Is the News Always Negative? Using Deep Learning to Track News Sentiment During the Covid-19 Pandemic
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
The news is notorious for its tendency to focus on negative events, giving rise to the adage, "If it bleeds, it leads." In this study, I examine whether news coverage during the Covid-19 pandemic mainly focused on negative events while downplaying positive developments such as decreasing Covid-19 cases. Utilizing a state-of-the-art fine-tuned language model, I analyzed the sentiment of over 900,000 Covid-19 related news articles from March 2020 to April 2022 across the United States, Canada, and the United Kingdom. The results indicate that the news is far more negative than positive—even when Covid-19 cases and hospitalizations are decreasing. This negativity is most pronounced in Op-Ed articles, front-page news articles, and articles published by large news organizations (e.g., New York Times, BBC, Fox News). However, non-Op-Ed news articles do become more positive as Covid-19 cases decrease, contradicting previous research findings. These discrepancies can be attributed, in part, to differences in model accuracy, as the model I trained is approximately 20% more accurate than other models used in the literature. Further, when dividing U.S. news by the publisher's political ideology, clear differences emerge: both left-wing and right-wing sources are much more negative than centrist news sources. Surprisingly, these differences in sentiment are about as large as the difference between regular news and Covid-19 news sentiment, indicating substantial differences in news reporting across political lines. These findings provide new insights into news reporting patterns during the pandemic and carry important implications for public health messaging and news reporting practices.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.002 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it