The simple performance data shows the homeschooled averaging mean percentile ranks of:

• 81.5 in reading; 
• 81.3 in language; 
• 74.6 in math; 
• 79.5 in social studies, and;
• 80.4 in science. 

This ranks them comfortably in the top fifth of students nationwide. But curiously, only five of the 28 predictors were significantly associated with two or more subject area scores, meaning that most factors simply aren’t important.

The Dogs That Didn’t Bark

Some of the factors considered in this study were shown to not have an impact on test results.

Race Doesn’t Matter 

There was no remarkable pattern regarding racial differences across the five subjects, though homeschooled Asian and Hispanic students outperformed White and Black students in math. This “dog that didn’t bark” is important in the public school context, where Hispanic and Black students consistently trail a half-year or more behind Asian and White students.

Income Doesn’t Matter 

The homeschooling median family income was $75,000 to $125,000 USD, compared to the median national household of $103,000 or married-couple household of $119,000. Fully 22 percent of homeschoolers earned under $75,000, possibly qualifying for federal lunch support in a public school. Yet, for the homeschooled, there is a lack of significant findings for income, the survey finds. This is another “dog that didn’t bark” compared to public school students where conventional students from lower income families trail more than a year behind their affluent classmates.

Parental Education Doesn’t Matter 

The parental education ranged from 22 percent of dads and 18 percent of moms with a high school diploma, 38 percent of dads and 46 percent of moms with a bachelor’s degree, and 29 percent of dads and 26 percent of moms with advanced degrees, yet no significant relations emerged. Strangely though, previous national surveys have shown that the parents’ educational level does constitute a significant factor for students in brick-and-mortar schools.

So, another non-barking dog. Just five percent of homeschooling parents had any formal teaching certificate, calling into question the educational establishment’s pretense about the importance of their professional pedagogy.

The Student’s Sex Doesn’t Matter 

Not surprisingly, being female was positively associated with reading and language scores, being male was positively associated with math, and there was no sex difference with social studies and science. Still, the homeschooled enjoyed an advantage across the board, regardless of their sex.

The Years Homeschooled Doesn’t Matter

The strangest negative finding is that there is no association between test scores and the proportion of years spent homeschooling or the number of years a child has been home-educated. What does this mean? Some homeschooling proponents suggest that, if homeschooling confers an advantage for students in their early years, this effect should be cumulative, magnifying their advantage in each successive grade. Yet their advantage remains consistent, grade after grade, without pushing them up into ever-higher percentiles, compared to brick-and-mortar school students. Why is the advantage not cumulative?

The authors wonder, as children get older, it might be that homeschool parents focus with their children on learning topics and activities not related to standardized testing. All sorts of beneficial factors like personalized instruction and customized curricula might not show up as ever-better performance on standardized tests, while nevertheless nurturing ever-better educated students. In any event, the non-significant association argues against the conclusion that homeschooling negatively affects standardized test performance — in other words, homeschooling does not, year-after-year, contribute to ever-worse performance on standardized tests. So more study is needed.

The Dogs That Barked

1. The younger children are when they begin to receive formal instruction, the higher their reading, language, and math scores on these standardized tests.
2. The more time they spend in structured learning activities, the higher their reading, language, and math scores. The authors call these results reasonable, but hesitate to lean into structured early instruction. Since standardized tests were developed for institutional schools, they say, it’s hard to know what long-lasting effects structure and early instruction will have for the home educated. So again, further studies are needed.
3. Some might expect that more state regulation of homeschooling would produce a greater sense of accountability and effort on the part of parents. But in fact, more state regulation of homeschooling was associated with lower math and social studies scores. The authors resist the temptation to speculate on why this might be the case. One might wonder whether state regulation subtly coerces parents into using the standard public curricula, particularly with math and social studies. Those subjects are where Progressive or Post-modern ideology can have the most deleterious effects.

Why Parents Homeschool

Most parents (57 percent) begin homeschooling their children at age five or six, though almost all the rest (40 percent) begin even earlier, age four or less. The vast majority (59 percent) spend three to five hours daily in structured learning, while 17 percent spend up to two hours, and 14 percent six or more.

On a 1-to-5 unimportant-important scale, 80 percent of the parents said their reasons for doing so were: 

• Emphasize family life (4.49); 
• Provide moral instruction (4.45); 
• Customized educational experiences (4.35); 
• Dissatisfaction with public school academics (4.34);
• Concerns about school safety, drugs or peer pressure (4.32), or;
• Provide religious instructions (4.07). 

Interestingly, the primary issues for Black parents were providing moral or religious instruction.

About a third of parents said their child has special needs that the school can’t, or won’t, address, which was somewhat (16 percent), quite (10 percent), or extremely (10 percent) important.

Ray and Hoelzle note their survey’s limitations. First, there’s no reasonable way to get a random sample of homeschooling families, so the results they did get were limited to the families who stepped forward when contacted by their state associations. Second, with an 83 percent White cohort, minorities were underrepresented. And third, the results were a snapshot in time, and there were a dozen other factors not considered (e.g. pedagogical approach). So, as good sociologists, they say considerable care should be taken in making generalizations.

That said, the issues of structure and early instruction are worth pondering, particularly with the foundational skills of literacy and numeracy — the primary issues with standardized testing. Reading and calculating are after all foundational, language-learning, and perhaps dependent on developing neural pathways or a kind of muscle memory, unlike more developed skills like personal or causal narrative, poetic analogy, or speculative inference. If this were so, it might account for the mystery that the homeschooling advantage demonstrated in standardized testing alone would remain consistent, not cumulative over successive grades. Their advantage would lie in their ability to access the material, but there would be no standardized measure of their ability to manipulate or work with it.

But this, as the authors say, would require further study.

Joseph K. Woodard earned a BA at the University of Alberta (political philosophy), MAs at Dalhousie and St. John’s (Great Books), diplomas in Greek and Latin at Berkeley, and a PhD at Claremont. He spent fifteen years as an academic (Bethany, Brock), fifteen as a journalist (Alberta Report, Calgary Herald), and eleven as a tribunal judge (Citizenship Canada). He has helped his wife, Kathy, raise ten children through homeschooling and parent-run school start-ups, and has always taught as an adjunct at independent high schools and colleges. He writes for websites such as The Imaginative Conservative and Public Discourse, serves as director of research for the Canadian Centre for Home Education, and moderates Great Books seminars online for the Angelicum Academy.