A role for reconsolidation in the testing effect

The act of recalling information from memory results in a higher probability of successfully remembering that information later than does an additional study opportunity. This testing effect is both powerful and well documented, yet there are only a few suggestions for why testing benefits memory. The three existing theoretical positions that have often been brought to bear on the testing effect are the desirable difficulties hypothesis, the semantic mediator hypothesis, and transfer-appropriate processing hypothesis. Recently, a new explanation for why the testing effect may occur has been put forth based on the idea of reconsolidation. Following results in the neuroscience of animal learning and human learning, the reconsolidation hypothesis assumes the act of retrieval renders the original memory for the retrieved event more malleable than a re-study opportunity. We expand on this hypothesis, proposing that this malleability results in a “generalized” contextual trace that includes elements of the original study episode and of the circumstances of testing. If the final criterion test is performed in a different context, then this more generalized context engenders a higher probability of successful recall. In the current experiments, we evaluate two predictions derived from this hypothesis. The first is that individuals may lose details of the original study context following retrieval practice (Experiment 1a, 1b, and 2). Second, there should be more interference between the original study context and a retrieval practice context than during the original study context and a restudy context (Experiment 3). The results did not support these predictions and thus do not support the reconsolidation hypothesis