New analysis examines how legislators set agendas on social media. Researchers analyzed fine-grained temporal patterns in tweets sent by members of Congress and citizens during the 113th U.S. Congress.

Data & Methods: Twitter messages were classified into topics using unsupervised methods, then vector autoregression models assessed issue prediction dynamics between public and politicians.

Key Findings: Legislators follow, rather than lead, public discussions online; this holds despite media influence. However, they show greater responsiveness to their supporters compared to the general public.

Real-World Significance: The findings suggest social media data can help clarify political agenda-setting processes beyond traditional metrics.