The Nones Project: Marriage

One of the biggest societal changes in the United States over the last fifty years is the simple fact that a growing share of Americans just don’t get married. It’s something that Derek Thompson wrote about in The Atlantic in February (America’s ‘Marriage Material’ Shortage). He links that, at least in part, to the fact that the average American is significantly less social today than their counterparts from just a couple of decades ago — it’s hard to find a spouse if you never leave the house.

Of course, someone reading this article is going to say, Who cares if people get married or not? Well, the data says unequivocally that married people are just happier than the unmarried portion of the population. But there’s a pretty significant knock-on effect when the marriage rate drops — the fertility rate declines as well. And, from a purely financial perspective, having fewer babies is going to make it a lot harder to balance budgets and take care of senior citizens in countries all over the world.

What does the marriage rate look like in the United States? The General Social Survey has been asking about marital status since its inception back in 1972. Respondents are given a range of options (married, widowed, divorced, separated). But the last option is “never been married.” Here’s the share of the sample who have been permanently single over the last 50+ years.

In the 1972 sample, singleness was pretty rare — just 15% of the population. But from that point forward you see the kind of trend line that makes every demographer smile in satisfaction — slow and steady in essentially one direction. There are no “inflection points” in this graph – it didn’t even spike (or drop) during COVID, like so many other social measures. The share of respondents who have never walked down the aisle just creeps up consistently over time. By 1990, it was above 20%. It reached 25% by 2010, and in the most recent data, about 31% of adults who took the General Social Survey had never been married.

The rate of singleness in the United States has more than doubled in the last half-century. And that’s happening as the age of the average American rose from 44 in the 1972 data to 48 in the most recent survey. So, even as folks are getting older, the share of adults who have never been married continues to rise.

Why? There are myriad of reasons — social atomization, the collapse of institutions, polarization. But our Making Meaning survey had a short section in which we tried to figure out if the rapid rise of the nones in the United States may be one of the proximal causes of the rise in singleness.

Let’s be honest about this — every religious tradition in the West encourages young people to get married, be fruitful, and multiply. I always have this sneaking suspicion that a lot of non-religious people just want to go a different direction in life. They enjoy deviating from the “normal” path, and being intentionally single in middle age would certainly be part of that.

So we asked our sample a question about their marital status using basically the same setup as the General Social Survey. I just restricted the sample to those who were at least 40 years old because I figured that if someone was going to tie the knot, they would have probably done it by then.

We do find some evidence here that non-religious people are more likely never to be married after 40 compared to Catholics and Protestants. I was kind of surprised to see how atheists, agnostics, and nothing-in-particulars were pretty tightly clustered in our sample. About a quarter of each have never been married, and the differences between those three groups are not statistically significant.

In contrast, our survey does seem to point toward the fact that Christians are more likely to marry in the front half of their lives. For Catholics, only 18% had never been married among those who were at least 40 years old. For Protestants, it was slightly lower at 16%. But from a purely statistical perspective, there’s no difference.

So, we can say pretty clearly that Christians were more likely to get married than the non-religious. But the differences between those two groups are not that large. Among the non-religious who were at least forty years old, about 24% had never been married. It was 17% among Christians in the same age category.

Let me show you this another way — here’s the entire range of responses to the marital status question. I divided the sample into Christians and atheists/agnostics and then into five age buckets.

I like this graph because you can really see when marital status begins to shift between these two groups. For the youngest adults in the sample (18–29), the differences are really minimal. The Christian group is about six points more likely to be married compared to the atheist/agnostics, but it’s really the only major difference. Even when you look at the folks in their thirties, there aren’t these huge gaps. Christians are four points more likely to be married, and atheist/agnostics are about three points more likely to have never walked down the aisle.

But when you consider respondents in their forties, that’s when the gaps really start to open up. For Christians, 58% were currently married, 23% had never been married, and 7% were divorced. For the atheists/agnostics, 41% were currently married, 33% had never been married, and 14% had been divorced.

