A Text Mining Approach to Determinants of Attitude Towards Syrian Immigration in the Turkish Twittersphere

Abstract

This study uses novel deep learning-based language models to extract meaningful information from vast chunks of textual data from Twitter on the competing narratives of the recent Syrian immigration to Turkey. Our analysis identifies five main topics in the framing of Syrian immigration in Turkish Twittersphere. In this paper, we demonstrate correlational links between the timing of landmark events and change in the percent share of trends in those topics across time. We highlight two important observations: (a) Social benefit demands of natives on Twitter rose sharply with the COVID-19 pandemic, leading to ever more widespread sentiments of welfare chauvinism and (b) Patriotic feelings and the implementation of an interventionist foreign policy agenda in the immigrants’ country of origin created a relatively tolerant yet patronizing attitude towards migrants. As the COVID-19 pandemic and immigration frequently occupy the center stage in politics of immigrant-hosting societies, our research has international appeal beyond its specific geographical context.