Retrieving a subset of previously studied material can impair later recognition of related items. Using the remember-know procedure (Experiment 1) and the receiver operating characteristic procedure (Experiment 2), the authors examined how such retrieval-induced forgetting can be explained in terms of single-process and dual-process accounts of recognition memory. Consistent across the 2 experiments, dual-process analysis suggested that retrieval practice reduces unpracticed items' familiarity but leaves their recollection largely unaffected, a finding that disagrees with prior work that points to recollective deficits in the forgotten items. Assuming that recognition is entirely based on a single source of memorial information, single-process analysis led to an excellent description of the data and suggested that retrieval practice reduces unpracticed items' general memory strength. This suggestion is consistent with prior work on free recall, cued recall, associative recognition, and response latencies and agrees with the inhibitory account of retrieval-induced forgetting. The authors argue that retrieval-induced forgetting in item recognition is caused by a reduction in general memory strength.