How Combining Terrorism, Muslim, and Refugee Topics Drives Emotional Tone in Online News: A Six-Country Cross-Cultural Sentiment Analysis

Chung-Hong Chan, Hartmut Wessler, Eike Mark Rinke, Kasper Welbers, Wouter van Atteveldt, Scott Althaus


This study looks into how the combination of Islam, refugees, and terrorism topics leads to text-internal changes in the emotional tone of news articles and how these vary across countries and media outlets. Using a multilingual human-validated sentiment analysis, we compare fear and pity in more than 560,000 articles from the most important online news sources in six countries (U.S., Australia, Germany, Switzerland, Turkey, and Lebanon). We observe that fear and pity work antagonistically—that is, the more articles in a particular topical category contain fear, the less pity they will feature. The coverage of refugees without mentioning terrorists and Muslims/Islam featured the lowest fear and highest pity levels of all topical categories studied here. However, when refugees were covered in combination with terrorism and/or Islam, fear increased and pity decreased in Christian-majority countries, whereas no such pattern appeared in Muslim-majority countries (Lebanon, Turkey). Variations in emotions are generally driven more by country-level differences than by the political alignment of individual outlets.


terrorism, refugee, Muslim, sentiment analysis, multicultural analysis

Full Text: