An Interesting Canadian Homeschooling Study

I started working with homeschoolers because while I was on the faculty at Ball State University, my best chemistry and physics students were homeschool graduates. At the time, I knew almost nothing about homeschooling, but since it was sending me my best university students, I thought I should investigate it a bit. As I looked through the academic literature that was available at the time, I saw that the few studies which had been done on homeschooling were in accord with my observation: homeschooled students are academically superior to their publicly- and privately-schooled counterparts.

Since then, a wealth of studies have been done on homeschooling, and the results are essentially the same. In 1999, for example, Rudner showed that homeschooled students academically outperform their publicly- and privately-schooled counterparts at every grade level.1 In 2010, a nationwide study of more than 11,000 homeschooled students showed that homeschooled students score, on average, more than 30 percentage points above the national average in all subject areas tested (which included math, science, social studies, reading, and language).2 In general, when you attempt to measure academic performance, the average homeschooled student beats the average privately-schooled student, and the average privately-schooled student beats the average publicly-schooled student.

Well, a recent study of homeschooled elementary students in Canada has produced very similar results, but in a different way. In addition, it actually compared two different homeschooling styles, and the results are very intriguing.

The authors (Sandra Martin-Chang, Odette N. Gould, and Reanne E. Meuse) of this new study3 decided to take a more “personal” approach than many of the previous homeschooling studies. Instead of relying on third parties to administer standardized tests in a wide variety of settings, for example, they administered the tests themselves. This allowed them to control the setting of the test, making sure that each student was tested in the same way. In addition, rather than comparing the homeschooled students to the national average (as Ray did in his 2010 study), they actually selected a comparable group of publicly-schooled students and tested them as well. As a result, they compared homeschooled students who they had tested to similar publicly-schooled students who they had also tested.

The authors also spent some time learning about the methods the parents used to homeschool their children, and that’s what makes this study intriguing. They found that slightly more than two-thirds of the students were homeschooled using a structured environment, where the parents tended to use premade curricula and some sort of lesson plan. However, the others were homeschooled in a much less structured way. The children’s day-to-day activities were used as the guide for their education. When they went to the store with their parents, for example, they learned about math. When they watched shows like Little House on the Prairie, they learned about history. For lack of a better word, this is often called an “unschooling” approach to home education.

Since the authors felt that these two methods of education are quite different, they decided that it would be unwise to group all the homeschooled students together. Instead, they compared the “structured” homeschooled students to the publicly-schooled students, and then they compared the “unstructured” homeschooled students to the publicly-schooled students.

What were the results? When comparing the “structured” homeschooled students to the publicly-schooled students, they found that the homeschooled students were academically superior in all subjects tested (letter-word recognition, comprehension, “word attack,” science, social science, humanities, and calculation). As the authors note:

…the children who received structured homeschooling were superior to the children enrolled in public school across all seven subtests.

Even though they tried to pick comparable publicly-schooled students, such factors as parental income and education were not quite equal. As the authors note:

Although we made efforts to ensure that the two groups were drawn from similar populations…mothers’ education and median income were slightly higher for the public school group. It should be noted, however, that this would have been expected to bias the study against finding a homeschool advantage. This was clearly not the case.

So despite the fact that the structured homeschooled students came from families with lower incomes and mothers with less education, they were still academically quite superior to their publicly-schooled counterparts. That says quite a lot!

Now for the bad news. In the authors’ analysis, the “unstructured” homeschoolers underperformed compared to the publicly-schooled students. As they note:

…the children in public school had a higher mean grade level for all seven measures compared with the unstructured homeschoolers.

So while “structured” homeschooling produced students who scored significantly higher than comparable publicly-schooled students, “unstructured” homeschooling produced students who scored lower than the same publicly-schooled students.

Now while these results are intriguing, I don’t see them as the final word by any means. This was a very small study. It had to be, given the fact that the authors tested each student. The “structured” group of homeschoolers had 25 students, and the “unstructured” group had only 12. There were 37 students in the publicly-schooled group. Such small studies make it hard to draw serious conclusions, unless the measured differences are very large. In the case of the comparison between “structured” homeschooled students and publicly-schooled students, the differences were large enough to produce a solid conclusion, despite the small size of the study. However, in the comparison between “unstructured” homeschooled students and publicly-schooled students, the authors caution:

Given the small sample size, none of the comparisons were statistically significant…

So the authors acknowledge that the “unstructured” homeschoolers’ poor performance compared to publicly-schooled students might not be real – it might be a statistical artifact.

