Million Homeschooled Students in the United States in 2003
Until 1999, little empirical information existed about the prevalence of homeschooling nationally (Bielick, Chandler, and Broughman 2001). In 1999 and 2003, the National Household Education Surveys Program (NHES) collected nationally representative data that can be used to estimate the number of homeschooled students in the United States. Data from the 1999 NHES showed that there were an estimated 850,000 homeschoolers in the United States?about 1.7 percent of the school-age population (Bielick, Chandler, and Broughman 2001).
This Issue Brief provides estimates of the number and percentage of homeschooled students in the United States in 2003, and compares these estimates to those from 1999. In addition, parents? primary reasons for homeschooling their children are described. Estimates of homeschooling in 2003 are based on data from the Parent and Family Involvement in Education Survey (PFI) of the 2003 NHES.
Students are defined as children ages 5 through 17 with a grade equivalent of kindergarten through grade 12 (K?12)1. Interviews were conducted with the parents of 11,994 students (239 of whom were homeschooled). When weighted properly, these data represent approximately 50 million students ages 5?17 with a grade equivalent of K?12 in the United States in 2003.
Students are considered to be homeschooled if their parents reported them as being schooled at home instead of at a public or private school for at least part of their education and if their part-time enrollment in public or private schools did not exceed 25 hours a week. Students who were schooled at home only because of a temporary illness were not included as homeschoolers. All differences discussed in this Brief are statistically significant at the .05 level according to two-tailed student?s t-tests.
Number and Percentage of Homeschooled Students in the United States
About 1.1 million students (1,096,000) were being homeschooled in the United States in the spring of 2003 (figure 1). This represents an increase from the estimated 850,000 students who were being homeschooled in the spring of 1999. In addition, the estimated homeschooling rate?the percentage of the school-age population that was being homeschooled?increased from 1.7 percent in 1999 to 2.2 percent in 2003 (not shown in tables or figures).
As with results from any sample survey, the numbers and percentages discussed in this Issue Brief are estimates of the actual numbers and percentages of homeschooled students in the population. Although 1,096,000 is the best estimate available from the 2003 NHES, another similar sample survey might produce a different estimate. A 95 percent confidence interval defines a range of values around an estimate, within which 95 percent of the estimates from all possible similar sample surveys are expected to fall. The 95 percent confidence interval for the number of students who were homeschooled in spring 2003 is 915,000 to 1,277,000 (figure 1)2. The best estimate provided here?1,096,000?is the midpoint of that interval.
The estimates of homeschooling discussed in this Issue Brief include students who were homeschooled only and students who were homeschooled while also enrolled in school for 25 hours or less per week. As shown in table 1, in both 1999 and 2003, about 4 out of 5 homeschoolers were homeschooled only (82 percent) while about 1 out of 5 homeschoolers were enrolled in public or private schools part time (18 percent).
Parents? Most Important Reasons for Homeschooling Their Children
In the 2003 NHES, parents were asked whether particular reasons for homeschooling their children applied to them. Parents were then asked which one of those applicable reasons was their most important reason for homeschooling.
Thirty-one percent of homeschoolers had parents who said the most important reason for homeschooling was concern about the environment of other schools (figure 2). Thirty percent said the most important reason was to provide religious or moral instruction. The next reason was given about half as often; 16 percent of homeschooled students had parents who said dissatisfaction with the academic instruction available at other schools was their most important reason for homeschooling.
From 1999 to 2003, the number of homeschooled students in the United States increased, as did the homeschooling rate. The increase in the homeschooling rate (from 1.7 percent to 2.2 percent) represents about 0.5 percent of the 2002?03 school-age population and a 29 percent relative increase over the 4-year period. While data from the NHES cannot explain why homeschooling was more prevalent in 2003 than in 1999, it can provide insight into why parents homeschooled their children in 2003.3 Parents may have homeschooled their children for a variety of reasons, but certain factors appear to have been more influential than others. Nearly two-thirds of homeschooled students had parents who said that their primary reason for homeschooling was either concern about the environment of other schools or a desire to provide religious or moral instruction.
Data from NHES can also be used to examine the student, family, and household characteristics of homeschoolers. Upcoming reports will use these data to study the characteristics of homeschoolers, to compare the characteristics of homeschooled students to those of public and private school students, and to see how homeschooling rates may have changed between 1999 and 2003 for different segments of the student population.
Bielick, S., Chandler, K., and Broughman, S.P. (2001). Homeschooling in the United States: 1999 (NCES 2001?033). U.S. Department of Education. Washington, DC: National Center for Education Statistics.
Schenker, N., and Gentleman, J.F. (2001). On Judging the Significance of Differences by Examining the Overlap Between Confidence Intervals. The American Statistician, 55(3): 182?186.
Tables and Figures
1 Students who were homeschooled or enrolled in an ungraded elementary/secondary school or special education program were considered to have a grade equivalent of K?12 if their grade-level equivalent was K?12 or if their grade-level equivalent was ?ungraded? and they were ages 5?17. Ages 5?17 represent the modal age range for grades K?12.
2 Although the confidence intervals surrounding the estimated number of homeschooled students in the United States in 1999 and 2003 overlap somewhat, the differences between the estimates are still statistically significant. Differences between estimates with overlapping confidence intervals can be statistically significant (Schenker and Gentleman 2001).
3 Questions about reasons for homeschooling were asked differently in 1999 and 2003, and thus are not comparable.
The Issue Brief series presents information on education topics of
current interest. All estimates shown are based on samples and are subject
to sampling variability. All differences are statistically significant at the .05
level. In the design, conduct, and data processing of National Center for
Education Statistics (NCES) surveys, efforts are made to minimize the
effects of nonsampling errors, such as item nonresponse, measurement
error, data processing error, or other systematic error. For more information
on the National Household Education Surveys Program, visit