Manipulating Prior Beliefs Causally Induces Under- and Overconfidence.

Area of Science:

• Cognitive psychology
• Neuroscience
• Decision science

Background:

• Individual differences in decision confidence are significant but lack a clear computational explanation.
• Under- and overconfidence in decision-making have profound impacts on life outcomes.

Purpose of the Study:

• To investigate the role of prior beliefs in shaping decision confidence, independent of objective performance.
• To develop a computational model explaining confidence variations based on prior task-related beliefs.

Main Methods:

• Two experiments involving perceptual decision-making in healthy adults (N=50 each).
• Prior beliefs were manipulated through comparative feedback and training difficulty.
• Objective performance and subjective confidence were measured during a test phase.

Main Results:

• Manipulated prior beliefs causally influenced decision confidence without altering objective performance.
• Findings were consistent across different methods of belief induction.
• An accumulation-to-bound model incorporating prior beliefs successfully accounted for the observed confidence levels.

Conclusions:

• Prior beliefs, shaped by early task exposure, are a key determinant of decision confidence.
• Decision confidence is computed as the probability of correctness, conditional on prior beliefs.
• This provides a mechanistic understanding of the computational basis for under- and overconfidence.