Social Learning and Mental-Model Revision

Abstract

In Progress. We experimentally study how people revise their perceived causal models in environments where they interact with others. We consider a setting where individuals first form mental models in isolation based on observational data. We then examine when and how exposure to others’ predictions—cast by people who may have extracted differing mental models—influences the model-updating process. We find that peer information does surprisingly little: returning to the original data proves far more effective.