Different surveys can produce different valid estimates
Pew's 2019 cohabitation analysis cited a 7% Current Population Survey estimate. Pew's July 2022 survey found that 11% of adults said they lived with a partner. The dates, questions, and classification methods differ.
Do not choose the larger figure. Name the source and definition beside each number.
Unmarried does not mean unpartnered
Someone can be legally unmarried while living with a partner or being in another committed relationship. Conversely, a married person may be separated.
This is why a dating-pool model should label a marital-status filter as ‘not currently married’ instead of calling everyone in that category single or available.
Household surveys miss partners who live in different homes
A household relationship field can pair partners who live together but cannot identify a committed couple who maintain separate homes.
Survey questions that ask directly about committed relationships capture some of that group, which helps explain why a broader partnered category can exceed household cohabitation.
Start with three different relationship groups
Legal marital status, who shares a home, and a person's stated relationship status are different facts. Someone can be unmarried and live with a partner. Someone can be married but separated. Someone can have a committed partner in another home.
Here, 'unmarried couple' means two people recorded as partners who are not legally married. It does not mean every home with two unmarried adults. Roommates, relatives, and housemates are not automatically a couple.
Why 7% and 11% can both be reasonable
The 7% figure comes from the Current Population Survey measure cited in Pew's 2019 analysis. The 11% figure comes from Pew's July 2022 survey, where adults described themselves as living with a partner. Different years, samples, questions, and classification rules can produce different estimates.
Do not average the estimates. They are not two readings of the same measure. Use 7% for the earlier household estimate and 11% for the 2022 self-reported relationship measure.
Current cohabitation and ever cohabited answer different questions
A current measure asks about the person's situation on the survey date. A past-living-together measure asks whether the experience happened at any time. Pew's NSFG analysis found that 59% of adults ages 18–44 had lived with an unmarried partner. That past history will naturally be much larger than a snapshot of who lives together now.
Someone can contribute to the 59% after marrying, separating, moving into a solo household, or beginning another relationship. The figure therefore describes how common an experience was in a specific age range, not the share of adults available to date. Keeping 'ever,' 'ages 18–44,' and the relationship definition in the sentence prevents the most common misreading.
The age range and counted group change the interpretation
The 59% ever-cohabited statistic covers ages 18–44, while the 7% and 11% current estimates are described across adults in their respective sources. Younger adults have had fewer years in which cohabitation could occur, and older adults have had more time to accumulate relationship histories. A cumulative percentage is especially sensitive to the age limits of the sample.
This is why an article should not splice the 59% history measure into a sentence about all current U.S. households. It can be useful for understanding how common cohabitation has become across early and middle adulthood. It cannot tell us how many cohabiting households exist today without a current household count and matching definition of who is counted.
Couples who live apart are outside the household count
Some committed partners keep separate homes because of children, work, caregiving, money, health, or preference. Others are apart for school, deployment, immigration, or housing. A household survey cannot pair them unless it asks a separate relationship question.
These people can be unmarried and in separate homes while firmly partnered. Treating every unmarried person as available exaggerates a dating pool. Use only the label the data support: not currently married, living with a partner, or self-described as partnered.
Cohabitation is a status, not a single intention
A shared home can be a step toward marriage, a long-term alternative to marriage, a practical response to rent, or part of a relationship whose future is uncertain. The household definition alone does not reveal which explanation applies. It also cannot show relationship quality, commitment, exclusivity, or whether both partners describe the arrangement in the same way.
Readers making personal decisions should ask directly about expectations for money, housing, legal protections, children, caregiving, and marriage rather than substituting a national percentage. Population statistics can show that an arrangement is common; they cannot supply the agreement a couple needs about what the arrangement means for them.
Unmarried-partner homes are not the same as dating-app users
A cohabiting population estimate describes an established household arrangement. Dating-app studies describe people who have ever used a service, are using one now, or report an app-mediated relationship. The populations overlap, but neither contains the other. Some cohabiting couples met offline; some app users are single; some are already partnered.
Using household figures to estimate a dating app's active pool would therefore add a false bridge between data sources. A credible dating-market estimate needs a defined geography, age range, relationship-status rule, orientation and interest assumptions, and a current activity measure. The cohabitation statistics answer a narrower and more reliable question about relationships and households.
Survey language has changed with relationship measurement
Older family surveys sometimes analyzed only opposite-sex cohabitation histories, including the NSFG comparison highlighted by Pew. That limitation must remain visible when the result is discussed. More inclusive household measures can identify same-sex spouses and unmarried partners, but historical series may not always offer an identical classification across years.
A trend is strongest when wording, age range, counted group, and coding remain comparable. When they change, researchers may still learn something, but the break belongs in the interpretation. Readers should prefer a clearly labeled recent estimate over a seamless-looking time series assembled from categories that did not mean the same thing in every year.
How to use the numbers responsibly
For a general snapshot, cite one current estimate and its instrument: 11% of adults described themselves as living with a partner in Pew's July 2022 survey. For household-history context, add that 59% of adults ages 18–44 had ever cohabited in the separate NSFG analysis. Use the 7% CPS figure only with its source and earlier context.
Do not subtract any of these percentages from 100% and label the remainder single. The remainder includes married adults, people in relationships across households, and people whose status is outside the measured category. The most useful lesson is methodological: relationship status must be asked, not guessed from legal marital status or a household statistic alone.
People, couples, and homes are different things to count
A percentage of adults and a count of unmarried-partner households cannot be substituted for one another. An adult survey gives each person a status, while a household table identifies a pair inside one address. One couple generally contains two partners, but converting between counts requires adjustment rules, household rules, and attention to whether both partners are in the survey group.
This matters when headlines estimate how many unmarried couples exist. Multiplying an adult percentage by the total population can fail if the survey excludes ages, institutions, or people outside households. Dividing by two assumes every identified adult pairs perfectly with another person surveyed. Use the source's published unit and estimate rather than creating a national count from a rounded percentage.
Trend claims require the same question over time
Cohabitation has become more common in the historical research Pew summarizes, but a careful trend holds the age range, relationship definition, survey, and coding as constant as possible. Comparing a 7% current CPS estimate with a 59% ever-cohabited NSFG estimate does not show growth; it compares two different outcomes and counted groups.
When a later release becomes available, compare it with the earlier wave of the same series and examine documentation for breaks. If only unlike sources exist, describe them side by side rather than drawing a line between them. The discipline may produce a less dramatic chart, but it prevents a change in measurement from being mistaken for a change in relationships.
A publication checklist for unmarried-couple statistics
Keep five labels near every headline number: survey or data system, survey year, age range, whether the number is current or past, and whether it counts adults or households. State whether the analysis includes all couple types represented today or uses a historically narrower opposite-sex relationship history. Link the original methods rather than only a secondary summary.
Finally, scan for words the data do not support: single, available, dating, engaged, stable, or intending to marry. Those may be valid findings in another study, but cohabitation status alone cannot supply them. The strongest article answers the searcher's numerical question quickly and then protects the reader from turning a household label into a complete relationship profile.