A reevaluation of semantic versus nonsemantic processing in implicit memory.

A number of investigations have purported to demonstrate that semantic processing does not produce a memorial advantage over nonsemantic processing on implicit tests, as contrasted with the typical advantage of semantic over nonsemantic processing on explicit tests. A review of 166 outcomes from 38 studies that have manipulated processing on implicit tests reveals that on 131 occasions (79%), priming was greater following semantic than it was following nonsemantic processing. This difference was found in both perceptual and conceptual implicit memory tests, as well as in within- and between-subjects designs. It suggests that implicit tests reflect the involvement of both conceptual and perceptual processes. Although explicit contamination may account for some of the processing difference in implicit tests, the pervasiveness of the phenomenon, especially in perceptual implicit tests, makes it an unlikely account for the entire effect.