The status of implicit and explicit memory research: is a comprehensive theory of performance achievable?

The status of Implicit and Explicit memory research was examined from a predominantly cognitive perspective. Across ten experiments, predictions from three extant models were tested: Roediger and colleagues' Transfer Appropriate Processing (TAP) framework, Graf and Mandler's Dual Processing (DP) framework and Tulving's Serial, Parallel, Independent (SPI) model. The inclusion of Jacoby and colleagues' Process Dissociation Procedure supplemented analyses for Experiments 7 to 10. At study, participants performed one of several 'depth-of-processing' tasks. For tests employing conceptual cues, an advantage for 'deeper' encoding conditions was consistently observed, regardless of test instructions (implicit or explicit). Conversely, for tests employing perceptual cues, test instructions resulted in a dissociation in performance. Despite evidence supporting the phenomenon of 'conceptual automaticity', PDP estimates were indicative of 'contamination' by controlled processes on the implicit conceptual tests. Issues concerning the validity of the PDP estimates were subsequently raised. Comparisons of performance within 'equivalent' levels of encoding, resulted in greater consistency with implicit, relative to explicit instructions. Across the experiments, explanations included reference to 'involuntary conscious memory' and utilisation of a 'generate-recognise' strategy. In general, the TAP framework was less successful in accounting for performance than the DP and SPI models. The latter were viewed as complementary models that conjointly provided a sound framework for understanding memory performance.