Maybe this has happened to you, or something close to it.
It is a Saturday morning in late September. You are on the sideline of a soccer field in a folding chair, holding a coffee that stopped being hot at the start of the second half.
Next to you is a mother you have stood beside all season. You know her son’s name and her dog’s name. You know she always remembers the orange slices, and that she laughs at the same parts of the game you do.
Between goals, she says it the way people mention a move. “We’re pulling Sam out after winter break. We’re going to homeschool.”
You say, “Oh, wow,” in the voice people use for news like that.
And a picture arrives. You did not ask for it. A large family. A long kitchen table. A van. A family more religious than yours, maybe, or more political, or with more money, or with less. A family that does things differently on purpose.
With the picture comes a small sentence: Huh. I didn’t know she was one of those.
Then a quieter one, about you: Anyway. That isn’t us.
You drive home. Your daughter is in the back seat, quiet in the way she has been quiet since spring. You have wondered about that quiet more than you have said out loud. You have wondered about a lot of things. Homeschooling was never one of them. It belonged to a different kind of family.
Nobody here is being unkind. The picture did what pictures do. It sorted an option before you had to think about it. The only question is whether the picture is still accurate.
Where the picture came from
Nobody sits down and builds a picture of “homeschoolers.” It collects on its own, from a few families you have met and a few you have read about.
For most of the years you were growing up, there were not many to see. In 1999, the federal government’s household education survey counted about 850,000 homeschooled children. That was 1.7 percent of school-age children, fewer than two in a hundred. If you were in school then, you may never have known one.
A small group is easy to imagine as one kind of family. The few you happen to see can stand in for all of them.
And once that image forms, it tends to stay put. It does not update itself when the numbers change. Nobody sends you a notice that the group you pictured back then has grown, or moved to the suburbs, or changed who is in it. You keep the one you have until something makes you look again.
Parents who homeschool know that image well, because they run into it. One wrote on a public forum: “When you say the word homeschool, their minds think something very different…” Another wrote a sentence you could imagine saying yourself: “We aren’t anti public school… not good fit.”
Here is what the picture costs you. It makes a decision before you do.
The quiet in the back seat is a real question about a real child. The question it raises is what would help her? The picture swaps in a different question, are we those people?, and answers it on the spot. One possible answer to the first question leaves the table before anyone has looked at it.
Nothing here says your answer should be homeschooling. For most families it will not be. The point is smaller: the picture should not get to answer for you without being checked.
The part of the picture that is true
First, the belief, said plainly, because it deserves that:
Homeschooling is for a certain kind of family, and that isn’t us.
It is not a silly belief, and it is not a mean one. It protects something. Most people do not want to become someone else in order to help their child. They do not want to explain themselves at every family dinner, or be the ones everyone quietly has an opinion about. Parents who already homeschool describe that outside-ness from the other side. One forum post said: “get used to swimming against the tide.”
So that isn’t us is partly a way of saying I don’t want to swim against the tide. That is a reasonable thing to want.
The belief also comes from noticing a pattern. And part of the pattern is real.
In the federal survey for 2016, the mother was the main teacher for 78 percent of homeschooled students. If your picture has a mother at the kitchen table, that part holds for most homeschooled students.
In the federal survey for 2022–23, White children were homeschooled at higher rates than Black or Hispanic children: 5.1 percent, against 1.7 and 1.8 percent.
Children in rural areas and small towns were homeschooled at more than twice the rate of children in cities: 5.2 and 5.1 percent, against 2.1 percent.
And in a 2024 national survey of parents, homeschooling parents leaned further to the political right than parents overall, and went to religious services more often. The numbers come later.
So if your picture is of a mother teaching, in a small town, in a family more religious than most and leaning right, the numbers do not call it false. Some families look just like it. Many of them are doing careful, loving work, and nothing here is aimed at them.
The belief has a second half, though: that isn’t us. That half depends on how much of homeschooling the picture covers. That is the part worth testing.
First, the doubling
You may have read that homeschooling doubled. It did. But it matters which count you mean, because there are two doublings in the numbers, and one comparison that only looks like one.
The slow doubling. The federal household education survey has counted homeschooled children since 1999. The Census Bureau ran its 2022–23 round for the Education Department. The share of school-age children being homeschooled went from 1.7 percent in 1999 to 2.2 percent in 2003, 3.0 percent in 2007 and 3.4 percent in 2012. That is a doubling in thirteen years, on one survey.
Then it stopped. The same survey found 3.3 percent in 2016 and 2.8 percent in 2019.
