The research story

How the labs tested psi, from Rhine’s cards to a preregistered test in 2023.

CHANCE

  1. 1 · The 1930s

    Rhine’s cards at Duke

    J. B. Rhine brought psi into a university lab at Duke and made the phrase “extra-sensory perception” famous with his 1934 book of that name. He turned the question into statistics: hits against a known chance rate, over thousands of guesses. Every later psi experiment inherits that frame. More on clairvoyance.

  2. 2 · The 1970s

    Remote viewing at SRI

    At SRI in California, Russell Targ and Harold Puthoff tested people describing distant sites and published in Nature in 1974, alongside an editorial that set out its doubts. More on remote viewing.

  3. 3 · 1986

    The ganzfeld and a joint communiqué

    After years of argument over the ganzfeld, the critic Ray Hyman and the researcher Charles Honorton did something rare: they wrote a paper together, agreeing on the standards a good ganzfeld study must meet and that new studies should be judged by them.

    The automated studies that followed reported hit rates above chance. More on telepathy.

  4. 4 · 1995

    Star Gate is evaluated

    The CIA commissioned an evaluation of two decades of government remote viewing. Its two expert reviewers, Jessica Utts and Ray Hyman, read the same data and reached opposite conclusions. The report concluded that using it for intelligence was not warranted.

  5. 5 · 1979 to 2007

    PEAR and the random machines

    Princeton’s PEAR lab spent nearly three decades on random number generators. A three-lab replication failed on its main measure, and a 2006 meta-analysis of 380 studies found a tiny effect that publication bias could explain. More on psychokinesis.

  6. 6 · 2011

    Bem and the replication crisis

    Daryl Bem’s “Feeling the future” passed peer review at a leading journal using the standard methods of the day. Critics asked: if those methods could show precognition, what else had they let through? Failed replications followed, and the question spread across psychology. More on precognition.

  7. 7 · 2023

    The Transparent Psi Project

    A test designed with supporters and sceptics together, registered in advance, with its data published live. 2,115 people took part, and the hit rate was 49.89% against 50% chance: no support for precognition in that design.

  8. 8 · Where things stand

    Two views, side by side

    The evidence is there

    Etzel Cardeña reviewed the meta-analyses across psi research and argued that the evidence is comparable to that for accepted findings in psychology, and deserves serious study.

    It cannot be right

    Arthur Reber and James Alcock argued that psi would break well-established laws of physics, so positive results must come from flaws, whatever they show.

    Why they disagree: Cardeña weighs the evidence as it stands, while Reber and Alcock argue that it cannot be right whatever it shows.

    How to read a psi study

    Chance level
    What random guessing would score: 25% with four choices, 20% with five Zener symbols. A result means little until you know it.
    Preregistration
    Was the plan, including how the data would be analysed, written down and published before the data came in?
    Sensory leakage
    Could the answer have reached the person by an ordinary route: a glimpse, a sound, a cue in the transcript?
    File drawer
    Studies that find nothing are less often published. Pooling only the published ones can make a small effect look real.
    Independent replication
    Has a different team, ideally a sceptical one, got the same result using the same method?
Sources

1 · The 1930s

2 · The 1970s

3 · 1986

4 · 1995

5 · 1979 to 2007

  • Jahn and others, 2000: the replication, in the Journal of Scientific Exploration 14 (not online)
  • Bösch and others, 2006: the meta-analysis

6 · 2011

7 · 2023

8 · Where things stand