Social Influence as Reinforcement Learning
Technical Report,15 Oct 2013,14 Oct 2015
Stanford University Palo Alto United States
Pagination or Media Count:
This project examined a reinforcement learning model of conformity and social influence. Under this model, individuals value consensusor agreement with groupsand conformity represents a strategy for maximizing consensus with a group. Under this project, we examined three predictions that emerge from a reinforcement model. First, people effects of social influence should generalize, such that people should agree with groups not only about particular exemplars e.g., foods or paintings, but also about rules governing group behavior. Second, people should be most inclined to conform when they experience social hunger, such as isolation or rejection. Third, people should act more prosocially towards individuals who conformed with them i.e., offer consensus. Predictions 1 and 3, but not 2, were borne out by our research.