The fast doubling. In spring 2020, the Census Bureau started a quick weekly survey of households to track the pandemic. In its first week, about 5.4 percent of households with school-age children reported homeschooling. By the start of the next school year, it was 11.1 percent. The Census researchers called the change “a doubling,” though a clarification had been added to the question in between. That count is of households, not children, and it happened between spring and fall 2020.
Where it settled. Here the counts part ways.
The federal household survey for 2022–23 found 3.4 percent of students homeschooled, roughly 1.8 million children.
The Census weekly survey, as analyzed by the Johns Hopkins Homeschool Hub, put the share of school-age children at about 5.8 percent that same year and 5.9 percent the year after. A separate 2024 national survey of parents found about 5.7 percent of the children in its parents’ homes were homeschooled. It was led by researchers at Johns Hopkins and Kennesaw State and published by EdChoice, a school-choice advocacy group.
The surveys ask different questions, and none of them can settle which number is right. So when you read that homeschooling “doubled from about 3 percent to about 6,” the two ends of that sentence come from different surveys. On the federal survey alone, its own figure moved from 2.8 percent in 2019 to 3.4 percent in 2022–23.
Take whichever count you trust. It is somewhere between about one child in thirty and one in seventeen. If your town has 500 school-age children, that is somewhere between about 17 and 30 of them.
And it grew again in 2024–25 where states keep count. Twenty-eight states, counting the District of Columbia, reported homeschool numbers for that year. A Johns Hopkins researcher found they grew by an average of 4.9 percent, “more than double the pre-pandemic homeschooling growth rate of around 2%.” Those are children registered with their state, so they are a floor, not a full count.
That is a lot of children to fit into one picture.
Who they are
One tool first, because the rest depends on it. There are two ways to ask who homeschools, and they give different answers.
A rate asks: of all the children in a group, how many are homeschooled? Of all children in rural areas, 5.2 percent.
A share asks: of all the homeschooled children, how many are in a group? Of all homeschooled children, what fraction live in rural areas?
A group can have a high rate and still be a small share, if the group itself is small. And a group with a low rate can still be a large share, if the group is big.
The picture from the sideline works like a rate. It fills the frame with the families where homeschooling is most common. The question is it for families like us?, the one that got in front of your daughter’s quiet on the drive home, is really a question about shares: how many of the families doing this look like you. So below you get both, and each one is named.
Money
What you imagined on the sideline may have had money in it, more than you have or less. Here is the share of children homeschooled at each household income, from the federal survey for 2022–23:
The rate rises from the lowest incomes to the middle, then falls. It is highest in homes earning $50,001 to $75,000.
Count adults instead of rates and the spread is wide. In the Census weekly survey for 2023–24, among adults with a homeschooled child at home, about 28 percent reported household incomes under $50,000, 25 percent between $50,000 and $100,000, and 23 percent of $100,000 or more. About a quarter did not say.
The federal survey for 2016 counted poverty too. That year, 21 percent of homeschooled students lived in households below the federal poverty threshold.
The 2024 parent survey compared homeschooling parents with all parents. It found homeschooling parents reported lower incomes, not higher: 46 percent under $50,000, against 36 percent of parents overall, and 18 percent over $100,000, against 28 percent.
None of these counts found homeschooling most common in the highest-income homes.
Race and background
The family you pictured may also have looked a certain way. You have already seen that White children are homeschooled at higher rates. Count heads instead of rates, and a second fact sits beside the first.
In 1999, about three in four homeschooled children in the federal count were White. In the federal counts from 2012 to 2019, between about three in ten and four in ten were not: 32 percent in 2012, 41 percent in 2016, 30 percent in 2019. In 2016, 26 percent of homeschooled students were Hispanic, 8 percent Black, and 3 percent Asian or Pacific Islander.
The newer surveys land in the same range. In the Census weekly survey for 2023–24, 63 percent of adults with a homeschooled child were White and not Hispanic, 18 percent Hispanic, 10 percent Black and 2 percent Asian. In the 2024 parent survey, 69 percent of homeschooling parents were White, against 56 percent of all parents.
Politics
You may also have imagined a family that votes differently from yours. The 2024 parent survey asked 3,191 parents about their politics. Two hundred forty-nine of them were homeschooling at least one child.
Among the homeschooling parents, 44 percent identified as Republican, against 36 percent of all parents. Asked about views rather than party, 43 percent called themselves conservative, against 32 percent.
Religion
Or one more religious than yours. The same survey asked about religious services. Forty-four percent of homeschooling parents said they attend at least weekly, against 33 percent of parents overall.
The federal survey for 2022–23 asked parents for their single most important reason for homeschooling. The most common answer, for 28 percent of homeschooled students, was concern about the environment at other schools, such as safety, drugs or negative peer pressure. Next came dissatisfaction with academic instruction, at 17 percent. A desire to provide religious instruction was the most important reason for 11 percent.
Work
Then there is the question you may ask first if you work: who would be home? For children living with two parents in 2022–23, the federal table sorts homeschooling by how the parents work. Where both parents worked full time, 1.1 percent of children were homeschooled. Where one worked full time and the other part time, 4.5 percent, above the national rate of 3.4. Where one worked full time and the other was not in the labor force, 8.8 percent.
For children with a single parent, the rates were 1.2 percent where the parent worked full time, 2.9 percent where the parent worked part time, and 1.8 percent where the parent was not in the labor force.
Count homeschooled children instead of rates. In the 2019 federal count, about 27 percent of homeschooled students, roughly one in four, lived with two parents who were both in the labor force. About 58 percent lived with two parents of whom only one was.
The table shows who homeschools, not why. It cannot say which came first: a parent leaving a job in order to homeschool, or a parent already at home deciding to.
Where they live
You have already seen the rates for small towns and rural areas. In the suburbs, the rate for 2022–23 was 3.1 percent, a little below the national 3.4. In cities it was 2.1.
Count homeschooled children instead. In the federal survey for 2016, 39 percent of homeschooled students lived in suburbs and 29 percent in cities. Towns held 10 percent and rural areas 22 percent. Four years earlier, rural areas had held 31 percent and suburbs 34.
The 2024 parent survey found 41 percent of homeschooling parents living in the suburbs, against 43 percent of all parents; about a quarter in cities, against a third; and a quarter in rural areas, against 15 percent.
Region and state
By region, the researchers behind the 2024 survey found homeschoolers spread across the Midwest, South and West in about the same proportions as other families, with homeschooling less common in the Northeast.
Among the states that reported 2024–25 numbers, ten hit their highest homeschool counts ever: Georgia, Louisiana, Minnesota, Nebraska, North Dakota, Ohio, Pennsylvania, South Carolina, South Dakota and Wyoming. Five states and the District of Columbia went down.
Counting the families
Go back to the picture from the sideline: a family more religious than yours, maybe, or more political. The 2024 survey that found the lean also found the parents the picture leaves out.
Thirty-one percent of homeschooling parents said they never go to religious services. Among all parents it was 30 percent, about the same. Another 25 percent went, but less than weekly, against 38 percent of all parents. Twenty-seven percent of homeschooling parents were independents, the same share as parents overall. Twenty-nine percent, close to three in ten, were Democrats, against 37 percent of all parents. And about two in three lived in suburbs or cities, against about three in four of all parents.
Whatever your politics, and whether or not you go to services, parents who gave your answer to each question were among the homeschooling parents in that survey.
In a 2024 survey, homeschooling parents leaned one way but were not one type.
The picture was right that there is a lean. It was wrong to make the lean a type, and then to leave you out of it. The mother at the kitchen table is real. So are the parents it did not show.
So the question from the sideline was never the real one. The real one is not are we those people? It is what would help her? And if homeschooling is one answer worth a look, what would our weeks have to look like for that to happen? For some families the week can be arranged differently, and for some it cannot, and neither is a verdict on who they are.
That is a question about a child and a calendar. Homeschoolers in these surveys came from every income, party, background and kind of neighborhood the surveys measured. You are allowed to ask it too.
The sideline, a year from now
It is a Saturday in late September again. Same field, same folding chair. The coffee went cold at the same point in the game.
This time it is a father from the other team who says it, standing behind the bench: they are homeschooling their two youngest this year. A picture arrives again, the way pictures do. But it is a different one. It is not a van or a long table. It is a calendar. The first thing you wonder is practical: How did they arrange their weeks? Who does what, on which days? You might even ask him.
Your daughter is on the field. Maybe the quiet from last spring had another answer: a new teacher, a friend, a year that simply went better. Maybe one night in January you sat at the kitchen table, counted your weeks, and decided not now. Maybe she is home this year, and Thursdays are for a class across town.
Whichever it was, you decided it by looking at her and at your calendar. The picture did not decide it for you.
Somewhere in there you also noticed something about the families you already knew. The one two doors down whose oldest did a year at home in middle school and then went back. The coworker who teaches her son at home and works her shifts in the evenings. The cousin nobody thinks of as “a homeschooler” at all. They were not a type. They were families who had looked at a child and a calendar, the same as you.
Sam’s mother is in the chair beside you. She is exactly who she was last year. She still remembers the orange slices. She hands you one.
You are the same family you were a year ago too. What changed first was which questions you let yourself ask.