Context change and retrieval difficulty in the list-before-last paradigm.

Using the "list-before-last" paradigm (Jang & Huber, Journal of Experimental Psychology: Learning, Memory, and Cognition, 34, 112-127, 2008; Shiffrin, Science, 168, 1601-1603, 1970), we examined whether the difficulty involved in retrieving a previously studied list affects the recall of the current list. Participants studied three lists (L1, L2, and L3), and between L2 and L3 study they either engaged in retrieval of L1 or solved math problems for the same duration of time. After L3 encoding, all participants recalled L2. We examined accurate recall of L2, intrusions from L3, and also the first-response function from L2 across all four experiments. In Experiments 1 and 2, retrieval difficulty was manipulated by presenting participants with variable numbers of letter cues for the L1 words, with some participants receiving more cues than others. In Experiment 1, L1 and L2 words shared the first two letters in common to create potential item interference, whereas in Experiment 2, they did not share common cues. In Experiments 3 and 4, we manipulated retrieval difficulty using a delay manipulation-L1 was encoded 1, 24, or 72 h prior to the session during which L2 and L3 were encoded. In Experiment 3, L1 retrieval involved a cued recall test, whereas in Experiment 4, it involved a free recall test. The results of all of these experiments showed that, as compared to solving math problems, retrieving L1 led to forgetting of L2, reduced intrusions from L3, and reduced first-response functions from L2. However, all of the dependent measures were invariant across the various manipulations of retrieval difficulty. We discuss the results in terms of the context-change interpretation of prior-list retrieval.