A Mixed-Methods Framework for Sentiment Analysis in News Discourse: Integrating Computational Lexicons with Appraisal Theory

# OSF File Upload Description: Emotional Expression Analysis Framework for Structured Journalism Research This project presents a comprehensive mixed-methods framework specifically designed for analyzing emotional expression within the structured journalism register. The methodology addresses the unique challenges of studying affect in professional news discourse by integrating computational lexicons (NRC EmoLex and VAD models) with Appraisal Theory-guided qualitative analysis. The framework introduces innovative analytical techniques including arousal-dominance integration and emotional displacement tracking, providing researchers with replicable protocols for sentiment analysis in journalistic writing. Key affect expressions identified include trust (establishing source credibility and institutional authority), anticipation (generating narrative momentum and framing future developments), fear (amplifying risk perception and justifying policy urgency), and anger (signaling moral boundaries and attributing responsibility). Each emotional category serves distinct journalistic functions, from accountability reporting to solution journalism. The arousal-valence dynamics reveal critical impacts on reader engagement: high-arousal negative emotions maximize attention and urgency, while low-arousal positive emotions stabilize credibility and enable complex information processing. Emotional displacement effects—temporal, metaphorical, and attributional—demonstrate how journalists recontextualize affect to maintain objectivity while enabling evaluative framing. This framework offers theoretical tools for cross-cultural communication studies and enables systematic comparison of emotional strategies across publications, topics, and cultural contexts.