Paper Submission Temporal Analysis and User Characteristics of Internet Censorship on Sina Weibo

This research investigates features used to identify surveillance targets and the probability of censored posts over time on the Chinese Sina Weibo social media platform. Targets include the recency of content chosen for censorship, frequency, and the users surveilled. The analysis consists of 14,000 censored posts on Sina Weibo collected over 3 months from August to November of 2021. Results, demonstrate that during the past 10 years the rate of censorship has increased and verified users (paying customers) who have a high user ranking in the system are censored more frequently than unverified (non-paying customers) low ranking users. In time T1 there is an 80% chance of a post to be censored if it is less than 50 days old while in time T2 there is an 80% probability (dashed red line) of a post being censored occurs after 1000 days (3100-2100).