Large language models for newspaper sentiment analysis during COVID-19: The Guardian
Rohitash Chandra, Baicheng Zhu, Qingying Fang, Eka Shinjikashvili
Abstract
During the COVID-19 pandemic, news media coverage encompassed topics such as viral transmission, allocation of medical resources, and government response measures. Studies have analysed sentiment on social media platforms during COVID-19 to understand the public’s response to the rising cases and the government strategies implemented to control the spread of the virus. Sentiment analysis can enable an understanding of the dynamics in the social and psychological well-being of a group during the pandemic. Apart from social media, newspapers have played a vital role in disseminating information, including information from the government, experts, and the public about various topics. A study of sentiment analysis of newspaper sources during COVID-19 for selected countries can give an overview of how the media covered the pandemic. In this study, we selected The Guardian newspaper and conducted a sentiment analysis during various stages of COVID-19 including initial transmission, lockdowns and vaccination. We employed novel large language models (LLMs) and refined them with expert-labelled sentiment analysis data. We also provide an analysis of sentiments experienced pre-pandemic for comparison. The results indicate that during the early pandemic stages, public sentiment prioritised urgent crisis response, later shifting focus to addressing the impact on health and the economy. In comparison with related studies about social media sentiment analyses, we found a discrepancy between The Guardian, with a dominance of negative sentiments (sad, annoyed, anxious and denial), suggesting that social media offered a more diverse emotional expression during the pandemic. We found a grim narrative in The Guardian with overall dominance of negative sentiments, pre- and during COVID-19 across news sections including Australia, UK, World News, and Opinion. • We select The Guardian newspaper and provide a sentiment analysis during various stages of COVID-19. • We employ large language models (LLMs) and refine them with expert labelled sentiment analysis data. • The results indicate that during the early pandemic stages, public sentiment prioritised urgent crisis response. • In comparison with social media , we found dominance of negative sentiments in The Guardian.