Skip to content
Above Top Secret
ESP

Presentiment: the experiment that broke psychology instead

A respected psychologist published nine experiments appearing to show people respond to events before they happen. The lasting consequence was not what he found — it was what the attempt to check it revealed about his entire field.

ATS Field DeskSites and witnesses
3 min read
Laboratory research equipment

Presentiment studies measure physiological response before a randomly selected stimulus.

Presentiment is the claim that the body responds to a future stimulus before it occurs — that skin conductance, pupil dilation or heart rate shift a few seconds ahead of a randomly selected emotional image.

It is an unusually good thing to test, because it requires no interpretation of what a subject says. The measurement is physiological and the stimulus is chosen by a random generator after the response has been recorded.

What Bem did

Illustration

In 2011 Daryl Bem, a well-regarded social psychologist at Cornell, published nine experiments in the Journal of Personality and Social Psychology, one of the field's leading journals.

The design took standard, uncontroversial psychology experiments and reversed them in time. In one, subjects were shown a list of words and tested on recall — and then, after the test, were given practice on a randomly selected subset. Subjects recalled the words they would later practise better than the ones they would not.

Eight of the nine produced statistically significant results.

What happened next

Replication attempts failed. Stuart Ritchie, Richard Wiseman and Christopher French ran three independent replications of the retroactive recall experiment and found nothing.

The journal that had published Bem declined to consider their paper, on the stated ground that it did not publish replications.

That refusal is the moment the story stopped being about psi.

What preregistration fixes

The reforms that followed Bem's paper are worth setting out, because they are the practical legacy.

Preregistration means the hypothesis, sample size, exclusion criteria and analysis are lodged publicly before data collection. It removes the ability to decide, after seeing the results, which comparison to report.

Registered reports go further: the journal reviews and accepts the method before the study runs, and publishes whatever it finds. That removes the incentive to produce a positive result.

What Bem himself said

Bem has been consistent that he used standard methods and that this was part of his point. He argued the experiments were conducted the way his field conducted experiments. Critics agreed with that description and drew the opposite conclusion from it — and that shared premise is why the episode produced methodological reform rather than a dispute about one laboratory.

The consequence

Psychologists looking at the Bem paper found they could not identify anything he had done that was outside normal practice — flexible stopping rules, multiple measures, analysis decisions made after seeing data, and a publication system with no route for negative findings.

If those methods could produce nine significant results for something as implausible as precognition, they could produce significant results for anything.

The paper is now routinely cited as one of the triggers of psychology's replication crisis, and the reforms that followed — preregistration, registered reports, mandatory power analysis, data sharing — are in substantial part a response to it.

Join the discussion

Still open on this one:

  1. 01

    Has any presentiment result survived preregistration?

    Preregistered work in this area exists and is sparse. It is the only category of evidence now worth attending to, and there is not much of it.

Join the discussion

Be the first to comment

Sources

Open them and check for yourself. Where a record is public, it is linked.

  1. Society for Psychical Research

    SPR

  2. Research guides and catalogue

    The National Archives

0 comments