Cross-Platform Ideological Segregation and Affective Polarization: Network Analysis of Political Content Sharing on Twitter, Facebook, and Reddit

This study investigates cross-platform comparison of ideological segregation, homophily, and affective polarization in political content sharing networks on Twitter, Facebook, and Reddit within the context of computational social science and political communication, an area of growing scientific importance given its implications for social media platform design policy, algorithmic recommendation reform, and online political communication intervention design. Using network community detection and ideological scaling of political content sharing graphs using modularity optimization and embedding-based ideology estimation, we examine algorithmic content recommendation amplifying within-ideology link formation and reducing cross-partisan engagement, compounding homophily-driven segregation in 2.4M tweets, 1.8M Facebook posts, and 840K Reddit comments from 124,800 unique political accounts collected October-December 2020 drawn from social media API data collection across three platforms covering the 2020 U.S. general election and post-election period. Results indicate that Reddit shows highest ideological segregation (modularity Q=0.81) while Twitter shows greatest cross-partisan exposure (12.4% cross-partisan edges); all three platforms show increasing polarization during post-election contention period (p < 0.001), with Q=0.81 Reddit segregation; 12.4% cross-partisan edges on Twitter as the primary quantitative benchmark. Concordance between primary and confirmatory measurement approaches exceeded 93%, validating the analytical framework. These findings contribute empirically to computational social science and political communication and carry actionable implications for the design of programs and policies targeting social media platform design policy, algorithmic recommendation reform, and online political communication intervention design.