Abstract
In a typical visual working memory task participants study an array of colored items and must report the color of an item at some probed location. In some recent studies, however, a color is probed and participants must report the location of that item. A striking difference between these tasks is a preponderance of so called "swap errors" in location reports, whereby participants almost never guess randomly as they do when reporting color, but instead mistakenly report the location of non-probed items. This finding has been taken as evidence for feature binding errors in memory, and evidence against the prediction of discrete capacity models that guessing should occur at high set sizes. We propose an alternative interpretation: That "swap errors" in location memory are in fact random guesses, but smart ones. In particular, when asked to report the location of a color for which participants have no memory, they may guess at locations where they know some item was presented, and avoid guessing at locations where they know no items were presented. In Experiment 1 we find evidence of such informed guessing by examining confidence ratings: When participants respond near non-studied items they have low confidence, indicating guessing rather than a true swapping of colors and locations in memory. In Experiment 2 we asked participants to report the locations of colors that were not in the study array. The errors and confidence ratings on these false-probe trials look exactly like swap errors — reports are centered around non-studied items and confidence is low — even though there was no studied item that could have possibly been swapped. These results suggest that the tendency to report features of non-probed items reflects a good guessing strategy; not something fundamental about how features and objects are represented in working memory.
Meeting abstract presented at VSS 2018