Abstract
A commonly-used task for investigating perception is visual search. However, much of the basic nature of search remains unclear. For example, what is the nature of the items on which it operates? Are stimuli perceived in terms of visual angle, or inferred size in the world? To investigate, performance for 16 observers was assessed on displays containing 8, 16, or 24 black vertical bars against a white background; the task was to detect the presence of a longer item among shorter ones. In Condition 1, search for this display was compared against that for a version which had been compressed by a factor of 2, as well as against both of these when viewing distance was doubled. Results showed a clear effect of compression but not distance. Similar behavior was encountered for Condition 2, which was much the same except that width was held constant. Condition 3 kept all positions but used only full-sized items, so that density alone varied. Results showed no effect of density or distance. Consequently, it appears that size alone was the relevant feature in these conditions, and reflected physical rather than retinal measures. To explore the mechanisms involved, Condition 4 used a full-size display, but flickered it using various combinations of on- and off-times. Results indicated that usability of iconic memory was limited, suggesting the involvement of feedback. Condition 5 tested whether performance was affected by the wearing of an eye-patch. No reliable effect of this was found. These results provide further support for the suggestion that visual search operates over a representation that describes the physical extent of world being viewed; they also indicate that the relevant scene-based properties can be derived entirely on the basis of static global / oculomotor considerations.