Abstract
Metamemory is the cognitive capacity to make decisions about the existence and fidelity of one's own memories. Here, we describe two experiments that demonstrate high-quality visual working metamemory. Participants were asked to remember the colors of a set of colorful dots. In the first experiment, on half of the trials participants reported the color of the item they remembered best, while on the other half of the trials they reported the color of a randomly-selected item (these two conditions were interleaved and revealed only after removal of the stimulus). We found that observers showed considerably less error for the best-remembered item than for one selected at random (SD of error distribution: 21° vs. 15°, p < 0.005). Using extreme order statistics, we show that this improvement implies that metamemory is of high quality. In a second experiment, participants gave a confidence rating on each trial in the form of an arc drawn over the color wheel, centered on the reported color, which represented the smallest range that the participant was fairly certain would contain the true color. We found that when participants drew larger arcs, their memories were less precise and they were more likely to guess blindly. The relationship between drawn arc size and imprecision (i.e., the standard deviation of the error distribution) was well fit by a linear function; similarly, the relationship between drawn arc size and guess rate was well fit by a logistic function. Together, the results of these experiments reveal high-quality visual working metamemory: participants can make decisions about the existence and fidelity of their working memories.
Meeting abstract presented at VSS 2012