The internet is often hailed as a liberating technology. No matter who you are or what kind of country you live in, your voice can be amplified online and heard around the world. But that assumes that people can get on the internet in the first place. Research has shown that poverty and remoteness can prevent people from getting online, but a new study out today also shows that just belonging to a politically marginalized group can translate to poorer access. The study, published online today in Science, provides the first global map of the people being left behind by the internet revolution.
Mapping the internet is hard. Although it is true that every computer with a connection has a real-world location, no one actually knows where they all are. Rather than being organized top-down, the world's computers are connected to each other by a bushy, redundant network of servers. Each country builds and maintains its own infrastructure for connecting citizens to the wider internet. The decision to expand and maintain the infrastructure in one region and not another is up to those in power. And therein lies the problem: Ethnic and religious minorities who are excluded from their country's political process may also be systematically excluded from the global internet.
The new study began as an examination of protest movements and how they depend on internet connectivity, says lead author Nils Weidmann, a political scientist at the University of Konstanz in Germany. "I have fine-grained protest data, but no fine-grained estimates for internet penetration." So he struck up a collaboration with computer scientists at the science and technology university ETH Zurich in Switzerland who were working on a solution to exactly this problem. Their method worked so well that he realized it could be used to map the internet access gap across the world.
Their method requires no travel at all. Instead, they passively sample the internet by collecting information from the packets zipping through the servers of a single internet service provider in Switzerland. Like pieces of mail going through a postal system, internet packets are stamped with "from" and "to" addresses. Those so-called internet protocol addresses include something like a zip code, the "subnets" that provide access points to actual people.
The ETH Zurich team scoured a massive sample of internet traffic between 2004 and 2012, finding more than 19 million unique active subnet addresses. To find where in the world those internet connections are, they used a commercial database commonly used by researchers to geolocate online users. Then they laid this internet connectivity map on top of a map of the world's ethnic groups to find out who is missing out.
The good news is that internet access generally improved for everyone over those 8 years. But it improved far more for those living in democratic countries, and some groups fared far better than others. The places where politically excluded groups live tended to be the very places that internet connections were the most sparse. The net result is that politically excluded groups had 30% less internet connectivity compared with others in their own country. And for the most excluded groups, that number plummeted as low as 1% (see ranking, above). The team performed a statistical analysis of other factors—general infrastructure development, poverty, distance to the capital, terrain ruggedness, and urbanization—and none could explain the gap, leading them to conclude that social policies are to blame.
The results are "contrary to a lot of the rhetoric you hear about liberation technology," says Marshall Burke, an economist at Stanford University in Palo Alto, California. He wonders how crucial internet connectivity is for political empowerment, particularly in the developing world. "Perhaps all you need is a few people with internet access to learn about a particular grievance." So the next step is to measure the impact of the internet gap using natural experiments. If people's access to online computers influences the outcomes of elections or the frequency of political riots, for example, then it would show that the old adage is truer than ever: Information really is power.