Overview

Skeptical critiques of parapsychology cover a wide range. They touch on how studies are designed, how data are analyzed, and what assumptions researchers bring to the table.

The leading skeptics are not simply dismissive. Ray Hyman, James Alcock, Richard Wiseman, and Susan Blackmore have each looked closely at specific experiments. They have pointed to real sources of error: information leaking through normal senses, results that never get published, and flawed randomization. These critiques have had a real effect. Many of the controls used in today’s psi research were added because skeptics demanded them. In some cases, skeptics and parapsychologists worked together to set shared standards.

Core Claims and Frameworks

Skeptical arguments fall into three groups: methodological, statistical, and philosophical.

Methodological concerns ask whether the studies were run properly. The main worries are:

• Sensory leakage: Information reaches the participant through normal channels (sound, sight, smell) rather than through any paranormal means.
• The file-drawer problem: Studies that find nothing tend to stay unpublished. Only positive results get into journals. This makes the evidence look stronger than it is.
• Optional stopping: A researcher keeps collecting data until the numbers look good, then stops. This inflates the chance of a false positive.
• Weak blinding or randomization: If the person running the study knows which target was chosen, they might give subtle cues to the participant without meaning to.

Statistical concerns ask whether the numbers are being read correctly. Skeptics point out that small effect sizes can look impressive but may reflect bias rather than a real phenomenon. They also note that running many statistical tests at once raises the odds of finding a “significant” result by pure chance. And they argue that psi is so unlikely given what we know about physics that even a positive result should be treated with great caution.

Philosophical concerns ask whether psi is even a reasonable hypothesis. The principle of parsimony says we should prefer the simplest explanation that fits the data. Psi requires entirely new physical mechanisms. Skeptics argue that the bar for accepting such a claim must be very high.

Not all skeptics agree on emphasis or tone. Some focus tightly on flaws in specific studies. Others take a broader stance about what counts as good science. What they share is a demand for rigorous evidence before accepting any psi claim.

Evidence Base: Key Skeptical Critiques

The Hyman Ganzfeld Critique. In 1985, Ray Hyman published a detailed critical review of ganzfeld experiments.^[1] The ganzfeld method was the most widely used protocol in psi research at the time. In a typical ganzfeld study, a “receiver” sits in a relaxed state while a “sender” in another room concentrates on a randomly chosen image. The receiver then tries to identify which image was sent, choosing from a small set of options. Hyman found multiple problems. Target sequences were not always randomized properly. Sound could leak between rooms. And researchers tended to report positive results while leaving negative ones unpublished. He argued these flaws were enough to explain the above-chance hit rates without invoking psi.

The Hyman-Honorton Joint Communique. Instead of simply trading attacks, Hyman and parapsychologist Charles Honorton worked together. In 1986 they published a joint statement in the Journal of Parapsychology.^[2] Neither side declared victory. Instead, they agreed on what future ganzfeld studies would need: automated randomization, proper acoustic isolation, and pre-registered analysis plans (meaning researchers had to say in advance what they would measure and how). This showed that serious disagreement does not have to block productive collaboration.

Methodological Concerns from Alcock and Wiseman. James Alcock has spent years documenting how flaws in study design can inflate apparent results across parapsychology as a whole.^[3] He has shown how file-drawer effects, optional stopping, and experimenter bias can each push results in a positive direction. Richard Wiseman has done detailed technical work on specific experiments. For example, he investigated whether sound could leak from the sender to the experimenter in automated ganzfeld systems, and whether an experimenter’s expectations could subtly influence participants.^[4]

Meta-Analytic Challenges. In 1999, Milton and Wiseman combined results from many ganzfeld studies into a single statistical review.^[5] When they applied proper statistical controls, they found no reliable evidence for psi. They also noticed that effect sizes varied a lot from lab to lab. Better-designed studies tended to show smaller effects. That pattern fits what you would expect if publication bias and selective reporting were driving the results, not a genuine psi signal.

Expectancy Effects. In 1997, Wiseman and Schlitz ran a study together on remote staring: the idea that people can sense when someone is looking at them from a distance.^[6] They found that the experimenter’s own beliefs about psi seemed to affect the results. When Schlitz (a psi proponent) ran the sessions, results were positive. When Wiseman (a skeptic) ran them, results were not. This showed that even subtle, unintended cues from the experimenter can shape outcomes. It made the case for double-blind protocols even stronger.

The Blackmore Database and Berger’s Audit. In 1987, Blackmore published an essay in Skeptical Inquirer titled “The Elusive Open Mind: Ten Years of Negative Research in Parapsychology”^[7]—a first-person account framing her parapsychological research as a decade of null results. Berger (1989) ^[8] published a critical examination in the Journal of the American Society for Psychical Research that documented the actual experimental record. Per Blackmore’s own dissertation reporting and Berger’s audit, the experimental work consisted of 29 psi experiments conducted between October 1976 and December 1978 (approximately 26 months) — 21 of which were published as separate experiments in five peer-refereed parapsychology journal papers. Blackmore’s 1987 first-person account confirms the paradigm breakdown: ESP/memory correlations (Blackmore 1980a), ESP testing with preschool children (Blackmore 1980b), target-material variations (Blackmore 1981a), error-pattern analyses (Blackmore 1981b), a single formal ganzfeld study with ten trained subjects and ten controls, and Tarot card readings (Blackmore 1983). Berger’s audit identified discrepancies between Blackmore’s unpublished dissertation and her published reports; “flaws” invoked to dismiss significant results while ignored when studies produced non-significant results; and two instances in which study chronology was reordered. Per the 29 experiments documented in Blackmore (1987) ^[7] and Berger (1989) ^[8], seven produced statistically significant results — a figure Blackmore’s 1987 framing of “the only noteworthy thing about all of the results was the number that were not significant” does not engage. Berger concluded that “no conclusions should be drawn from this database.” Post-1978, Blackmore’s research turned to out-of-body experiences, which she classifies in the same 1987 essay as non-paranormal. The “ten years of ganzfeld” framing that has circulated in secondary sources is not supported by either the cited literature or Blackmore’s own first-person account.

Counterarguments and Debate

Current Discourse and Productive Outcomes

Modern psi experiments routinely use automated randomization, acoustic isolation, pre-registered protocols, and independent outcome assessment. Hyman 1985 ^[1], Alcock 1987 ^[3], Wiseman et al. 1996 ^[4], and Milton & Wiseman 1999 ^[5] each name specific procedural gaps that subsequent protocols addressed.

The Hyman-Honorton communique is the clearest example of how serious disagreement can still produce useful results.^[2] Two people who disagreed sharply about whether psi exists agreed on what a well-run study would look like. That is a real scientific achievement.

Skeptical engagement with parapsychology is still active. Skeptics keep examining new claims and new methods. Parapsychologists keep trying to address the objections. The exchange is sometimes heated. But it has been more useful than either side simply ignoring the other.

The Hyman-Honorton communiqué ^[2] specifies what cited authors on both sides agreed would constitute methodologically adequate evidence: automated randomization, proper acoustic isolation, and pre-registered analysis plans. The standards published in 1986 and modeled in subsequent protocols are more specific than the standards described in the corpus that predates the communiqué.