Nine Experiments Run Backwards

The design of these experiments is their whole rhetorical point. Bem took procedures every social psychologist recognises, took nothing else from parapsychology, and reversed the time order. In one, participants studied a list of words, were tested on recall, and only afterwards did a computer select a random subset of those words for them to practise; the reported finding was that they had recalled more of the words they would later practise. In another, a variant on a standard priming task, the prime appeared after the response. In another, participants guessed which of two curtains on a screen concealed an image, with the image's position selected by the computer only after the guess. Across nine experiments and more than a thousand participants, eight returned statistically significant results in the direction predicted.

The paper appeared in the Journal of Personality and Social Psychology in 2011, one of the most selective outlets in the discipline, after normal peer review. Bem's framing was that if these effects are real they matter enormously, and that if the methods are sound the results must be reported regardless of prior plausibility. The journal's editor later defended the decision to publish on the ground that the paper met the field's standards.