Daniel Mirza

(Université de Tours, CEPII)

 

 

Résumé: This article is interested in media attitudes towards migration, in times of terrorism. We first study changes in media reportings on migration, in terms of salience and tone. Second, we assess by how much they matter in voting outcomes. We use natural language processing (NLP) methods to analyze French news articles about immigration before and after successive waves of terrorist attacks. We find that during 3 to 8 weeks after a significant attack, a shift of around 30-50% in both the intensity and net negative tone of reporting about migration is clearly observed. These effects happen to be robust across French media political orientations. We also find that during a nationwide election in France that happens to take place, incidently, 3 weeks after the Bataclan November 13th attack, votes in favor of the main anti-migrant party were higher in municipalities which were more exposed to media adopting negative tones.