In three experiments, we examined the performance of patients with schizophrenia on implicit and explicit memory tests that have been shown to involve predominantly data-driven or predominantly conceptually driven processes. In Experiment 1, we compared the implicit tests of category production (conceptually driven) and word identification (data driven) and found that schizophrenic patients' performance on these tests did not differ from that of normal subjects. In Experiment 2, a comparison of the category-production and explicit cued-recall tests, both of which involve conceptual processes, showed that schizophrenic patients were impaired on the cued-recall test but not on the category-production test. In Experiment 3, a comparison of the word-identification and explicit graphemic cued-recall tests, both of which involve data-driven processes, showed that patients were impaired on the cued-recall test but not on the wordidentification test. The results of both Experiments 2 and 3 revealed a dissociation between implicit and explicit test performance under conditions in which the two tests involve similar types of processes. These results support theoretical views that distinguish implicit from explicit modes of retrieval.Over the past decade, many studies have revealed striking dissociations in performance between explicit memory tests such as recall and recognition and implicit memory tests such as word-fragment or word-stem completion, in which performance changes are measured by repetitionpriming effects (for review, see Richardson-Klavehn & Bjork, 1988;Schacter, 1987). One interpretation of findings of dissociations is that there is a fundamental difference in the type of memory involved in explicit and implicit memory tests (Graf &