China’s Social-Media Smoke Screen

Illustration by Whooli Chen
Illustration by Whooli Chen

It has long been suspected that the Chinese government, as part of its effort to control the Internet within its borders, surreptitiously floods social media with fake posts written by a vast army of hired promoters posing as ordinary people. The “50-cent party,” it’s called, because each fake post supposedly earns its author 50 cents. 

The phenomenon has been talked and written about widely by journalists, academics, activists, other social-media users, but evidence for these claims has been hard to find—until recently. In a study (to be published this year in the American Political Science Review) that has already prompted a startled response from Beijing, Weatherhead University Professor Gary King, the director of Harvard’s Institute for Quantitative Social Science, confirmed the suspicion: the 50-cent party, he says, is real, although much of the rest of what everyone believed about it is wrong. For one thing, the fake posters likely aren’t paid 50 cents. Most aren’t independent contractors: they’re government employees writing online comments on their off time, and there’s no evidence they earn extra money for it. 

More surprising, the purpose of these fabricated posts is not to argue with other social-media users, but to distract them. To perform the study, King and his two coauthors—Jennifer Pan, Ph.D. ’15, and Margaret Roberts, Ph.D. ’14—analyzed a trove of leaked emails sent between local government offices and the propaganda department in one county in southeastern China. “A big giant mess of a dataset,” King recalls, from which the researchers harvested nearly 44,000 fabricated social-media posts from 2013 and 2014. Across all of China, they calculated, that suggests about 450 million posts per year. In those King and his team read, 50-cent party members “are not arguing with anybody at all,” he says. They don’t jump into fights when other users complain about the regime’s repressions or corruption among local officials. 

Instead, they change the subject. “They’ll say, ‘I woke up this morning and thought about how important our martyrs were to the history of China,’” King says. “Or, ‘What a beautiful day it is today.’ Lots and lots of these—and not just randomly. They’ll post them in big bursts when they need them.” King’s team found large batches of fake posts turning up around the same time as crises, holidays, and other events that might stir up public action: the Shanshan riots in June 2013, the Urumqi Railway explosion in April 2014, Martyr’s Day, Tomb Sweeping Day, Communist Party meetings to discuss national policies. “It’s almost like when you’re having an all-out fight about something with your spouse or your kids,” King points out, “and you want to end the argument, and so you say, ‘Hey, why don’t we go get ice cream?’” 

This finding—that 50-cent party members are less interested in controversy than in cheerleading—fits with King’s previous research on China’s social-media control (see, in which he found that the government would ignore comments disparaging the regime or local leaders, while posts about organizing protests, or even pro-government rallies, were invariably censored. “They don’t care what you say or what you think,” King says. “They only care what you can do. They don’t want people in the streets.” 

Last spring came an unexpected twist, when a Western reporter got hold of an unfinished draft of the 50-cent party research paper and called King with some questions for an article. King answered them and then, realizing that his research would be going public ahead of schedule, posted the paper on his website. The reporter published his article, and about an hour and a half later, another publication picked up the story; 72 hours after that, some 5,000 articles had appeared worldwide. 

That’s when the Chinese government responded. In an editorial in the pro-government Global Times, the regime “for the first time admitted the existence of the 50-cent party,” King says, and attempted to explain to its citizens the reason for this “‘public opinion guidance,’ which is their term of art for information control.” Basically, the government argued that without such control, the country would fall into strife and chaos. “And,” King adds, “they said that the Chinese people are in agreement about the necessity of this public opinion guidance.” 

As it happens, that was an assertion King could check. After the international blizzard of attention, there was enormous discussion on Chinese social media about the paper and the government’s answer to it. “So we downloaded all the posts commenting on it,” King says. The finding? He smiles. “Well, it turns out that the Chinese government’s claim in their editorial is incorrect. Eighty percent of the people, at least on social media, think it’s not a good idea to be censoring and fabricating posts.”

Read more articles by: Lydialyle Gibson

You might also like

“The Promise of This New Presidency”

Harvard officially installs Claudine Gay, its thirtieth leader.

Installation Academics

A half-dozen symposiums feature Harvard research on AI, climate change, inequality, and more

Performing for the President

A musical and poetic “prelude” to Claudine Gay’s installation festivities

Most popular

The Speeches 2022

President Bacow on humility, Prime Minister Jacinda Ardern on democracy and social media, and Attorney General Merrick Garland on civic engagement and threats to U.S. democracy

Victor Clay

The new police chief introduces a new policing culture.

Off the Shelf

Recent books with Harvard connections

More to explore

Picking Team Players

A test can identify these productivity-boosting personnel.

Irene Soto Marín

Ancient history professor studies coins, ceramics, and Zelda.

Getting His Reps in

Anwar Floyd-Pruitt’s wildly profuse art