Dissociating measures of associative memory: Evidence and theoretical implications

In four experiments, the authors investigated whether two measures of associative recognition memory (associative identification and associative reinstatement) are dissociable from one-another on the basis of their reliance on strategic retrieval and are dissociable from item recognition memory. Experiment 1 showed that deep encoding of relational information, but not of individual items, increased both types of associative memory significantly, as indexed by both measures, while it only marginally increased item memory. Experiments 2–4 showed that a short response deadline, a speeded recognition and an overlapping pairing condition interfered with associative identification, but left associative reinstatement unaffected. Associative reinstatement provides a measure of associative memory, but unlike associative identification, it is less reliant on strategic retrieval processes. We propose that associative familiarity underlies this measure. This process may index binding of information at encoding without involving the vivid, conscious re-experiencing characteristic of recollection at retrieval.