If you’re at all connected to the world of homeschooling, you may have seen this snazzy new infographic on homeschooling floating around recently.
Unfortunately, this infographic is highly misleading in its content and assertions. Using data in this way, though, is sadly common in the homeschooling world, and that’s something that is worth addressing. In the name of good information and accurate data, I am going to run down this infographic and offer some pointers on where the information is incorrect or misleading. My commentary here may also be of interest in pointing to what we do and do not know about homeschooling, and to the limitations of much of the data we currently have.
The first date on the list states that in 1840 some 55% of children went to school while the other 45% were educated in the home or by tutors. This is misleading. For one thing, school then was not like school today—most children who went to school did so for only a few years, and school terms were often only a few months long. For another thing, very few parents in 1840 could hope to afford tutors. In reality, most children in 1840 started working at an early age, some in brutal conditions in factories and mines, and most children never learned much in the way of academics beyond reading, writing, and basic arithmetic. The homeschool lobby often paints a picture of modern homeschooling as natural and public education as some sort of historical aberration. This is both incorrect and irrelevant.
Next, it’s factually untrue that in 1980 homeschooling was only legal in 20 states. Homeschooling was on the books either implicitly or explicitly in 36 states in 1980, and had been for some time. In the remaining fourteen states, parents wishing to homeschool often found that they could do so by operating their homeschools as private schools or by getting permission from the local superintendent. Yes, some states had requirements that went too far, and yes, some local superintendents were difficult to work with. Some states required parents who wished to homeschool to have teaching licenses, and some superintendents ignored the law and unfairly targeted homeschoolers. However, the statement that homeschooling was only “legal” in 20 states is false. Milton Gaither goes over all of this in his chapter “Making it Legal” in his book Homeschool: An American History, which is conspicuously absent from the list of sources at the end of the infographic.
The last thing I’ll mention about the timeline is the listing for 1983. What were the “changes in tax laws” that which supposedly forced many Christian schools to close and caused homeschooling rates to soar? The answer is Bob Jones University v. United States, in which the Supreme Court held that if a religious university practiced racial discrimination, the federal government could revoke its tax exempt status. This was about whether Christian schools could be racist and retain their tax exempt status. Regardless, it is untrue that this decision “forced many Christian schools to close.” Many Christian schools have closed since the 1980s, which was the heyday of the Christian school movement, but the reasons have to so with the challenges involved in running a school and the fact that the rise of homeschooling decreased enrollment in Christian schools, not changes in tax law. It is odd that this innacurate entry would be included on this infographic, especially when something hugely important to the history of homeschooling did happen in 1983—that was the year James Dobson had Raymond Moore on his radio show, which provided many Christian parents with the endorsement they needed to make the jump to homeschooling and directly contributed to the growth of Christian homeschooling.
This section is fairly accurate, with a few small caveats. First, Iowa’s law changed several months before this infographic came out, and it no longer requires either assessments or notification but is still listed here as such. Second, the criteria used to highlight six states and list them as states with additional requirements such as curriculum approval, parental notification, or home visits” is a bit unclear. For one thing, it ignores the fact that a full 11 states have some form of parent qualification requirements. For another thing, in spite of its suggestion no states uniformly require home visits, and even those that require home visits in case of a problem stipulate that those home visits must not be unannounced. The left side also has a some inaccuracies. Indiana has a tax credit for homeschoolers, but this isn’t listed.
There’s one more thing to add regarding this section. It is true that 11 states require no notification or assessments, 14 states require only notification, and 25 states require notification and assessment. However, it must be borne in mind that most of the states that require assessments have additional options—such as operating as a private school—that allow many homeschoolers to bypass these requirements. Further, many of these states do not do much in the way of ensuring that their assessment requirements are actually followed. Finally, several of the states that require standardized tests do not require students to hit any particular threshold. Unfortunately, it is really hard to put this sort of nuance into a quick chart for easy reference.
This data is accurate but is slightly outdated given the release of the NCES data for 2011 (for an up to date listing, see here). Still, it’s a fairly good picture of the current number of homeschooled students, and of the growth of homeschooling in the past decade and a half.
