An ambitious new project, the American National Social Network Survey, looks at the social networks of thousands of Americans over time, exploring how who people associate with affects their worldview.

The project is sponsored by AEI and the Knight Foundation, carried out in partnership with NORC (who administer the General Social Survey).

The latest report from the project, “Socially Distant: How Our Divided Social Networks Explain Our Politics,” highlights how polarized Americans have become from those who hold different political views and provides insights on how forging relationships across partisan lines changes our orientation towards politics.


Americans overwhelmingly choose to surround themselves with others who share the same political views as themselves. For Democrats and Republicans alike, roughly three-quarters of their core networks are co-partisans, with less than a fifth of relationships crossing partisan lines. Independents are split pretty evenly in terms of associating with Democrats versus Republicans:

Moreover, the more strongly one identifies with a particular party or ideology, the more likely it becomes that one’s core social network is comprised entirely of people who share one’s own political leanings:

Consistent with a good deal of other research, the American National Social Network Survey report finds that negative affect across party lines seems to be very high, and that negative partisanship seems to play a growing role in political decision-making.

However, inter-political friendships were shown to help mitigate these trends. Having a politically diverse core social network corresponded with a 12-15 percentage point reduction in highly unfavorable views about the other party.

Moreover, people with politically diverse social networks also seem to possess more epistemic humility with respect to their own political views. Less than half (47%) of Americans with politically homogeneous networks say they sometimes doubt whether their political positions are correct. The number is 10 percentage points higher (57%) among those with politically-diverse social networks.

People with more diverse social networks are also far more open to voting for the other party’s political candidates:


One question that may be nagging the reader (which was nagging me as I read the report) is the question of what causes these patterns? Is it that people who are naturally more open-minded, comfortable with disagreements, less vitriolic, etc. are just more likely to forge relationships across political divides? Or is it the networks themselves that push people towards greater humility and away from hostility?

The report provides some clues to these questions. For instance, it does not seem to be the case that people who form inter-political relationships are significantly more comfortable with disagreement. Nearly the same percentages of respondents with politically diverse networks (57%) versus political homogeneous networks (60%) find it to be more stressful and frustrating than informative or interesting to engage with people of the other party.

Nonetheless, people with politically-diverse social networks are much more likely to report having political disagreements than people with more homogeneous social networks (71% versus 56%) — and are much more likely to report being criticized for their views as well:

That is, although they find political arguments just as unpleasant as people with more homogeneous networks, the relationships of people with politically diverse networks nonetheless expose them to more disagreement.

Regardless of social network composition, people tend to walk away from such encounters feeling as though they have less in common, rather than more in common, with their interlocutors (60% among those with diverse networks feel greater distance after a political argument compared to 63% of those with more homogeneous networks).

Nonetheless, it seems as though the increased prevalence of these encounters may help soften political positions. People with more people of the other party in their core social networks are significantly more likely to report doubting their own views as a consequence of these disputes:

In other words, although inter-political relationships are often raucous and messy, sometimes unpleasant, they do seem to moderate positions, increase openness and soften hostility. These findings are very much in line with research by Interfaith Youth Core’s IDEALS team on the effects of interfaith friendships.


There is abundant research suggesting that more intelligent and highly-educated people tend to be more ideological, extreme and dogmatic in their views than most. One open question is whether or not universities are simply filtering the population for these people, or whether these tendencies are instead inculcated into students as a result of their education.

Here, the American National Social Network Survey provides some good news: it appears that education may help moderate views rather than polarize people. The more education people had, the more likely they were to express political self-doubt.

Social networks may play a significant role here. Americans with a college degree were less likely to experience social isolation (one person or fewer in their core network), and were much more likely to have four people or more in their core networks. Education was also correlated with an increase in inter-political relationships – at least for Republicans.

Higher education is associated with a 19 percentage point drop (!!) in the share of Republicans who have homogeneous networks (64% of Republicans without college education have politically homogeneous networks, as compared to 45% of Republican college graduates). Most college-educated Republicans have at least one person in their core network who supports the other party, whereas most college-educated Democrats (54%) have completely homogenous networks. Indeed, education seems to have no effect on increasing political diversity among Democrats’ core networks.

The political skew of institutions of higher learning probably plays a significant role in these disparities. At college, students from conservative political backgrounds are consistently forced to confront views that run contrary to their own, and are surrounded by – and form relationships with — many more people who hold different perspectives from themselves. For Democrat-leaning students, this is much less the case.

It may therefore be possible to enhance the positive effects of a college education – and thereby ameliorate America’s civic crises — by increasing political diversity within institutions of higher learning and facilitating inter-political relationships among students on campus.


The full report contains much more information on Americans’ social networks – including their racial diversity, and how inter-party and inter-racial relationships affect perceptions on race, inequality and social justice. It also includes a battery of questions about the state of the contemporary U.S. and the upcoming elections.

I will be giving some remarks, and participating on a panel, for a launch event introducing the project on Tuesday, October 13th, from 10:30 am – noon EDT. I’ll be joined by HxA member and political scientist Sam Abrams, Camille Busette of the Brookings Institution, plus Daniel Cox and Ryan Streeter from AEI. The event is free and open to the public.