FINDING · EVALUATION

The study collected behavioral data from approximately 3.9 million Facebook users by instrumenting client-side JavaScript to detect composition-then-abandonment events, without capturing the suppressed text itself; this passive measurement approach allowed population-scale inference of suppression rates without content-level surveillance.

From 2013-das-self-censorship — Self-Censorship on Facebook · §2 Methods · 2013 · Conference on Weblogs and Social Media

Implications

Side-channel behavioral telemetry (timing, keystroke presence/absence) can reveal censorship-induced suppression at scale without accessing content; circumvention systems should minimize metadata leakage about editing/composition activity, not only payload content.
Privacy-preserving instrumentation designs (event counts without text) can serve as a template for measuring censorship impact without exposing user content to the measurement platform.