One of the homeschool lobby’s most touted claims is that homeschooled students score higher than public school students on the SAT. This claim, highlighted in a June 2016 news release by the National Home Education Research Institute (NHERI), is based on a self-selected sample of homeschooled students—those who choose to take the SAT—and is not representative of the entire homeschool population. Nonetheless, NHERI’s report contains an interesting point that is briefly acknowledged but not fully explored. Using data from the College Board, NHERI reports that 13,549 seniors who graduated in 2014 and took the SAT at some point during their high school years identified as homeschooled. This number is alarmingly low, and suggests that homeschooled students may be taking the SAT—and attending college—at much lower rates than their traditionally schooled counterparts.
An estimated 3.2 million students graduated from high school in 2014. Of these, 1.7 million students, or 53%, took the SAT at some point before graduating. In 2014, an estimated 136,000 high school seniors were homeschooled. The numbers reported by NHERI indicate that roughly 10% of these students (or 13,549 individuals) took the SAT before graduating. This data is not the first to indicate a low level of homeschool SAT-taking. Analyzing data from 2001, Clive Belfield, an economics professor at Queens College, found that “home-schoolers made up only 0.5% of all SAT test-takers, a proportion considerably below their representation in the student population and lower than any other school type.” At the time, roughly 2.0% of students in grades K-12 were being homeschooled.
Low homeschool SAT-taking raises concern because the SAT is considered a proxy for intention to attend college. Eighty percent of four-year colleges and universities require all applicants to submit SAT or ACT scores. If homeschooled students take SAT at lower rates than other students, they likely also intend to apply for college at lower rates than other students. Because attending college is a sign of upward mobility, low college attendance rates could raise concern about homeschool outcomes.
While our analysis suggests that homeschool SAT-taking rates are lower than would be expected for this population, the rate is almost certainly higher than 10%. As we will explore, it is probable that as many as 50% of homeschooled students, possibly more, have access to non-homeschool SAT codes that they may use when taking the SAT. These students may be homeschooled through online charter schools, private “umbrella” schools, or correspondence schools. Assuming that these students use these non-homeschool SAT codes brings the rate of SAT-taking for the homeschooled students who remain up to 20%, a number still far lower than the 53% of public school students who take this same test—and concerningly low given the central role these scores play in college entrance requirements.
The remainder of this essay will address a variety of factors that help explain or provide context for homeschool SAT-taking rates. We will touch on the similarities we see in ACT-taking rates; homeschooled students with non-homeschool SAT codes; the role the SAT and ACT play for homeschoolers; the effect of background factors on SAT-taking rates; and whether community college attendance could help explain low SAT-taking. We will finish by discussing the implications of this data for both our current understanding of homeschooling and future research on homeschooling.
Homeschool ACT-Taking Rates
While the SAT is not the only test used to determine college admissions, homeschooled students’ low SAT-taking is not offset by a correspondingly higher rate of ACT-taking. The ACT website reports that 57% of the 2014 graduating class took the ACT, slightly higher than the 53% of graduating 2014 seniors who took the SAT. That same year, 13,435 ACT-takers indicated that they were homeschooled. This number is nearly identical to the number of 2014 homeschool seniors who took the SAT before graduating, and amounts to 10% of all homeschooled seniors. As with the SAT code, it is likely that as many as 50% of homeschooled students, possibly more, had access to non-homeschool ACT codes.
As with SAT-taking, low homeschool ACT-taking appears to be consistent over time. In 2001, when 2% of students were homeschooled, 0.5% of ACT-takers indicated that they were homeschooled. Recall that Belfield found that homeschoolers made up 0.5% of SAT-takers that same year, a proportion he noted was lower than that of any other group.
Some students take both the SAT and the ACT, though it is difficult to tell how many do so; most colleges will accept scores from either the SAT or the ACT. Removing the roughly 50% of homeschooled students who likely have access to non-homeschool SAT and ACT codes, 20% of the remaining students take the SAT, and 20% take the ACT. Even if no student took both tests—which is not the case—the combined rate of homeschooled students taking one of these two tests (40%) would be lower than the rate of students nationwide taking each test independently (53% for the SAT; 57% for the ACT). ACT data does not explain low homeschool SAT-taking. Instead, it adds a second concern—low homeschool ACT-taking.
Using the SAT and ACT Homeschool Codes
The College Board (which administers the SAT) and the administrators of the ACT determine the number of homeschooled test-takers by asking homeschooled students to use a specific homeschool code when they fill out their testing information. While some students move in and out of homeschooling, data from the NCES suggest that high school students are homeschooled at a rate similar to younger students. In other words, low SAT-taking rates are not a symptom of lower numbers of high school students being homeschooled.
Both national and state-level homeschool organizations advise homeschool parents to have their children use the homeschool codes when taking these tests, and the College Board advises public high schools that allow homeschooled students to take the SAT in their facilities to have these students use the homeschool code. There is no evidence of any systematic effort to undermine the use of the homeschool codes provided by the College Board and the administrators of the ACT. However, some students are homeschooled through programs that give them alternate codes to use when taking the SAT.
Twelve states either allow or require homeschooling to take place through enrollment in private “umbrella” schools or homeschool associations, which frequently have their own SAT codes. Other students are homeschooled through cyber charters or independent study programs run by public school districts, and would also not use the homeschool SAT code when taking the SAT. How popular these programs are is difficult to say. According to data from the National Center for Education Statistics’ 2012 National Home Education Survey, as many as 50% of students being homeschooled in 2012 were homeschooled but enrolled in a school (public, private, or charter) part-time, or were enrolled in school (public, private, or charter) but homeschooled part-time. Many of these students, though not all, likely had access to a non-homeschool SAT code.
