A common complaint of recent years is that social networks whose revenue is based upon segmenting users and delivering both adverts and content specifically tailored to their tastes has resulted in echo chambers emerging online that have contributed to the polarization of political discourse.

A new study from the University of Pennsylvania suggests things might not be as bad as we feared however.  It suggests that even in the most politically homogenous groups, the collective intelligence of the group can increase.  The findings run counter to the perceived wisdom that homogenous groups merely amplify existing biases rather than make the group smarter.

Previous research has shown that social networks can reduce polarization and improve the collective wisdom of those groups when they’re properly designed.  Indeed, a properly designed social network can even improve understanding of contentious topics within an echo chamber.

Smart crowds

The researchers conducted an experiment whereby volunteers had to answer a series of questions that were designed to stir up their political leanings.  They were then placed into groups of likeminded peers, which in this instance were either fellow Republicans or Democrats.  At this point, they were given the chance to change their answers based upon the answers of their peers.

The data showed that even in homogenous groups, the answers were changed to become 35% more accurate than they were originally, even whilst at the same time they became more similar to their peers.  It emerged that they also became more similar to those of opposing views, thus meaning that both groups tended to converge towards the truth.

“Our results cast doubt on some of the gravest concerns about the role of echo chambers in contemporary democracy,” the authors explain. “When it comes to factual matters, political echo chambers need not necessarily reduce accuracy or increase polarization. Indeed, we find them doing the opposite.”

It should be said that this was conducted in an experimental environment whereby participants were incentivized to be accurate, which is probably not the case in normal social networks, even when discounting the role of those who deliberately strive to misinform.  Nonetheless, it does show that social networks don’t have to be the negative force they appear to be today.

“Many political theorists and practitioners have advocated for the value of deliberative democracy, which has as its cornerstone the ability to learn from one another,” the authors conclude. “But there’s been a longstanding question of whether deliberation actually works as intended. This study is a good step toward answering that question.”