Abstract
Confidence measures and other rating scales are often used to index metacognition and visual awareness/consciousness. However, confidence ratings do not always reliably track performance or awareness, suggesting that confidence ratings should not be used in place of performance or subjective awareness measures. To assess the relationships between confidence, subjective awareness, and objective performance, we conducted an experiment to evaluate discrimination performance on shape-majority decisions (a fixed ratio of two groups of shapes) while also measuring confidence ratings and subjective reports of awareness of a near-threshold color difference that was present on some trials to facilitate the shape-majority decisions. We found that the near-threshold color difference improved performance on the shape discrimination task even when the color difference was not consciously perceived. However, confidence ratings reflected performance accuracy in the shape-majority decision task only when participants were subjectively aware of the color difference or when no color difference was present. Importantly, this relationship between confidence ratings and performance was not significant when participants were unaware of the near-threshold color difference. These results demonstrate a dissociation between confidence ratings, performance, and awareness, and indicate that these different types of measures cannot be used interchangeably, as many studies on metacognition and visual awareness have been conflating.