Another way to investigate the time to shift attention is to display an array and ask subjects to search among objects for a specific object. If one can manipulate the number of objects that a subject must search through, one can manipulate search time. The slope of the function relating number of objects attended to search time gives an estimate of time to move attention. This straight-forward logic is complicated by the fact that subjects can select which objects to attend to on the basis of their features. Thus, for instance, in looking for a red object, subjects will not be affected by the number of green objects in the array.
An example which reflects such a paradigm and its complexities is Schneider and Shiffrin's (1977) study of visual search. In their Experiment 2, subjects had to detect a target item when it was presented in a visual display of 1 to 4 items (frame size). The target letter was in a memory set of 1 to 4 (memory set size). For instance, subjects might hold a memory set of B and K and be asked if either element occurred in a visual array that contained G, K, M, and F (in which case they would respond yes). Subjects were either in what was called the varied-mapping or the consistent-mapping condition. In the varied-mapping condition both distractors and the memory set items were letters (drawn from the same pool on each trial) and in the consistent-mapping condition the memory set was composed of numbers and the distractors were letters (therefore, they were always drawn from different pools).