Inference from social evaluation

Area of Science:

• Cognitive Psychology
• Social Psychology
• Computational Social Science

Background:

• Humans infer hidden causes from physical evidence.
• Social evaluations, such as praise and blame, are common in daily interactions.
• These evaluations offer clues to understanding events and actions.

Purpose of the Study:

• Investigate how people use social evaluations as evidence to infer hidden causes.
• Determine the systematic inferences drawn from praise and blame.
• Model the generative process of praise and blame judgments.

Main Methods:

• Conducted three experiments presenting scenarios with social evaluations.
• Participants inferred situational factors, actions, capabilities, and social roles.
• Developed computational models for praise/blame generation and used Bayesian inference.

Main Results:

• People systematically infer situational factors, actions, capabilities, and social roles from social evaluations.
• Computational models accurately predicted participant inferences.
• Bayesian inversion of generative models explained observed inference patterns.

Conclusions:

• Social evaluations are a rich source of information for causal inference.
• Computational models can capture the reasoning behind praise and blame.
• Bayesian inference provides a framework for understanding how social feedback informs causal judgments.