Different Effects of Level-of-Processing and Self-Generation in Implicit Tests

Abstract

Three experiments were conducted to assess effects of level-of-processing and self-generation on conscious and unconscious components of memory within an implicit test of word-stem completion. The two components of memory were dissociated by a post-test dissociation procedure. Equations for estimating these two components of memory are formulated by excluding recognition memory that is irrelevant to repetition priming and guessing effects from estimates of memory contributing to completion. Results of Experiments 1 and 3 show that level-of-processing produced a positive effect on overall completion associated with a positive effect on the estimate of conscious memory and a null effect on the estimate of unconscious memory. This pattern of results was due to repetition-priming effects on both estimates of conscious and unconscious memories under the two levels of processing. Re-suits of Experiments 2 and 3 show that generation produced a null effect on overall completion, accompanied by a positive effect on the estimate of conscious memory and a reverse effect on the estimate of unconscious memory. This pattern of results was due to a repetition-priming effect on the estimate of conscious memory under each study condition and a repetition-priming effect under the shallow and a repetition-inhibition effect under the generation condition on the estimate of unconscious memory. The present results are discussed with an encoding-facilitation/ inhibition theory.