The coexistence of enduring authoritarianism and the Internet has drawn wide academic attention. Why are some authoritarian regimes, previously believed to be inherently incompatible with free information, surviving and even thriving despite the expansion of the Internet? In Censored, Margaret E. Roberts develops a novel theory of censorship to explain authoritarian resilience in the digital age. She argues that strong authoritarian governments such as China have the capacity to enforce censorship more forcefully, but choose not to do so; they instead embrace a strategy of porous censorship. Such a strategy combines fear, diversion, and distraction mechanisms, allowing autocrats to differentiate politically active citizens from the majority of the public and shape the latter’s information-seeking behavior by altering the costs to accessing information. In this way, authoritarian governments manage to maintain sufficient control over information while avoiding the repercussions that are associated with heavy-handed repression. The book is an impressive display of theoretical originality, methodological sophistication, and empirical richness.
The book advances a theory of censorship explaining the strategic interplay between government control and citizens’ production and consumption of information (Chapter 2). Roberts first describes the censorship incentives of the government, as well as how the media and citizens, respectively, interact with information, arguing that both the media and the public are sensitive to the costs of information production, distribution, and consumption. From there, she elucidates the three main mechanisms of censorship: fear, friction, and flooding. While fear works through deterrence, both friction and flooding reprioritize the production and consumption of information by affecting its relative price: Friction raises the absolute cost; flooding raises the relative cost (by lowering the costs of information from alternative sources). An example of friction is the “Great Firewall” of China, which is not impossible to circumvent, but prevents the majority of citizens from accessing banned sites by costing them more time, money, and energy. An example of flooding is “Twitter armies” amassed by governments to push for their views. Since the media and the public are cost sensitive, censorship is not just fear based but can work through friction and flooding tactics that do not stop information flow completely, but rather inconvenience users. Moreover, Roberts argues that in the Internet age, fear-based censorship has become more costly, while friction has become more impactful and flooding much cheaper.
After outlining the theoretical framework, Roberts tests her theory empirically using evidence primarily from China. In Chapter 3, she provides an overview of censorship in Chinese history as well as today. In Chapters 4 through 6, she discusses in detail how Chinese citizens react when they observe censorship, how their information-seeking behavior may be affected by even small and less observable frictions, as well as how flooding may shape information supply and demand. Roberts demonstrates that typical netizens in China are angered rather than intimidated after experiencing censorship, showing that fear-based censorship is ineffective; speed of censorship bears implications for the spread of information, while small impediments can generate an immediate impact on users’ information-access behavior; and coordinated government propaganda can make certain types of information more likely to be shared in social media spheres.
It is in the empirical chapters that Roberts has fully demonstrated her mastery of rich data and sophisticated methods. She has truly done an admirable job testing her theory using experimental methods, national survey data, large-scale social media data sets, leaked propaganda archives, and data sets of newspapers. In a nutshell, she conducts mixed-method research long advocated in the field in an exemplary way, and her contributions are especially worth highlighting in two respects. First, the innovative approach and the data strategy are particularly inspiring for scholars working on sensitive topics in a constrained political environment. It is challenging to conduct research in a repressive authoritarian regime like China. Despite overall improvements in the reform era, there are still many obstacles, and the situation has only deteriorated in the past few years. What Roberts has done can serve as a road map for doing research in China and perhaps other authoritarian regimes in terms of a) how to conduct research without triggering repercussions for the researcher, the project, and most importantly our subjects; and b) how to make good use of available data. Second, while many of us still worry about data availability, we are at the same time in an age of data explosion. The Internet is providing so much more data in so many more forms and from so many more sources today that social scientists have to learn to deal with data overabundance and keep up with methodological innovations to process such data. Roberts’s adept use of social media data sets, digitalized newspaper sources, and leaked propaganda archives is illuminating, especially for junior scholars, in this regard.
Another strength of Censored is the generalizability of the theory. Roberts argues in a quite convincing fashion that her theory applies not just to China, where much of the empirical evidence is from, or even just authoritarian regimes, but also to liberal democracies. While porous censorship clearly works to the advantage of autocrats, the theory, especially the functions of friction and flooding, may well travel into democracies. This is a timely warning for all of us, not just social scientists who care about the values of freedom and democracy. The discussion of the theory’s implications for free speech in democracies in Chapter 7 thus deserves our attention.
In addition to its contributions, the book invites debate from fellow scholars and provides directions for future research. First, is porous censorship a strategic choice by the state, or is it simply that the state lacks the capacity to impose perfect control? I ask this question because under President Xi Jinping, China is relying on more fear-based censorship than before. This question also has theoretical implications in that China is the clearest case of strong authoritarianism in the world. Whether China is enforcing such a strategy or not, we should continue looking for alternative explanations of censorship outcomes as well as authoritarian resilience (even though that search does not invalidate Roberts’s theory). Second, by interpreting coordinated media campaigns as information distraction and diversion, the book implies that authoritarian propaganda is less about promoting what the state likes, and more about crowding out what it dislikes. While this idea is innovative and quite reasonable, it runs the risk of oversimplifying the goals, functions, and implications of authoritarian propaganda. Further research is thus needed. Third, the role of the audience also deserves more careful examination. The book makes a contribution by highlighting that censorship is customized. But by categorizing citizens into two groups, with one actively seeking information and the other being rationally ignorant, Roberts implicitly assumes that all citizens are at least potentially against authoritarian rule. However, recent studies, including my own, have suggested otherwise. While the studies do not hurt Roberts’s overall argument, they may provide alternative explanations for censorship effectiveness and authoritarian resilience.
The points made in the previous paragraph are not quibbles, but rather suggestions for future research. As Jonathan Sullivan puts it in the blurb, “Sometimes you read a book where all you can do is hold your hand up and recognize that this is as good as it gets.” This accurately describes what I feel after reading the book. The mix of robust theory, rich data, and careful analysis makes Censored an important work and a new standard for scholarship. Researchers, policymakers and all observers of authoritarianism, media/Internet politics, and China should read the book.
