By Tabitha Tally • Wednesday, November 18, 2026
Seventy-one percent of children live with two parents in the Census Bureau’s newly released 2026 family and living-arrangement estimates, compared with 80 percent in 1976. The November 17 release also reports 40.5 million one-person households, representing 30 percent of all households. Together, the figures describe several kinds of change in American homes, but they count different populations and should not be combined as if they were shares of the same total.
Among children living with one parent, 19 percent lived with their father in 2026, compared with 7 percent five decades earlier. That denominator is children living with one parent, rather than all children. The distinction substantially changes what the number means. The release also puts the share of householders age 65 or older at nearly 31 percent, up from 20 percent in 1976, adding an age dimension to the household picture.
A companion Census analysis shows why household and family statistics differ. Average household size was 2.5 people in 2026, while average family size was 3.1. The bureau’s definitions count everyone occupying a housing unit in the household, including unrelated residents. A family instead includes the householder and people related by birth, marriage or adoption who live together. A person living alone can therefore form a household without forming a family under this statistical definition.
The companion analysis also finds that household size peaked at 3.2 people when the householder was 35 to 44, then declined at older ages. Families with their own children under 18 averaged 4.0 people. These comparisons help explain why an overall average can conceal very different living arrangements. They do not show that a particular household became smaller, because a comparison across categories is different from following the same people over time.
Census subject definitions add a further caution: family household members can include nonrelatives, while family members include the householder’s relatives. Family groups are another category, covering related people living together even when the householder is not among them. These distinctions may seem technical, but they determine which people enter each count. A chart labeled families should therefore be checked before it is used to make a claim about every person sharing a home.
Most statistics in the release come from the Current Population Survey Annual Social and Economic Supplement. Survey estimates are subject to error, and Census says the release’s comparative statements were statistically tested. Unmarried college students in dormitories are counted as living in their parents’ home in these CPS data. That convention matters when interpreting young-adult living arrangements. The new tables provide a documented national portrait; they do not assign causes to family changes or measure the quality of relationships within the homes they count.
Sources
FOLLOW THE ENTIRE CREW
Daily reporting and newsroom updates from FrontPage Crew.
