For a better browsing experience, please upgrade your browser.

Skip to main content
Marriage Isn’t Dead — Yet

Fewer Americans of all stripes are getting married. But beneath that overarching trend lies an important distinction: Some groups are merely delaying marriage, while others are skipping it entirely.

The Pew Research Center last week released a big report on the decline in marriage in the United States. In 1960, nearly 90 percent of 25- to 34-year olds had been married; today, that figure is barely 50 percent.1 The trend is remarkably broad, cutting across race, education, income and other characteristics.2

More Culture


But dig deeper into the numbers and they reveal a hidden divide: Affluent, college-educated Americans are increasingly delaying marriage until their 30s. But they aren’t abandoning marriage altogether; in fact, they appear likely to get married at close to the same rate as past generations. They rarely have children outside of marriage, and they are relatively unlikely to get divorced.

For poorer and less-educated Americans, the story is different. They, too, are getting married later, but many also aren’t getting married at all. They are more likely to have children outside of marriage, are more likely to say they don’t plan to marry, and, when they do marry, are more likely to get divorced.

The marriage divide mirrors other trends in the economy, particularly the stagnation of incomes and the disappearance of jobs for those without a college degree. The trends may be related: For men especially, marriage and employment tend to go hand in hand.

In its report last week, Pew estimated that a quarter of today’s young people — those between ages 25 and 34 — will never marry, or at least still won’t have married by age 55. But as I wrote on Wednesday, that estimate is based on an extrapolation from past generations’ behavior. This is a fundamental problem in all marriage forecasts: At any given moment in time, it’s impossible to say whether an unmarried person will never marry, or just hasn’t done so yet.

There’s no question Americans are getting married later in life. But most still get married eventually, or at least they have so far. More than 70 percent of Americans have been married at least once by age 35, down from more than 90 percent 30 years ago but still a solid majority.3

Moreover, most of the trends in marriage have been fairly steady. So it’s probably reasonable to assume that today’s 25- to 34-year-olds will follow the same general path as those a decade older. Focusing on those two age groups reveals some interesting patterns.

The chart below shows the marriage rate,4 broken down by education. For Americans ages 25 to 34, the rate has fallen faster for less-educated men than for those with a college degree, but there is a clear downward trend for all groups that has accelerated since about 2005.

For people a decade older, however, the story is significantly different. The decline is much steeper among the less educated. Among college graduates, in fact, marriage rates have hardly fallen at all since 2000, and are only modestly lower than they were in 1980.5


Again, it’s possible that this generation will turn out to be different. But there’s one piece of evidence that the pattern will continue: College-educated people aren’t having children out of wedlock.

College graduates tend to wait longer than non-graduates to have children. About half of women in their late 20s and early 30s have children, but among women with bachelor’s degrees, it’s only about one in three.6 But college-educated women who do decide to have children in their 20s are nearly all married: 90 percent, versus about 36 percent for college educated women without children in the same age range. Young college graduates, in other words, still see marriage as something close to a requirement for raising children; that suggests that when they decide they’re ready for children, they’ll get married, too.

The marriage-children link still exists for less-educated Americans, but it isn’t as strong. The marriage rate for young mothers with at most a high school diploma is about 70 percent, much higher than the rate for those without children, but far below the rate for college-educated mothers.


There is also evidence that the divergence in who gets married and when reflects attitudes toward marriage that are also shifting. In a 2010 Pew survey, 45 percent of respondents with at most a high school diploma agreed that “marriage is becoming obsolete,” versus just 27 percent of college graduates. The survey results suggest the division is primarily along socioeconomic lines: There was a similar split by income, while the divisions by race, political orientation and even age were significantly smaller.7

In its report last week, Pew suggested that one reason for falling marriage rates is the decline in employment among young men. That may also help explain the education gap in marriage. Put simply, men without jobs are much less likely to get married, and men without a college degree are much less likely to get jobs.

