Some Developmental Trends in Metacognition

By Chris Was, PhD; Kent State University

Recently, I have conducted some experiments with K – 6 grade students related to children’s ability to predict their ability to recall simple items. Although a simple measure, this form of calibration is a measure of a child’s knowledge of the own memory abilities. This is, at its most basic level, metacognition.

The work in which my collaborators and I are currently engaged builds on the work of Amanda Lipko and colleagues (e.g., Lipko, Dunlosky, & Merriman, 2009). What was most striking about Lipko’s work was the robust overconfidence displayed by preschool children. Granted, there is a large body of literature that demonstrates young children are overconfident in both their physical abilities (e.g., Plumert, 1995) as well as their cognitive abilities (e.g., Cunningham & Weaver, 1989; Flavell, Friedrichs, & Hoyt, 1970). Much of this work indicates that with preschool children this overconfidence is quite persistent. But Lipko et al.’s (2009) work found that even following repeated practice and feedback, specifically salient feedback when children recalled their own previous performance, this overconfidence remained.

There are several hypotheses, both tested and untested, as to why this overconfidence exists and why it is robust against correction. Perhaps it is wishful thinking (a hypothesis test by Lipko et al.), perhaps it is a developmental issue, or perhaps it serves as a learning mechanism (children who give up the first time they fail may not learn to do succeed at much). In any case, I became interested in circumstances in which young children are capable of making accurate predictions of their cognitive abilities.

A review of the experimental methodology used by Lipko et al. is warranted. In their 2009 study Lipko et al. presented young children (mean age of approximately 5 years 0 months) with pictures of common items. As children were presented with pictures they were asked to name them. If correctly named the picture was placed on a board until 10 pictures were on the board. The experimenter then said to the children, “I am going to cover up the pictures,“ and asked, “how many do you think you will remember after I cover them?” The children then made a prediction of how many pictures they would remember. Finally, the children attempted to recall the pictures. In a series of experiments, children were overconfident in their ability even after repeated trials and even after correctly recalling their poor performance on previous trials.

Are there circumstances when children are more accurate? The simple answer is, “yes.” In a recent experiment (Was & Al-Harthy, 2015) we found that when children complete the Lipko task with unfamiliar items, their predictions of how many items they might remember are significantly lower than for familiar items. This familiarity overconfidence bias is likely due something similar to the fluency effect. That is, when the pictures are familiar to children, they seem easy to remember, but when the pictures are unfamiliar, children understand that they might be hard to recall later.

We are also investigating the developmental trends of the ability to predict recall. Our most interesting finding to date, is calibration (accuracy of recall predictions) is strongly related to the increase in working memory capacity. Put differently, as the number of items children are able to recall increases, so does their ability to accurately predict the number of items they will recall. Some will argue that this is not an unsuspected finding. The argument being that as working memory capacity increases, the ability to think about one’s own memory should also increase. My response is that it is not clear if metacognition is directly related to working memory or executive functions. Perhaps a mediating relationship exists. Recent investigations have suggested that performance on many measures of working memory are more dependent on strategy than they are on cognitive ability. Perhaps, metacognition is just good strategy use, or perhaps it is a cognitive ability.

The finding of the relationship between recall performance and calibration (the difference between predicted performance and actual performance) supports the hypothesis that metacognition is not a single skill that children have or not, but rather it is a complex of many skills and processes the children acquire through experiences and maturation. I suggest that developmental research in metacognition need focus on aptitude-by-treatment interactions. Questions such as, “What variety of academic activities contribute to the development of metacognition at different stages or levels of cognitive development?” will not only forward our understanding of metacognition, but perhaps also how to help young students develop metacognitive strategies and perhaps metacognitive performance.

Cunningham, J. G., & Weaver, S. L. (1989). Young children’s knowledge of their memory             span: Effects of task and experience. Journal of Experimental Child Psychology, 48,   32–44.

Flavell, J. H., Friedrichs, A. G., & Hoyt, J. D. (1970). Developmental changes in memorization    processes. Cognitive Psychology, 1,324–340.

Lipko, A. R., Dunlosky, J., & Merriman, W. E. (2009). Persistent overconfidence despite practice: The role of task experience in preschoolers’ recall predictions. Journal of Experimental Child Psychology103(2), 152-166.

Plumert, J. M. (1995). Relations between children’s overestimation of their physical abilities and accident proneness. Developmental Psychology31(5), 866-876. doi: http://dx.doi.org/10.1037/0012-1649.31.5.866

Was, C. A., & Al-Harthy, I. (2015). Developmental differences in overconfidence: When do children understand that attempting to recall predicts memory performance? The Researcher, 27(1), 1-5, Conference Proceedings of the 32nd Annual Conference of the Northern Rocky Mountain Education Research Association.