Examining the mechanism and influential extent of retrieval-induced forgetting

Three assumptions of the pattern suppression model of retrieval-induced forgetting were examined, with a view to apply results toward making effective use of reminders. These assumptions included: (1) that retrieval-induced forgetting is retrieval specific, (2) that it results from retrieval competition and (3) that access to affected items is inhibited at the level of the item representation. In Experiment 1, the retrieval-specific nature of this effect was examined by contrasting the effects of read and generate tasks, conducted in the retrieval practice phase of the retrieval practice paradigm, on later free recall. Subjects showed retrieval-induced forgetting only when the generate task was employed, suggesting that the retrieval process is necessary to produce a forgetting effect. The scope of retrieval competition was tested in both Experiments 2 and 3. Experiment 2 focused upon whether items become competitors simply by sharing semantic features with retrieval practiced targets, or whether competitors also must be subsumed within the domain of the retrieval practice cue. Items in the overlap between two categories were retrieval practiced as members of one of the paired categories (e.g., "lemon" was practiced as a "YELLOW" item, but was also a "FRUIT"). Subsequent recall of critical items from either the retrieval practiced (YELLOW) or overlapping (FRUIT) category was examined. Subjects showed retrieval-induced forgetting only for items that were members of the retrieval practiced category, suggesting that mere semantic relatedness is not sufficient to make an item a competitor. Experiment 3 was designed to examine: (1) whether nonstudied items in the semantic domain of an episodic retrieval cue act as competitors, and (2) whether item representations are inhibited in retrieval-induced forgetting. A perceptual identification task was administered as the last segment of the retrieval practice paradigm. These results were inconclusive, as no effect of retrieval practice was found either for reaction times or numbers of errors. Possible interpretations of this result in light of current literature are discussed, as are potential applications and future directions for this line of research, particularly in cognitive rehabilitation