The information we receive from our friends through social networks and emails is much more homogenous than what we derive from search engines.

That’s the conclusion of a recent study, “Measuring Online Social Bubbles,” published on the open-access journal PeerJ. The findings contradict conventional wisdom that Google would one day control our thoughts and deeds while social networks would allow uncomfortable messages to pierce censorship and unleash a “Twitter revolution.”

Instead, it’s the opposite, say the authors, a research group consisting of Dimitar Nicolov and Diego Oliveira at Indiana University and Italians Filippo Menczer and Alessandro Flammini. They analyzed web traffic filtered by how users arrived on the sites — whether from social media, email or search engines — using data servers of their universities.

They found that the sources of news accessed through social networks and email are much less diverse than those reached via search engines. To measure diversification, the researchers used a formula representing the principle of entropy: the greater the number of possible outcomes, the greater the entropy. In this case, the greater the number of possible links, the greater the diversity of information sources.

The study is the first quantifiable confirmation of the “information bubble,” which theorizes that we are increasingly only exposed to information that resembles and reinforces our own views. The bubbles are the cause of the poor quality of online discussions, from which we all suffer to varying degrees. On social media, we often agree with each other, and furious virtual quarrels erupt if we don’t.

The main consequence of losing information diversity is a reduced perspective; whole chunks of reality disappear from our consciousness. And since it is the social networks that the search engines base their advertising sales, in the end it is mainly advertisers who choose what to show or hide.

These mechanisms are strengthened on social networks, where users form “friendships” based on personal affinities. In addition, Facebook regularly updates its algorithms to show users information that benefits itself. By changing a few variables in the algorithm, it can move hundreds of millions of Internet users to a particular corner of its network.

The biggest losers, in addition to the users themselves, are news sites. Traffic from Facebook to the 30 major news sites dropped by 32 percent during 2015, according to an analysis by Digiday, which covers the latest trends in publishing and online advertising. For the top 10, the decline was as high as 40 percent.

Analysts observe that Facebook has done this intentionally to confine users within its own platform and to persuade news sites to publish their articles directly on Facebook — for a price, of course.

Marketing strategizing is also pervasive at search engines, where Google reigns. But search engines exist to direct traffic to websites and do not have a broader ecosystem and engagement like Facebook does. The purpose of a search engine is still to answer our questions, to satisfy our curiosity. Few would have predicted that the last bastion of pluralism would be Google.

–> Originally published in Italian at il manifesto on Dec. 8, 2015