Category: Web/Tech

It’s long been known that the Chinese government hires people to support the government with fabricated posts on social media. In China these people are known as the “50c party”, so called because the posters were rumored to be paid 50 cents (5 jiao or about $.08) to write the posts. The precise nature and extent of the 50c party has heretofore been unknown. But in an amazing new paper, Gary King, Jennifer Pan and Margaret Roberts (KPR) uncover a lot of new information using statistical sleuthing and some unusual and controversial real world sleuthing.

KPR’s data-lever is an archive of leaked emails from the Propaganda Office of Zhanggong. The archive included many 50c posters who were sending links and screenshots of their posts to the central office as evidence of their good work. Using these posts, KPR are able to trace the posters though many social media accounts and discover who the posters are and what they are posting about. Both pieces of information reveal surprises.

First, the posters are government workers paid on salary not, as the 50c phrase suggests, piece-rate workers. Second, and more importantly, it has long been assumed that propaganda posts would support the government with praise or criticize critics of the government. Not so. In fact, propaganda posts actively steer away from controversial issues. Instead, the effort appears to be to distract (especially to distract the people from organizing collective action; thus distraction campaigns peak around times and places where collective action like marches and protests might become focal). KPR write:

Distraction is a clever and useful strategy in information control in that an argument in almost any human discussion is rarely an effective way to put an end to an opposing argument. Letting an argument die, or changing the subject, usually works much better than picking an argument and getting someone’s back up…

Debate is about appealing to an individual’s reason; debate is thus implicitly individualistic, respectful of rights and epistemically egalitarian. (As I argued earlier, respect for the truth is tied to individualism because any person may have truth and reason on their side.) Authoritarians don’t care about these things and so they lie and distract with impunity and without shame. In this case, the distraction is done subtly.

From the initial archive, KPR are able to create a statistical picture of 50c posters. In one of the most remarkable parts of the paper they use this picture to identify many other plausible 50c posters not in the original archive. Then KPR test their identification with a kind of academic catfish–essentially they trick the 50c posters into self-identifying. It’s at this point that KPR’s paper begins to read more like the description of a CIA op than a standard academic paper.

We began by creating a large number of pseudonymous social media accounts. This required many research assistants and volunteers, having a presence on the ground in China at many locations across the country, among many other logistically challenging complications. We conducted the survey via “direct messaging” on Sina Weibo, which enables private communication from one account to another. With IRB permission, we do not identify ourselves as researchers and instead pose, like our respondents, as ordinary citizens.

Using their own fake accounts, KPR directly message people they think are 50c posters with a message along the lines of:

I saw your comment, it’s really inspiring, I want to ask, do you have any public opinion guidance management, or online commenting experience?

The question is phrased in a positive way and it uses the official term “public opinion guidance” rather than the 50c term which has a negative connotation. Amazingly, 59% of the people KPR identify as 50c posters answer yes, essentially outing themselves.

Now, one might wonder whether such a question has evidentiary value but KPR do a clever validation exercise. First, they ask the same question to people from the original leaked archive, people whom KPR know are actual 50c posters. Second, they ask the same question of people who are very unlikely to be 50c posters. The result is that 57% of the known 50c posters answer the question, yes. Almost exactly the same percentage (59%) as in the predicted 50c sample. At the same time, only 19% of the posters known not to be 50c answer yes (that doesn’t mean that 19% are 50c but rather that 19% is a measure of the noise created by asking the question in a subtle way). What’s important is that the large 40 point difference gives good statistical grounds for validating the predicted 50c sample.

Using this kind of analysis and careful, documented, extrapolation, KPR:

…find a massive government effort, where every year the 50c party writes approximately 448 million social media posts nationwide. About 52.7% of these posts appear on government sites. The remaining 212 million posts are inserted into the stream of approximately 80 billion total posts on commercial social media sites, all in real time. If these estimates are correct, a large proportion of government web site comments, and about one of every 178 social media posts on commercial sites, are fabricated by the government. The posts are not randomly distributed but, as we show in Figure 2, are highly focused and directed, all with specific intent and content.

As if this weren’t enough, an early version of KPR’s paper leaked and when the Chinese government responded, KPR became part of the story that they had meant to observe. The government’s response is now in turn used in this paper to verify some of KPR’s arguments. Very meta.

It took courage to write this paper. I do not think any of the authors will be traveling to China any time soon.