Cascade-based censorship (ICM model) and uniform random deletion produce measurably different topological signatures: cascade removal causes greater increases in network diameter and radius as the censorship fraction γ increases and a substantial increase in assortativity at mid-removal levels (γ=0.2–0.5), whereas uniform deletion shows slower, more gradual changes across these same metrics.
From 2014-morrison-toward — Toward automatic censorship detection in microblogs
· §4.1
· 2014
· Data Mining in Social Networks
Implications
Censorship monitoring tools can distinguish targeted cascade suppression from broad indiscriminate deletion using topology alone, enabling finer-grained characterization of censor strategy without content access.
Propagating content as independent posts rather than repost chains produces topological signatures closer to uniform-deletion networks, making cascade-targeting by censors structurally harder and censor-strategy inference less reliable.