This section relies on data from a study Brian Ray conducted for HSLDA, a homeschool advocacy group, in 2009, and here’s where we come to a problem. Ray’s study does not actually tell us anything about the average homeschool parent or the average homeschooled child. Ray recruited participants through major testing services, meaning that his study only included homeschooling parents who choose to have their students tested, which self-selects for a certain type of parents. Further, Ray had an absolutely terrible response rate; no social scientist would ever dare to suggest that a study with such a low response rate was representative of the population it was targeting. This means that only the most dedicated and involved homeschool parents volunteered to participate in Ray’s study. In other words, Ray was studying what one might call the cream of the crop of homeschooling families. It is incredibly deceptive to suggest that Ray’s results could be generalized to the homeschooling population as a whole, and yet that is what this infographic does. This would be like passing out surveys at suburban PTA meetings and then passing off the results as though they are representative of the average public school family.
So here’s a question. Do we actually know the education levels of homeschool parents? Yes, actually. We have data on this point from the National Center for Education Statistics, which conducts studies on random samples of the population. Unfortunately, while the NCES sample size is large, some of the numbers bounce around enough to suggest that it may not be quite large enough. In addition, a significant population of homeschooling parents are anti-establishment, and would not participate in the NCES should they receive one of its surveys in the mail. So the NCES numbers are not perfect. However, they still offer a better idea of homeschool parents’ educational backgrounds than does any other piece of data we have. And what do those numbers say, exactly?
That looks a bit different, doesn’t it? According to the infographic, only 1.4% of homeschool fathers and 0.5% of homeschool mothers do not have high school diplomas. But according to the NCES data, a full 11% of all homeschooling takes place in families where neither parent has a high school diploma. Further, according to the NCES data, only 39% of homeschooling takes place in families where where at least one parent has a bachelor’s degree. This contrasts with the infographic’s claims that a 66.3% of homeschool fathers and 62.5% of homeschool mothers have bachelor’s degrees. While the NCES data isn’t perfect, it is the best data we have and the only data on this subject that uses a randomly selected sample and therefore the only data that can be argued to be representative of homeschoolers as a whole. Interestingly, the NCES data suggests that homeschooling parents have, on average, essentially the same level of education as other parents (compare the two here and here).
Rather than relying at the NCES data, which is the best data we have, this infographic relies on the data from Ray’s descriptive study. Remember that Ray was studying the cream of the crop of the homeschooling population. It shouldn’t be surprising, then, that his sample was much more highly educated than the average parent. What should be surprising is that this infographic would be so misleading as to claim that those numbers represent the homeschool average. It’s especially odd given that the creators of the infographic were obviously aware of the existence of the NCES data, because they used it as their source for the information in the very next section.
This section uses the NCES data from 2007. While this data has some limitations, as I’ve already mentioned, it is nevertheless the best data we have. Like the section on the number of homeschooled children, this section is a breath of fresh air. I would however offer two small caveats, however: First, the questions the NCES uses to ask about parental motivation for homeschooling are not well designed, which can be challenging for researchers. Second, the infographic only lists what each family put as their number one reason for homeschooling, and leaves out how many parents chose each option when they were allowed to mark more than one reason. You can read that information here, and you can see how parents reasons for homeschooling have changed over time here.
In contrast, this section is a disaster. Once again, the infographic relies on data from Ray’s 2009 study, and once again that is a problem. Let’s talk just a little more about this study. Ray released his findings in an easy-to-read summary in a highly deceptive manner. In other words, Ray’s findings do not actually say what he says in that summary that they say. Ray’s full study, which he published in 2010, can be read here. In his published study, Ray explains that “this is a nationwide, cross-sectional, descriptive study . . . . It is not an experiment and readers should be careful about assigning causation to anything.” He also states that “it was not possible within the constraints of this study to confirm whether this sample is representative of the population of home-educated students.” In other words, Ray was not studying a random sample and his results cannot be assumed to be representative of homeschooled students as a whole. Yet, both the summary of his findings and this infographic state that Ray’s findings are representative. I am at a loss as to whether this is the result of ignorance or malicious intent.
