Evaluation of the fixation duration in visual search.

Abstract

Vaughan and Graefe (1977) proposed that visual fixation duration measures dunng visual search are confounded by two factors, cognitive processing ti.~e and time required by the oculomotor system to trntiate the next eye movement, the oculomotor latency. It was concluded that, under some conditions, the fixation duration is independent of the visual information available to the eye during the fixation. That is the oculomotor latency exceeds the information p;ocessing time. This independe~~e was ~ound. to occur in a relatively simple recogrution task involving one display item per fixation. In a visual search involving relatively high information .display it~ms.' or where several items are included m each fixation, the oculomotor latency would not be expected to determine the fixation duration. This proved to be the case in a study by Mackworth (1976) in which several items were included in each fixation. Subjects searched a strip of several rows of circles for a square target. Display density increased with the number of circles in the strip. By counting the number of eye movements made on each strip, the number of circles per fixation was estimated. Above four circles per fixation, both the fixation duration and the number of circlesper fixation tended to increase with display density. Mackworth attributed the increase in fixation duration to the extra cognitive demand added by the need to process more nontarget circles during each fixation. An alternate method of estimating the number of display items per fixation is to construct a display that controls this factor. Mocharnuk (1978) had subjects search a display consisting of clusters of letters. It was likely that each fixation would accommodate one cluster. As expected, fixation duration increased with the number of letters per cluster. A fixation duration effect was also found by Moffitt (Note I), who varied the information content of each display item while controlling the number of items per fixation. Subjects scanned back and forth be~ween two changing display items, located 10 deg of visual angle apart, searching for a member of the memory. set. Given the size of each display item, .22 deg of visual