Language & LLMs
Fear in English COVID-19 tweets fell after chloroquine was called a cure and rose again after the FDA rejected it
In a 2021 paper in the International Journal on Semantic Web and Information Systems, researchers ran sentiment analysis, a method that identifies emotions in text, on a large set of English-language tweets about COVID-19. They looked for ten emotional themes. Their findings show a clear shift in the emotions people expressed. After top officials, including Trump, called chloroquine a cure, fear and anxiety in the tweets fell while gratitude, happiness and hope rose. After the FDA said chloroquine was not an effective cure, fear climbed again.
What did the researchers do?
The researchers used natural language processing and related text-analysis methods to identify emotions in a large Twitter dataset of English tweets. They investigated ten emotional themes, including fear and anxiety, gratitude, happiness and hope, and other mixed emotions.
What did the tweet analysis find?
The analysis found that COVID-19 had spread fear and anxiety among people, but it also found gratitude, happiness, hope and other mixed emotions. The clearest pattern was tied to chloroquine. Positive statements from top officials were followed by lower fear and more positive emotions. The FDA's statement that chloroquine was not effective was followed by renewed fear.
Frequently asked questions
Which platform and language did the study use?
The study used a large dataset of English tweets from Twitter.
What emotions did the researchers track?
The researchers investigated ten emotional themes, including fear and anxiety, gratitude, happiness and hope, and other mixed emotions.
What was the chloroquine pattern?
Fear fell and positive emotions rose after officials called chloroquine a cure. After the FDA said it was not effective, fear rose again.
Limits and what's next
The abstract describes one sentiment-analysis study of a large Twitter dataset of English tweets. Because the data come only from Twitter and only from English-language posts, the findings may not apply to other platforms or languages. The abstract reports no independent replication. It does not say what follow-up research the authors plan.
Source: Jalal Suliman Alowibdi et al., “Coronavirus Pandemic (COVID-19),” International Journal on Semantic Web and Information Systems, 2021.