Search for multiple targets: Remember the targets, forget the search

Models of visual search performance typically assume that search proceeds b y sampling without replacement. This requires memory for each deployment of attention. We tested this assumption of memory-driven search using a multi ple-target search paradigm. We held total set size constant, varied the num ber of targets in the display, and asked subjects to report whether or not there were at least n targets present, where n was varied by block. This al lowed us to measure the time to find each subsequent target. Memory-driven search predicts that reaction time should be a linear function of n. The al ternative memory-free search hypothesis predicts an accelerating function. The data falsify the memory-driven hypothesis. They were consistent with th e memory-free search hypothesis but would also be consistent with memory fo r a small number of previously attended locations.