| Collaborators | David Domingo (ULB, professor in journalism), Ingrid Van Marion (ULB, expert in science communication), Sébastien de Valeriola |
This project aims at studying how journalists invoked academic experts in Belgian media during the peak of the COVID-19 crisis. To do this, we use a corpus of approximately 225,000 press articles.
Since the focus is on experts, we have applied named entity recognition models to identify mentions of people’s names in texts. These mentions were then analysed manually, to identify those referring to experts.
The experts’ specialties and institutions (or at least those mentioned with their names by the journalists writing the articles) were also extracted, enabling us to study when experts from which disciplines and universities are invoked.
The texts were also subjected to topic models, in order to determine what exactly they were talking about (vaccination, hospitalizations, school closures, etc.).
The initial results we’ve obtained are very promising, allowing us to make a detailed analysis of the phenomenon of expert citation.