Confidence weighting (CW) is concerned with measuring two variables: (1) what a respondent believes is a correct answer to a question and (2) what degree of certainty the respondent has toward the correctness of this belief.[1]
Confidence weighting when applied to a specific answer selection for a particular test or exam question is referred to in the literature from cognitive psychology as item-specific confidence, a term typically used by researchers who investigate metamemory or metacognition, comprehension monitoring, or feeling-of-knowing.[2] Item-specific confidence is defined as calibrating the relationship between an objective performance of accuracy (e.g., a test answer selection) with the subjective measure of confidence, (e.g., a numeric value assigned to the selection).[3] Studies on self-confidence and metacognition during test taking (e.g.,[4][5]) have used item-specific confidence as a way to assess the accuracy and confidence underlying knowledge judgments.
Researchers outside of the field of cognitive psychology have used confidence weighting as applied to item-specific judgments in assessing alternative conceptions of difficult concepts in high school biology and physics (e.g.,[6][7]), developing and evaluating computerized adaptive testing (e.g.,[8]), testing computerized assessments of learning and understanding (e.g.,[9][10][11][12]), and developing and testing formative and summative classroom assessments (e.g.,[13]). Confidence weighting is one of three components of the Risk Inclination Model.
