scaleMeta-analysis of forced-choice ESP studies for the period 1987 to 2010: 91 studies in 65 papers by 96 investigators, totaling 812,626 trials and 221,034 hits
Reported resultA homogeneous dataset of 72 studies yielded a weak but significant mean effect size of 0.01 (Stouffer Z = 4.86, p = 5.90 × 10−7); the authors report no evidence that the results were due to low-quality design or selective reporting, no significant difference between clairvoyance and precognition studies, and suggestive evidence that the number of choices per trial is inversely related to effect size
Mossbridge, Tressoldi, and Utts (2012)[29]
Design and scaleMeta-analysis of 26 reports published between 1978 and 2010 on predictive physiological anticipation of seemingly unpredictable stimuli (presentiment), 21 of them based at least partially on electrodermal data
Reported resultA significant overall effect with a small effect size: fixed-effect ES = 0.21 (95% CI 0.15–0.27, z = 6.9, p < 2.7 × 10−12) and random-effects ES = 0.21 (95% CI 0.13–0.29, z = 5.3, p < 5.7 × 10−8); higher-quality experiments produced a larger effect size and greater significance than lower-quality studies, and the authors conservatively calculate that 87 contrary unpublished reports would be needed to reduce the significance to chance
Storm and Tressoldi (2020)[30]
Design and scaleTen-year update to Storm, Tressoldi, and Di Risio (2010), collecting free-response studies for 2009 to 2018 in the same three categories
Reported resultNine new ganzfeld studies yielded a mean effect size of 0.119, 19 new nonganzfeld noise-reduction studies 0.045, and 15 new standard free-response studies 0.050, each database with a significant Stouffer Z; no new database differed significantly from its counterpart in the 2010 analysis, and the enlarged ganzfeld database of 38 studies yielded a mean effect size of 0.133
Tressoldi and Storm (2024)[31]
Design and scaleStage 2 registered report: a pre-registered meta-analysis of anomalous perception in a ganzfeld condition covering more than 40 years of investigation, with a final database of 78 studies (113 effect sizes) by 46 principal investigators, analyzed with frequentist random-effects and Bayesian models and three publication-bias tests (p-uniform*, three-parameter selection model, robust Bayesian meta-analysis)
Reported resultMean effect size .074 (95% CI .03–.12), p = .0009; Bayesian weighted effect size .084 with a Bayes factor of 89.5; between-study heterogeneity I² = 63.8; the publication-bias-adjusted estimates are .12 (p-uniform*), .15 (three-parameter selection model), and .074 (robust Bayesian meta-analysis)

Analysis. The entries show the range of outcomes the method produces on the same or overlapping databases. Milton and Wiseman’s thirty-study set returned an effect size of 0.013 where Storm and Ertel’s 79-study recombination returned 0.138,[3][22] and Bem, Palmer, and Broughton attributed the gap to protocol adherence, with standard replications at ES = .096 and non-standard replications at ES = −.10.[24] Across paradigms, the pooled values reported in these papers run from 0.01 for forced-choice ESP[28] to 0.21 for presentiment,[29] and the pre-registered 2024 ganzfeld estimate of .074[31] is smaller than the 0.142 the same research line reported for the 1997–2008 period.[6] On the random number generator literature, Radin and Nelson report unequivocal non-chance experimental effects with no quality–effect size relation,[4] while Bösch, Steinkamp, and Boller report that the same class of effect could in principle be a product of publication bias;[27] both readings rest on meta-analyses of overlapping databases.

Methodological critiques

The central dispute in RNG meta-analysis concerns the relationship between effect size and sample size. A 2006 critical meta-analysis published in a mainstream psychology journal concluded that psychokinesis was “not proven.”[27] A formal reply argued that this verdict rested on an assumption that effect size should be independent of sample size, and that correcting it left the cumulative RNG data supporting a genuine effect.[13] A companion paper by the same authors made the same case in a specialist venue.[10]

A second line of criticism focuses on questionable research practices and the file drawer. Defenders have responded that simulations meant to show such practices can manufacture ganzfeld effects rely on flawed prevalence estimates and unsupported assumptions.[14] Earlier replies to ganzfeld critiques argued that file-drawer concerns and meta-analytic methods had been mishandled by the critics.[15] A separate critique aimed at an umbrella review of anomalous-cognition meta-analyses flagged problems in inclusion criteria, heterogeneity reporting, and effect-size metrics, urging cautious reading of its conclusions.[16]

Some analysts within the field have also warned against over-reading patterns. One argument holds that supposed long-term experimenter effects and chronological declines often fail to survive rigorous meta-analysis and reflect a tendency to err in interpretation.[17] A further consideration is statistical power: replication failures for weak effects may stem from underpowered studies rather than from the phenomenon vanishing.[18]

Current practice

Contemporary psi meta-analysis increasingly mirrors reforms in mainstream psychology: pre-registration of protocols, explicit publication-bias tests, and shared databases. The episode that pushed this hardest followed a 2011 report of anomalous anticipation effects, after which dozens of laboratories across many countries posted replication attempts. A 2016 meta-analysis of ninety such experiments pooled this collected record into a single estimate,[12] and the scale of the replication effort, some thirty-three laboratories in fourteen countries, became a reference point for how the field now organizes large-scale tests.

