Editorial correction — September 20, 2026. The headline wrongly predicted that 25% would never marry. The study measures marriage history at age 40 in 2021, not lifetime outcomes. Undated figures and unsupported claims of continuously accelerating decline have been removed. Expanded September 20, 2026 with additional primary-source data and a Sources list.
One in four American 40-year-olds had never been married as of 2021. That is a record, and it is a real change. It is also one of the most misquoted statistics in the debate over family life, because it describes people at a single age, not a lifetime outcome. This article sets out what the Pew Research Center found, how Census and CDC data fit around it, and which questions each number can and cannot answer.
The finding
Pew’s June 28, 2023 analysis, written by Richard Fry using Census Bureau data, found that 25% of 40-year-olds in the United States had never married as of 2021. The share was 20% in 2010 and 6% in 1980. Pew describes the decline as decades long, with people born from the 1960s onward increasingly delaying marriage and a growing share forgoing it altogether.
| Year | Share of 40-year-olds who had never married |
|---|---|
| 1980 | 6% |
| 2010 | 20% |
| 2021 | 25% |
Source: Pew Research Center analysis of Census Bureau data.
Being unmarried at 40 is not the same as being alone. Pew reports that 22% of never-married adults aged 40 to 44 were cohabiting with a partner in 2022, so most were not living with a partner. Note that this uses a different year and a slightly wider age band than the 40-year-old figure, so it should be read as context and not as a share of the 25%.
Who is more likely to be unmarried at 40
Pew found that the share varied by group in 2021. A higher share of men than women had never married. Black 40-year-olds were much more likely to have never married than Hispanic, White or Asian 40-year-olds. Education showed the clearest gradient, as the table shows.
| Education of 40-year-olds (2021) | Share who had never married |
|---|---|
| High school diploma or less | 33% |
| Some college | 26% |
| Bachelor’s degree or more | 18% |
There is a twist. The share of 40-year-olds with at least a bachelor’s degree rose from 18% in 1980 to 39% in 2021. Since more educated adults are more likely to marry, you might expect that shift to have raised the overall marriage share. It did not: Pew says the growth of that group has not reversed the overall trend of delaying or forgoing marriage. Pew’s short analysis does not identify why, and this article does not guess.
Why “never married by 40” is not “never will marry”
Pew states plainly that we cannot assume someone unmarried at 40 will never marry. About one in four 40-year-olds who had not married in 2001 had done so by age 60. If that pattern held, a similar share of today’s never-married 40-year-olds would marry in the coming decades. That is Pew’s conditional projection, not a forecast, and the 2001 cohort’s experience may not repeat. But it is enough to rule out treating the 25% as a count of people who will stay single for life.
Delay is part of the story
The Census Bureau’s December 2, 2025 release on families and living arrangements, based on the Current Population Survey’s Annual Social and Economic Supplement, adds the timing dimension. The estimated median age at first marriage in 2025 was 30.8 for men and 28.4 for women, up from 23.5 and 21.1 in 1975. When the typical first marriage happens about seven years later for both men and women, a snapshot at a fixed age will show more unmarried people even if many go on to marry later. The 40-year-old measure captures both delay and forgoing marriage, and by itself it cannot separate them.
The wider household picture
The same Census release shows how much household life has changed since 1975.
| Measure (Census CPS ASEC) | 1975 | 2025 |
|---|---|---|
| Households that are married couples | 66% | 47% |
| Married-couple households with their own children under 18 | 54% | about 37% |
| Families with their own children under 18 | 54% | 39% |
| One-person households, share of all households | 20% | 29% |
Census also reports 39.7 million one-person households in 2025 and that 58% of adults aged 18 to 24 lived in their parents’ home, compared with 16% of those 25 to 34. These are descriptions of living arrangements. They do not on their own show whether people are lonelier, happier or worse off, and they do not measure marriage itself.
Marriage and divorce rates answer a different question
The CDC’s National Center for Health Statistics reports events, not the stock of married people. Its FastStats page, drawing on provisional 2023 data, lists 2,041,926 marriages, a rate of 6.1 per 1,000 total population, and 672,502 divorces, a rate of 2.4 per 1,000 population. The divorce figure covers 45 reporting states and the District of Columbia, so it is not a complete national count.
Be careful with these two rates. Dividing the divorce rate by the marriage rate is a common way to claim that a large share of marriages end in divorce. It is not valid: the marriages and the divorces in a given year come from different groups of couples, and the divorce data are incomplete. A separate cohort analysis would be needed to estimate the chance that a marriage ends, and the sources used here do not provide one.
Three questions, three different numbers
- How often do people marry in a given year? Answered by the marriage rate (6.1 per 1,000 in provisional 2023 data).
- Who is married right now? Answered by the share currently married, including the household figures above.
- Who has ever married by a given age? Answered by measures such as Pew’s 25% of 40-year-olds.
Mixing these produces false conclusions. A rise in the never-married share at 40 can coexist with a stable marriage rate, and a fall in married-couple households can reflect more people living alone or other changes in who heads a household, as well as fewer marriages.
What this evidence does not establish
None of these sources explains why marriage patterns changed, and none measures the effects on adults or children. Claims that the trend proves a moral collapse, or that it is harmless, go beyond the data. What the record supports is more modest: marriage is happening later, a larger share of adults have not married by 40 than in 1980, married-couple households are a smaller share of all households, and the trend is uneven across education and race.
How to read the next release
- Ask which age, year and population the figure describes.
- Keep “by age 40” separate from “ever.”
- Do not turn a marriage rate and a divorce rate into a divorce probability.
- Look for a cohort or age breakdown before generalizing.
- Read causes as hypotheses unless the source tests them.
Sources
- Pew Research Center, Richard Fry, “A record-high share of 40-year-olds in the U.S. have never been married,” June 28, 2023: pewresearch.org short read
- U.S. Census Bureau, “Census Bureau Releases New Estimates on America’s Families and Living Arrangements,” December 2, 2025 (CPS ASEC): census.gov press release
- CDC, National Center for Health Statistics, FastStats: Marriage and Divorce (provisional 2023 data; last reviewed March 17, 2025): cdc.gov/nchs/fastats/marriage-divorce.htm