Opinion Formation by Social Influence: From Experiments to Modeling.

Area of Science:

• Social Sciences
• Computational Social Science
• Behavioral Economics

Background:

• Collective behavior prediction is crucial in social sciences.
• Opinion formation models often lack empirical validation.
• Existing models draw parallels with physical system dynamics.

Purpose of the Study:

• Quantify opinion and confidence changes under social influence.
• Investigate factors modulating opinion shifts.
• Empirically inform agent-based models of opinion dynamics.

Main Methods:

• Experimental design with 85 subjects answering 20 questions.
• Measurement of opinion and confidence before and after social influence.
• Parameter estimation for a dynamical agent-based model.

Main Results:

• Confidence changes are largely independent of other factors.
• Opinion changes are significantly influenced by initial confidence.
• Opinion changes are not affected by the initial difference from a reference opinion.
• Opinion dynamics occur over longer timescales than confidence changes.

Conclusions:

• Experimental data provide empirical grounding for opinion formation models.
• Agent-based models can simulate consensus and the impact of opinion leaders.
• Findings highlight the interplay between individual confidence and collective opinion shifts.