Corpus Analysis of Coronavirus Disease-19 (COVID-19)-Related Loneliness in Twitter (Preprint)

Abstract

BACKGROUND Loneliness is a complex mental health issue involving feelings of isolation, disconnectedness, and a lack of purpose in life. The coronavirus disease (COVID-19) pandemic and lockdowns significantly increased social isolation, contributing to a global rise in loneliness on social online platforms. However, research on mental health expressions in Arabic on social media is limited, exacerbating the lack of understanding of loneliness and mental health issues in these communities. Stigma and societal resistance in the Middle East hinder individuals from seeking mental health support, a trend which was particularly noticeable amidst the challenges of the pandemic. OBJECTIVE This study aimed to conduct a corpus analysis of an open-source dataset to gain insights into the origin of how loneliness was expressed in Arabic on Twitter during the initial months of COVID-19 regulations. Additionally, it explored temporal trends during the same time period to enhance comprehension of the prevalence of loneliness and mental health struggles during the initial months of the pandemic. METHODS Grounded in corpus linguistic methodology, the study also employs thematic analysis, explores temporal trends, and conducts Inter-Annotator Agreement (IAA) to comprehend the prevalence of loneliness and mental health struggles during the pandemic's early stages. RESULTS Our analysis of Twitter data using AntConc revealed crucial insights into the origins of mental health challenges and loneliness during the early days of the COVID-19 pandemic. Employing sophisticated linguistic techniques, we identified over 18,000 instances of loneliness-related terms, which deepened our understanding of the issue. Emotions were expressed through various themes, with a significant focus on COVID-19 news and opinions. The advanced search features of AntConc were instrumental in this process, allowing us to gauge word frequency and locate relevant tweets. The association between loneliness and the pandemic was clear, especially with the word "corona" being the most frequent collocate. Thematic analysis identified the seven key themes that were despair in the face of uncertainty, grief for the loss of life, hope through faith, opinions on COVID-19, opinions on isolation, COVID-19 news, and fear of illness which characterized the emotional landscape during this time. Inter-Annotator Agreement highlighted the diversity of interpretations. CONCLUSIONS Our study demonstrates the crucial role social media plays in capturing and understanding human emotions during major global events like the COVID-19 pandemic. Our AntConc analysis offers insights into mental health and loneliness during the early COVID-19 pandemic. It highlights the Arabic-speaking online community's experiences and contributes to global understanding of mental health and social media dynamics. This approach can be applied to different languages, enriching our understanding of loneliness and mental health during crises.