4. Discussion
In the present work, we compared the metacognitive abilities of children and adults using the same perceptual type I task and type II task. The protocol and tasks were built to be age-appropriate for both groups. This was crucial given that we aimed to measure metacognitive differences between the two groups and not differences in their cognitive performance. In doing so, we strengthen the result that metacognition, as measured through our perceptual metacognitive task, is already developed in young children, as well as the notion that children are overconfident. In addition, we show here that this tendency may not be due to a poor metacognitive access (d’), which is not significantly different from adults, but instead to a difference in the decision criteria (c) employed to make their choices. Our results suggest that when middle childhood begins, metacognitive access is already at an adult level; however, the output of metacognitive monitoring to regulate these operations, metacognitive control, may not be fully developed yet. Furthermore, the results from the adult group seem to indicate that an optimal level of metacognitive control may never be achieved, given that they still present a tendency to be suboptimally overconfident.