Evolution of Public Opinion under Conformist and Manipulative Behaviors

In a prior work, we investigated how the outcome of a tracking poll can be influenced by two contradictory behaviors of participants, namely conformity and manipulation. A common characteristic of opinion polls is that a participant has only a few options to express her opinion, thus such polls do not capture the opinion dynamics of social networks where the stated opinions are deemed to belong to a continuum. Inspired by this argument, we address in this paper the conformist and manipulative behaviors of individuals in a realistic opinion network. Such an opinion network can indeed be viewed as a tracking poll where a participant casts her vote by selecting from a continuum rather than a discrete set. To carry out the analysis, we model an opinion network as a game among the individuals and define their payoff functions appropriately in the sense that optimizing them captures the conformist and manipulative behaviors. We then investigate the dynamics of the game and its behavior as time grows.