How overconfidence bias influences suboptimality in perceptual decision making

Area of Science:

• Cognitive Psychology
• Neuroscience
• Decision Science

Background:

• Human observers often combine sensory information and prior knowledge suboptimally in perceptual decision-making.
• A 'conservative decision bias' occurs when individuals do not sufficiently adjust their decisions based on prior probabilities.
• Overconfidence in interpreting sensory information is a proposed explanation for this bias.

Purpose of the Study:

• To formally derive and test a model linking overconfidence and conservative decision bias.
• To investigate how overconfidence influences the integration of sensory information and prior probabilities in a visual discrimination task.
• To examine the relationship between overconfidence bias and conservative decision bias in human observers.

Main Methods:

• Developed a formal model linking overconfidence and conservative decision bias.
• Conducted a visual discrimination experiment manipulating prior stimulus probabilities.
• Measured decision confidence and criterion placement in 69 participants across two sessions.

Main Results:

• Confirmed the presence of both overconfidence bias and conservative decision bias in the data.
• Found no significant link between overconfidence bias and conservative decision bias as predicted by the model.
• Observed that overconfident participants reduced sensory information processing when prior probabilities were known.

Conclusions:

• The study challenges the direct link between overconfidence and conservative decision bias in explaining suboptimal decisions.
• Findings suggest overconfidence may lead to reduced effort in sensory processing when prior information is available.
• Offers new insights into the mechanisms underlying overconfidence and its impact on decision-making strategies.