Now I am inclined to think that “unstructured” homeschooling will not produce the strong academic benefits that “structured” homeschooling will. Even though this study hints that my inclination may be correct, the result is anything but conclusive.


1. Rudner, L., “Scholastic Achievement and Demographic Characteristics of Home School Students in 1998,” Education Policy Analysis Archives, 7:1–38, 1999. (available online)
Return to Text

2. Ray, B. D., “Academic achievement and demographic traits of homeschool students: A nationwide study,” Academic Leadership: The Online Journal, 8:1–31, 2010. (available online)
Return to Text

3. Sandra Martin-Chang, Odette N. Gould, and Reanne E. Meuse, “The Impact of Schooling on Academic Achievement: Evidence From Homeschooled and Traditionally Schooled Students,” Canadian Journal of Behavioural Science, 43(3):195-202, 2011.
Return to Text

25 thoughts on “An Interesting Canadian Homeschooling Study”

  1. As a structured homeschool grad, it(any difference between unschooling and homeschooling) is something that I always wondered about, although as you say, it’s certainly not definite that there’s a difference.

    1. Vivielle, when I speak at homeschooling conventions, I am often asked if studies have been done contrasting unschooling with structured homeschooling. This is the first study in which I have seen that done. It is not definitively, but it does seem to be the first attempt to answer the question.

  2. You know, talking to my other homeschooling mom friends who have actually graduated kids from high school… They say you really don’t see the test score differences that much until high school. In our state, we have to test in 3rd, 5th, 7th, 9th, and 11th. In 3rd, 5th, and 7th, homeschool kids run about the average. But, in 9th and 11th the test scores really diverge. But, I only know of one mom that “unschools” and haven’t really talked to her much about that choice. Most moms I know use some curriculum, even if they’re eclectic, and use maybe Mystery of History for history, Apologia for science, Math-U-See for Math, etc. I’ve seen that structure is better for my kids than unstructured, that’s for sure. I even do “school light” during the summer (math, spelling and reading, since 2 of my 3 kids are dyslexic) so we have some structure to our days.

    It’s always interesting to see how these studies turn out.

    1. Thanks for your comment, Trish. Most studies see a difference for all grade levels, so it is interesting that you and your homeschooling friends don’t see that. Now, most studies do see that the gains made by homeschoolers grow in the high school years. Thus, high school homeschooled students score even higher compared to publicly-schooled students than do elementary homeschooled students. Nevertheless, the studies see higher scores for the average homeschooler at all levels.

      As you say, it is interesting to see how the studies compare to our individual (and small group) experiences.

  3. Fascinating reading. All the “unschooler” families I have met and spent significant time with have made me seriously worry about the education their children were receiving, as well as the future of freedoms we homeschoolers are used to.

    1. Thanks for your comment, Leigh. I guess one thing to keep in mind is that parents are rather good at determining what works best for their children. As you can see in any school setting, there are some students who simply do not do well with a structured approach. As a result, they tend to be low achievers in a standard school setting. It is possible that many unschoolers have such children, and they are simply choosing the best situation for those kids. It is possible that even though they score lower than the average publicly-schooled student (if this study is accurate), they are still doing better than if they were forced to do a structured approach.

  4. I think part of it, whether structured or unstructured, homeschooling parents are just more involved in their kids lives, and that makes a huge difference. We look at the abilities and weaknesses of our kids, and adjust accordingly. We know we might have to work more on math with one, and more on reading with another, and so we do it.

    I know my parents were very involved in my life, even as a public school kid. My mom helped me study for tests, proof read term papers, and listened to my complaints. My dad bought me my first computer in 9th grade, coached my soccer teams, and yelled at teachers who crossed the line (there was one in particular…). My parents never missed a band performance or concert. And, both of my parents worked. I was one of those kids who did very well on standardized tests (90th percentiles).

    So, I really see it as how involved the parents are that makes the difference. But of course, what they teach in public school now and what I was taught 20-25 years ago is different, too.

    1. Trish, you are most certainly correct that parental involvement is a huge factor. Studies show that the more educated a parent is, the better the student does in public school. However, homeschooled student outcomes are mostly independent of parent education, and they are higher (on average) than the publicly-schooled students whose parents have the highest education levels. Most people interpret these results as showing just how important parental involvement is. Publicly-schooled students benefit from their parents’ high education level not because of their parents’ high level of education, but because their parents’ high level education makes them more likely to be involved in the student’s education. Homeschooled students don’t benefit from their parents’ high level of education, because the parents are already heavily involved in the students’ education.

