Metacognition and Control in Human Learning Memory

Columbia Univ New York Morningside, New York NY

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): This research is directed at the relation between people's metacognitions and the efficacy of subsequent control processes targeting learning. Control of learning is a central question both in understanding human learning amd memory, but also in devising efficient therapeutic interventions. The experiments test a Region of Proximal Learning model of how people's metacognitions are used to apportion study time and effort. There are two components to the model: choice and perseverance. The first series of experiments investigates how people make choices of what to study, and to the fundamental question of whether these choices are efficacious. Model predictions are tested by allowing people to choose, and then either honoring or dishonoring their choices. A second series of experiments investigates when people should persevere and when they should decline further study. The model indicates perseverance when people's jROLs--the judged rate of learning-are high, and stopping when jROLs approach zero. This stop rule allows the model to escape the 'labor in vain' paradox of other models. People's perseverance will be compared to model predictions, and related to data determining optimal time allocations and spacings. A third series of experiments investigates the criterion for study-choice related to calibration of judgments of learning. The model indicates that when overconfident people will decline study, even though study is warranted; when underconfident, they will study, even when study is unecessary-and time is stolen from other materials that would benefit. These predictions will be tested under systematic variations in confidence, and the reasons for those variations investigated. The goal of this research-to test the Region of Proximal Learning model-will contribute to understanding of how people use their metacognitions to control learning. Control of learning is central for understanding human executive processes, and fundamental for human cognition. This research will also contribute to the foundations of a scientific study of dysfunctions of executive processes. Metacognitive and control impairments may be especially important in OCD, where the setting of both the choice and perseverance components of the model may be too stringent, in depression, where significant underconfidence has been observed, and in impulsive disorders, such as ADHD, where overconfidence and too lenient a stop rule may be critical.