Interdependent Diffusion: The social contagion of interacting beliefs.

Social contagion is a well-studied phenomenon in which people adopt beliefs that they are exposed to by their neighbors, and then pass those beliefs along to others. Research (and daily life) shows that people prefer to adopt beliefs that are consistent with those they already hold. However, scholars do not often account for interactions between beliefs in their models of social contagion. Instead, they assume that beliefs spread independently of one another. Is this a harmless simplification? Or does omitting interdependence between beliefs suppress important dynamics, and change the outcome of social contagion? This paper performs a head-to-head comparison between independent and interdependent diffusion. Simulations identify two social processes that emerge when diffusants interact, and predict that as a result of interdependent diffusion, worldviews will emerge that are unconstrained by external truth, and polarization will develop in homogenous populations. A controlled laboratory experiment confirms these predictions with 2400 participants in 120 artificial social networks. I conclude that the assumption of independence between diffusants is not as universally appropriate as its ubiquity would suggest. Instead, interdependence between diffusants is likely to be both common and consequential.