The necessity of equivalent comparisons for interpreting masked priming effects: Implications for studies of unconscious perceptual processing.

Abstract

Many masked priming studies compare measurements from two conditions: Direct reports of judging heavily masked, briefly presented prime stimuli versus indirect priming effects collected from judgments of easily visible targets. However, comparisons between unequal conditions may introduce a significant interpretational weakness for inferring unconscious processing. Examples of unequal conditions may include differences in tasks, stimulus visibility, difficulty (hard vs. easy), and statistical properties (unequal metrics). We addressed these problems by aiming for methodologically equivalent direct and indirect task conditions. Lexical decisions (words vs. nonwords) of prime and target stimuli were collected, with masked prime visibility being manipulated (short = 16 ms, medium = 51 ms, and long = 103 ms prime durations). For short duration primes, no statistically significant prime discrimination or priming effects were found, with Bayesian analyses being inconclusive. In contrast, the long duration prime conditions showed significant prime discrimination and a priming effect. A sensitivity analysis showed that direct sensitivity greatly exceeded indirect sensitivity to the prime stimuli in the medium and the long prime duration conditions. The importance of equivalent testing conditions and statistical analysis is discussed.