The power of the unexpected: Prediction errors enhance stereotype-based learning

Area of Science:

• Social Psychology
• Cognitive Science
• Computational Neuroscience

Background:

• Stereotyping is common in social cognition.
• Little is known about how group beliefs affect learning about individuals.

Purpose of the Study:

• Investigate the impact of gender stereotypes on instrumental learning.
• Explore how beliefs influence the acquisition of person knowledge.

Main Methods:

• Probabilistic selection task.
• Computational modeling using Reinforcement Learning Drift Diffusion Model analysis.

Main Results:

• Observed an expectancy-based preference for stereotype-consistent responses.
• Found faster learning rates for counter-stereotypic individuals due to unexpected information.
• Both positive and negative prediction errors were learned faster for counter-stereotypic individuals.

Conclusions:

• Findings align with predictive accounts of social perception.
• Suggests that unexpected information can accelerate learning about individuals.
• Provides insights into reducing stereotyping in social cognition.