Abstract
During visual search for two mutually exclusive targets, attention is biased towards whichever of the targets the search is framed as "for". For example, when looking for a P that can be either green or red, if participants have to answer the question "is the P red?", they will tend to search through red items. Green items, by comparison, are neglected, even though they provide information; on the half of trials the P is green, and therefore not red. Does this confirmation bias occur simply because one target's color is held in working memory? In the present study, two groups of participants completed visual searches for one of two possible colored targets letters in displays of stimuli drawn in two colors. One group was asked to search for the target letter while maintaining a particular color in memory, and after each search their memory was tested to confirm that the color was remembered. A second group was instead asked whether the target letter was a particular color, with no additional memory load. By varying the ratio of the number of stimuli drawn in these two colors, we tracked which subset was being prioritized. Search frames (is the p red?) produced a robust bias towards the subset matching one target type. However, requiring participants to remember a color (e.g., red) did not produce biases towards subsets matching the color held in memory, despite showing a recognition advantage for targets matching the color held in memory. These data show that visual confirmation bias is not merely a result of the contents of working memory. More broadly, these results support the notion that there is more to guiding visual attention to stimulus features than simply holding features in memory.
Meeting abstract presented at VSS 2017