Questions and answers about
the UK economy.

Viral narratives: evidence from media outlets during Covid

Culture provides heuristics for decision-making that are developed as part of the evolutionary process (Richardson and Boyd, 1985, Nunn 2015). In this project, we aim to trace the evolution of narratives as a particular form of heuristic regarding the roles of different agents. We seek to show that the narratives or mental shortcuts people rely on persist, are applied to various different contexts, and can be shifted by economic shocks. We apply this to the current coronavirus pandemic and seek to test whether the China trade shock (Autor, Dorn and Hanson, 2013) influences the mental model people use to understand this pandemic. In particular, we hypothesize that areas more negatively impacted historically by Chinese trade, today hold a worldview where China plays an active role in creating or facilitating the pandemic. We do this by using newspaper and Twitter data from January-March 2020, matched to Chinese trade data from Autor, Dorn and Hanson. Methodologically, we apply quantitative narrative analysis techniques to create a representation of the subjective causal networks of different actors relevant to a given topic. We measure how central different actors are in the network implied by the text analysis. This allows us to measure regional variation in a worldview in which China caused the damaging coronavirus pandemic - either directly or indirectly. The project makes the conceptual point that narratives act as heuristics or mental models that apply across different contexts. In doing so, it introduces an empirical methodology to quantify types of mental models held by individuals.

Lead investigator:

Arthur Blouin


University of Toronto

Primary topic:

Attitudes, media & governance

Secondary topic:

Lessons from history

Region of data collection:

North America

Country of data collection


Status of data collection


Type of data being collected:

From private company

Unit of real-time data collection


Start date


End date