Social Science Research Council Research AMP Just Tech
Citation

Politically Motivated Retroactive Self-Censorship on Social Media: Managing Political Impression in a Politically Polarized Environment

Author:
Zhang, Xinzhi
Publication:
Mass Communication and Society

Social media offers individuals sophisticated communicative tactics to manage politically motivated interpersonal disagreement. This study investigates an underexplored behavior in this context: retroactive self-censorship, which includes actions such as deleting previously published posts or untagging oneself from others’ timelines or photos for political reasons. These behaviors are conceptualized as a form of selective self-presentation within the framework of the spiral of silence, triggered by interpersonal political disagreement. The study also examines the role of discussion network heterogeneity in shaping this process, hypothesizing that higher levels of heterogeneity make disagreement-induced self-censorship more likely. A two-wave online panel survey was implemented in Hong Kong, one of the most politically polarized societies globally. Findings reveal a sequence in which political expression on social media leads to disagreement, which then prompts ad hoc self-censorship. The mediation effect of opinion expression on self-censorship via disagreement becomes stronger as network heterogeneity increases from low to moderate levels, but this pattern does not extend to highly heterogeneous networks. The results suggest that ad hoc content manipulation on social media would aggravate political polarization by making people’s networks more homogeneous. The democratic implications of managing political impressions in a politically polarized digital public sphere are discussed.