The Social Costs of Digital vs. In-person Surveillance Xu Xu∗ Abstract The world is witnessing an explosion of digital surveillance in recent years. This paper examines the social costs of digital surveillance versus in-person surveil- lance in dictatorships. I argue that both types of surveillance deter political participation because citizens fear targeted repression. However, digital surveil- lance does not entail human-agent intrusion into private lives and therefore is less likely to undermine interpersonal trust and regime legitimacy than in-person surveillance. I manipulate information about surveillance operations in an in- the-field survey experiment on college students in two regions of China and find supportive evidence. I further establish the external validity of the experimental findings on digital surveillance using a nationally representative survey and a natural experiment caused by the 2015 Tianjin explosion. Overall, the findings that digital surveillance deters political participation without some of the costly byproducts of in-person surveillance have board implications for surveillance in dictatorships and even in democracies. ∗Postdoctoral Research Fellow, Department of Political Science and Department of Communication, Stan- ford University.
[email protected], https://xu-xu.net. For helpful comments, I would like to thank Xun Cao, Volha Charnysh, Martin Dimitrov, Matt Golder, Sona Golder, Sheena Greitens, Luke Keele, Genia Kostka, Jennifer Pan, Molly Roberts, Suzanne Scoggins, Joseph Wright, Yiqing Xu, Boliang Zhu; partici- pants at the MPSA conference, the annual meeting of the Society for Political Methodology, and department seminars at Penn State. \WIR SIND UBERALL¨ (WE ARE EVERYWHERE)" | Stasi, German Democratic Republic 1 Introduction Rarely in history could any autocratic regimes surveil citizens at a scale as large as those achieved by today's digital surveillance states.