Remember that Ray found that the participants in his study had parents who were significantly more educated than the average parent. Remember that the NCES data suggests that homeschool parents are not actually on average more educated than the average parent. This alone makes it crystal clear that Ray is not studying the average homeschooler, and that his results cannot be generalized to the homeschooling population as a whole. More than this, Ray does not correct for background factors. His participants are from families that are more educated, more financially well off, and more married than the average family. We know that public schooled children from families that are more educated, more finically well off, and more married outperform other public schooled children. Because Ray does not correct for background factors, which would involve comparing his participants to their demographically matched peers rather than to the national average, what he is actually studying is not the effects of homeschooling but rather the effects of growing up in families that are more educated, more financially well off, and more married. And yet, this infographic suggests not only that Ray’s results reflect the average homeschooler (they don’t) but also that their higher level of performance can be credited to homeschooling (without correcting for background factors, which Ray does not do, it cannot).
So, what do we actually know about homeschooled students academic achievement? A number of things, actually. The state of Arkansas tests homeschooled students annually, and releases those findings. It finds that homeschooled students score in the low 60th percentiles, and score far better in reading than in math (in math, they are only slightly above the public school average, and are sometimes below it). Unfortunately, we do not have the data to correct for background factors, and therefore cannot say whether or not these students outperform their public school peers. By this I mean that, for example, students whose parents are better educated do better in than other students regardless of whether they are educated in public, private, or homeschools. If these Arkansas homeschool students have parents who are, on average, better educated than their peers, that may explain their slightly higher test scores. The Rudner study, conducted in 1997, found a full 20 percentile point difference between the children of homeschool parents with college degrees and the children of homeschool parents without high school diplomas, indicating that parental education does affect homeschool performance. We also know that homeschooled students are more likely to have two parents living in their home than are public schooled students, and studies suggest that having two parents living in the home improves students test scores. In other words, we do not know whether Arkansas homeschoolers’ slightly higher test scores are a result of homeschooling, or of factors such as parental education and family status. Regardless, the Arkansas numbers show that this infographic’s claims that homeschooled students score on average in the 87th percentile are flat out false, and demonstrably so.
We also have some interesting numbers from Alaska, which I’ve played around with a bit. Most homeschool students in Alaska enroll in public school extension programs designed to serve homeschoolers. These programs offer students monthly contact with certified teachers and require parents to create an educational plan for each child each year and to file quarterly progress reports. The students are tested at the end of each year, and the parents receive around $2000 per student per year for use on textbooks, computers, and tutoring. Each of these extension programs must make its test scores public (here is an example), and I’ve compiled the scores for the largest of these programs as follows. In reading, 86% of homeschooled students were proficient or advanced as compared to 80.1% of public schooled students. In writing, 77.3% of homeschooled students were proficient or advanced as compared to 74.2% of public schooled students. In mathematics, 64.3% of homeschooled students were proficient or advanced as compared to 68.6% of public schooled students. In other words, these Alaska homeschool students outperformed the public school average in reading and writing and underperformed the public school average in math. This data does come with background factors, though the results of that analysis are too complex to discuss here. Of course, it should be remembered that this is only homeschooled students enrolled in the public school extension programs, and does not include those who homeschooled under Alaska’s minimalist homeschool statute. Still, the results are intriguing.
What else do we know? Well, we know that homeschooled students score slightly above average on the SAT and the ACT. However, a study of homeschooled students who take the SATs found that their parents were significantly better educated than were the parents of public schooled students taking the SAT, suggesting that homeschooled students’ slightly better performance on college entrance examinations like the SATs and ACTs may be largely attributable to their family backgrounds rather than to homeschooling. Further, homeschooled students are less likely than public schooled students to take the SATs and ACTs, suggesting that homeschool graduates may be less likely to attend college than public school graduates and that, as a result, homeschooled students SAT and ACT scores may be inflated by the fact that only the best of the best of homeschoolers take these tests.
Can homeschooled students succeed academically? Absolutely. Do all homeschooled students succeed academically? No. One question that never seems to be brought up is why successful homeschoolers succeed. First, homeschooled children in well educated middle class families would likely score well above average even if they attended public school simply because of their socio-economic status. Second, It takes hard work and dedication to succeed at homeschooling, it’s not like you can just apply the label “homeschooling” and get instant success. There’s also the fact that some students who would have wilted in public school thrive at home, and some students who would have thrived in public school wilt at home. Which educational method works best for which child may vary greatly, but many homeschooling advocates act as if homeschooling is some sort of universal formula that works automatically for every child.
