The Washington PostDemocracy Dies in Darkness

Study: China’s ‘Great Firewall’ may not actually isolate Internet users

Placeholder while article actions load

Everyone from Hillary Clinton to Amnesty International has bashed China’s “Great Firewall” as an impediment to free speech and democracy. That’s what makes a new study out of Northwestern University both odd and intriguing: According to its authors, media researchers Harsh Taneja and Angela Xiao Wu, Chinese censorship actually has little impact on what people there read online, and Chinese Internet users aren’t particularly isolated, even vis-a-vis users in countries with unrestricted access.

What’s more, they essentially argue, the Internet isn’t that grand, global community that connects and equalizes everyone. Instead, users self-select the cultural communities and content they’re most comfortable with -- rather like real life.

Claims that firewalls like China’s isolate users, they write, “are based on an assumption that Internet users would use all websites if given access. On the contrary, a large body of research on global cultural consumption shows that audiences prefer products that are closer to their culture, even when they have access to products from abroad.”

But let’s step back for a moment and consider the methodology behind these surprising findings. Taneja and Wu came to their conclusions by analyzing data from the Internet’s 1,000 most popular Web sites, according to traffic data from ComScore. Using this data, they looked for Web sites with audiences that overlapped more than was typical: Sina Weibo and Renren, the Chinese Twitter and Facebook equivalents, would be expected to overlap, for instance. English-language versions of those social networks would presumably overlap, as well.

Those overlaps gradually surfaced a number of interconnected clusters, each corresponding to a corner of the Internet most popular among a certain group. The researchers call these clusters “culturally defined markets,” or CDMs, and conclude that they generally correspond to languages or geographic areas.

Nothing to this point should seem surprising; after all, the Internet is a big place, and plenty of researchers have demonstrated already that surfing habits vary based on what language you speak and where you live. (You’ll remember this map of which countries tend to e-mail each other most, which found -- surprise! -- that countries with cultural or geographic similarities often had the most to send.)

The unusual part of this study is that, when Taneja and Wu went on to measure the isolation of each CDM, they found that China was no more isolated than any other group. In fact, the opposite was true -- despite the state’s blocking of certain popular foreign Web sites, including social networks and media sites like, Chinese users were more connected to the greater Internet than people in Turkey, Vietnam, Italy and Poland.

The researchers conclude, essentially, that even if they could access these sites, few Chinese would choose to; that, in other words, “rather than assuming that people use indiscriminatingly [sic] all the websites they have access to, we believe that they mainly prefer culturally proximate websites on the global Internet” -- i.e., Renren and Weibo to the blocked Facebook and Twitter.

The conclusion is a bit of a bombshell, given popular wisdom on the Great Firewall (and the great myth of the Internet as a borderless, democratizing utopia). At times, the authors seem a little too forgiving -- even glib -- about the state of Chinese censorship. At one point, for instance, they argue that “only a small number of foreign sites are blocked,” when their own research found that one of every 10 sites they tested was not available in China. And their attempts to reduce the Great Firewall to mere politics on the part of pesky Americans seems to underestimate the principle of the thing -- the idea that Chinese users should have the choice to access the greater web, even if they (like us) would usually choose to access local sites.

Still, Taneja and Wu make a compelling point that state censorship of messages within the Chinese Internet, such as government filtering on Weibo, could ultimately impact freedom of speech far more than the highly-hyped content blocking. Basically, Chinese Internet users would be more likely to encounter that material, so its absence is more conspicuous. Considering the volume of Chinese censorship, it would be conspicuous anyway: A recent study found the government takes down 12 percent of all messages sent on Weibo, with an eye for political subjects.

“Compared to removing the [Great Firewall] of China, on which most policy, popular, and scholarly discourse tends to concentrate, battling against content censorship over domestic websites may bring about much more substantial changes in what Chinese people make use of on the Internet,” they write.

But it may be harder for outside advocates to fight that battle between China and Chinese Web sites -- particularly since Chinese Internet users don’t really seem to mind. A poll by the Pew Internet & American Life Project found that more than 80 percent of Chinese users think the Internet should be censored, and that the government should be the one to do it.