Searching for conjunctively defined targets

A. M. Treisman and G. Gelade proposed that in searching for a target defined as a conjunction of 2 or more separable features, attention must be paid serially to each stimulus in a display. Support for this comes from studies in which Ss searched for a target that shared a single feature with each of 2 different kinds of distractor items (e.g., a red O in a field of black O 's and red N 's). RT increased linearly with display size. It is argued that this design may obscure evidence of selectivity in search. The present 2 experiments, with 24 undergraduates, were modeled after the experiment by Treisman et al (1977) with the exception that the numbers of the 2 distractors were unconfounded. Results indicate that Ss can search through specified subsets of stimuli. For example, Ss told to search through just the O 's to find the red O 's target do so without searching through N 's. It appears that figure-ground segregations may occur when there is a large asymmetry in the number of 2 distractor types, with all of the elements forming the ground being rejected in parallel. (12 ref)