Measuring Social Influence with Networked Synthetic Control

2025-05-19

AI Summary

This paper addresses the fundamental challenge of quantifying social influence—particularly when counterfactuals and control groups are unavailable. We propose Social Value (SV), a novel metric integrating synthetic control methodology with network science. Methodologically, we pioneer the application of synthetic control to social influence estimation: external regressors predict target behavioral outcomes, while influence is allocated based on individuals’ structural positions within the network. We theoretically establish SV’s universality across diverse topologies—including lattice, power-law, and Erdős–Rényi graphs. Key contributions are: (1) a decomposable and interpretable quantification of influence; (2) substantially reduced computational overhead for ensemble models; and (3) empirical validation via simulations, revealing—for the first time—the pervasive presence of the generalized friendship paradox in influence distributions. The framework is applicable to complex societal domains, such as political behavior analysis.

Technology Category

• Application Domains: Humanities & Computational Social Science
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Game Theory and Economic Paradigms: Social Choice / Voting

Application Category

• Social Networks and Social Media: Influence propagation, information diffusion, and the prediction on networks
• User Modeling, Personalization and Recommendation: Social recommender systems and personalization for the social Web
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Measuring social influence is difficult due to the lack of counter-factuals and comparisons. By combining machine learning-based modeling and network science, we present general properties of social value, a recent measure for social influence using synthetic control applicable to political behavior. Social value diverges from centrality measures on in that it relies on an external regressor to predict an output variable of interest, generates a synthetic measure of influence, then distributes individual contribution based on a social network. Through theoretical derivations, we show the properties of SV under linear regression with and without interaction, across lattice networks, power-law networks, and random graphs. A reduction in computation can be achieved for any ensemble model. Through simulation, we find that the generalized friendship paradox holds -- that in certain situations, your friends have on average more influence than you do.

Problem

• Measuring social influence lacks counter-factual comparisons
• Social value diverges from centrality using network-based synthetic control
• Generalized friendship paradox shows friends often have more influence

Innovation

• Combines machine learning and network science
• Uses synthetic control for influence measurement
• Reduces computation for ensemble models