Along with Justin Lane, Ivan Puga-Gonzalez, Roger Normann, and Jan Pastorek, I’ve just published an article in PsychArXiv, which is open access and available for download here.

Abstract:

In this article we present the findings of social media analysis of the spread of misinformation in the wake of the COVID-19 pandemic and outline how analyses of the psychological properties of a text can be used to optimize strategic messaging online. Our data used Twitter data, collected during the COVID-19 pandemic and analyzed using a suite of AI based analytical tools, which provided data for further empirical analysis. The analysis yielded insights related to the differences in the dynamics of the spread of misinformation within (and outside of) Scandinavian countries. Analysing this data enabled us to explore three hypotheses: (1) Misinformation will be associated with specific moral signatures, which will differ between Scandinavian and non-Scandinavian samples, (2) Levels of engagement will be associated with specific themes and moral concerns, which will differ between Scandinavian and non-Scandinavian samples, and (3) Within Scandinavia, similar unique signatures will be discernible at the country level, with Sweden driving significant differences. These specific results provide guidance for healthcare professionals responsible for communicating information and crafting messages that are more resonant with their target population and more generally demonstrate the ability for social media analysis to be useful in strategic decision making when going beyond focusing on engagement metrics or sentiment alone.