Understanding the Gap Between Stated and Revealed Preferences in News Curation: A Study of Young Adult Social Media Users
New research shows users engage with low-quality content they don't endorse, while craving high-quality information.
A new study from researchers Do Won Kim, Cody Buntain, and Giovanni Luca Ciampaglia, to be published at CSCW '26, investigates a critical flaw in how social media algorithms understand users. The research focuses on the gap between 'stated preferences' (what users say they want) and 'revealed preferences' (what their engagement behavior suggests they want) among young adults. Using a mixed-methods approach, the team found that participants frequently found themselves engaging with low-quality content they did not personally endorse, despite explicitly stating a desire for high-quality, accurate information. This highlights a fundamental misalignment between user values and the signals that drive algorithmic curation.
The study's key experiment asked participants to curate an ideal social media news feed for a hypothetical persona. When given this agency, users created feeds they considered significantly more satisfying and higher in quality. They did this by consciously prioritizing values like accuracy, diversity, and context, while navigating trade-offs with other factors like social relationships. This demonstrates that feed curation is not a passive act but a 'socially situated process' of judging what is appropriate and valuable in shared information spaces. Based on these insights, the authors propose new design directions for social platforms, suggesting systems could be built to better reflect user-stated values rather than purely optimizing for engagement metrics that capture revealed—and often contradictory—preferences.
- Users frequently engage with ('reveal' a preference for) low-quality content they do not personally endorse or value.
- When given control to curate an ideal feed for a persona, users prioritize accuracy and diversity, creating feeds they rate as more satisfying.
- The research identifies feed curation as a socially situated process and offers concrete design directions to help algorithms bridge the value gap.
Why It Matters
This research provides a blueprint for building social media algorithms that align with user values, not just engagement, potentially reducing misinformation and low-quality content.