Echo Chambers

by The New Republic Staff | January 26, 2007

by Cass Sunstein

In recent years, many people have been concerned about the risk that citizens will use the Internet to construct echo chambers, or information cocoons, in which their own views are constantly confirmed or reinforced. A real problem with echo chambers is that those who live in them tend to become more extreme. In an echo chamber, people are likely to hear arguments that support their antecedent views, and such support is a recipe for extremism. The underlying phenomenon, well-known in social science, is called "group polarization," which refers to the tendency of like-minded people, engaged in deliberation with one another, to end up in a more extreme position in line with their antecedent tendencies. (Corporate boards, labor unions, students organizations, and future presidents beware.) The literature on group polarization is more illuminating than the corresponding literature on "groupthink," a phenomenon that is less well-defined, and that offers much less in the way of testable hypotheses. There is now a website,, that is specifically and unself-consciously designed to enable people to create echo chambers. In its own cheerful words, "NewsTrove employs patent-pending neural lingual technology to filter viewpoints, making it as easy as a single mouse-click to focus news and blogs to sources that agree with your viewpoint. NewsTrove filters out the rest." You can choose, for example, Liberal, or conservative, or Jewish, or Islamic, or Evangelical, or Military, or Catholic. It's not clear that the "neural lingual technology" is yet perfected. My various searches for the Chicago Bears turned up only a Canadian article on Soldier Field; and I couldn't find a single item for Barack Obama under the Evangelical, Catholic, or Jewish categories. (Maybe NewsTrove has a small data base thus far, or maybe its filters are too powerful, or maybe I couldn't figure out how to use it.) But the much broader question, on which the jury is still out, is the extent to which is merely describing what many people are already doing, even without the aid of "neural lingual technology."

Source URL: