Narratives and counter-narratives in religious responses to COVID-19: A computational text analysis.

Religious responses to COVID-19 as portrayed in a major news source raise the issue of conflict or cooperation between religious bodies and public health authorities. We compared articles in the New York Times relating to religion and COVID-19 with the COVID-19 statements posted on 63 faith-based organizations' web sites, and with the guidance documents published by the Centers for Disease Control and Prevention (CDC) and World Health Organization (WHO) specifically for religious bodies. We used computational text analysis to identify and compare sentiments and topics in the three bodies of text. Sentiment analysis showed consistent positive values for faith-based organizations' texts throughout the period. The initial negative sentiment of religion-COVID-19 coverage in the New York Times rose over the period and eventually converged with the consistently positive sentiment of faith-based documents. In our topic modelling analysis, rank order and regression analysis showed that topic prevalence was similar in the faith-based and public health sources, and both showed statistically significant differences from the New York Times. We conclude that there is evidence of both narratives and counter-narratives, and that these showed demonstrable shifts over time. Text analysis of public documents shows alignment of the interests of public health and religious bodies, which can be discerned for the benefit of communities if parties are trusted and religious messages are consistent with public health communications.