11:20 - 13:00
P7-S175
Room: 0A.07
Chair/s:
Nicolai Berk
Discussant/s:
Javier Lorenzo-Rodriguez
Online political expression under censorship: when backlash is short-lived
P7-S175-3
Presented by: Kristina Aleksandrovna Pedersen
Kristina Aleksandrovna Pedersen
Copenhagen Business SchoolUniversity of Copenhagen
The implications of censorship are often boiled down to two overall mechanisms by scholars: either the regime succeeds in silencing dissent or will endure the cost of backfire in the form of circumvention or public backlash. Much existing scholarship debates the contextual characteristics and settings in which each mechanism prevails. This paper analyzes how overt and repressive censorship shapes online discourse in times of political unrest. Using Russia's full-scale invasion of Ukraine and subsequent authoritarian consolidation as a case, this paper studies over 4,500,000 comments across different pro and anti-regime-leaning YouTube channels targeted toward Russian-speaking audiences. I use an embedding-based topic model to analyze the changes in online discourse in the days following the invasion as well as the introduction of Russia's censorship law in March 2022. The results suggest that discourse on collective action nearly doubled at the start of the invasion. However, both discourse on collective action and criticism of the war sharply dropped as the law was introduced. In sum, the findings indicate that Russia’s invasion of Ukraine initially triggered (online) backlash but that the regime swiftly curtailed the dissent through the implementation of overt and repressive censorship. This showcases how regimes can successfully balance sparks of unrest by doubling down on repression. In light of the findings, the paper contributes to the theoretical discussion on the complicated dynamics of backlash and deterrence. It demonstrates how the impermanent nature of backlash patterns can make the associated costs worthwhile for autocrats hoping to reduce long-term risks.
Keywords: political expression online, self-censorship, Russia, information control, NLP

Sponsors