Computational Analysis of Linguistic Markers Reveals Group Identification Patterns During a Global Public Health Crisis

Abstract

Using a computational social science approach integrating evolutionary psychology, social psychology, and intergroup relations, this research examined associations between disease threats and group identification across cultures by analyzing the relationship between COVID‐19 threats (actual and perceived) and first‐person plural versus singular pronoun usage—a linguistic marker of collective identity. Two studies analyzed epidemiological and Google Trends search data over 154 weeks: Study 1 established baseline patterns using U.S. data, while Study 2 investigated cross‐cultural variations globally. Through transfer function methodology to pre‐whiten time series data for cross‐correlation function analysis and random‐slope‐random‐intercept multilevel modeling, the research explored temporal dynamics of threat‐identification relationships. Results indicated that perceived COVID‐19 threat, measured through pandemic‐related search volumes, exhibited consistent correlations with increased first‐person plural pronoun usage across both studies. While COVID‐19 cases demonstrated non‐significant relationships globally but positive associations in the U.S., mortality rates showed negative correlations with first‐person plural pronoun usage globally, differing from U.S. patterns. Cultural collectivism emerged as a moderating factor in the global analysis, with individualistic cultures displaying stronger associations between perceived threat and collective linguistic markers. These findings contribute to our understanding of the relationship between collective threats and group identification processes across cultures while demonstrating the potential of computational linguistic analysis for monitoring real‐time collective responses to pandemics, offering insights relevant to preparedness strategies for future global challenges.