This paper challenges assumptions about citizen self-censorship in authoritarian regimes.

➡️ The Core Question: Do citizens in authoritarian systems truly self-censor when asked about their governments?

The authors introduce a novel approach: calculating a self-censorship index by comparing nonresponse rates to regime assessment questions with less sensitive ones. This method allows for cross-country and over-time comparisons without relying on potentially unreliable direct responses.

🔍 Data & Methods: Analyzed item nonresponse patterns across diverse political systems.

Used survey data from multiple countries, ensuring anonymity in all examples.

Constructed a self-censorship index based on differential response to regime-related questions versus neutral ones.

📊 Key Findings: Contrary to expectations, authoritarian citizens don't show statistically significant higher nonresponse on regime assessments than those in democracies. This suggests that these questions may lack sufficient sensitivity to capture genuine self-censorship if it exists. The exception appears to be regimes without executive electoral competition — lower competition correlates with reduced self-censorship detection.

💡 Why It Matters: These findings call into question the reliability of using regime assessment surveys as indicators of authoritarianism or popular satisfaction. They challenge the assumption that high nonresponse necessarily signals deception, opening up new avenues for comparative political analysis and measurement.