Dual strategies in human confidence judgments

Area of Science:

• Cognitive Psychology
• Neuroscience
• Decision Science

Background:

• Confidence is crucial for decision-making, but its origins are debated.
• Bayesian theories suggest confidence stems from sensory evidence uncertainty.
• Alternative theories propose heuristics based on stimulus properties.

Purpose of the Study:

• To investigate the basis of confidence in perceptual decision-making.
• To compare Bayesian models of confidence with heuristic-based models.
• To determine if confidence relies on internal uncertainty or external stimulus cues.

Main Methods:

• Developed ideal observer models for Bayesian and heuristic confidence.
• Conducted psychophysical experiments where participants reported stimulus orientation and confidence.
• Manipulated internal and external noise levels and stimulus orientations (cardinal vs. oblique).

Main Results:

• Confidence correlated with behavioral variability (precision) for a fixed orientation, supporting Bayesian predictions.
• However, observers were more confident in oblique orientations despite less precise estimates (oblique effect).
• This inverse pattern was explained by a heuristic strategy based on perceived stimulus noise.

Conclusions:

• Confidence is not solely computed from the degree of uncertainty in perceptual evidence.
• Visual cues, acting as heuristics, significantly influence confidence judgments.
• Perceived stimulus noise plays a key role in confidence, independent of internal evidence precision.