Recent research has shown that divided attention at retrieval, in contrast to encoding, affected memory performance only minimally. This immunity at retrieval was associated with a significant secondary task cost. In this article the authors further investigated these effects employing a cued-recall task and a multimeasure approach with accuracy, latency, overall attentional costs, and the temporal distribution of attentional costs associated with the encoding and retrieval of low- and high-frequency words. The results of 2 experiments yielded a complex pattern of both similarities and differences between encoding and retrieval. Simultaneous inspection of the different measures of performance was instrumental in identifying 3 major types of retrieval (unsuccessful, slow, and fast), as well as different phases of the retrieval process, each of which was characterized by a different demand for attentional resources.
