Our tendency to surround ourselves with people who have beliefs or backgrounds similar to our own is well-established in social science, and what’s more, it is not inherently a bad thing. It would be difficult, for instance, to pass on our religious values and traditions or to maintain a shared cultural heritage if our social groups were completely random. Historically, this has had benefits, too, with some research finding that people with politically like-minded friends and family are more likely to participate in politics. But our recent survey casts doubt on how true that still is, finding that Americans with politically uniform social networks are not any more politically active than those with less uniform connections.

It’s true that interacting with people who disagree with us politically can be difficult. In our social network survey, a majority (57 percent) of Americans say it is a stressful and frustrating experience to talk about politics with people who disagree with them. But some of this anxiety may have to do with the fact that we’re just really bad at understanding our political opponents. A study from 2015, for instance, found that Americans routinely overestimated just how divided the Democratic and Republican parties were. Across a range of issues, such as taxes, immigration and trade, people perceived their political opponents as being more extreme than they actually were. And a 2018 survey of 2,100 Americans conducted by More in Common,1 an international think tank, revealed that Democrats and Republicans widely misunderstood the views of the opposing party on basic questions of race, immigration and gender equality. This misperception even extends to how we think about our political opponents’ wealth, sexual and gender identity, religious beliefs or racial identity. A 2017 study found that Americans held consistently biased perceptions about the type of people who identified as Democrat or Republican.