Subliminal priming as a task-characteristic artifact

Abstract

Demonstrations of subliminal priming rely on a dual task design: A target identification task in which the priming effect is measured, and a prime identification task in which visibility of the primes is assessed. The validity of this design relies on the critical assumption that the estimates of prime visibility accurately reflect prime visibility in the target identification task. Here it is suggested that the difference in difficulty between the two tasks results in a violation of this assumption. Specifically, the target identification task is easy while the prime identification task is extremely difficult. It is shown that decreasing the overall difficulty of the prime identification task results in increased prime identifiably. It is also shown that primes which are unable to be identified in a task which accurately estimates prime identifiably do not elicit a priming effect. Hence, we conclude that demonstrations of subliminal priming are an artifact of this violation rather than a real phenomenon.