In the Pew survey, 78 percent of never-married women said it was “very important” for a prospective spouse (in most cases, a husband) to have a steady job. That ranked above any other requirement, including “same moral and religious beliefs” (38 percent), “at least as much education” (28 percent) and even “similar ideas about having and raising children” (70 percent). The survey results are borne out by women’s actual behavior. About half of men ages 25 to 34 with a steady job have been married, compared to just a third of those without a steady job.8

For men without a steady job, having more education doesn’t help much in terms of finding a spouse — marriage rates are nearly identical regardless of education. But having a degree makes men much more likely to be employed — and therefore more likely to get married. According to the Current Population Survey, more than 20 percent of men ages 25 to 34 with a high school diploma are out of work, versus 10 percent of young men with a college degree.9 And when they do have jobs, less-educated men earn less and are more likely to be laid off.

There is also another key distinction between educated and less-educated Americans when it comes to marriage: divorce rates. In 1980, about 10 percent of Americans ages 35 to 44 who had ever been married were divorced, and that rate was more or less the same regardless of education.


Three decades later, the divorce rate for Americans without a college degree has doubled, to about 20 percent. But the divorce rate for Americans with a bachelor’s degree has held more or less steady.10 Once again, a similar pattern exists by income, with the divorce rate rising much less for more affluent families than for poorer ones.

Taken together, the numbers paint a picture of two very different Americas when it comes to marriage. But while the divide is real, it is important not to overstate the decline in marriage for less-educated and poorer Americans. Even among those with at most a high school degree, nearly two-thirds are still married by age 35. For those with children, marriage rates at age 35 top 85 percent. Marriage is declining in the U.S. But it isn’t dead yet.


  1. Pew’s report focused on the share of adults who have never been married, which means all of their charts slope upward. I find that confusing, so in this article I’ll focus on the inverse: the share of the population that has ever been married, whether or not they’re married now. ^
  2. Same-sex couples who are married, or who identify as married, are included here but almost certainly make up too small a share of the total to have a significant impact on the overall figures. Moreover, recent research has called into question the Census Bureau’s data on same-sex marriages, finding that a large percentage of married couples identified in the data as same-sex are actually opposite-sex couples who were misclassified. ^
  3. Unless otherwise noted, all data in this article is from the decennial census (for all years through 2000) and the American Community Survey (years 2001 to present), via the Minnesota Population Center’s IPUMs series, which asks to be cited as: Steven Ruggles, J. Trent Alexander, Katie Genadek, Ronald Goeken, Matthew B. Schroeder, and Matthew Sobek. Integrated Public Use Microdata Series: Version 5.0 [Machine-readable database]. Minneapolis: University of Minnesota, 2010. ^
  4. I’ll use the term “marriage rate” in this article to refer to the share of the population that has ever been married, whatever their current marital status. ^
  5. The same is true, to some degree, by race and ethnicity. For whites and Hispanics, marriage rates are falling much faster for the young (people ages 25 to 34) than for those ages 35 to 44. Black marriage rates are falling steeply for both age groups. We’ll post all our data, including breakdowns by sex and regions, on our GitHub page. ^
  6. The American Community Survey doesn’t directly ask respondents if they have children, only if they have children of their own living at home. As a result, the data doesn’t do a good job distinguishing between non-parents and non-custodial parents. Since women remain far more likely to retain custody of children in a divorce or breakup, I’ll focus on women and use children living at home as a proxy for parenthood. ^
  7. The same survey found that young college graduates were significantly less likely than high school graduates to say they “don’t want to marry.” But there was very little distinction between college graduates and those with some college but no degree. The sample sizes were small, however. ^
  8. I’m defining a “steady job” as having been employed at least 50 weeks the prior year. In a regression analysis, steady employment was more important than any other single factor as a predictor of marriage. ^
  9. This is based on data from June through August. “Out of work” includes everyone who isn’t employed, whether or not they are officially unemployed. ^
  10. I’m using the term “divorce rate” a bit loosely here. Technically, this is the share of adults who have ever been married who are currently divorced. People who divorced and remarried are considered “married” rather than “divorced.” (People who have never married are left out of the calculation entirely.)

    One note: I’m not controlling for duration of marriage, which means it’s possible that the divorce rate is low among college-educated couples because they tend to get married later, and therefore haven’t had as much time to get divorced. But the same basic pattern holds for people ages 45 to 54, who have presumably had plenty of time to have their marriages turn sour. ^

Ben Casselman is FiveThirtyEight’s chief economics writer.

Filed under , ,

Comments Add Comment