What drives engagement with disinformation on social media? A Social Identity Approach

ACL ARR 2025 February Submission1494 Authors

Abstract: Research on user engagement with disinformation has expanded significantly in recent years, yet the mechanisms that make implausible narratives persuasive remain unclear. While emotion, source credibility, and information structure have been found to influence disinformation endorsement, this study argues that linguistic and collective identity plays a significant role in shaping engagement. It posits that sharing and interacting with disinformation—via comments, reactions, and reposts—functions as a discursive practice encoding both collective agency and cultural-linguistic identity. Using a dataset of 3,885 tweets in English (EN), French (FR), and Spanish (ES), the study examines whether language choice and call-to-action expressions (e.g., mobilization phrases) drive engagement. Statistical analysis, topic modelling, and lingistic analyses reveal notable cross-linguistic differences: Spanish-speaking users engage significantly more with directive language, while English and French users exhibit distinct, less predictable interaction patterns. These findings challenge the assumption that mobilization rhetoric universally boosts engagement and suggest that audience responsiveness to disinformation-based engagement is shaped by linguistic and cultural context. These results therefore underscore the need to consider linguistic and cultural variations in disinformation research as well as tailored communication strategies in disinformation mitigation and digital content dissemination.

Paper Type: Long

Research Area: Computational Social Science and Cultural Analytics

Research Area Keywords: misinformation detection and analysis,NLP tools for social analysis,quantitative analyses of news and/or social media,human behavior analysis

Contribution Types: Data analysis

Languages Studied: English, French, Spanish

Submission Number: 1494