Relation between belief and performance in perceptual decision making.

Area of Science:

• Cognitive Neuroscience
• Decision Science
• Psychophysics

Background:

• Effective decision-making in uncertain environments necessitates metacognition, aligning belief with performance.
• Accurate self-monitoring and outcome prediction depend on this belief-performance congruence.
• Researchers often assume belief and performance are directly correlated, crucial for inferring internal states.

Purpose of the Study:

• To investigate the relationship between belief and performance in decision-making under varying conditions.
• To identify circumstances where belief and performance equality holds or diverges.
• To address implications for experimental design and theories of metacognition.

Main Methods:

• Analysis of decision-making models where performance is defined as accuracy for a fixed stimulus.
• Examination of belief-performance correlations across different task difficulties.
• Modeling of certainty judgments to explain miscalibration and the hard-easy effect.

Main Results:

• Belief and performance equality is rare when performance is defined as response accuracy for a fixed stimulus.
• Belief and performance correlations vary significantly across tasks and are weak in many standard psychophysical settings.
• Miscalibration and the hard-easy effect can arise from a mismatch in expected task difficulty distributions between the decision-maker and the experimenter.

Conclusions:

• Standard definitions of performance in psychophysics can lead to a divergence between subjective belief and objective performance.
• This divergence poses challenges for interpreting metacognitive states and requires careful consideration in experimental design.
• Understanding the conditions that cause belief-performance mismatch is crucial for advancing theories of metacognition and certainty judgments.