This section relies on Brian Ray’s study of homeschooled adults. It should be noted that Ray is associated with HSLDA, the country’s leading homeschool advocacy group, and is not an unbiased researcher. For more, see this article. Anyway, in his study of young adults who were homeschooled, Ray did not use a random sample or correct for background factors. Further, his study was composed of highly leading questions and was not scientific in any sense (for more on Ray’s habitual use of leading and unscientific questions, see this critique of new study of homeschooled young adults that Ray is currently spearheading). In other words, Ray asked current homeschoolers to help him find homeschool graduates who would be willing to fill out a survey on their experiences. It’s not surprising that the homeschool graduates who would volunteer to participate in such a study would be those who most positive about their experiences. For a study conducted by Homeschool Alumni Reaching Out, a group dedicated to helping those homeschool graduates who experienced some form of abuse or educational neglect, see here. The group’s 2013 Basic Survey found that its participants ranked their academic experience in the home fairly low and experienced high rates of physical and emotional abuse. Neither Ray’s study nor HARO’s study used a random sample and neither can be assumed to be representative, but then, that’s kind of my point.
The only random sample we have of adults who were homeschooled is in the Cardus Education Survey, which corrected for background factors like family income or parental education. The Cardus found that being homeschooled negatively impacted a student’s future educational attainment and earning potential, and made students less likely to attend college and, for those who did attend college, more likely to attend a community college and less likely to attend a prestigious university. Further, they were less likely to be politically engaged as adults. Homeschool graduates were also significantly more likely to report “lack of clarity of goals and sense of direction” and “feelings of helplessness in dealing with life’s problems” than any other group. Finally, they were also more likely to marry early, to have fewer children, and to divorce. Again, this study actually corrected for background factors and it used a random sample. While further research is badly needed, the Cardus suggests that Ray’s congratulatory survey is far from the whole truth when it comes to homeschooling outcomes.
Finally, the section on how much it costs to homeschool a child is highly misleading. If I were to leave my career and homeschool my own two children, that would cost our family tens of thousands of dollars a year in lost income, not the $500 per child listed in the infographic. When a family homeschools, they give up an income, because, in effect, they have to have a teacher—i.e., one of the two parents—employed full time teaching the children. And yet, that cost, that sacrifice, is routinely left out of infographics like yours. If a woman who made a salary of $50,000 chooses to quit her job and stays home to homeschool her single child, it costs that family $50,000 to homeschool that child, even before counting in books and supplies. If that woman quits her job and stays home to homeschool her five children, it costs that family $10,000 per year to educate each of those children, even before counting in books and supplies. Suggesting that homeschool parents’ average of $500 per child in school books is a figure you can compare to the average $9,963 public schools spend on educating children is highly disingenuous, and that’s me being gracious.
Notice something about these footnotes? Very few are to actual good academic research on homeschooling. The history of homeschooling links are lay articles, not scholarly pieces. Lay articles have their place and are sometimes very good, but for something like creating an infographic it’s important to use sound academic data—and, well, this infographic is an example of what happens when you don’t do that. Now there is a link to the NCES data, which is some of the best data we have. However, the infographic relies on that data selectively and completely ignores it when it comes to areas like parental education (if you have any skill at reading charts, you might find these summaries of the NCES data interesting: here, here, and here). Further, there is not a single link to the International Center for Home Education Research, which was founded by Milton Gaither and Robert Kunzman and is probably the best resource currently out there on research on homeschooling. Finally, this infographic relies on Wikipedia for the history of education, rather than scholarly books and articles on the subject. This might help explain why its history section was so terrible, and it should be a sign that the people who put this infographic together did so without putting in a lot of effort.
This infographic does nothing but mislead readers and misrepresent data. It may be a nice pat on the back for those homeschoolers who want to make themselves feel good, but it’s not good research or accurate information. And you know what? This sort of terrible scholarship does nothing but hurt the image of homeschooling. These sort of infographics make homeschoolers look either ignorant or willfully deceptive. If you want facts about homeschooling, I suggest you go elsewhere.