Research / TikTok recommendations and health misinformation
What TikTok shows, and what people watch
Studies of recommendation algorithms usually look either at what a platform serves, by running test accounts, or at what users say they saw, by asking them. Neither tells you what a real person actually watched.
In HARMONY, young adults donate their TikTok history, which records every video the feed served and how long it stayed on screen. Linked to surveys, this lets us ask how much health content and health misinformation reaches people, how much of it they watch rather than scroll past, and whether it relates to their wellbeing.
How the data comes together
- Participants answer surveys about their health and wellbeing.
- They download their own TikTok data and donate it to the study.
- The donated history shows every video the feed served and how long it stayed on screen.
- Linking the two lets us compare what the platform sent a person with what that person watched.
Projects
HARMONY: Health-Related (Mis-)Information on Social Media and Its Impact on Young Adults
Talks
1 Oct 2026 · Lightning talk at Trust & Safety Research Conference, Stanford University: Inside the For You Page: algorithmic content exposure, health misinformation and well-being among young adults on TikTok
7–8 Dec 2026 · Poster at New Directions in Social Media Research and Policy, University of Cambridge: TRACE: a four-layer design for reconstructing algorithmic content environments
Publications
Papers from this work will be listed here as they are published. See all publications.