Chinese government censorship of social media constitutes the largest selective suppression of human communication in the history of the world. Three approaches are taken to learn about this system. First is an observational study where millions of social media posts are downloaded before the Chinese government can read and censor those they deem objectionable. Second, to make causal inferences, a large scale randomized experimental study is conducted. And finally, for descriptive inferences, the current approach of conducting confidential interviews is supplemented by setting up a own social media site in China, contracting with Chinese firms to install the same censoring technologies as existing sites, and reverse engineering how it all works. Results offer unambiguous support for, and clarification of, the view that criticism of the state, its leaders, and their policies are routinely published whereas posts with collective action potential are much more likely to be censored. This approach also clarifies the internal mechanisms of the Chinese censorship apparatus and show that local social media sites have far more flexibility than was previously understood in how (but not what) they censor. This talk is based on two papers, available at http://bit.ly/CensorObsn and http://bit.ly/CensorExpt.