Participants studied category-exemplar pairs ("FRUIT Cherry," "FRUIT Grape") and then practiced some of the items ("Cherry"). In Experiment 1, practice that involved retrieving the item from memory suppressed recall of related items ("Grape"), a finding known as the retrieval-induced forgetting (RIF) effect. In Experiment 2, practice that involved studying the item without retrieval produced no RIF effect. Both retrieval and nonretrieval practice facilitated the subsequent recall of practiced items ("Cherry"). The dissociation between "strengthening" of practiced items and forgetting of related items is thought to be evidence that RIF is the result of inhibition during earlier retrieval attempts rather than interference from competing memories at retrieval. However, simulations of the SAM-REM model show that competitor interference can account for this dissociation. Experiments 3-6 supported the predictions of the model by demonstrating that nonretrieval practice can produce the RIF effect under conditions that emphasize context encoding or increase the number of competitors.