Media coverage of disasters frequently frames self-interested behavior in contrast to collective responsibility and coordinated response. This study aims to explore how such behavior is emotionally constructed in disaster-related media, using a carefully selected corpus of 12 text-centered news articles focusing on selfish behavior. The analysis combines transformer-based sentence-level emotion classification using the tweetnlp RoBERTa model, which predicts 11 emotion categories, with Latent Dirichlet Allocation topic modeling across single-sentence and three-sentence windows in a small purposively selected corpus. Emotion–topic relationships are quantified by weighting emotion probabilities by topic distributions and visualized using bar charts, network graphs, and heatmaps. The findings suggest that fear and disgust dominate portrayals of self-interested behavior, while anticipation appears in projections of harm and anger is linked to inequality and institutional accountability. Two discursive configurations emerge: Responsibility Across Individuals and Institutions, emphasizing public accountability and authority, and Collective Fear and Self-Protective Practices, reflecting affect-driven responses under uncertainty. Although negative emotions predominate, optimism appears conditionally, signaling coordination and recovery. Overall, disaster reporting constructs selfishness through integrated emotional–semantic patterns that position individual actions within broader social risk and collective responsibility.
Building similarity graph...
Analyzing shared references across papers
Loading...
Soyoung Kim
Seoul National University of Science and Technology
Christopher Stream
Suyeon Lee
Behavioral Sciences
University of Nevada, Las Vegas
Seoul National University of Science and Technology
Building similarity graph...
Analyzing shared references across papers
Loading...
Kim et al. (Tue,) studied this question.
synapsesocial.com/papers/69eb0bfa553a5433e34b56d9 — DOI: https://doi.org/10.3390/bs16040621