Fairness in Opinion Dynamics

TL;DR

The paper investigates fairness in opinion-dynamics prediction by examining how demographic and network-topology features influence misclassification by the CoDiNG model on the NetSense dataset. It deploys three classifier pipelines—survey-based, topology-based, and hybrid—to predict when CoDiNG will err, enabling a nuanced analysis of bias across minority subgroups and question types. The study identifies four key patterns of algorithmic bias, showing that no single modelling paradigm universally outperforms others; effectiveness is highly context-dependent on the minority group and the issue type. The results advocate context-aware, multifaceted modelling approaches to reduce bias and promote inclusive decision-making in social networks, with implications for the design of fairness-aware opinion-dynamics tools.

Abstract

Ways in which people's opinions change are, without a doubt, subject to a rich tapestry of differing influences. Factors that affect how one arrives at an opinion reflect how they have been shaped by their environment throughout their lives, education, material status, what belief systems are they subscribed to, and what socio-economic minorities are they a part of. This already complex system is further expanded by the ever-changing nature of one's social network. It is therefore no surprise that many models have a tendency to perform best for the majority of the population and discriminating those people who are members of various marginalized groups . This bias and the study of how to counter it are subject to a rapidly developing field of Fairness in Social Network Analysis (SNA). The focus of this work is to look into how a state-of-the-art model discriminates certain minority groups and whether it is possible to reliably predict for whom it will perform worse. Moreover, is such prediction possible based solely on one's demographic or topological features? To this end, the NetSense dataset, together with a state-of-the-art CoDiNG model for opinion prediction have been employed. Our work explores how three classifier models (Demography-Based, Topology-Based, and Hybrid) perform when assessing for whom this algorithm will provide inaccurate predictions. Finally, through a comprehensive analysis of these experimental results, we identify four key patterns of algorithmic bias. Our findings suggest that no single paradigm provides the best results and that there is a real need for context-aware strategies in fairness-oriented social network analysis. We conclude that a multi-faceted approach, incorporating both individual attributes and network structures, is essential for reducing algorithmic bias and promoting inclusive decision-making.

Figures (6)

• Figure 1: Matrix of correlations between membership in each of the corresponding minorities. Moreover, entries for the node degree, number of the question and how well the CoDiNG algorithm has performed for each of the minorities have been included.
• Figure 2: Diagrams of F1 scores for classifiers predicting CoDiNG model misclassifications. Results are shown for each of the six survey questions and across seven identified minority groups. Dashed lines indicate the F1 score for the general population for Survey-Based (Blue), Topology-Based (Red), and Hybrid (Green) approaches.
• Figure 3: Diagram of the workflow together with the structure of the data used for training of each of the classifier models.
• Figure 4: Graphs obtained via the t-SNE dimension reduction method illustrating distribution of the agents in relation to a set of their chosen characteristics.
• Figure 5: Representation of the student network with division into each of the minorities together with of a node degree and pagerank value color map - blue indicating a higher value and red - a lower one.