Factors Influencing the Prevalence of PTSD Tendencies on Social Media During Emergencies

Post-traumatic stress disorder (PTSD) is a psychological condition that may develop after exposure to severe traumatic events, particularly in unexpected situations characterized by high uncertainty and stress. As social media has become an integral part of daily life, it serves as a platform for sharing personal experiences, emotional expressions, and public responses during emergencies. Grounded in Protection Motivation Theory, this study explores the factors associated with PTSD tendencies expressed on social media during emergencies. Using large-scale user-generated data from social media, posts exhibiting PTSD tendencies were identified through a machine-learning classifier based on semantic embeddings, and the Latent Dirichlet Allocation topic model was applied to extract relevant themes. A multiple linear regression model with Newey-West standard errors was then constructed to examine the factors related to the prevalence of PTSD tendencies. The results indicate that epidemic containment, protective measures, and health concern are negatively associated with the prevalence of PTSD tendencies on social media, whereas expressive negativity is positively associated with trauma-related expression. Focusing on the context of COVID-19, this study highlights the factors associated with PTSD tendencies on social media and underscores the importance of well-organized epidemic prevention and control measures and balanced information dissemination in mitigating the psychological impact during emergencies.