Related Experiment Video
Updated: Aug 9, 2025

Comparing Objective Conjunctival Hyperemia Grading and the Ocular Surface Disease Index Score in Dry Eye Syndrome During COVID-19
Published on: May 25, 2022
Emotional Distress During COVID-19 by Mental Health Conditions and Economic Vulnerability: Retrospective Analysis of
Michiko Ueda1,2, Kohei Watanabe3, Hajime Sueki4
1Department of Public Administration and International Affairs, The Maxwell School of Citizenship and Public Affairs, Syracuse University, Syracuse, NY, United States.
Social media analysis during public health crises can monitor emotional distress, especially in vulnerable populations. This study developed a machine learning framework to track mental health using tweets without extensive training data.
Area of Science:
- Public Health
- Computational Social Science
- Mental Health Surveillance
Background:
- Social media analysis offers a cost-effective method for monitoring psychological conditions during public health crises like COVID-19.
- Limited understanding of user characteristics and lack of large annotated datasets hinder mental health surveillance using social media data.
- Supervised machine learning approaches for mental health are often infeasible due to data limitations.
Purpose of the Study:
- To propose a machine learning framework for real-time mental health surveillance that minimizes the need for extensive training data.
- To track emotional distress levels among Japanese social media users during the COVID-19 pandemic using survey-linked tweets.
- To analyze the relationship between user attributes, psychological conditions, and emotional distress.
Main Methods:
- Conducted online surveys (N=2432) of Japanese adults, collecting demographic, socioeconomic, and mental health data, linked to Twitter handles.
- Calculated emotional distress scores for over 2.4 million tweets using a semisupervised algorithm (latent semantic scaling).
- Analyzed 495,021 tweets from 560 users (aged 18-49) using fixed-effect regression models to compare distress levels in 2020 versus 2019.
Main Results:
- Emotional distress peaked in early April 2020, coinciding with the state of emergency declaration, but was unrelated to COVID-19 case numbers.
- Government-induced restrictions disproportionately impacted vulnerable individuals, including those with low income, precarious employment, depressive symptoms, and suicidal ideation.
- The study identified specific demographic and psychological factors associated with increased emotional distress during the pandemic.
Conclusions:
- Established a framework for near-real-time monitoring of social media users' emotional distress, complementing traditional data sources.
- Demonstrated the potential of survey-linked social media posts for continuous well-being monitoring.
- Highlighted the framework's adaptability for detecting suicidality and measuring population sentiment in real-time.
Related Concept Videos
Stress and Mental Health
Individuals with depression often experience challenges in both their personal and professional...
Psychological Responses to Stress
Introduction to Stress and Lifestyle
Coping Strategies: Emotion Focused
Steps in Outbreak Investigation
Long-term Depression
Calcium Ion Concentration Mechanism
If over...

