Specific Guidance
Rules for 21 different living situations, including a backup rule for people without a usual home.
Not every child lives in one home with one family.
Some children live with foster parents or grandparents. Some move between two parents’ homes. Others stay in shelters, group homes, treatment programs, or share housing with another family because they have nowhere else to go.
These are already some of the children the census has the hardest time counting accurately. A proposed change to the census rules could make that even harder by removing decades of guidance for people whose lives do not fit neatly at one address.
For the 2020 Census, the Census Bureau published detailed rules for counting people in 21 different living situations. Those rules helped census workers, shelter staff, facility operators, families, and communities understand where people should be counted.
They addressed foster children, people staying in domestic violence and homeless shelters, families doubled up with relatives, youth in group homes and treatment centers, boarding school students, farm workers living in employer housing, and more.
There was also a simple backup rule: if someone did not have a usual home, they were counted wherever they were staying on Census Day.
These rules were developed over decades, updated through public comment, and used as recently as 2020. States, local governments, shelters, child welfare agencies, and other organizations built their census outreach and planning around them.
Rules for 21 different living situations, including a backup rule for people without a usual home.
Greater reliance on tax records and days spent at an address between January and April.
The new proposal keeps only a handful of those rules.
For everyone else, the Bureau would use a general test based partly on tax records and the number of days a person spent at each address between January and April. Even spending a single hour at a location could count as a full day.
That may sound straightforward on paper. For children whose lives span multiple homes, temporary placements, shelters, or other living arrangements, it is anything but.
In September 2026, 6,163 Virginia children and youth were in foster care. Thousands more live with grandparents and other relatives through informal kinship arrangements that never involve a court or social services.
Children in foster care can move between placements during the same months the census is taken.
Under the 2020 rules, foster children were counted where they lived and slept most of the time. The new proposal instead says children should be counted with their “provider” without defining who that provider is.
For a child in foster care, that could mean the foster family they live with, the local department of social services that has legal custody, a group home, or even a birth parent who still claims them on their taxes.
The proposal does not say.
Children who split time between two parents’ homes face a similar problem. If a child spends the same number of days at each home, the new day-counting test could result in a tie, with no clear rule for resolving it.
That creates more opportunities for a child to be counted twice, counted in the wrong place, or missed altogether.
The uncertainty becomes even more serious for children and families fleeing domestic violence.
Under the 2020 rules, people staying in a domestic violence shelter on Census Day were counted at the shelter. The new proposal removes that guidance.
Imagine a parent who leaves an abusive home with their children in February or March. They may have spent more days earlier that year at the home they fled, and their tax records may still list that address.
Under the proposed test, that family could be counted on paper at the very home they had to leave, or not counted correctly at all.
There are safety concerns, too. Using tax, benefit, or other government records to determine where shelter residents “belong” raises serious questions when survivors’ locations need to remain private.
And this is not a small population.
On a single day in September 2025, 820 adults and children found safety through emergency shelter, transitional housing, or other housing provided by Virginia domestic violence programs, and not every program reported. In previous counts, about half of those sheltered on the survey day were children.
Clear census rules should help protect families in crisis, not create another layer of uncertainty.
The same problem extends to children without stable housing.
Virginia schools identified 20,942 students experiencing homelessness during the 2024–2025 school year. About 7 in 10 were living doubled up, sharing housing with another family after losing their own.
The 2020 Census rules included guidance for counting people in shelters, doubled-up households, soup kitchens and mobile food locations, and outdoor locations.
The new proposal provides none of that specific guidance.
It also says only days “lawfully spent” at a location would count. That could leave families sleeping in cars or in places where local laws prohibit camping with no clear qualifying location at all. Following a 2024 Supreme Court decision, more localities may penalize sleeping outdoors, potentially making that gap even wider.
Unaccompanied youth may be especially vulnerable. Many have no fixed address and little or no tax-record trail to point the Census Bureau toward where they actually live.
Without clear guidance, the children already easiest to overlook become even easier to miss.
In September 2026, 933 Virginia children in foster care were living in group homes or institutions.
The 2020 rules treated juvenile detention facilities, group homes, and residential treatment centers as different living situations, giving census workers and facilities clear direction about how people in each setting should be counted.
The new proposal instead refers broadly to “correctional or detention facilities, including juvenile facilities.”
Does that include a group home? A residential treatment program? A child placed in treatment for a few weeks?
The proposal does not make that clear.
Those distinctions matter because temporary placement in a treatment program or group home is not the same as detention.
We have also asked the Census Bureau to consider where these young people are most meaningfully connected. Youth in facilities are often there temporarily while remaining connected to their families and home communities.
Virginia already counts people in state prisons at their home address when drawing its own legislative districts. We have asked the Bureau to study and test whether a similar approach could work for youth in juvenile facilities, group homes, and treatment centers when their home address is known.
The 2020 rules also provided clear guidance for minors attending boarding schools, including Bureau of Indian Education schools: they were counted at their family home.
The proposed rules could change that.
They also say nothing specifically about children attending residential schools for students with disabilities.
Without clear guidance, one child could be counted at school while another in a similar situation is counted at home, or missed entirely.
That matters because census numbers help communities understand who lives there and plan the services children need, including education and disability supports.
The children most affected by unclear counting rules are not evenly distributed across communities.
LGBTQ young people make up a much larger share of youth in foster care and unstable housing than of youth overall.
One study of nearly 900,000 California students found that nearly 1 in 3 youth in foster care and 1 in 4 youth experiencing unstable housing identified as LGBTQ, compared with about 1 in 9 of their peers. National research from Chapin Hall has also found that LGBTQ young adults face a 120 percent higher risk of homelessness than their peers.
Many of these young people move among group homes, shelters, relatives’ homes, and friends’ couches. These are exactly the kinds of living situations that are least likely to fit cleanly into a tax record or day-counting formula.
At the same time, the proposal would permanently prohibit census questions about sexual orientation.
The result is a troubling combination: LGBTQ youth could be more likely to be missed while the census gives up a tool that could help communities understand whether that is happening.
When young people disappear from the data, it becomes harder for states and communities to see the need and harder to plan affirming placements, housing, services, and support.
We have asked the Census Bureau to:
These may sound like technical census rules. Their consequences are not technical.
The children affected are often the same children and families who rely most on the schools, housing programs, health services, child welfare supports, and community resources that census data helps shape.
The children who are hardest to count should not be the ones given the least clarity about how they will be counted.
Foster and kinship caregivers, caseworkers, shelter staff, service providers, and youth advocates see every day how complicated children’s living situations can be.
That experience is exactly what the Census Bureau needs to hear.
Submit a public comment by November 2 under docket number USBC-2026-0628 at Regulations.gov.
You do not have to be a census expert. Tell the Bureau why clear rules matter for the children and families you know and serve. Ask it to restore specific guidance for children in foster care, shelters, group homes, treatment programs, and other complex living situations.
Submit a Public Comment ↗If you can, share a real-world example of a situation these proposed rules would get wrong.
Need help getting started? Use one of our templates for individuals or organizations as a starting point. Public comments are posted online, but commenters may submit anonymously.
Then share this post with others working in child welfare, housing, education, domestic violence prevention, and youth services. Sign up for our updates to follow this work.
A child’s life may not fit neatly at one address. That should not determine whether they count.
October 19, 2018