Skip to content
Stories

Relating Digital Trace Data to User-Level Perceptions: Satire, Misinformation, and Source Cues

Partially supported by a NULab Seedling Grant.

Many large-scale studies of digital trace data involve domain lists. An individual is associated with a set of websites that they have viewed or shared, and a construct of interest is operationalized by looking for the presence of specific domains within these user-level sets. This approach is widely used in the study of misinformation sharing or exposure. Although this method provides a parsimonious and scalable approach to studying misinformation, its utility depends on the domain list used. A domain list can bias researchers’ results if it does not properly operationalize the construct of interest.

This project argues that some widely used misinformation-related domain lists suffer from this problem. Using a dataset of URL-sharing events on Facebook, this research argues that satirical websites, especially right-leaning ones like the Babylon Bee, are treated by researchers as misinformation websites although they are not seen as such by their audience (i.e., they are received as humor rather than news). By fielding an online survey examining the reception of satirical and news articles by actual people, it will attempt to demonstrate two things. First, it attempts to confirm that articles received as humorous on Facebook are understand by readers to be satire rather than true news stories with humorous content. Second, it explores the role Facebook’s user interface plays in determining an article’s reception, distinguishing source cues (the ability to see an article’s publisher) from reception cues (seeing how other readers have perceived the article). It examines under what conditions source and reception cues affect the perception of in-group and out-group satire as misinformation.

This project will also contain replications of previously published results on the prevalence of right-wing misinformation, exploring how estimates of conservatives’ preference for misinformation are sensitive to the use of these domain lists.

Principal Investigator

Stefan McCabe, PhD Candidate, Network Science

More Stories

Mexican Textile Factories 1843-1857

06.07.2022

Exploring GPT3 Training for ‘Black’ Neural-Network Text Generation

05.31.2022

Power’s Inversions

07.04.22
Research Projects