Tracking The Rise and Fall of Mask-Related COVID-19 Theories
Andrew Beers, 2nd-year PhD student in HCDE
Sarah Nguyen, 1st-year PhD student in the iSchool
With guidance from faculty advisers Kate Starbird, HCDE, and Emma Spiro, iSchool.
The Center for an Informed Public at UW has been archiving posts from Twitter users engaging in arguments about whether or not to wear a mask. We’ve spent this last year qualitatively analyzing a subset of ~5,000 posts from these arguments, to understand the many different theories that Twitter users are employing to argue for or against public mask mandates. We found that users have a variety of theories about masks — about their ability to block virus particles, their potential harm to users, when they should be used, and more — and defend these theories aggressively using the language of science and a wide variety of external links, images, and videos.
This Winter quarter, we would like to expand this project to our full dataset, which currently numbers in the tens of millions of posts and is growing every day. Specifically, we’re aiming to use automated techniques to classify this larger tweet dataset of arguments into the theories we identified in the first stage of this project. We want to see how the popularity of certain theories about masks changed over time, and whether they responded to external events, such as the publication of a new scientific paper, a change in the severity of the pandemic, or a comment by a politician.
We’re looking for up to three students with an interest in public health communication, misinformation, and/or natural language processing. We are interested in students familiar with Python and with the Twitter platform, but we strongly encourage students without prior experience in either of these areas to also apply.
What You’ll Be Working On
We’re looking for a relatively small team, and we can see several tasks for this project over the course of the quarter:
- Qualitative coding of tweets according to a pre-existing protocol
- Case-study analysis of events causing changes in the prevalence of different theories
- Data analysis and visualization of data produced during the course of this project
Attend either one 2-hour or two 1-hour meetings each week, time TBD upon registrants schedules.
Work 6 hours outside of the class meeting.
Register for 3 credits of HCDE 496/596.
If you are interested in this DRG, please apply by Friday, December 18 using this Google Form. We’ll then interview some applicants as soon as possible after their application comes in, and will notify all applicants by January 1st, 2021. Please contact email@example.com with any questions you may have.