ATTENTION IN VISUAL-SEARCH - DISTINGUISHING 4 CAUSES OF A SET-SIZE EFFECT

Visual search is a common task both in naturalistic settings and in the laboratory. Outside the laboratory, one might look for a car in a parking lot, a name in text, or a navigation marker on the horizon. In the laboratory, search is simplified in several ways; commonly, the subject views a set of distinct objects and is asked to detect the presence of a particular object (the target) among a set of distracters. Two examples are shown in Figure 1. The top two panels illustrate the contrast increment task, in which the target is a disk of high luminance and the distractors are disks of lower luminance; the bottom two panels illustrate the line bisection task, in which the target is a rotated L and the distractors are rotated Ts. One of the most studied aspects of visual search is the effect on performance of the number of objects, here referred to as the display set size. Display set sizes of 2 and 24 are illustrated in Figure 1. In the top panels, the target ''pops out''-even for a large display set size. More precisely, display set size has little or no effect on search time or accuracy when the target is much brighter than the distracters. In contrast, in the bottom panels, finding the target requires ''scrutiny'' for the large set size. Display set size has a large effect on both search time and accuracy when the target and the distracters are these different rotated characters. These variations in the magnitude of set-size effects pose a central question for research on visual search. In this article, I use signal detection theory to analyze set size effects on search accuracy. It remains to be seen how this analysis will extend to the more commonly studied set-size effect on search time.