  5. The hard part about these studies is there are always exceptions. My family is probably one of those. (Neither of my parents had college degrees, although they each had some college. My dad is dyslexic, and so is my brother. Neither my mom or I are dyslexic. But, it turns out I married a dyslexic and 2 of my 3 kids are dyslexic.) God has been gracious to me in overcoming what would be considered by these “studies” as road-blocks to being successfully educated. I don’t think you can ever rule out the God factor when it comes to individuals.

    1. That’s certainly true, Trish, but it is true of all studies that involve people. God has designed a huge variety in the human population, and as a result, studies can only talk about the average person. This is why these kinds of studies need to be larger than what was done in this particular case. The more people you have in a study, the more those differences average out, and the more clear the results of the study become. Even then, there will always be some exceptions.

      This is true of education studies, sociological studies, and even medical studies. A medical study that indicates a particular drug is safe might have hundreds of thousands of subjects, but that just means is it is safe for the vast majority of people. There will always be some who are allergic to the drug, have a bad reaction to the drug, etc. The key is to determine (educationally, socially, or medically) whether or not you are an exception to the norm. That way, you know whether or not the study applies to you. Just as people who are allergic to eggs should not get the flu shot, there are some children who should not be homeschooled in a structured way, even if structured homeschooling produces the best academic results for the vast majority of homeschoolers.

  6. I was a mostly-structured home-schooled student. I really wish I had been more unschooled in the sense of learning outside the box. There is definitely a place for structure, but it must go along with an ability to explore and discuss in a free-will setting. Otherwise, either the will to learn will die or self-discipline will suffer. I thoroughly enjoyed using your Exploring Creation books. Your books were one of the best things that happened to my science education. They allow for both structure and creativity. Your recommended further reading, as well as the conversational style itself allowed for both aspects very well. I would love to see an Exploring Creation with Mathematics series as well. I used Saxon from kindergarten through calculus and do not feel that mindless repetition really helped me learn in this way.

    1. Thanks so much, WSH. I am glad that my books were helpful for you. If done right, I think homeschooling can combine the best of structured and unstructured education. For example, you can start out with a structured curriculum and a lesson plan, but if the student wants to explore a new concept in more depth, you can ignore the lesson plan for a certain amount of time and let the student explore on his or her own. This requires the parent to be willing to ignore some of the lesson plan goals while at the same time make sure the student is productively learning while he or she is exploring.

  7. Saxon is part of why I can sit and do Jackson Electrodynamics integrals without resorting to Mathematica – unlike my friends. Most college students I’ve seen (I’m teaching now) have an unreasonable amount of trouble trying to do anything beyond the simplest algebraic manipulations, and when the problem gets big they just give up. This is even true among physics graduate students.

    Saxon teaches discipline. It makes you do tons of problems, and the problems are hard given the concepts they’re trying to teach. They seem mindless only after you’ve done five hundred of them. And that’s the point: to make the calculations mindless. Then you can go on to do real things with them (given a cylindrically symmetric nucleus, its mean radius, and its quadrupole moment, calculate the difference between its semimajor and semiminor axes).

    1. That’s a great skill, Jake. Being able to do integrals analytically will help you no matter what physical science related field you choose!

  8. Jay –

    We do structured homeschooling, and I do think that it is generally more beneficial than someone doing a complete unschooling. Nonetheless, I would imagine that unschoolers actually have a different set of values, and what is reflected in the study is the comparison to the standard set of values. For instance, (just guessing – I really don’t know) unschoolers *may* be better at things such as community life, helping others, and pursuing their own dreams. I don’t know if this is true or not, but this is one of the reasons I believe in educational freedom – I don’t expect everyone else to share in my educational values.

    In fact, one of the main reasons I started homeschooling was for the connection the kids would have to the family and community – something I simply didn’t pick up on in school, but had to learn later in life, as well as how to integrate my faith into all of my knowledge. I was aware of the academic benefits of homeschooling, but would have done it even if homeschooling was the lesser option academically, because I believe there are values that are more important than the education, and those are the ones I want my children to grow up with more than anything.

    Anyway, all that to say that I imagine that unschoolers value different things in education and childrearing. Whether or not they are achieving those, I don’t know, but I would imagine that whatever it is, standardized tests probably don’t cover it appropriately.

