Response latencies in visual search involving redundant or irrelevant information

Three properties of models for comparison of multiattribute visual stimuli were considered: parallel vs serial processing, efficient vs exhaustive comparisons, and biased vs unbiased acceptance of attributes for processing. Ss performed two comparison tasks, matching-to-sample (M) and identification of odd stimuli (0), with color and form attributes presented singly and in redundant and nonredundant pairings. Analyses of means and of total distributions of response latencies supported the conclusion that parallel and efficient comparisons were the rule, along with a kind of partial selection of attribute to be processed. Ss differed in their relative speed of processing form and color attributes, and these differences accounted for most, but not all, of the differences among them in processing multiattribute stimuli.