New method quantifies political partisanship across social media platforms

Researchers have developed a new text-based methodology to measure political partisanship across different social media platforms. This approach uses transformer-based sentence encoders to embed posts and cluster them by topic, with labels derived from aggregated media bias scores of cited news outlets. The system was applied to posts from Bluesky and Truth Social, offering the first cross-platform comparison of partisanship on these ideologically distinct platforms. The scores showed significant correlation with external media bias data and revealed partisan dynamics not apparent from platform identity alone. AI

IMPACT Enables more robust cross-platform analysis of political discourse and polarization.

RANK_REASON Academic paper detailing a new methodology for analyzing social media content. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• AllSides
• alphaXiv
• arXiv
• Bluesky
• CatalyzeX
• DagsHub
• Gotit.pub
• Hugging Face
• ScienceCast
• Truth Social
• X

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →