On this page

Disclosure. Some links on this page are affiliate links. Every Homeschool may earn a small commission if you purchase through them, at no extra cost to you. Editorial picks are not influenced by commissions; see how we make money.

As an Amazon Associate I earn from qualifying purchases.

Introduction

Ask how many children are homeschooled in the United States and the answer depends on who is counting. The National Center for Education Statistics (NCES) reported that 3.4 percent of students ages 5 through 17 in kindergarten through grade 12 or the equivalent were homeschooled in 2022-23 (NCES, Parent and Family Involvement in Education: 2023). The National Home Education Research Institute (NHERI) estimates that 6.262 percent of the school-age population was homeschooled in 2024-25 (NHERI). The two figures cover different school years and use different definitions and methods. In our assessment, the gap between them is one of the more useful facts about the homeschool data landscape.

This page is the hub. It lays out the size of the movement, how the rate has changed since 1999, who is doing it, and why, then points to deeper treatments of demographics, academic outcomes, and socialization.

Summary

  • The headline count depends on the counter. NCES reports 3.4 percent of students ages 5 through 17 homeschooled in 2022-23 (NCES), while NHERI estimates about 3.408 million students, or 6.262 percent of the school-age population, in 2024-25 (NHERI).
  • Census Pulse data show a doubling in 2020. The Census Bureau’s Household Pulse Survey found homeschooling among households with school-age children at 5.4 percent in the first week of data collection (April 23 to May 5, 2020) and 11.1 percent in September 30 to October 12, 2020 (US Census Bureau).
  • Johns Hopkins’ Pulse-based estimates for 2022-23 and 2023-24 were largely similar. Johns Hopkins reports that 5.92 percent of school-aged children were homeschooled in 2023-24 and 5.82 percent in 2022-23 (JHU Homeschool Hub).
  • School environment is the most frequently cited reason. NCES reports that 83 percent of parents of homeschooled children rated concern about the environment of other schools as important, and 53 percent rated a desire to provide religious instruction as important (NCES).
  • Race and ethnicity of homeschooled students. In NCES’s 2016 survey, 59 percent of homeschooled students were White non-Hispanic and the other 41 percent were Hispanic, Black, Asian or Pacific Islander, or another non-Hispanic group (NCES). The Census Bureau reported that the 2020 rate among households with Black or African American respondents rose by five times, while the increases for the other race and Hispanic-origin groups were not statistically different from one another (Census).
  • Outcome claims need caveats. Kunzman and Gaither’s review of the research reports that many of the largest and most widely cited homeschool achievement studies have design flaws that limit their generalizability, including volunteer samples recruited for the purpose (Kunzman & Gaither). The Coalition for Responsible Home Education makes a similar critique of a 2003 survey of homeschool graduates (CRHE).

How many: federal vs NHERI

Two institutions anchor the count. The first is NCES, whose National Household Education Surveys Program collects data on homeschooled students in its Parent and Family Involvement in Education (PFI) Survey (NCES homeschooling page). A parent or guardian who knew about the sampled child completes the PFI questionnaire (NCES, 2023 PFI First Look). NCES reported that 3.4 percent of students ages 5 through 17 with a grade equivalent of kindergarten through grade 12 were homeschooled in 2022-23 (NCES, 2023 PFI First Look), compared with 2.8 percent in 2018-19 (NCES, 2019 homeschooling rates). NCES notes that changes in the sample frame, data collection methods, and wording of some items over time mean its results should be compared with earlier collections only with caution. The homeschooling page also lists NCES reports on the topic.

The second is NHERI, a 501(c)(3) nonprofit that specializes in research and statistics on home education (NHERI). Its 2024-25 estimate draws on homeschool registration and enrollment data from 23 states and on Census Household Pulse data, and it assumes that homeschooling is underrepresented by 20 percent in the state data and by 15 percent in the Pulse data. NHERI reports about 3.408 million homeschool students in 2024-25, an estimated 6.262 percent of the school-age population, with a range of 3.067 to 3.749 million (plus or minus 10 percent) (NHERI). A third reference point sits between them: the Johns Hopkins University Homeschool Hub, drawing on the Census Household Pulse Survey, reports that 5.92 percent of school-aged children were homeschooled in 2023-24 (JHU Homeschool Hub).

Three estimates from three sources. The measures, denominators, and school years differ.
SourceMeasureRateSchool year
NCES PFI-NHES (federal survey)Students ages 5 to 17, grades K-12 or equivalent, reported as homeschooled3.4%2022-23
Johns Hopkins (Census Pulse)School-aged children reported in responding households5.92%2023-24
NHERI (state data and Census Pulse, adjusted)Share of school-age population6.262%2024-25

The three estimates differ in what they count and when. NCES counts students whose parents report homeschooling and reports full-time virtual education separately (NCES). The Johns Hopkins figure is the share of children reported in responding Pulse households (JHU). NHERI’s figure applies assumed underrepresentation adjustments to state and Census data (NHERI). The section on why the surveys disagree sets out each difference with its source.

How the rate has changed

The Census Bureau’s experimental Household Pulse Survey found that homeschooling among households with school-age children rose from 5.4 percent in the first week of Phase 1 (April 23 to May 5, 2020) to 11.1 percent (September 30 to October 12, 2020), a doubling. The Census Bureau notes that a clarification was added to the school enrollment question between those two points to separate homeschooling from virtual schooling (Census, published March 22, 2021). In the Census state tables, the three largest increases among households were Alaska, from 9.6 to 27.5 percent (up 17.9 points), Florida, from 5.0 to 18.1 percent (up 13.1 points), and Vermont, from 4.1 to 16.9 percent (up 12.8 points) (Census, Table 1).

NCES data show the homeschooling rate at 1.7 percent in 1999, 2.2 percent in 2003, and 2.9 percent in 2007, then leveling off at 3.4 percent in 2012 and 3.3 percent in 2016, which were not significantly different from the 2007 rate. The 2018-19 rate was 2.8 percent (NCES, 2019 homeschooling rates). NCES reported 3.4 percent for 2022-23 (NCES, 2023 PFI First Look). The Johns Hopkins Homeschool Hub reports 5.82 percent of school-aged children homeschooled in 2022-23 and 5.92 percent in 2023-24 in Household Pulse data, which it describes as largely similar (JHU).

Rates across rows are not directly comparable because the sources, denominators, and definitions differ.
PeriodRateSourceNote
19991.7%NCESStudents ages 5 to 17
20072.9%NCESStudents ages 5 to 17
20123.4%NCESStudents ages 5 to 17
20163.3%NCESStudents ages 5 to 17
2018-192.8%NCESStudents ages 5 to 17
Spring 20205.4%Census PulseHouseholds with school-age children, April 23 to May 5
Fall 202011.1%Census PulseHouseholds with school-age children, September 30 to October 12
2022-233.4%NCESStudents ages 5 to 17
2022-235.82%JHUSchool-aged children in Pulse households
2023-245.92%JHUSchool-aged children in Pulse households

Readers tracking education savings account programs can check the state-by-state ESA guide.

Who homeschools

NCES’s 2016 survey found that 59 percent of homeschooled students ages 5 through 17 were White non-Hispanic, 26 percent were Hispanic, 8 percent were Black non-Hispanic, 3 percent were Asian or Pacific Islander non-Hispanic, and 4 percent were other non-Hispanic (NCES, Homeschooling in the United States: Results from the 2012 and 2016 PFI). During the pandemic, the Census Bureau reported that the share of households with Black or African American respondents homeschooling increased by five times, from 3.3 percent (April 23 to May 5, 2020) to 16.1 percent (September 30 to October 12, 2020), while the increases for the other race and Hispanic-origin groups were not statistically different from one another (Census).

Race, income, geography, and motivation are covered in the companion guide on who homeschools and why. Families weighing whether the model fits their own situation can use the decision guide is homeschooling right for your family.

Why families homeschool

NCES asked parents of homeschooled children which reasons were important to their decision, and parents could choose more than one. The most frequent was a concern about the environment of other schools, such as safety, drugs, or negative peer pressure, rated important by 83 percent of parents and named the most important reason by 28 percent. Moral instruction followed at 75 percent, with a desire to emphasize family life together and dissatisfaction with academic instruction at other schools each at 72 percent. A desire to provide religious instruction was rated important by 53 percent of parents and named the most important reason by 11 percent (NCES, 2023 PFI First Look, table A-7).

Reasons parents of homeschooled children gave, 2022-23. Parents could rate more than one reason important, so shares do not sum to 100 (NCES, PFI-NHES:2023, table A-7).
Reason givenRated importantNamed most important
Concern about the environment of other schools (such as safety, drugs, or negative peer pressure)83%28%
Desire to provide moral instruction75%6%
Dissatisfaction with academic instruction at other schools72%17%
Desire to emphasize family life together72%8%
Desire to provide religious instruction53%11%
Desire to provide a nontraditional approach to the child’s education50%8%
Child has other special needs21%7%

Because parents could rate multiple reasons important, those figures overlap rather than partition the population. No single reason was named the most important by more than 28 percent of parents in the NCES survey. EdChoice’s 2024 Schooling in America survey asked current parents of each school type for their top three reasons for choosing a school. Among homeschool parents, a safe environment (53 percent) and individual or one-on-one attention (38 percent) were the most frequent (EdChoice, 2024).

What research says about outcomes

This is where careful reading matters most. NHERI’s fact sheet reports that home-educated students typically score 15 to 25 percentile points above public-school students on standardized academic achievement tests (NHERI). A 2003 survey by Brian Ray, commissioned by the Home School Legal Defense Association (HSLDA), covered more than 7,300 adults who had been homeschooled. Among respondents who had been home educated at least seven years, whose responses the synopsis reports, over 74 percent of those ages 18 to 24 had taken college-level courses (HSLDA, “Homeschooling Grows Up,” synopsis of the survey).

Kunzman and Gaither’s review reports that the five large achievement studies conducted under HSLDA sponsorship from 1990 to 2010 relied on samples of homeschoolers recruited for the purpose, with volunteers submitting their own test scores. The review adds that scholars have emphasized that studies of this kind do not use random sampling or control for confounding variables (Kunzman & Gaither). The Coalition for Responsible Home Education makes a similar point about the 2003 survey, stating that because of the way participants were recruited, its results cannot be generalized to the homeschool population as a whole (Coalition for Responsible Home Education).

NHERI’s own fact sheet states that research designs to date do not conclusively show that homeschooling causes the positive outcomes it reports, and that one hypothesis is that the positive findings might be due to the demographics of the homeschool students and families in the studies (NHERI). Kunzman and Gaither report that many smaller studies find homeschooling does not have much of an effect on student achievement once family background variables are controlled (Kunzman & Gaither). The companion guide on research on outcomes presents the favorable figures alongside the self-selection and volunteer-sample critiques.

Socialization, in brief

Kunzman and Gaither reviewed nearly 100 empirical studies of homeschooler socialization. A majority evaluated children’s social skills, and in most of those studies homeschoolers do not seem to suffer in comparison with conventionally schooled peers. The review notes that most of these studies rely almost entirely on self-report by students or their parents, and that the studies using other data sources still used convenience samples. It also notes that the most widely publicized socialization studies, by Ray, drew on non-random samples of homeschool graduates and relied on self-reporting (Kunzman & Gaither). In our assessment, those findings describe the samples studied and cannot be generalized to all homeschoolers. The dedicated guide on what the research says about socialization presents the positive findings alongside the sampling limits.

What it costs

NHERI reports that homeschool families spend an average of $600 per student annually for their education (NHERI, retrieved October 2026).

For families managing the line item rather than the average, the full cost guide covers what homeschooling costs per child per year and why the NHERI average and the 2024 HSLDA survey figures differ, and the budget-under-$200 picks list curriculum options chosen to keep the annual per-child cost under $200. The curriculum finder asks about budget alongside method and worldview.

Why the surveys disagree

In our assessment, the gap between 3.4 percent and 6.3 percent comes from the following differences between the estimates.

  • Definition. NCES counts a student as homeschooled when the student receives instruction at home instead of at a public or private school either all or most of the time, excluding students enrolled in public or private school more than 25 hours per week and students homeschooled only because of a temporary illness. Students in full-time virtual education as students of public or private schools (1.8 percent of students in 2022-23) are counted separately, and NCES’s combined measure of instruction at home, covering homeschooling and full-time virtual education, was 5.2 percent in 2022-23 (NCES).
  • Unit of measurement. The Census Bureau’s 2020 figures are shares of households with school-age children that reported homeschooling (Census). The Johns Hopkins figure is the share of school-aged children reported in responding Household Pulse households, excluding responses of “none” (JHU). The NCES figure is the share of students, based on a parent or guardian’s answers about one sampled child (NCES), and the NHERI figure is a share of the school-age population (NHERI).
  • Sampling and adjustment. The 2023 NCES survey used a nationally representative address-based sample, with an overall estimated weighted unit response rate of 53.3 percent (NCES). NHERI builds its estimate from state data and Census Household Pulse data. Its assumed underrepresentation adjustments raise the 23-state figure from 4.18 to 5.016 percent and the Pulse figure from 6.530 to 7.509 percent, and its estimate of 6.262 percent is the mean of the two adjusted figures (NHERI).
  • Base year. The NCES, Johns Hopkins, and NHERI figures cover 2022-23, 2023-24, and 2024-25 respectively.

In our assessment, a citation of these figures should name the source, the year, and what the rate counts. From here, the demographic, outcome, and socialization spokes carry the analysis further, and anyone moving from data to decisions can start with how to start homeschooling or the curriculum selection guide.