A quantitative digest of the ESP-Nexus structured study database for Presentiment: 14 studies spanning 1997–2023, with every figure traceable to its source. ESP-Nexus reports what the studies found and takes no position on whether the effects are genuine.

A 2012 meta-analysis by Mossbridge pooling 26 studies found a small but positive pooled effect size of Cohen's d = 0.21 for predictive physiological anticipation preceding emotional stimuli. Individual studies spanning 1997 to 2023 have frequently reported statistically significant results in the same direction: z-scores across Radin studies range from 2.72 to 5.0 in earlier work, with more recent values of z = -3.087 and z = -3.825, while Mossbridge studies report p-values as low as p < 1×10⁻⁶. Effect sizes where reported vary considerably, from 0.064 to 0.21, suggesting meaningful heterogeneity in magnitude.

Despite this pattern of positive findings, several methodological concerns complicate interpretation. The effect sizes are uniformly small, and the literature shows variation in analytic approaches, physiological channels measured, and statistical thresholds applied, as seen in Kittenis (2011) and the multi-channel Radin (2011) results. Whether this consistency across labs reflects a genuine anticipatory signal, shared methodological artifacts, or selective reporting remains a subject of active disagreement among researchers.

Pooled meta-analytic effects

The evidence

PaperReported findingEffect / significanceBasis
Radin (2023), World Futures [source]Combined permuted-slope Stouffer Z across all 10 languages.z = -3.087, p = .001k = 10 events/tests
Mossbridge (2023), Journal of Anomalous Experience and Cognition [source]EEG presentiment – pre-stimulus prediction of future button-press response.p < 1 × 10−6N = 4000
Hamelin et al. (2022), Journal of Behavioral and Experimental Finance [source]Pre-choice GSR difference across valid quizzes.p = 3.3 × 10−4N = 72
Radin (2022), DRAFT (unpublished manuscript)Retrospective: all 10 languages combined, Stouffer Z for sad-day slopes.z = -3.825, p = 6.5 × 10−5k = 10
Mossbridge (2017), Lecture Notes in Artificial Intelligence (AC 2017, Part I, LNAI 10284) [source]Random forest classification: original vs.p < 2.5 × 10−6N = 4000
Mossbridge et al. (2012), Frontiers in Psychology [source]Overall pooled effect – fixed-effect model.ES 0.21, z = 6.9, p < 2.7 × 10−1226 studies
Kittenis (2011), Journal of ParapsychologyPre-stimulus phase-synchronisation difference, New vs Old faces.N = 20
Radin (2011), AIP Conference Proceedings [source]Two most-recent unconscious classes combined — sign test.101 studies
Tressoldi et al. (2010), Journal of Scientific ExplorationThis study – hits of high-Absorption-level groups.ES 0.11, z = 2.43, hit rate 0.56N = 440
McCraty et al. (2010)Win vs loss pre-stimulus physiological difference.N = 13
Radin et al. (2009), Explore [source]Experiment 1 – Pupillary Dilation: Emotional vs Calm.z = 3.17, p = 8 × 10−4N = 1438
Radin et al. (2007), The Journal of Alternative and Complementary Medicine [source]Females – prestimulus SCP differentiation, flash vs no-flash.z = 2.72, p = .007N = 1300
Radin (2004), Journal of Scientific ExplorationAll four experiments combined – weighted mean pre-stimulus effect size.ES 0.064, z = 4.04, p = 1.3 × 10−5k = 4; N = 4569
Radin (1997), Journal of Scientific ExplorationCombined autonomic presponse – before display.z = 5.0k = 2
Source: ESP-Nexus structured study database (14 studies). ESP-Nexus reports what each study found and takes no position on whether the effects are genuine.

Source: ESP-Nexus structured study database, 14 studies. 14 born-digital figures; OCR-derived figures awaiting source-verification are excluded from this digest. Effect sizes are shown only where a standardized metric is recorded.