Thematic and sentiment analysis on X: a multi-method approach to examine the refugee crisis in Türkiye / Análisis temático y de sentimiento en X: un enfoque multimétodo para examinar la crisis de refugiados en Turquía

Abstract

In this research, we present findings from two studies analysing how the major refugee groups in Türkiye, i.e., Syrians, Afghans and Ukrainians, are represented on social media. In Study 1, we performed a sentiment analysis to explore how three groups are portrayed on social media using the BERT-based Turkish Sentiment Model. The results revealed that Afghans were the targets of the most prominent negative sentiment, followed by Syrians, whereas the posts about Ukrainians exhibited the highest positive sentiment. Secondly, to get a more nuanced view of the differential representation of refugee groups, we conducted a thematic content analysis on a randomly drawn subset of 300 posts from the datasets used in Study 1. The overall representation of refugees was negative and aligned with the Integrated Threat Theory. Moreover, negative stereotypes, realistic, symbolic and demographic threats, and intergroup anxiety were almost exclusively about Syrians and Afghans, except for negative stereotypes about Ukrainian women. In contrast, the positive stereotypes were predominantly about Ukrainian refugees. Also, Afghan refugees were represented as more threatening than Syrians.