Half of all published psychology papers that use significance testing contain at least one p-value that does not match its own test statistic (Nuijten…
In 1982, two psychologists resubmitted 12 already-published papers to the journals that had printed them, under made-up names. Peer review spotted onl…