Toward specifying the attentional demands of recognition memory

Previous studies have concluded that recognition memory is immune to disrup tion from divided attention and therefore is a relatively automatic process (A. Baddeley, V. Lewis, M. Eldridge, & N. Thomson, 1984; F. I. M. Craik, R . Govoni, M. Naveh-Benjamin, & N. D. Anderson, 1996). Because costs have be en found on the concurrent task used to divide attention, recognition may n evertheless require some attentional resources (M. Naveh-Benjamin, F. I. M. Craik, J. Guez, & H. Dori, 1998). The present authors used attention-deman ding concurrent tasks to demonstrate significant costs on both the concurre nt task and recognition memory performance. Decrements in recognition accur acy were found for classes of items that were studied deeply but not for mo re shallowly learned materials. The present findings suggest that recogniti on processes can require significant attentional resources when tested unde r the appropriate conditions. The results are discussed in terms of the req uirements both at encoding and at test that are needed to observe dual-task decrements to recognition accuracy.