Biased Assimilation, Homophily and the Dynamics of Polarization (1209.5998v1)

Abstract: Are we as a society getting more polarized, and if so, why? We try to answer this question through a model of opinion formation. Empirical studies have shown that homophily results in polarization. However, we show that DeGroot's well-known model of opinion formation based on repeated averaging can never be polarizing, even if individuals are arbitrarily homophilous. We generalize DeGroot's model to account for a phenomenon well-known in social psychology as biased assimilation: when presented with mixed or inconclusive evidence on a complex issue, individuals draw undue support for their initial position thereby arriving at a more extreme opinion. We show that in a simple model of homophilous networks, our biased opinion formation process results in either polarization, persistent disagreement or consensus depending on how biased individuals are. In other words, homophily alone, without biased assimilation, is not sufficient to polarize society. Quite interestingly, biased assimilation also provides insight into the following related question: do internet based recommender algorithms that show us personalized content contribute to polarization? We make a connection between biased assimilation and the polarizing effects of some random-walk based recommender algorithms that are similar in spirit to some commonly used recommender algorithms.

• The paper introduces a generalized DeGroot model incorporating biased assimilation that shifts opinion formation from consensus to polarization.
• It shows that homophily alone is insufficient for polarization, with a bias threshold in blended interactions determining the outcome.
• The analysis reveals that recommender systems amplify pre-existing biases, thereby intensifying polarization and echo chamber effects in digital networks.

Analysis of "Biased Assimilation, Homophily and the Dynamics of Polarization"

The paper "Biased Assimilation, Homophily and the Dynamics of Polarization" proposes a model of opinion formation that extends the DeGroot model by integrating the psychological phenomenon of biased assimilation. This work addresses the observation that society is perceived as increasingly polarized and provides a mathematical framework to understand the contributing factors.

Overview

The authors begin by questioning the societal polarization trends and build their analysis on the empirical observation that like-minded interaction (homophily) is a driver for polarization. They distinguish their approach from DeGroot's model, a well-known opinion formation technique that relies on individuals adopting the weighted average of their neighbors' opinions. The DeGroot model, by construction, tends to facilitate consensus rather than polarization, a limitation the proposed model seeks to overcome.

Key Contributions

1. Generalized DeGroot Model: The proposed model introduces biased assimilation, where an individual's tendency to favor evidence supporting their initial belief is factored into opinion updates. This generalization allows the process to be polarizing under certain conditions.
2. Biased Assimilation: This social psychology phenomenon is modeled mathematically, providing insights into the opinion formation dynamics. The model demonstrates that biased assimilation leads to polarization, persistent disagreement, or consensus based on the bias parameter's threshold in the context of homophilous networks.
3. Polarization Metrics: The concept of Network Disagreement Index (NDI) is utilized to quantify the divergence among individuals' opinions in a network over time. The paper defines an opinion formation process as polarizing if the NDI is greater at equilibrium compared to initial conditions.
4. Role of Recommender Systems: The paper makes an important connection between biased assimilation and online information personalization. It analyzes three recommender algorithms—SimpleSALSA, SimplePPR, and SimpleICF—highlighting how these may inadvertently contribute to societal polarization by reinforcing pre-existing biases, particularly with biased individuals.

Findings

• Homophily and Polarization: A key result is that homophily alone does not suffice for polarization in the absence of biased assimilation. The degree of bias directly influences whether society becomes polarized.
• Individual's Bias and Dynamics: In settings where individuals are exposed to mixed evidence, their tendency for biased assimilation—modeled as a bias parameter—determines whether opinions become more extreme or moderate over time.
• Recommender Systems' Impact: The analysis reveals that personalization algorithms can be polarizing, reinforcing the existing preferences and possibly leading to echo chambers, especially when combined with individual biases.

Implications and Future Directions

The paper's insights have significant implications for both theoretical understanding and practical applications. Theoretically, it advances the knowledge of opinion dynamics under realistic social psychology considerations. Practically, it suggests caution in the deployment of personalized content algorithms that may inadvertently amplify societal divisions.

Future research may explore ways to design systems that mitigate the effects of biased assimilation in networked interactions, potentially paving the path for algorithms and interventions that promote consensus and minimize polarization. Additionally, human subject experiments could further validate the model's applicability to real-world dynamics.

In conclusion, this paper provides a robust mathematical framework to understand polarization, explaining how personal bias and network structures interact to influence societal opinion dynamics. This understanding is crucial in the digital age, where information dissemination rapidly evolves and increasingly influences public discourse.