Almost a coin flip: Chance-like performance on TOMM Trial 1 using a sham extrasensory perception condition

Binomial probabilities are used to interpret below chance performance on forced-choice validity tests, but this does not account for variance inherent in true random responding. This study examined random responding on Trial 1 of the Test of Memory Malingering (TOMM) using a novel experimental condition. Fifty-five undergraduates completed TOMM Trial 1 under sham instructions to induce random guessing. Total scores and item-level responses were tested against exact binomial probability assumptions. Mean Trial 1 performance was 25.96/50 (SD = 3.38, range = 17-33), significantly above nominal chance, t(54) = 2.12, p = .039. A 95% probability range would be 19-33/50, with the theoretical exact range being 18-32/50. Twelve TOMM 1 items deviated significantly from .50 after correction. Sham random responding on TOMM Trial 1 closely approximated the binomial model at the total-score level, with only small item-level departures from .50 producing a slight, clinically trivial upward shift in the mean. These data empirically support the continued use of binomial-based total-score interpretation for below-chance responding, while indicating that item-level probabilities are not perfectly uniform.