The evidence, weighed

What the research actually shows.

For each research program: what proponents claim, the best meta-analytic effect sizes and citations, the principal skeptical critique, and where replication stands. The bar shows our read on how seriously mainstream science treats the program – not whether the claim is true.

NDE & resuscitation

Near-death experiences

Rejected / fringeStudied in mainstream
Studied within mainstream medicine; the paranormal interpretation is unconfirmed.
The claim
Vivid, structured experiences are sometimes reported from periods when the brain is severely compromised (e.g., cardiac arrest), and a minority describe seemingly veridical out-of-body perception.
Best evidence
AWARE-II (Parnia et al., 2023, Resuscitation 191:109903) was a prospective 25-site in-hospital study using hidden visual targets and audio cues during CPR. Of 567 arrests, 53 survived and 28 were interviewed; none recalled the hidden visual targets and one identified an audio cue. EEG showed gamma and other lucid-pattern activity during CPR in some patients – offered as a “biomarker of consciousness” hypothesis, not a paranormal finding. Bruce Greyson’s NDE Scale (1983) is the field-standard instrument.
Skeptical read
No veridical perception of a hidden target has been confirmed. Lucid EEG activity is not evidence of anything paranormal; expectancy, residual brain function, and reconstruction after the fact are sufficient explanations.
Replication
AWARE-III / COOL studies are ongoing. As of 2026 no verified hidden-target perception exists.
Telepathy / ESP

Ganzfeld telepathy

Rejected / fringeStudied in mainstream
A small, contested meta-analytic effect.
The claim
A “receiver” in mild sensory isolation can identify a target image being viewed by a distant “sender” at rates above chance.
Best evidence
Storm, Tressoldi & Di Risio (2010, Psychological Bulletin) reported an effect size ≈ 0.14 across 29 studies; Storm & Tressoldi (2020) ≈ 0.13 for 2008–2018. Rouder, Morey & Province (2013) found a Bayes factor ≈ 330 favoring an effect – while arguing it could stem from randomization weaknesses.
Skeptical read
Milton & Wiseman (1999) found no significant effect in a separate database. Effect sizes are small and may reflect randomization flaws, optional stopping, and selective reporting.
Replication
Preregistered ganzfeld studies tend to report smaller, less consistent effects than the older literature.
Precognition

Presentiment & precognition

Rejected / fringeStudied in mainstream
A small meta-analytic effect; the flagship modern study failed to replicate.
The claim
Physiology (and sometimes behavior) anticipates a randomly chosen future stimulus before it occurs.
Best evidence
Mossbridge, Tressoldi & Utts (2012, Frontiers in Psychology 3:390) meta-analyzed 26 experiments from 7 labs, ES ≈ 0.21. Duggan & Tressoldi (2018, F1000Research) added 19 studies, ES ≈ 0.28, reporting no publication-bias contamination by a Copas model.
Skeptical read
Daryl Bem’s 2011 “Feeling the Future” (JPSP) failed to replicate (Ritchie, Wiseman & French, 2012; Galak et al., 2012) and became a catalyst for psychology’s replication crisis. Large preregistered confirmatory replications remain scarce.
Replication
Bem’s flagship paradigm did not replicate; the surviving effect is small and contested.
Psychokinesis

Mind–matter interaction (RNG / micro-PK)

Rejected / fringeStudied in mainstream
A tiny effect that meta-analysts consider consistent with publication bias.
The claim
Human intention can bias the output of electronic random number / event generators.
Best evidence
Princeton’s PEAR lab (1979–2007) reported effects on the order of 0.1%. Bösch, Steinkamp & Boller (2006, Psychological Bulletin) meta-analyzed 380 studies and found a tiny but significant effect.
Skeptical read
The same meta-analysis judged the effect consistent with publication bias. PEAR’s own consortium replications with German labs (Giessen, Freiburg) failed to confirm the original results.
Replication
Multi-lab confirmation has failed.
Field RNG

Global Consciousness Project

Rejected / fringeStudied in mainstream
Rejected by most statisticians outside the field.
The claim
A worldwide network of random generators shows correlated departures from chance around major global events – a cumulative ~7-sigma result.
Best evidence
Roger Nelson’s project (hosted at noosphere.princeton.edu) reports ~500 formally defined events over ~20 years, framed as trillions-to-one against chance.
Skeptical read
Edwin May, Robert Carroll and others argue the results are artifacts of post-hoc event selection and pattern-matching; alternative event-selection rules yield chance-level results. The cumulative “trillion-to-one” framing is misleading.
Replication
Not accepted by mainstream statistics; no independent confirmation of the event-selection method.
Survival research

