Crowd control: Reducing individual estimation bias by sharing biased social information.

Area of Science:

• Cognitive psychology
• Behavioral economics
• Social decision-making

Background:

• Cognitive biases are prevalent in humans and animals, sometimes amplified by social interactions.
• A common bias is the tendency to underestimate large quantities, which persists even with random social information sharing.

Purpose of the Study:

• To investigate if restructuring social information sharing can reduce the underestimation bias.
• To determine how the number of shared estimates impacts social influence and accuracy.
• To explore the mechanisms behind integrating multiple social estimates.

Main Methods:

• Estimation experiments involving participants estimating quantities before and after receiving group estimates.
• Testing a novel social information sharing method exploiting the value-underestimation relationship.
• Computational modeling to analyze integration mechanisms and social influence.

Main Results:

• Restructured social interactions successfully countered the underestimation bias.
• Sharing more than one estimate significantly reduced the underestimation bias.
• Observed tendencies include herding, trusting larger estimates more, and reduced influence from disparate information.

Conclusions:

• Knowledge of cognitive biases can be leveraged to mitigate their negative impacts and enhance judgment accuracy.
• This approach offers a method for combating biases that threaten collective systems.
• Understanding social information integration is key to improving decision-making accuracy.