It is impossible, even undesirable, for us to remember everything, thus intentional forgetting is an essential process for cognitive efficiency. Decades of research using an item-method directed forgetting paradigm demonstrate our ability to intentionally forget information. To-be-remembered (R) and to-be-forgotten (F) items are encoded differently, with R items better represented than F items in long-term memory. Consequently, we predicted that the well-established part-list cuing impairment of recall caused by re-exposure of some of the studied items prior to test should be greater when R items rather than F items serve as the part-list cues. We found that, relative to no part-list cues, R items reduced recall performance for the remaining R items, but F items improved recall performance for the remaining F items. We interpret this finding as evidence that F items disrupt the retrieval strategy less than R items do, and propose a context differentiation account.