NCES Data Points to Changing Homeschool Demographics

Every four years, the National Center for Education Statistics (NCES) conducts its National Household Education Survey. Beginning in 1999, the NCES has collected data on homeschooling as part of this survey. This data offers the most comprehensive statistical information we have on homeschooling in the United States. Last month, the NCES released its most recent findings, which covered data collected during 2015-2016.

Unlike studies of homeschooling that draw from convenience samples gathered via individuals a researcher knows or emails sent out by homeschool organizations, the NCES uses a nationally representative random sample for this survey. Even so, this data is not without its challenges. The NCES switched from phone surveys to mail-in surveys when conducting the 2011-2012 survey, and some researchers have questioned whether this or that method may favor different types of respondents.

The NCES does not collect information on students’ test scores; as a result, their surveys primarily provide demographic data. Below, we report some of the key findings of the 2015-2016 survey; we cover numbers, language, income, parental education, grade level, and parents’ reasons for homeschooling. Remember, as with any survey data, this information should be approached with some degree of caution.

1. The homeschool rate has stopped increasing, ending a long period of homeschool growth. 

One of the most significant findings of the 2015-2016 National Household Education Survey is that the homeschool rate has ceased growing and has perhaps even declined slightly. In 2011-2012, 3.4% of school-aged children were homeschooled; in 2015-2016, that number dipped slightly to 3.3%. Overall, the number of homeschooled children declined from 1,770,000 in 2011-2012 to 1,704,000 in 2015-2016. This finding is not as surprising as it might initially seem; state level homeschool enrollment data has, in recent years, been mixed, with some states showing declines while others have seen growth. Additionally, the pace of growth had already been slowing; while the number of children being homeschooled increased by 38% between 2003 and 2007, it grew by only 17% between 2007 and 2011-2012.

2. One in ten homeschooled children does not have a parent/guardian who speaks English.

According to the 2015-2016 data, 11% of homeschooled students live in household where no parent or guardian speaks English. This is the first year the National Household Education Survey asked this question, so we cannot compare this finding to previous years. However, the finding may be related to a dramatic rise in the number of Hispanic students being homeschooled: in 2015-2016, 26% of homeschooled students were Hispanic, up from 15% in 2011-2012. While white, non-Hispanic students were still homeschooled at the highest rate, fewer white, non-Hispanic students were homeschooled in 2015-2016 than in 2011-2012. (White, non-Hispanic students made up 59% of homeschooled students in 2015-2016, down from 68% in 2011-2012 and 77% in 2007.)

3. Children living below the poverty line are more likely than other children to be homeschooled.

The most recent NCES survey found that children living below the poverty line were more likely to be homeschooled than other children. This appears to have shifted over time. In 2007, just 1.8% of school-age children in poverty were homeschooled, compared with roughly 3.2% of those not in poverty. In 2011-2012, in the midst of the Great Recession, the rate was practically identical for each group—3.5% of children in poverty and 3.4% of children not in poverty were homeschooled. In the latest data, however, 3.9% of children in poverty were homeschooled compared to 3.1% of children not in poverty. Children living in poverty were 26% more likely to be homeschooled than children not living in poverty.

4. Parents without a high school diploma or GED homeschool at a higher rate than other parents.

According to the most recent data, 4.4% of children whose parents do not have a high school diploma or GED are homeschooled as compared to only 3.6% of children with at least one parent with a bachelor’s degree. To break it down another way, 15% of homeschooled children do not have a parent who has completed high school. Nearly one-third of homeschooled children are being educated by parents with a high school diploma or below. Less than half of homeschooled children (45%) have a parent with a bachelor’s degree or above.

In 2015-2016, for the first time since the NCES began collecting data, the rate of homeschooling among children whose parents have a bachelor’s degree or above declined. The high rate of homeschooling among children whose parents have not completed high school is also a new thing; in 2007, the first year to release data on homeschooling rates among parents who had not completed high school, only 0.4% of parents without a high school diploma or GED homeschooled; in 2015-2016, that number was 4.4%.

5. High school students are homeschooled at a higher rate than elementary school students.

In 2007, the homeschooling rate was slightly higher for elementary school children than for middle or high school students. More recent data suggests that this has changed. In 2011-2012, high school students were 19% more likely to be homeschooled than young elementary students; in 2015-2016, this number changed to 31%. It is difficult to tell what is driving this shift. While the overall rate of homeschooling declined slightly between 2011-2012 and 2015-2016, the rate of high school students being homeschooled actually increased slightly.

It is possible that the rise in virtual education has led to an increase in the number of high school students being homeschooled; in 2011-2012, at lest one-fifth of all homeschooled students were enrolled in online courses through a public school. It is also possible that changes taking place in the nation’s elementary schools over the past few years have made them more attractive to parents who might otherwise have homeschooled their children.

6. The percentage of parents homeschooling for religious or academic reasons has been declining.

On the 2015-2016 survey, 17% of parents listed a dissatisfaction with academic instruction in other schools as their most important reason for homeschooling, down slightly from 2011-2012 when 19% of parents selected this reason. Overall, 61% of parents listed a concern about academic instruction as one of their reasons for homeschooling, down from 74%. The percentage of parents who listed a desire to provide religious instruction as their most important reason for homeschooling declined similarly from 17% to 16% between 2011-2012 and 2015-2016; those who selected it as a reason overall declined from 64% to 51%.

The percentage of parents selecting a concern about the environment in other schools as their most important motivation for homeschooling increased from 25% to 34% during this period, though but the percentage of parents selecting it as a reason overall declined from 91% to 80%. The percentage of parents homeschooling to provide a nontraditional education or because a child has special needs, meanwhile, remained largely unchanged.


What does all of this mean? For one thing, it means that homeschool demographics are beginning to more closely resemble those of the nation as a whole. Before the last decade or so, families that homeschooled tended to be more likely to be white, less poor, and better educated than other families. Data from the most recent National Household Education Survey suggests that this may no longer be the case. These changes may point to increasing use of online public school programs and other virtual school options, which could be attracting students with a different set of demographics than traditional homeschooling.

In 2016, the NCES released a full analysis of the homeschooling data it had collected in 2011-2012. This report explained that, when initially contacted, some of the respondents had asked for the homeschool survey while others had asked for the enrolled in school survey and then marked that their children were homeschooled part-time. This latter group had a distinct demographic profile. These individuals were more likely to be Hispanic, far more likely to be poor, and much more likely to not have a high school diploma or GED. While the full analysis of the 2015-2016 data will not be released for some time, we may be seeing a similar effect.

These changing demographics throw into stark relief the limitations of current research on homeschooled students’ academic achievement. Studies of homeschooled students’ performance have typically drawn on convenience samples that skew heavily toward white children with college educated parents. These findings cannot be generalized to homeschooled students with vastly different demographic factors. Our lack of knowledge about these students is especially concerning in the wake of a 2015 Stanford study which found that, over the course of a year, virtual charter school students lagged behind their demographically matched peers in reading and showed no progress whatsoever in math.

Are public schools offloading problem students onto online programs? Are virtual charter school programs attracting a new demographic of students to home education? Or are the demographics of homeschooling changing across the board, regardless of the method or means of instruction? The 2015-2016 NCES data leaves these and many other questions unanswered.

Skip to content