Evaluating Content Exposure Bias in Social Networks

Online social platforms employ personalized feed algorithms to gather and prioritize messages from accounts followed by users, which distorts content's perceived popularity prior to personalization. We call this "exposure bias," and our research focuses on quantifying it using diverse exposure bias metrics, and we evaluate recommendation algorithms through various content ranking heuristics. Similarly we simulate activity in a network to assess the influence of such ranking heuristics on exposure bias. Furthermore, we are working on agent-based model simulations to comprehend the impact of ranking schemes, with the ultimate goal of exploring intervention effects over time. Our empirical findings reveal that users exposed to popularity-based feeds experience significantly lower exposure bias compared to chronologically-ordered feeds.