Moral Outrage Shapes Commitments Beyond Attention: Multimodal Moral Emotions on YouTube in Korea and the US

TL;DR

This study investigates how multimodal moral emotions in YouTube news content shape audience engagement across Korea and the United States. It introduces a cross-lingual multimodal classifier trained on Korean and English thumbnail–title pairs to detect four moral emotion categories, validated on ~1,276 Korean and ~900 English samples, and applied to nearly 400k videos. Through negative binomial regression, the authors show that other-condemning rhetoric increases views, likes, and especially comments in both countries, indicating that moral outrage boosts commitment beyond mere attention. The work highlights polarization risks associated with moral outrage, provides cross-cultural validation, and contributes reproducible tools by releasing the classifiers for both languages. This has implications for understanding the role of emotion in the attention economy and for designing interventions to mitigate polarization on digital platforms.

Abstract

Understanding how media rhetoric shapes audience engagement is crucial in the attention economy. This study examines how moral emotional framing by mainstream news channels on YouTube influences user behavior across Korea and the United States. To capture the platform's multimodal nature, combining thumbnail images and video titles, we develop a multimodal moral emotion classifier by fine tuning a vision language model. The model is trained on human annotated multimodal datasets in both languages and applied to approximately 400,000 videos from major news outlets. We analyze engagement levels including views, likes, and comments, representing increasing degrees of commitment. The results show that other condemning rhetoric expressions of moral outrage that criticize others morally consistently increase all forms of engagement across cultures, with effects ranging from passive viewing to active commenting. These findings suggest that moral outrage is a particularly effective emotional strategy, attracting not only attention but also active participation. We discuss concerns about the potential misuse of other condemning rhetoric, as such practices may deepen polarization by reinforcing in group and out group divisions. To facilitate future research and ensure reproducibility, we publicly release our Korean and English multimodal moral emotion classifiers.

Figures (9)

• Figure 1: Overview of the research framework. Multimodal YouTube data and human-labeled moral emotions are used to examine the relationship between predicted moral emotions and user engagement (views, likes, and comments).
• Figure 2: Topic distributions derived from BERTopic for Korean and English YouTube news data (January–December 2024).
• Figure 3: Distribution of primary moral emotions in the Korean and U.S. datasets, categorized by the highest predicted probability from the fine-tuned models.
• Figure 4: Predicted engagement by other-condemning emotion probability. Top panels show fitted counts from the negative binomial models, and bottom panels show relative engagement (IRR) for Korea and the United States. As moral outrage intensity increases, engagement progressively rises from views to comments. Shaded areas denote 95% confidence intervals.
• Figure 5: Mean and median daily view growth rates (log-scaled) across 1,703 videos over 60 days after release.