“A tool for longitudinal assessment of individual differences in metacognition during category learning : A Pilot Study” (Preprint)

UNSTRUCTURED Accurately judging the quality of one’s estimates is a critical psychological process that varies across individuals, fluctuates over time, and is highly domain specific. While there are many methods for assessing metacognitive monitoring during basic sensory tasks and tests of general knowledge, the field lacks a rigorous, controlled tool for assessing it during category learning. This is especially important given how category learning strategies vary both in the general population and across psychiatric conditions. We therefore performed a pilot development and feasibility evaluation of a tool that pairs an established category learning paradigm with a standard metacognitive report, and interrogated the tool through longitudinal testing. A total of 80 adult subjects completed three sequential sessions online and answered a set of personality and psychiatric trait questionnaires. Test-retest results suggested that the initial version of our tool provides fair reliability in assessing Brier score, and resulted in calibration improvements across tests. We found preliminary evidence that select psychiatric symptom and personality trait measures were associated with metacognitive performance. In addition, attentional and decision strategies significantly interacted with metacognitive metrics: most participants reported narrowing attention to a subset of features while carefully evaluating available options during decision-making Our results suggested specific steps to improve this new tool’s retest stability and develop it for longitudinal assessment of individual differences in metacognitive monitoring during category learning.