Current meta-analyses also pay closer attention to moderators rather than headline pooled values, modeling how procedural details shape outcomes.[1] The forced-choice synthesis of 1987–2022 illustrates the modern two-stage approach, combining a review phase with a quantitative pooling phase.[8] The unresolved questions remain the ones meta-analysis was built to address: whether small pooled effects survive strict bias correction, and whether independent laboratories converge on the same estimate. On those points the literature still divides, and the method’s main value is to make the disagreement quantitative rather than rhetorical.

References
  1. Pooley, A., Murray, A., & Watt, C. (2023). Understanding the Factors at Play in the Sender-Receiver Dynamic During the Telepathy Ganzfeld: A Meta-Analysis. Journal of Anomalous Experience and Cognition, 3(1), 42–77. https://journals.lub.lu.se/jaex/article/view/23878 [PDF] R001 [Pooley 2023] ↩︎
  2. Bem, D., & Honorton, C. (1994). Does psi exist? Replicable evidence for an anomalous process of information transfer. Psychological Bulletin, 115(1), 4–18. https://doi.org/10.1037/0033-2909.115.1.4 [DOI] R002 [Bem 1994] ↩︎
  3. Milton, J., & Wiseman, R. (1999). Does psi exist? Lack of replication of an anomalous process of information transfer. Psychological Bulletin, 125(4), 387–391. https://psycnet.apa.org/record/1999-05876-001 [DOI] R003 [Milton 1999] ↩︎
  4. Radin, D., & Nelson, R. (1989). Evidence for consciousness-related anomalies in random physical systems. Foundations of Physics, 19(12), 1499–1514. https://doi.org/10.1007/bf00732509 [DOI] R004 [Radin 1989] ↩︎
  5. Radin, D., & Ferrari, D. C. (1991). Effects of consciousness on the fall of dice: A meta-analysis. Journal of Scientific Exploration, 5(1), 61–83. https://web.archive.org/web/20230331215150/https://www.scientificexploration.org/docs/5/jse_05_1_radin.pdf [archive.org] R005 [Radin 1991] ↩︎
  6. Storm, L., Tressoldi, P. E., & Di Risio, L. (2010). Meta-analysis of free-response studies, 1992–2008: Assessing the noise reduction model in parapsychology. Psychological Bulletin, 136(4), 471–485. https://doi.org/10.1037/a0019457 [DOI] R006 [Storm 2010] ↩︎
  7. Tressoldi, P., Radin, D., & Storm, L. (2010). Extrasensory Perception and Quantum Models of Cognition. NeuroQuantology, 8(4). https://doi.org/10.14704/nq.2010.8.4.353 [DOI] R007 [Tressoldi 2010] ↩︎
  8. Storm, L., & Tressoldi, P. (2023). Assessing 36 Years of the Forced Choice Design in Extra Sensory Perception Research: A Meta-Analysis, 1987 to 2022. Journal of Scientific Exploration, 37(3), 517–535. https://journalofscientificexploration.org/index.php/jse/article/view/2967 [PDF] R008 [Storm & Tressoldi 2023] ↩︎
  9. Tressoldi, P. (2011). Extraordinary claims require extraordinary evidence: the case of non-local perception, a classical and Bayesian review of evidences. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.2030810 [DOI] R009 [Tressoldi 2011] ↩︎
  10. Radin, D., Nelson, R., Dobyns, Y., & Houtkooper, J. (2006). Assessing the evidence for mind-matter interaction effects. Journal of Scientific Exploration, 20(3), 361–374. https://web.archive.org/web/20240712074323/https://www.scientificexploration.org/docs/20/jse_20_3_radin_1.pdf [archive.org] R010 [Radin 2006 JSE] ↩︎
  11. Radin, D. (2005). The sense of being stared at: A preliminary meta-analysis. Journal of Consciousness Studies, 12(6), 95–100. [citation pending verification] R011 [Radin 2005] ↩︎
  12. Bem, D., Tressoldi, P., Rabeyron, T., & Duggan, M. (2016). Feeling the future: A meta-analysis of 90 experiments on the anomalous anticipation of random future events. F1000Research, 4, article 1188. https://doi.org/10.12688/f1000research.7177.2 [PDF] R012 [Bem 2016] ↩︎
  13. Radin, D., Nelson, R., Dobyns, Y., & Houtkooper, J. (2006). Reexamining psychokinesis: Comment on Bösch, Steinkamp, and Boller (2006). Psychological Bulletin, 132(4), 529–532. https://doi.org/10.1037/0033-2909.132.4.529 [DOI] R013 [Radin 2006 PB] ↩︎
  14. Storm, L. (2026). Questions about Questionable Research Practices. Journal of Anomalous Experience and Cognition, 6(1), 92–102. https://doi.org/10.31156/jaex.27502 [PDF] R014 [Storm 2026] ↩︎
