Particularly Exciting Experiments in Psychology™

photo of stack of 6 APA journals that focus on experimental psychology

February 9, 2017

Metacognitive Judgments in People and Rats

young adult male student taking standardized paper and pencil testDuring midterms, students rely on judgments of learning to decide when to stop studying, and rely on confidence judgments to decide whether to double check an exam answer or move on to the next question. Such metacognitive judgments may be based on beliefs (e.g., the belief that studying an item twice will lead to better memory), or subjective experience (e.g., that the answer came easily therefore it must be correct).

Frank and Kuhlmann (online first, Journal of Experimental Psychology: Learning, Memory, and Cognition) (PDF, 163KB) tested how experience and beliefs jointly contribute to metacognitive judgments.

Participants were told they would hear words presented quietly or loudly. To measure pre-existing metacognitive beliefs, participants were asked to estimate the percentage of loud vs. quiet words they would remember, prior to hearing any of the actual words. Then, the words were presented and participants made judgments of learning (e.g., how likely there were to remember each word) immediately following each word. These item-level judgments could be based on subjective experience, beliefs, or some combination.

Overall, participants predicted they would remember a larger number of loud than quiet words, and loud words had higher average judgments of learning than quiet words, even for participants who did not believe that volume would affect their memory. Moreover, the more strongly people believed pre-study that loud words would be remembered better, the larger the judgments of learning for loud vs. quiet words during study. A subsequent experiment confirmed that beliefs and experience make independent contributions to metacognitive judgments, but only when the difference in volume for loud vs. quiet words during the study phase was perceptually large.

Frank and Kuhlmann measured metacognition by directly asking participants to make metacognitive judgments. Of course, such methods cannot be used to study metacognition in non-human animals. Perhaps for this reason, metacognition was initially thought to be a uniquely human ability.

Yuki and Okanoya (2017, Journal of Experimental Psychology: Animal Learning and Cognition) (PDF, 289KB) tested whether rats have metacognitive abilities using a task in which rats choose between making a memory choice or switching to an easier task that offers a smaller reward. The logic is that animals will choose the easier task with a smaller reward when they are less likely to succeed in the memory task; i.e., based on a metacognitive prediction of their likely memory test accuracy.

Results showed that rats increased their likelihood of receiving a reward by using the easy task option, as revealed by better performance when rats had the choice of an easy task vs. trials where they were forced to make a memory judgment. In other words, rats had higher memory accuracy when they chose to make a memory judgment. However, this was only the case for a 6-alternative memory test, not a 2-alternative memory test.

These results suggest that rats exhibit the use of metacognition to maximize performance when that task is sufficiently difficult, consistent with results obtained in some non-human primates. Because rats are genetically and environmentally controllable, establishing that they have metacognitive abilities has important implications for studying metacognition at the molecular and neural levels.


  • Frank, D. J., & Kuhlmann, B. G. (2016, October 6). More Than Just Beliefs: Experience and Beliefs Jointly Contribute to Volume Effects on Metacognitive Judgments. Journal of Experimental Psychology: Learning, Memory, and Cognition. Advance online publication. http://
  • Yuki, S., & Okanoya, K. (2017). Rats show adaptive choice in a metacognitive task with high uncertainty. Journal of Experimental Psychology: Animal Learning and Cognition, 43(1), 109–118.