Public Opinions And The Construction Of Self- And Other- Identity In Anti-Immigration Discourse Towards Ukrainians

Purpose One of the consequences of digital transformation is social polarisation, which has negative effects on discourse polarisation and radicalization, particularly targeting immigrants. These processes are also present in the case of Poland and a flood of Ukrainian immigrants. As a result, scholars seek novel tools and methodologies to grasp the scale of discourse radicalization and society polarization. The content online is expressed particularly freely on social media, which house distributed, decentralized information created in a bottom-up fashion (Yarchi et al., 2020). They have become a substitute for institutionalized public debate (Baden et al., 2025) and often allow the radical content to spread virally. Therefore, the purpose of this study is to demonstrate how this radical, polarized discourse can be analyzed. The onset of the Russian invasion of Ukraine in February 2022 precipitated massive emigration of Ukrainians to the neighbouring countries, with Poland emerging as the host of the most significant number of immigrants from the conflict-ridden nation. The unprecedented aid and hospitality extended to Ukrainians by Poles were lauded across European media. This initial positive sentiment and welcoming reception, however, have gradually shifted towards a growing dissatisfaction and even hostility towards Ukrainians over the ensuing three years, as evidenced on social media. The purpose of this research is thus to analyse the social climate of Poles discussing Ukrainian immigrants to Poland based on comments expressed on social media (YouTube) as a reaction to an interview with a Ukrainian leader of the Ukrainian diaspora in Poland. Design/methodology/approach This study aims to analyze comments posted in February 2025 on Polish YouTube regarding the situation and status of Ukrainians in Poland and the perception of Ukrainians by Poles following the widely publicized, controversial statements of Natalia Panchenko, the leader of the Ukrainian diaspora in Poland. The leader’s provocative opinions on Poles and the status of Ukrainians in Poland, presented during her interview on the Ukrainian TV Channel 5, sparked extensive commentary by Polish journalists, politicians, and social media users. Even though she later denied expressing these statements, attributing them to Russian manipulation and fake news in several interviews published on YouTube, her words nonetheless provoked a significant backlash, which manifested in an outpouring of grievances, expressing disapproval or even condemnation of some Ukrainians, as well as a severe criticism of the immigration policy adopted by the Polish government. Following the interview on the Ukrainian TV, she gave another interview on the Super Express YouTube channel (affiliated with one of Poland’s private television stations), where she attempted to clarify her views on Polish-Ukrainian relations and to fend off the severe criticism she had received after her controversial remarks on Ukrainian television. The critical comments analyzed in this study were posted by YouTube users in response to her interview with the Polish journalist. The comments were collected over a one-month period, resulting in a dataset of approximately 7,000 entries. A randomly selected sample of ca. 1200 sample of comments was extracted from the original set and analyzed both qualitatively and quantitatively. For qualitative analysis, Atlas.ti was used to enable research in line with the Grounded Theory (Strauss and Coblin, 1994). A carefully selected sample of the collected comments was next annotated by four expert annotators and a super-annotator after a series of training sessions (with inter-rater agreement oscillating around 0.8). The annotation-based qualitative study is followed by a computerized automatic examination of the data, particularly sentiment analysis, topic modelling, and trends. The tools used in this part of the study stem from corpus linguistics methodology and the latest, cutting-edge neural network-based tools (the so-called deep learning methods) known as Transformers (BERT; Devlin et al., 2019). Theoretically, the study is framed within theories of identity, Critical Discourse Analysis (e.g., van Leeuwen, 2008), and theories and models of evaluative language. The analysis had the following research questions: RQ (1) What is the general perception of Ukrainian residents in Poland by Poles 3 years after the war onset, according to the comments under inspection? RQ (2) How is the Ukrainian identity constructed by the Polish YT users? RQ (3) How is the Polish identity constructed by the Polish YT users in the context of the Ukrainian refugees’ flood? Findings The study has uncovered interesting tendencies and trends in how Polish social media users evaluate immigrants from the war-ridden Ukraine over time, and which factors are crucial in the user-generated, radically negative opinions. Our research also shows the effectiveness of computational methods when applied to media discourse in public opinion research. Originality/value The originality of our study is twofold. It resides in merging two scholarly approaches and traditions: media studies and linguistics. On the other hand, the use of computerized automatic data retrieval and analysis, particularly neural networks represented by deep learning, is a novel approach in media studies. The categories of description that researchers typically resort to in such studies are various types of discourse strategies, well-known categories in linguistic scholarship. In our study, however, we propose novel categories which combine methodological traditions originating in linguistic and media studies.