How Did the Refugees Become a Social Problem? 10 Years of Refugee Narratives in Polish Public Television with Latent Dirichlet Allocation (LDA)

A decade ago, in 2013, refugees were not really present in the Polish television landscape. Now, they appear almost every day, accompanying many other news stories. In this article, I identify the narratives surrounding refugees, examine how they may have been used to serve different agendas, and trace the long-term slow process of framing refugees as a social problem. I propose a mixed methods framework, combining latent Dirichlet allocation (LDA), a probabilistic topic modeling algorithm, to facilitate the analysis of large datasets of television recordings, and critical discourse analysis (CDA), to capture a deeper understanding of the changes. In the research, I explore how refugees have been represented in “Wiadomości”, the main evening news program of Polish public television, over the past decade. The corpus consists of almost 55 hours of audiovisual material, and 1064 news segments broadcast between 2013 and 2023, containing the term “refugee”. With latent Dirichlet allocation, I identify eight key thematic frames that emerge in the discussion of refugees: (1) Security Concerns, (2) Polish Parliamentary Election, (3) The Russia-Ukraine War, (4) Conflicts in the Middle East, (5) Polish Support for Ukrainian Refugees, (6) Impact of War on Families, (7) Polish Border Crisis, (8) European Union Policy on Refugees. Then, I use critical discourse analysis to identify the key geopolitical events underlying these narratives and determine which groups of refugees these narratives refer to. My findings offer insights into the media’s role in shaping public opinion and provide frameworks for future use of computational methods for media and discourse analysis.