How political symbols spread in online social networks: Using agent-based models to replicate the complex contagion of the yellow ribbon in Twitter

Abstract

This paper analyzes the diffusion of the yellow ribbon in Twitter, a political symbol that represents the demand for the release of Catalan prisoners. We gathered data on potential users of the symbol in Twitter (users that publicly backed the cause), including their social network of friendships, and built an agent-based simulation to replicate the diffusion of the symbol in a digital twin version of the observed network. Our hypothesis was that complex contagion is the best explanation of the observed statistical relation between the proportion of adopting neighbors and the probability of adoption. Results show that the complex contagion model outperforms the simple contagion model and generates a better fit between the observed and the simulated pattern when the typical conditions of a complex contagion process are added to the baseline model, that is, when agents are affected by their reference group behavior rather than by the most influential nodes of the network, and when we identify a peripherical and densely connected network community and trigger the process from there. These results widen the set of behaviors whose diffusion can be explained as complex contagion to include adoption in low-risk/low-cost behaviors among people who would usually not resist adoption.