Mapping Linguistic Shifts During Psychological Coping With the COVID-19 Pandemic

Abstract

How does language change reveal the psychological trajectories of people coping with a COVID-19 infection? This study examined writings on social media over 12 weeks from people who self-reported having tested positive for COVID-19. People used fewer words reflecting anxiety and distancing but more words indicating reinterpretation over time. The language patterns for describing the experience of COVID-19 infections differed from those for describing other unrelated topics. The findings reveal the temporal dynamics of psychological adjustment to an unfolding crisis.

Environmental Monitoring for Arctic Resiliency and Sustainability: An Integrated Approach with Topic Modeling and Network Analysis

“…Unlike traditional analyses that describe topics through archetypal examples, topic modeling identifies topics based on how words are distributed in the documents and the extent to which the cluster of words leads to clear and coherent semantic meaning [28]. Previous work has applied topic modeling to various areas of sustainability, such as smart factories [29] and logistics [30], and has used a diverse set of datasets, including scientific literature [29], patent applications [30], and social media posts [31].…”

“…Then, the probabilities can be included in statistical models predicting outcomes of interest. Indeed, research beyond environmental science has demonstrated the utility of incorporating text-mining results into predictive modeling [31,50]. For example, text-mining results were found to improve the accuracy of machine learning models predicting information utility in online platforms [50].…”

“…For example, text-mining results were found to improve the accuracy of machine learning models predicting information utility in online platforms [50]. Topic modeling also enables researchers to predict anxiety as people cope with traumatic events [31].…”