Metacognitive sensitivity moderates reactivity to confidence ratings
Year: 2025
Type: article
Abstract: Confidence ratings are frequently used to measure metacognition. However, a growing body of evidence has found that eliciting confidence ratings affects reasoning performance – a phenomenon known as reactivity. Additionally, recent research has suggested that there are individual differences in the direction and magnitude of reactivity effects. Here, in a preregistered experiment (N = 200), we examined whether eliciting confidence ratings during a fluid intelligence test (Raven’s Progressive Matrices) affected performance, and whether this effect was moderated by participants’ baseline metacognitive abilities captured using a perceptual discrimination task. We evaluated whether reactivity was moderated by 1) metacognitive sensitivity (the extent to which confidence judgments track objective performance), 2) metacognitive efficiency (the ratio of metacognitive sensitivity relative to cognitive performance), and 3) mean confidence on the perceptual discrimination task. Our findings suggest that only metacognitive sensitivity moderated reactivity, with participants who had relatively poor metacognitive sensitivity showing positive reactivity and those with relatively good metacognitive sensitivity not showing reactivity. This suggests that eliciting confidence ratings may have a compensatory effect – benefiting those who have poorer metacognition, which suggests they may be a useful intervention for promoting metacognitive processes.
Source: Thinking & Reasoning
Authors: Ryan Choo, Kit S. Double
Institution: The University of Sydney
Language: English
Subfield: Cognitive Neuroscience
Field: Neuroscience
Domain: Life Sciences
Open Access status: hybrid
Funder: Australian Research Council
OpenAlex ID: W4413104203