Evidence of Cyberbalkanization?

This new study by Adamic and Glance suggests that liberal and conservative bloggers are not linking to each other as much as they are linking to themselves, so that as a result the liberal and conservative spheres are polarized.

Is this inconsistent with my previous arguments about the blogosphere? Yes, but only in part. There are two questions: one is whether we will find ideological polarization in the blogosphere. This study says that we will. The second is whether the blogosphere (and the Internet generally) causes or facilitates this polarization, and whether the polarization that it causes or facilitates is substantially greater or more worrisome than polarization that occurs through other mass media. On this second question, the evidence remains mixed. It is still quite possible that linking and the culture of linking creates marginally more exposure to divergent ideas than people otherwise experience in real space, and thus, that it is not a contributing cause of existing political polarization. That is to say, the Internet creates two opposite effects. One is ease of searching for and finding information that confirms what you already believe. That would facilitate and enhance polarization. The other is serendipitious exposure to information that you disagree with or that you weren't looking for. That would work in the opposite direction. The question is which effect dominates the other.

Although this study is quite worrisome, I am not yet convinced that the Internet is a major contributing cause of the polarization we see in American society today. Indeed, I believe that it still holds out the hope of breaking down ideological barriers.