This was a paper that I submitted for my “Simulating Society” class for the MSc in Oxford. The class taught agent-based modelling using NetLogo. Final grade: 75
Summary:
We live in an age where people are becoming more and more divided, be it in political ideologies, religious beliefs, social status, or core values. One repercussion of societal polarization is that it may reinforce “echo chambers,” where people are only exposed to information depending on their ideology, that defines the cluster to which they belong.
In this paper, I use a diffusion agent-based model to answer the question, “do echo chambers exist in polarized societies?” I present a simple two-cluster case of political polarization, modeled after the established and well-documented partisanship (Democrat-Republican ideological divide) in the United States.
The results of my study suggest that echo chambers may exist in polarized societies, but it largely depends on the network architecture and threshold assumptions within the population.
It showed some indication that information remains largely within their clusters of origination, and may suggest the presence of echo chambers. It may be noted however that this behaviour may depend on the network architecture, as it can be observed that various calibrations on the threshold distributions give inconclusive results on the diffusion behaviour.The study may be more useful and informative to policy if we can further validate and extend the model to accommodate other network structures and other parameters pertaining to information quality and the population characteristics, as these can confound agent behaviour. If we can achieve robust results from the model given its various extensions, then this can help justify the policy stance on echo chambers and misinformation spread in online societies. Lastly, the agent-based model could also be an important first step in simulating other forms of polarization and inequality, which could be sensitive and costly on social welfare and economic resources if addressed improperly. These simulations could provide policymakers with scenarios that can aid them in their crucial decisions, in operationalizing interventions and reforms.
The goal of this paper is observational rather than causal. It only aims to demonstrate counterfactuals regarding the flow of information in divided societies and does not establish a causal claim on how these echo chambers may contribute to further polarization.
Full version of the paper: Email me on LinkedIn!