    1. Thanks for your comment, Jonathan. That is a very good point. I am so focused on university preparation that I tend to look at education through such a lens. However, I specifically point out in my lectures that not even the majority of students should go on to university. Thus, it is rather short-sighted of me to look at academic preparation as the “end all, be all” of secondary education!

  9. I wonder if test taking skills have not yet been mastered by the ‘unschoolers.’ I know from my own experience home educating my children that test taking must be taught and simply cannot be ignored until high school.

    I like the idea of unschooling in that one does not simply pattern school after the PS model. There IS room for creativity and life while a schedule is in place. I think that’s what I like about the Charlotte Mason approach. We have our schedule after all because all of society is built upon making and keeping schedules and deadlines. It seems to me that those children taught early-on to organize and follow a set schedule display more self-discipline in the long run.

    Naturally, the flip side for the unschooling philosophy is that there is more ‘un’ than actual schooling. I think a more relaxed approach combined with a schedule that benefits each child might be the best model for this method.

    Just my .02 = )

    1. Thanks for your comment, Brenda. I think it was worth more than $0.02! The issue of test-taking skills is important, and once again, it is something I really didn’t think about. You are correct that taking tests does require practice, and the poor performance of the unstructured homeschoolers in this small study might be due, at least in part, to a lack of practice.

      I also agree that flexibility is very important in homeschooling. Diana Waring (a titan in the homeschooling community) tells the story of a homeschooling family that was traveling across the U.S. during the “school year.” They were planning to stop by Mount Rushmore as they passed it, but they canceled the stop because their children needed the time to finish their workbook assignments. Obviously, the family’s rigid adherence to educational structure resulted in less education on that day!

  10. I agree with the point about test taking. It stands to reason that the further your schooling model is from the convention, the more poorly tests designed to assess that convention can measure your performance. To some extent this reflects intrinsic failures in the examination structure itself, since results depend more on “exam technique” than on actual knowledge and understanding.

    Nonetheless I do believe that some structure is needed, whether or not an individual intends to try for University. Otherwise it is far too easy for candidates to have a speckled education, with whole fields in which they have virtually no understanding. This is especially dangerous if interest alone drives them, for interest is a fickle thing. Some sort of curriculum at least ensures foundational understanding around the board. At least for keen learners, ensuring that they study a low interest topic will not, even cannot, prevent them from learning about the areas in which they have real passion. From there they can supersede the structure of the school or home school curriculum.

    1. Thanks for commenting, Josiah. You are correct, especially about interest being a fickle thing. Everyone needs to learn things in which he or she is uninterested. First, it adds to educational discipline, and second, it can spark an interest that the student didn’t even know existed!

  11. Not to completely change the subject… But I have to ask… Why do scientists, psychologists, authors, etc. bother to conduct studies that they are just going to declare statistically insignificant? I mean, frankly in this case, it kind of looks like the results weren’t what they wanted, so they were written off as not significant. I know that’s not the case… even though it left me in tears on a regular basis… I did learn something in my statistics class. I guess I just don’t understand conducting a study if you can’t draw any “significant” conclusions from it.

    If the study is conducted and it produces results that are not significant… Okay. But here it all comes down to sample size. If you know your population is a certain size, then draw a big enough sample to get good results. If you can’t, save your time and money until you can. Am I crazy?

    1. I am not going to answer your last question, Black Sheep, as I think we all know the answer to that one! However, I will answer your first one. I don’t think researchers start out with a study they know will produce insignificant results.

      I think there are two reasons small studies like this are done. The first reason is that the goal is limited. In this study, for example, they had not planned to compare “structured” and “unstructured” homeschooling students. They were planning on just comparing publicly-schooled students to homeschooled students. I think their sample size was adequate (not great, but adequate) for that purpose. However, once they got the data in, they noticed that there was a natural distinction among the homeschoolers, so they decided to learn what they could from it. I would hope that this will lead to a larger study that is designed specifically to compare “structured” and “unstructured” homeschoolers.

      The second reason small studies are done is to see what is out there. Often, investigators don’t know enough about the details of a specific social phenomenon to know what they should look for. As a result, they do a small, cheap study just to learn more about the phenomenon, and then they design a larger study once they decide what they will specifically focus on. However, since they spent time and money on the small study, they feel like they need to publish something, even if the conclusion is not significant.

      I would also say that these investigators weren’t looking for a specific result. One comment they make in their paper is that too many homeschooling studies have been done by homeschool advocates. They say they have no position on whether or not homeschooling is good, so I doubt that they had a desired outcome in mind.

Comments are closed.