On the reliability of implicit and explicit memory measures

Functional dissociations between implicit and explicit memory tests often t ake the form of large differences between groups or experimental conditions (e.g., amnesics and controls, elderly and younger persons, or persons lear ning with and without a distracting secondary task) when performance is ass essed using explicit memory tests, whereas no difference is observed with i mplicit memory tests. We argue that the interpretation of such dissociation s in terms of the memory processes or systems involved in performance is pr oblematic because the same data pattern would emerge as a result of a mere methodological artifact, that is, the situation that implicit memory tests have low reliability whereas explicit memory tests are fairly reliable meas urement instruments. We present reasons for such a reliability difference, and we demonstrate it empirically in Experiments 1a, 1b, and 2. However, ou r analysis also shows, and Experiment 3 confirms empirically, that implicit memory tests need not necessarily be less reliable measurement instruments than explicit memory tests. (C) 2000 Academic Press.