Low involvement of preexisting associations makes retrieval-induced forgetting long lasting.

Research has shown that selective retrieval of episodic memories usually leads to forgetting of related memories that compete for retrieval [a phenomenon known as retrieval-induced forgetting (RIF)]. However, there are conflicting data regarding the duration of this incidental kind of forgetting. While some studies have found that this forgetting effect disappears within 24 h after selective retrieval, others suggest that it may last for as long as at least a week. In two experiments, we explored whether discrepancies in the durability of RIF may be due to variations in the type of relationships (preexisting vs. novel) that are present between items associated with a given cue. We explored this issue by manipulating the degree of involvement of preexisting/novel associations among competing items as well as the delay between retrieval practice and test (immediate in Experiment 1 and 24-h delay test in Experiment 2). The results suggest that forgetting lasts longer when the degree of preexisting associations among targets and competitors is low.