Exposure inequality beyond friends in out-of-network recommendations

Ascertain whether exposure inequalities documented for friends’ tweets on Twitter/X extend to the broader set of out-of-network recommended users in Twitter/X’s “For You” timelines, and quantify any differences in exposure inequality across users with different political leanings.

Background

Previous work has shown that popularity biases create uneven visibility among tweets when comparing personalized feeds with reverse-chronological timelines, and that users are disproportionately exposed to friends’ tweets. The authors point out that it is not yet established whether similar exposure inequalities exist outside the friend network among recommended users.

To address this gap, they propose measuring exposure inequality using the Gini coefficient applied to an exposure metric that weights tweet occurrences by rank-based visibility. This framing sets up the unresolved question of whether inequality observed within friends’ content also characterizes out-of-network recommendations.