What’s interesting to me, though, is the folks in their forties clearly stand out compared to the other categories. When comparing people in their fifties, Christians are about seven points more likely to be married and seven points less likely to have never gotten married compared to the atheists/agnostics — a gap about half the size of the previous age cohort. And when I looked at the oldest adults in the sample, the gap between the two groups is very negligible.

However, we can be even more granular in this regard. We asked everyone who had been married at least once, “How old were you when your first marriage began?”

I’ve got to admit that I was pretty surprised to see these results because of how small the differences are between each of these five groups. Agnostics got married at the most advanced age — 26.2 years. But atheists were right behind them at 26 years. The Roman Catholics and the Protestants were not statistically different from atheists or agnostics at 25.9 and 25.6 years old, respectively.

In fact, the only comparison group where the mean age of first marriage was statistically different was between agnostics and nothing-in-particulars. And even then, the actual difference in age was less than a single year.

But we have to keep something in mind here from a methodological perspective: we only asked people who had been married this question. Which means that there’s a self-selection effect. As we have seen previously, the nones tend to be less likely to get married than Christians. However, when a non-religious person does choose to get married, they tend to make that decision right around the same time as a Catholic or Protestant.

Just to check that out more fully, I calculated the share of each group that had their first marriage at 40 years old or later. It was between 4% and 6% for all five groups, and the differences weren’t statistically significant.

But then I had another thought — Catholics and Protestants tend to be older than the non-religious. So maybe what we are seeing above could be better understood if we broke the sample down into generations.

Again, I’m struck by how few differences we actually see across these five different groups. In fact, there’s not a single generation where we can clearly say that non-religious people get married older (or younger) compared to Protestants or Catholics. In every single generation, the actual range of values does not vary by more than 1.1 years. Even if the differences were statistically significant, that’s certainly not a substantive difference.

I also wanted to point out that there’s just not a whole lot of variation when comparing one generation to the next, regardless of religious affiliation. Here’s the mean age for each generation:

Silent: 24.5
Boomers: 25.9
Generation X: 26.9
Millennials: 25.7

In this sample, at least, there’s not a ton of evidence that younger generations are choosing to get married at a much older time in their lives. The average American gets married for the first time in a window between 25 and 27 years old.

We did have one more question about marriage in the survey: “I would prefer to marry someone who has the same view of religion as I do.” Notice how we didn’t say, “the same religion.” We wanted to make this more about a general perspective on matters of religion.

I didn’t really know what to expect when we put this in the field. I am aware of the concept of “homophily,” which is the idea that people tend to associate with folks who have the same view of the world as their own. In this context, we would expect Christians to want to marry Christians, and the non-religious would gravitate toward folks who aren’t part of a faith tradition.

The data indicates that Protestants are clearly the most likely to want to marry someone who has the same view of religion as they do — two-thirds agreed with this statement. In contrast, only 59% of Catholics wanted to “marry someone who has the same view of religion as I do.”

But I was interested to see that even the non-religious were inclined to marry someone with a similar view of religious matters. Among atheists, 56% agreed with the idea of marrying someone like-minded. It was 50% for agnostics as well. And it definitely tracks that the nothing-in-particulars are at the bottom of this graph — they are folks who just don’t have strong feelings about religion either way, so it makes sense that they aren’t strongly committed to the idea of marrying a person with a similar view.

There’s a lot here. So some takeaways:

1. The rate of singleness in the United States has doubled since the early 1970s.

2. Non-religious people are more likely to never be married later in life compared to Christians, but the differences aren’t that large (about 7 percentage points).

3. When the nones do choose to get married, they make that choice at almost exactly the same age as Protestants and Catholics.

4. People (religious or not) tend to want to marry someone with a similar perspective on matters of faith.

So, is the rise of the nones leading to the decline of marriage in this country? After looking at all this data, my opinion is that it might be driving the rate of singleness up a marginal amount. But I’m not convinced that a return to faith will also lead to a return to marriage.

Code for this post can be found here.

---

Ryan P. Burge is a professor of practice at the Danforth Center on Religion and Politics at Washington University.