Abstract
This study investigated information processing in visual search, by examining how eye movement patterns changed during search as a function of changes in the stimulus array and in performance variability. 3 Ss (2 naive) searched regular, dense linear arrays of Landolt-C-like boxes for one of 4 targets and reported its orientation (4AFC). A video eyetracker (EyeLinkII) recorded eye movements at 500Hz. We manipulated the orientation (Ss scanned left, right, up and down), width (1 or 3 columns), and target/distracter similarity (high, medium, low) of the search array across trials. We measured search speed (rows scanned per sec), perceptual span (number of items scanned divided by number of fixations) and fixation duration. All manipulations of the search array substantially influenced search speed as well as perceptual span, changing them by 15–55%. Fixation durations changed much less (5–15%). In particular, changing from easy to hard distracters decreased span by 29%, and changing from 1 column of easy distracters to 3 columns of hard distracters decreased span by 41%, but fixation durations increased by only 9% in both cases. There was also an extremely robust and consistent correlation between span and search speed within and across conditions (typical r = .8–.9). The correlation between fixation duration and search speed was much weaker and often not significant. Thus, the visual system appears to use a relatively set amount of time to process information at each fixation. Increasingly difficult searches rely on increased foveal proximity of stimuli, not longer exposure to them.