Does mass protest after the police killing of an unarmed Black civilian broadly mobilize anti-police opinion or instead deepen racial and political divides? This research leverages a natural timing shock to answer that question.

📆 Weekly National Surveys and a Natural Shock

A large dataset of weekly cross-sections of the American public is used alongside a regression discontinuity in time (RDiT) design that exploits the essentially random timing of George Floyd's killing and the subsequent nationwide protests.

📊 How the evidence was identified

• Weekly cross-sectional public-opinion data capturing attitudes toward police and perceptions of anti-Black discrimination
• Regression discontinuity in time (RDiT) approach that treats the timing of Floyd's killing and the start of nationwide protests as an exogenous shock to opinion

🔎 Key Findings

• The Floyd protests produced rapid decreases in favorability toward the police and increases in perceived anti-Black discrimination among Americans who are low in racial prejudice and who are politically liberal.
• By contrast, attitudes among high-prejudice and politically conservative Americans either remained unchanged or showed only small, short-lived shifts.

⚖️ Why it matters

The evidence indicates that the 2020 protests did not produce a uniform opinion-mobilizing backlash against police; instead, they further racialized and politicized public attitudes about race and law enforcement in the United States.