  15. Radin, D. (2007). Finding Or Imagining Flawed Research? The Humanistic Psychologist, 35(3), 297–299. https://doi.org/10.1080/08873260701578384 [DOI] R015 [Radin 2007] ↩︎
  16. Schmidt, S. (2021). Open Peer Comment to “Anomalous Cognition: An Umbrella Review of the Meta-Analytic Evidence”. Journal of Anomalous Experience and Cognition, 1(1-2), 73–75. https://journals.lub.lu.se/jaex/article/view/23439 [PDF] R016 [Schmidt 2021] ↩︎
  17. Storm, L. (2023). The Dark Spirit of the Trickster Archetype in Parapsychology. Journal of Scientific Exploration, 37(4), 665–682. https://journalofscientificexploration.org/index.php/jse/article/view/2715 [PDF] R017 [Storm 2023] ↩︎
  18. Tressoldi, P. (2012). Replication unreliability in psychology: elusive phenomena or “elusive” statistical power? Frontiers in Psychology, 3. https://doi.org/10.3389/fpsyg.2012.00218 [PDF] R018 [Tressoldi 2012] ↩︎
  19. Hyman, R., & Honorton, C. (1986). A joint communiqué: The psi ganzfeld controversy. Journal of Parapsychology, 50(4), 351–364. https://doi.org/10.30891/jopar.2018S.01.09 [DOI] R019 [Hyman 1986] ↩︎
  20. Honorton, C., & Ferrari, D. C. (1989). “Future telling”: A meta-analysis of forced-choice precognition experiments, 1935–1987. Journal of Parapsychology, 53. [citation pending verification] R020 [Honorton 1989] ↩︎
  21. Utts, J. (1991). Replication and meta-analysis in parapsychology. Statistical Science, 6(4), 363–403. https://doi.org/10.1214/ss/1177011577 [DOI] R021 [Utts 1991] ↩︎
  22. Storm, L., & Ertel, S. (2001). Does psi exist? Comments on Milton and Wiseman’s (1999) meta-analysis of ganzfeld research. Psychological Bulletin, 127(3), 424–433. https://doi.org/10.1037/0033-2909.127.3.424 [DOI] R022 [Storm 2001] ↩︎
  23. Milton, J., & Wiseman, R. (2001). Does psi exist? Reply to Storm and Ertel (2001). Psychological Bulletin, 127(3), 434–438. https://doi.org/10.1037/0033-2909.127.3.434 [DOI] R023 [Milton 2001] ↩︎
  24. Bem, D. J., Palmer, J., & Broughton, R. S. (2001). Updating the ganzfeld database: A victim of its own success? Journal of Parapsychology, 65. https://www.infoamerica.org/documentos_pdf/bem06.pdf [PDF] R024 [Bem 2001] ↩︎
  25. Sherwood, S. J., & Roe, C. A. (2003). A review of dream ESP studies conducted since the Maimonides dream ESP programme. Journal of Consciousness Studies, 10(6–7), 85–109. [citation pending verification] R025 [Sherwood 2003] ↩︎
  26. Schmidt, S., Schneider, R., Utts, J., & Walach, H. (2004). Distant intentionality and the feeling of being stared at: Two meta-analyses. British Journal of Psychology, 95, 235–247. [citation pending verification] R026 [Schmidt 2004] ↩︎
  27. Bösch, H., Steinkamp, F., & Boller, E. (2006). Examining psychokinesis: The interaction of human intention with random number generators—A meta-analysis. Psychological Bulletin, 132(4), 497–523. https://doi.org/10.1037/0033-2909.132.4.497 [DOI] R027 [Bösch 2006] ↩︎
  28. Storm, L., Tressoldi, P. E., & Di Risio, L. (2012). Meta-analysis of ESP studies, 1987–2010: Assessing the success of the forced-choice design in parapsychology. Journal of Parapsychology, 76, 243–273. [citation pending verification] R028 [Storm 2012] ↩︎
  29. Mossbridge, J., Tressoldi, P., & Utts, J. (2012). Predictive physiological anticipation preceding seemingly unpredictable stimuli: A meta-analysis. Frontiers in Psychology, 3, article 390. https://doi.org/10.3389/fpsyg.2012.00390 [PDF] R029 [Mossbridge 2012] ↩︎
  30. Storm, L., & Tressoldi, P. E. (2020). Meta-analysis of free-response studies 2009–2018: Assessing the noise-reduction model ten years on. Journal of the Society for Psychical Research, 84(4), 193–219. https://doi.org/10.31234/osf.io/3d7at [PDF] R030 [Storm 2020] ↩︎
  31. Tressoldi, P. E., & Storm, L. (2024). Stage 2 Registered Report: Anomalous perception in a Ganzfeld condition—A meta-analysis of more than 40 years investigation. F1000Research, 10, article 234. https://f1000research.com/articles/10-234/pdf [PDF] R031 [Tressoldi 2024] ↩︎

Last updated: 2026-09-08 02:58:09

Copyright © 2026 Innovative Software Design. All rights reserved.

ESP-Nexus reports published parapsychology research as written by its own authors. Source-fidelity to cited documents is audited on every page; the site takes no position on the epistemic status of parapsychological claims themselves.