Considering pivotal democratic elections like those of 2020 and 2024, understanding the role of news media in shaping political polarization is critical: media influence profoundly impacts democratic processes. Using agenda-setting theory and ingroup/outgroup dynamics from social identity theory as frameworks, this study examines the framing of Trump’s rally speeches and their coverage by liberal and conservative news sources. Study 1 uses a BERT-based LLM and the probabilistic LDA model to analyze the relationship between topics in liberal versus conservative media coverage, examining how media outlets' political alignments influence the alignment and selective emphasis of topics in relation to the topics present in Trump's rally speeches. This led to further inductive analysis of topics using a neural-network based topic modeling pipeline. In Study 2, the political alignment of media outlets in connection to the sentiment of news articles was examined using another BERT-based LLM, while also considering the sentiment expressed in Trump’s rally speeches. Our findings revealed that despite significant differences in topic alignment throughout sources, both liberal and conservative sources exhibited predominantly negative sentiments. Further analysis revealed that each outlet selectively criticizes the other to express ideological bias, rather than merely advocating its own views. This conclusion exemplifies a pervasive issue in contemporary media: the emphasis on adversarial coverage over constructive discourse. Such coverage inundates the public with divisive narratives, heightening political polarization and weakening trust in media institutions, thus impairing informed decision-making. Our research demonstrates the need for increased media literacy to safeguard against a misinformed electorate, reduce polarization, and strengthen democratic processes.

Keywords: topic modeling, sentiment analysis, agenda-setting theory, ingroup-favoritism, outgroup-discrimination, social identity theory, media literacy, polarization