Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Large language models for newspaper sentiment analysis during COVID-19: The Guardian

Rohitash Chandra, Baicheng Zhu, Qingying Fang, Eka Shinjikashvili

TL;DR

The study analyzes The Guardian's COVID-19 coverage to map sentiment dynamics using large language models refined with the SenWave dataset. By fine-tuning BERT and RoBERTa on multi-label sentiments and applying N-gram analyses, the work reveals a dominant negative framing in Guardian articles, with early pandemic focus on crisis response shifting toward health and economic concerns. RoBERTa generally outperforms BERT on this task, and the authors exclude the 'official report' sentiment to better reveal nuanced emotional patterns across Guardian sections and regions (Australia, UK, World). The work highlights differences between traditional media sentiment and social media, offers a detailed, reproducible framework (data, models, visualizations), and provides open-source code and data for further exploration and cross-outlet comparisons.

Abstract

During the COVID-19 pandemic, the news media coverage encompassed a wide range of topics that includes viral transmission, allocation of medical resources, and government response measures. There have been studies on sentiment analysis of social media platforms during COVID-19 to understand the public response given the rise of cases and government strategies implemented to control the spread of the virus. Sentiment analysis can provide a better understanding of changes in societal opinions and emotional trends during the pandemic. Apart from social media, newspapers have played a vital role in the dissemination of information, including information from the government, experts, and also 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 select The Guardian newspaper and provide a sentiment analysis during various stages of COVID-19 that includes initial transmission, lockdowns and vaccination. We employ novel large language models (LLMs) and refine 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 dominance of negative sentiments (sad, annoyed, anxious and denial), suggesting that social media offers a more diversified emotional reflection. 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

Large language models for newspaper sentiment analysis during COVID-19: The Guardian

TL;DR

The study analyzes The Guardian's COVID-19 coverage to map sentiment dynamics using large language models refined with the SenWave dataset. By fine-tuning BERT and RoBERTa on multi-label sentiments and applying N-gram analyses, the work reveals a dominant negative framing in Guardian articles, with early pandemic focus on crisis response shifting toward health and economic concerns. RoBERTa generally outperforms BERT on this task, and the authors exclude the 'official report' sentiment to better reveal nuanced emotional patterns across Guardian sections and regions (Australia, UK, World). The work highlights differences between traditional media sentiment and social media, offers a detailed, reproducible framework (data, models, visualizations), and provides open-source code and data for further exploration and cross-outlet comparisons.

Abstract

During the COVID-19 pandemic, the news media coverage encompassed a wide range of topics that includes viral transmission, allocation of medical resources, and government response measures. There have been studies on sentiment analysis of social media platforms during COVID-19 to understand the public response given the rise of cases and government strategies implemented to control the spread of the virus. Sentiment analysis can provide a better understanding of changes in societal opinions and emotional trends during the pandemic. Apart from social media, newspapers have played a vital role in the dissemination of information, including information from the government, experts, and also 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 select The Guardian newspaper and provide a sentiment analysis during various stages of COVID-19 that includes initial transmission, lockdowns and vaccination. We employ novel large language models (LLMs) and refine 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 dominance of negative sentiments (sad, annoyed, anxious and denial), suggesting that social media offers a more diversified emotional reflection. 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
Paper Structure (26 sections, 21 figures, 7 tables)

This paper contains 26 sections, 21 figures, 7 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Recurrent Neural Networks
  4. Transformer model
  5. BERT and RoBERTa models
  6. Methodology
  7. Data
  8. SenWave sentiment dataset
  9. The Guardian newspaper
  10. N-gram analysis
  11. Fine-tuning BERT and RoBERTa models
  12. Framework
  13. Technical details
  14. Results
  15. COVID-19 - cases and deaths
  16. ...and 11 more sections

Figures (21)

  • Figure 1: A sentiment analysis framework for The Guardian articles (01/01/2018 - 31/03/2022) utilising BERT and RoBERTa models trained and tested on the Senwave dataset.
  • Figure 2: Comparative analysis of article content and death cases in Australia and UK news sections of The Guardian.
  • Figure 3: Top 10 bigrams and trigrams for World News in 2020 first quarter (January - March) covering the beginning of COVID-19 pandemic.
  • Figure 4: Top 10 bigrams for Australia news and UK news in 2020 second quarter (April - June) covering the beginning of lockdowns of the COVID-19 pandemic
  • Figure 5: Sentiment distribution by the BERT and RoBERTa models from The Guardian (Australia, UK, World news and Opinion sections combined), spanning from 1st January, 2018, to 31st March, 2022.
  • ...and 16 more figures