Reincarnation-type cases

Rejected / fringeStudied in mainstream
Qualitative case investigation – not an experiment.
The claim
A minority of young children make statements and show behaviors that appear to match a specific deceased person they could not have known.
Best evidence
UVA’s Division of Perceptual Studies maintains a database of 2,200+ “cases of the reincarnation type” begun by Ian Stevenson; Jim Tucker’s Return to Life (2013) and Before (2021) describe American cases.
Skeptical read
Cases rest on interview accuracy, parental memory, and after-the-fact record-matching; they cannot be controlled experimentally and are not accepted by mainstream historians or psychologists.
Replication
Not applicable – qualitative cases produce no controlled, repeatable effect.
After-death communication

Mediumship

Rejected / fringeStudied in mainstream
Above chance in one lab; independent replication is limited.
The claim
Under blinding, some mediums report information about a deceased person more accurately than chance would predict.
Best evidence
Windbridge’s “quintuple-blind” protocol (Beischel et al., 2015, EXPLORE) reports above-chance accuracy in some studies with certified research mediums.
Skeptical read
Replication outside the Beischel/Boccuzzi lab is limited; cold reading, rater bias, and statement-fitting are the standard alternative explanations.
Replication
Limited independent replication to date.
Field investigation

Apparitions & hauntings

Rejected / fringeStudied in mainstream
Mostly fringe; a thin tradition of careful survey work.
The claim
People report recurring apparitional experiences, sometimes tied to particular places.
Best evidence
The SPR’s Census of Hallucinations (1894) collected 17,000+ accounts; careful modern work uses surveys and environmental measurement.
Skeptical read
The bulk of “paranormal investigation” – EMF meters, spirit boxes, orb photography – meets the standard of no peer-reviewed journal. Suggestion, infrasound, electromagnetic fields, and expectancy explain much of the rest.
Replication
Field investigation, not laboratory work; no controlled, repeatable effect.
The central scientific exchange

In 2018, Etzel Cardeña reviewed ~750 studies across 11 paradigms in American Psychologist and argued the cumulative evidence “cannot be readily explained away” by quality, fraud, or selective reporting. In 2020, Reber & Alcock replied – in the same journal – that the claims “cannot be true” on a priori physical grounds. Neither side has produced an empirical resolution. That unresolved tension is the honest state of the field.

Why the replication crisis matters here

Parapsychology sits at the center of psychology’s reckoning with replication. Daryl Bem’s 2011 “Feeling the Future” reported precognition using ordinary methods; the failures to replicate it became a canonical case for why the field needed preregistration and registered reports. The irony: parapsychologists were among the earliest adopters of those reforms – and the resulting preregistered studies tend to show smaller, less consistent effects, suggesting the older literature was inflated by the same questionable practices that affected mainstream psychology.

So when you read an effect size here – ganzfeld ≈ 0.13, presentiment ≈ 0.21 – read it with its skeptical reanalysis attached. Reasonable experts look at the same numbers and disagree.

Key sources

  1. Cardeña, E. (2018). The experimental evidence for parapsychological phenomena: A review. American Psychologist, 73(5), 663–677.
  2. Reber, A. S., & Alcock, J. E. (2020). Searching for the impossible. American Psychologist, 75(3), 391–399.
  3. Storm, L., Tressoldi, P., & Di Risio, L. (2010). Meta-analysis of free-response studies. Psychological Bulletin, 136(4), 471–485.
  4. Mossbridge, J., Tressoldi, P., & Utts, J. (2012). Predictive physiological anticipation. Frontiers in Psychology, 3:390.
  5. Parnia, S., et al. (2023). AWARE-II. Resuscitation, 191:109903.
  6. Bem, D. J. (2011). Feeling the future. JPSP, 100(3), 407–425; Ritchie, Wiseman & French (2012), PLoS ONE.
  7. Bösch, H., Steinkamp, F., & Boller, E. (2006). Examining psychokinesis. Psychological Bulletin, 132(4), 497–523.