Unrealistic optimism in advice taking: A computational account.

Area of Science:

• Cognitive Psychology
• Behavioral Economics
• Decision Science

Background:

• Expert advisors frequently make inaccurate predictions, yet their advice is often followed.
• Unrealistic optimism in advice-taking is a prevalent phenomenon.
• Existing models do not fully explain the cognitive mechanisms behind this bias.

Purpose of the Study:

• To provide a computational account of unrealistic optimism in advice-taking.
• To investigate the role of initial expectations and information processing in advice-taking.
• To explore the effectiveness of crowdsourced ratings in mitigating this bias.

Main Methods:

• Three studies involving participants observing advisors predict stock performance.
• Advisors varied in accuracy (above, at, or below chance).
• Development and testing of a computational model of advice-taking.
• Exploration of crowdsourced star ratings as a debiasing strategy.

Main Results:

• Participants consistently overestimated advisor accuracy and over-relied on their predictions.
• Decision-making aligned with a computational model assuming optimistic initial expectations and biased information incorporation.
• Manipulating initial expectations successfully altered optimism bias and advice-taking.
• Crowdsourced ratings amplified, rather than corrected, unrealistic optimism.

Conclusions:

• A computational model explains inflated perceptions of expertise through optimistic expectations and biased learning.
• Initial expectations are critical in shaping optimism bias and advice-taking behavior.
• Crowdsourced ratings may inadvertently exacerbate unrealistic optimism.
• The findings offer insights into cognitive processes underlying expertise overestimation and potential extensions to other domains.