Dynamic Homeschooling: Switching and Mixing Among Homeschooling Families

By Albert Cheng and Matthew H. Lee, September 2025

Homeschooling families are often stereotyped as exclusively choosing homeschooling. The “once homeschooled, always homeschooled” generalization suggests that homeschooled students will only ever be homeschooled for every year of their school-age years. The “homeschool one, homeschool all” generalization suggests that homeschooling families will choose to homeschool all their children. But these stereotypes are not supported by evidence.

Yet if these stereotypes are not entirely accurate, how have they ossified in the collective mind since the beginning of the modern homeschooling movement in the 1970s? In her seminal research of homeschooling families conducted in the 1980s, sociologist Jane Van Galen divided homeschooling families into ideologues (those who desired to instruct children according to a religious, often Christian, tradition) and pedagogues (those who implemented certain teaching practices, like unschooling, that better accorded with their beliefs about the nature of the child).

Though the two groups differed in motivations to homeschool, both shared the strong view that homeschooling was the best way to educate their children. The strength of their convictions paired with Van Galen’s labels then fed both the “once homeschooled, always homeschooled” and “homeschool one, homeschool all” beliefs. In short, if parents felt so strongly about homeschooling, why would they not homeschool all their children and for all the years of their schooling?

As other studies of homeschooling conducted up through the first decade of the 21st century indicate, those popular beliefs have some empirical support. But there are exceptions. Moreover, those exceptions appear to be more the rule, especially in the face of recent trends such as the growing interest in homeschooling, the diversification of the homeschooling population, the increasing availability of educational resources, and the enactment of homeschool-friendly policies that have legitimized the practice and made it more accessible.

“Once homeschooled, always homeschooled”?

Based on our analysis of the 2023 Cardus Education Survey data, about 36 percent of U.S. adults who were homeschooled at some point only did so for one or two years. Those adults, on average, also spent over 8 and 3 years in traditional public and charter or private schools, respectively, for the remainder of their schooling. In contrast, less than 20 percent of adults who were homeschooled did so for the entirety of their primary and secondary schooling.

Figure 1 displays sector “switching” for adults who were homeschooled at any point in their schooling. That is, in addition to showing the proportion of those adults who were homeschooled and not homeschooled for any given grade level, Figure 1 shows the proportion of adults who switched settings between grade levels. At any given grade level, about 40 percent of adults who were ever homeschooled were educated in that setting (the pink bars). The other approximately 60 percent of adults who were ever homeschooled were in a different educational setting for that grade level (the blue bars). However, these percentages mask the dynamic switching that occurs as students move in and out of homeschooling from one grade to the next. For instance, 9.9 percent of adults who were ever homeschooled did it for kindergarten but not for first grade. Conversely, 8.8 percent of ever-homeschooled adults were not homeschooled in kindergarten but switched to homeschooling in the first grade.

Figure 1: Switching In and Out of Homeschooling across Grade Levels

Note: Authors’ calculations based on data from the 2023 Cardus Education Survey.

This analysis of the 2023 CES data demonstrates that the “once homeschooled, always homeschooled” belief is not true. In fact, the findings from the 2023 CES are the same as ones of an analysis of the Understanding America Study, another representative sample of U.S. adults.

“Homeschool one, homeschool all”?

In another study, we found that homeschooling households engage in a significant amount of sector mixing, which challenges the notion that a child who was homeschooled also has siblings who were homeschooled. Another recent study asked a nationally-representative sample of over 3,000 parents of school-age children about the educational sector of each child. The vast majority of these parents (77 percent) reported having at least one child enrolled in a traditional public school (TPS – see column 1 of Figure 2). Other households had at least one child enrolled in a public charter school (5.6 percent), enrolled in a private school (10.8 percent), or homeschooled at least one child (7.8 percent).

Because the survey asked about each child within the household, the data provide insights into sector mixing— that is, the degree to which a household may choose different educational settings for different siblings. Households with at least one child enrolled in a charter school, enrolled in a private school, or homeschooled often had at least one sibling enrolled in some other educational sector. Specifically, 26% of homeschool households also use other types of education from other children in the home. Many of these households enrolled at least one child in a TPS— 21% of homeschool households had at least one TPS child.

Interestingly, the data show that TPS households are the most likely to be sector exclusive and the least likely to engage in sector mixing. Ninety-two percent of TPS households exclusively used traditional public schools for their children. Only 1% enrolled at least one child in a charter school, 3% used a private school, and 2% homeschooled. The data clearly show that “homeschool one, homeschool all” is a misconception of homeschooling families, who are more similar to charter schools and private schools in terms of sector mixing. In fact, if such a claim of exclusivity could be made about any educational sector, it should be made about TPS.

Figure 2: Sector Mixing within Household

Note: Author’s calculations based on data from Watson and Lee’s homeschooling study.

What We Can Conclude

Since the rise of modern homeschooling, two stereotypes have proliferated. One stereotype—“once homeschooled, always homeschooled”—suggests that those who choose to homeschool a child will make the same decision about that child’s education every year. Another stereotype—“homeschool one, homeschool all”—suggests that families that homeschool, whether ideologues or pedagogues, do so for reasons that would extend to all children in that household. Both stereotypes should rightly be considered misconceptions as our different analyses show.

This evidence about the ways families rely on different sectors to educate their children amplifies the importance of longitudinal data when considering the influence of homeschooling on American students. Since many homeschoolers switch in and out of homeschooling, data collected at a single point in time likely underestimates how many U.S. families choose to homeschool. Classifying the sector in which someone was educated based on a single point in time similarly may be inaccurate.

Policymakers should also restrain themselves from making homeschooling policy based on a specific segment of the homeschooling population. For instance, the rationale behind proposals for significant state oversight of homeschooling assumes that the” once homeschooled, always homeschooled” and “homeschool one, homeschool all” generalizations are true. But for the many families who only homeschool their child for one to two years or avail themselves of charter, private, or traditional public schools for some of their children, it is less clear that such oversight is needed or desirable.