ABSTRACT
A series of visual search experiments was carried out in order to verify the commonality of short-term search processes and selective attention. Experiment I used a multiple frame procedure and accuracy as a measure. It varied the relation of the target set to the distractor set, memory-set size, and frame size. To reach a given accuracy level, the presentation time per frame varied from 40 to 800 msec (or more) across conditions. The data showed evidence for slow serial search, termed controlled search, when the target and distractor sets varied from trial to trial, and fast efficient parallel search, termed automatic search, when the target and distractor sets were fixed from trial to trial. Experiment II used a single frame procedure and latency as a measure. Virtually all the accuracy effects of Experiment I were found in the latency results of Experiment II, and the two experiments were linked together. Experiment III was like Experiment I but presented more than one target per trial. The results qualitatively distinguished automatic from controlled search. First steps were taken toward a search model for selective attention.
