The search for memory: Visual short-term memory capacity predicts performance during visual search tasks

During the last decade, there have been repeated attempts to answer the question of whether visual search is supported by memory. Horowitz and Wolfe (1998) were the first to demonstrate that search with or without memory is equally efficient implying that visual search has little or no memory component. Since then, the evidence has been mixed. We argue that the inconsistent findings are a result of an insufficient operational definition of memory effects during visual search. To overcome this issue, we propose a simple formula to estimate the number of search items held in memory. Previously, we have applied this approach to a preview search task and demonstrated that the number of items that could be inhibited was roughly equal to the capacity of visual short-term memory (VSTM). Here, we revisit Horowitz and Wolfe's study to test whether VSTM supports visual search more generally. We presented participants with search displays in which either all of the search items were moved randomly around the display, thereby disrupting the potential effect of memory, or displays in which 0 to 7 of the total search items remained static, allowing the previously searched static stimuli to be stored in memory. At large set sizes, the efficiency of search in the static condition was similar to that of the dynamic condition, in line with the results of Horowitz and Wolfe (1998). However, at set sizes below the estimated capacity of visual short-term memory (i.e., 4), the search slope in the static condition was significantly smaller. Using our formula, the estimated number of items held in memory during visual search peaked at roughly four items. In addition, this estimate was moderately correlated with the capacity of VSTM obtained in a change-detection task. Thus, we provide further evidence that VSTM supports the representation of search items.