Narratives of COVID-19

In January 2020, while building a collection about Taiwan's presidential elections, Observatory researchers noticed something unusual: an online poster promoting a narrative that Taiwanese voters could become infected by a coronavirus coming out of China, and would be safer wearing a mask when voting. At a time in which there was no information about the virus, many interpreted this narrative as an attempt to keep Taiwanese citizens away from voting booths. We documented this instance of attempted voter manipulation through an information operation, and started looking for similar claims coming out of China.

Shortly after that, the disease that would later be named COVID-19 enveloped the world. By February we began building a COVID-19 Observatory, and launched our first full-scale multinational Observatory into COVID-19 narratives. We first focused on Hong Kong in March, and then expanded our research to Russia, Brazil, Nigeria and India, with the hypothesis that these four countries, all with large populations and landmasses, might experience significant viral outbreaks. Each of these countries has complex information ecosystems marked by active and often contentious social media spaces, a mix of independent and state-run or state-aligned media outlets, and politics that often generate false and misleading narratives. As COVID-19 spread throughout the world in 2020, India, Russia and Brazil for a time marked some of the fastest rates of infection, and experienced confusing and often false information claims, meant, in broad strokes, to protect the reputation of governments instead of safeguard the health of citizens.

We then expanded the research to include multinational narratives and especially interesting instances of information manipulation in a range of other countries, including Bangladesh, Bolivia, Venezuela, and Myanmar. Throughout 2020, the Global Voices’ COVID-19 Observatory documented and annotated nearly 700 media items coming from 272 media sources in 22 languages, identified and explained 192 narratives and 118 themes, and built a extensive set of analytics tools to inform our research for Global Voices stories, to share with partners and to warn social media platforms of the potentially harmful effects of false and misleading narratives, and identify unusual trends in information manipulation.

With this work, we also gained significant insight into how to build a transnational research project focused on careful and labor-intensive reading of media ecosystems, and explication of the possible benefits and harms of specific media items through a normative analysis of each. The resulting dataset is available for public use, and forms an annotated digital collection of media items, themes, narratives, and media sources. A non-public version also includes screen captures of media items and where possible, full text, available for use for qualified researchers. We also include extensive analysis of the data, in the form of an illustrated slide deck.