It is a truism that politics makes for strange bedfellows, but there is evidence that it also makes for normal bedfellows.
Research shows that married couples on average share similar political beliefs.
Political scientists and sociologists have sought to understand what drives this homogeneity. Do people seek partners who have similar political beliefs? Do couples’ political views coalesce over time? Are shared politics a side effect of other factors, such as shared religious beliefs?
A recent article in the Journal of Politics by Gregory Huber, Yale professor of political science, and Neil Malhotra, a professor of political economy at Stanford University, offers fresh insight into these questions.
They conducted two studies — one involving a survey using manipulated online dating profiles, and another using a trove of data from an online dating service —that measure people’s attitudes before they form relationships. The researchers found evidence that people are more inclined to seek dating partners who have similar political characteristics as them but that other factors, such as religion or race, are more significant in determining relationships than political similarity.
Huber, a resident fellow of the Institution for Social and Policy Studies and the Center for the Study of American Politics, recently spoke to YaleNews about his work. An edited version of the conversation follows.
Your article covers two studies, the first of which was based on a survey experiment. What did the survey involve?
In the first study we took real photos and profiles from online dating sites and randomly manipulated the religion and politics expressed in those profiles. Then we showed approximately 1,000 individuals a series of these manipulated profiles and asked them their interest in dating each person, whether they shared the individual’s values and whether the person was attractive to them.
We found that — even though politics is just one of several characteristics displayed in the profile — whether or not they shared politics with the person in the profile affected their level of interest in dating the person. That effect is substantial but not overwhelmingly large. People seem to generally prefer, and rate as more attractive dating partners, those who share their political characteristics.
How did political orientation compare to religious orientation in driving people’s interest in potential dates?
Religion matching is very important. Catholics want to date other Catholics. Jews want to date other Jews, and so on. That effect is actually quite a bit larger than the political effect, which is still reasonably significant.
Interestingly, disinterest in politics has an effect. People who aren’t interested in politics are not that excited about dating people who are really interested in politics. If you know people who are not interested in politics, then this strikes me as completely accurate.
The second study analyzed data from an existing online dating site. What kinds of data did you have access to?
The second study is in some ways the more novel of the two. We worked with an online dating service, which provided us access to the actual behaviors of the site’s users. The advantage of this is that people are not just answering a survey question but investing their scarce energy in trying to find dates.
At our request, the online dating company included seven questions about politics in the questions the site asks people. Then we observed the frequency with which people reached out to potential dates — the term used is “messaged.” We also observed the frequency with which they received responses on the basis of shared or not shared political orientations. We analyzed data from about 143,000 men and nearly 120,000 women.
What’s unique about the online dating data?
It’s really rich data that wasn’t possible in a large scale in prior research.
From a research perspective, this study is interesting because, while marriages on average are quite alike in a lot of characteristics, we often don’t know why they’re alike. If two white, evangelical Protestants marry, they may be both Republican, but they may not have started dating because they are Republicans, they may have started dating because they have a shared ethnic and religious orientation.
Usually we study couples after they have formed relationships so we can’t figure out whether they are alike because have learned to get along over the years or developed political preferences according what they perceive is best for the household.
The online dating data allowed us to see relationships before they formed. Is this person a liberal before he’s reaching out to a woman, and is that woman a liberal before the man reached out to her? It’s an unusual feature of these data: You can look at relationships as they are forming.
That’s the equivalent of being plopped down on Yale’s campus when the freshmen arrive and knowing everyone’s political views, and four years later seeing who became couples, comparing that to what we knew about them before they became a couple, and determining which characteristics explain who ended up being a couple.
What did you learn?
The online data provided evidence that at the earliest stage of dating, people are looking for potential partners who are like them politically. Even when you account for a lot of other characteristics on which people choose dating partners, people seem to be more likely to reach out to people who have a shared political orientation.
The term for this is “homophily.” It’s the Greek word roughly for “love of self.” It’s a widespread phenomenon that people are attracted to and find beauty in things that are like them: height, skin color, religion, all sorts of things. Politics seems to be one of those things that people are conditioning their social relationships on.
As with the first study, politics is by no means the biggest factor in how we select partners. To be blunt, the biggest factor in online dating is age. Men want to date women younger than them and women want to date men older than they are. That effect is gigantic. Race has a very big effect. Education level has a substantial effect. The effect of political orientation is not on the same level as those other things, but it is still a factor.
It seems that a conservative in a liberal place like Manhattan would have a limited pool of other conservatives to date. What role does geography play in choosing dating partners?
Our online dating research shows that a lot of sorting depends on where you live and whether the available partner market is relatively restricted. For example, if I had a friend who lived in Salt Lake City and was not a Mormon, you might imagine that the dating pool would provide less choice with regard to religious beliefs. You can imagine that there are a lot of parts of the country where marriages or dating couples were on average composed of two conservatives not because people picked conservatives, but because only politically conservative people were available.
Our study can account for a restrictive partner market, and still show that politics affects who people choose to ask for dates. The effects we show are above and beyond those that exist simply due to restrictive partner markets.
What do you consider the most important implications of your findings?
For one, researchers have known that marriages and dating appear to be more homogenous than you would expect by chance, and our work shows that it doesn’t appear that this homogeneity arises only from people selecting on politics. There are a lot of characteristics correlated with politics that cause couples to be alike. I think we should be careful about assuming that because two Democrats end up married on average or two Republicans end up married on average that politics plays an outsized role in how we select a potential spouse.
I think it’s encouraging that these other characteristics, such as age or race or religion, matter more and that people are not, by and large, sorting only based on politics.
Does the study open avenues for additional research?
One area for ongoing research would be to examine if this homogeneity affects how kids are raised. If you’re concerned about polarization in America today, you might want to know: If people are seeking out like-minded partners, does it mean that kids are growing up in homes where there is just one political point of view? What, in fact, these data suggest is that, yes, there is a little bit of that — people try to seek out a partner who shares their political views, but even if they weren’t doing that, it would happen quite a bit because of all of the other factors that drive our decision making.
Those factors are, in some ways, alongside politics but also prior to politics. We don’t quite know if it affects how children are being raised, but it would be concerning if, by and large, kids were being raised in households where they are only exposed to one political orientation.