Abstract
Despite the richness of research testing performance in visual tasks, and the extensive body of research exploring metacognitions about memory and language, very little research has explored subjects' beliefs about their visual capabilities. Here, I present research testing the basis of beliefs about the detectability of visual changes, and memorability of large numbers of pictures. In both cases, subjects made large metacognitive errors: they overestimated their ability to detect unexpected changes that occurred across cuts in movies (illustrating a phenomenon we refer to as “change blindness blindness”, or CBB), and they underestimated their ability to remember even relatively few (10–50) pictures. In addition, subjects completed a questionnaire asking them to estimate the percentage of objects they would look at in a complex scene, and the degree to which paying attention to one thing precludes “seeing” something else. These and other estimates were entered into two multiple regressions, one predicting CBB, and the other predicting estimates of pictures memory. Breadth of attention, and percentage of objects viewed predicted high CBB, while estimates of picture memory were unassociated with CBB. Breadth of attention was positively associated with optimistic estimates of picture memory, while percentage of objects viewed and magnitude of CBB were unassociated with estimated of picture memory. These results suggest that subjects have specific beliefs about the degree to which they countenance visual information, and that these beliefs influence their predictions of success for change detection and picture memory tasks. This finding will be discussed in the context of the broader hypothesis that subjects may generally understand the difference between representations and the real world but that they have difficulty predicting how the representational process interacts with attention-demanding real-world tasks, even when those tasks are very familiar.