Does this explain seemingly low SAT-taking? Not completely. If 50% of homeschooled students are enrolled in programs that give them access to non-homeschool SAT codes, 20% of the remaining students take the SAT, a percentage still far lower than that of students taking the SAT overall. It is possible, of course, that more than 50% of homeschooled students have access to non-homeschool SAT codes. However, to bring the SAT-taking rate up to its nationwide 53% among remaining homeschooled students, over 80% of homeschooled students would have to be enrolled in programs that would result in them being counted, for the purposes of taking the SAT, as public, private, or charter school students rather than as homeschooled students. While more research is merited, we find this unlikely.
Why are the SAT and ACT Important?
While tests such as the SAT and ACT have long been an important metric for all college-bound high school students, they have special importance for homeschooled students because they are frequently the only impartial third-party evaluation a homeschooled student can show a college admissions board. Indeed, colleges and universities place more weight on homeschooled students’ SAT and ACT scores than on those of other students.
The importance of the SAT for college-bound homeschooled students has been demonstrated by several researchers. A researcher who surveyed 51 colleges and universities in 1995 found that SAT or ACT scores were “the most common deciding factor” in homeschool graduates’ admission. Another researcher surveyed colleges and universities in Pennsylvania in 2003 and compared the weight admissions officers placed on various admissions criteria for both traditionally schooled and homeschooled graduates. He found that “while . . . SAT scores are important for both groups . . . the outcomes suggest that these scores are significantly more important information for making an admission decision regarding the former home schooled.” While admissions officers rated SAT scores as the third-most important admissions criterion for traditionally schooled students (less important than a student’s high school transcript and GPA), they rated SAT scores as the most important admissions criterion for graduates of homeschool programs.
While some universities and colleges no longer require SAT or ACT scores, the vast majority (80%) continue to make these scores a requirement for admission; others require these scores for homeschooled students only. Even Christian colleges, which often portray themselves as homeschool-friendly, tend to require SAT or ACT scores. Because homeschool diplomas and transcripts continue to be overwhelmingly created and issued by students’ parents, the external verification that SAT or ACT scores provide to admissions officers has continuing importance in demonstrating homeschooled students’ credentials.
Could Demographics Explain low SAT-taking?
We know that various demographic factors affect the rate at which students take the SAT. Do homeschooled students have demographic factors that would predict a lower rate of SAT-taking? While demographic data on homeschooled students is sparse, the NCES does include questions about homeschooling in its quadrennial National Household Education Survey. Using the NCES and College Board data, we can analyze what impact homeschooled students’ parental education, household income, or race should have on their SAT-taking.
In each demographic category, homeschooled students have factors that indicate more SAT-taking (more have parents with bachelor’s degrees, fewer live in low-income households, fewer are Hispanic) and factors that indicate less SAT-taking (fewer have parents with graduate degrees, fewer live in high-income households, fewer are Asian). In each category, these various factors even out, leaving homeschooled children with a roughly 1% increased likelihood of taking the SAT as compared to other students.
This is only part of the story, however. We carried out these calculations using the NCES adjusted demographic data for all homeschooled students. When we remove those students whose parents reported that they were homeschooled only part-time, the demographics change dramatically. Students homeschooled full-time—those most likely to be using the homeschool code rather than a code provided by another program—are substantially whiter and less poor than other students, and are less likely to have parents who have not completed high school and more likely to have parents with graduate degrees. Students with higher levels of parental education, in particular, take the SAT at a higher rate than other students. Given the demographics of full-time homeschooled students, we would expect them to be taking the SAT at a rate higher than the overall national SAT-taking rate of 53%.
What about Community College?
There is still one more potential explanation for low homeschool SAT-taking: community college. Community colleges do not generally require SAT or ACT scores for admission. If large numbers of homeschool graduates are attending community college and then transferring to four-year institutions with enough credits to be considered transfer students, these individuals would not need to take the SAT. Unfortunately, no data exists on the overall rate of homeschool attendance at community colleges. This information has simply not been collected. All we currently have is data on homeschool attendance rates at four community colleges studied by researchers interested in homeschool performance.
When researcher Jack N. Bagwell looked at a community college in South Carolina, he found that 3.6% of students enrolled at York Technical College (a medium-sized open enrollment community college in South Carolina) in 2007 had been (or were being) homeschooled. That rate is higher than the overall homeschool rate at the time (2.9% in 2007). However, Bagwell’s findings contrast with those of other researchers. When Molly H. Duggan looked at students at a multi-campus community college in 2010, she found that out of 39,000 students enrolled, only 171 students, or 0.4%, had previously been homeschooled. A large regional community college that responded to an exploratory query from CRHE in 2013 reported that only 0.1% of of its incoming students in 2012 were homeschooled. Based on data Benjamin G. Kramer reported in his 2012 study of a mid-sized Mid-Atlantic community college, an estimated 1% of students enrolled between 2004 and 2011 were homeschooled.
The research on homeschool community college enrollment is thus far limited to these four data points. More research is needed to determine whether homeschooled students are attending community colleges and transferring to four-year institutions in large enough numbers to offset their low SAT-taking. The information we currently have—the four data points provided by Bagwell, Duggan, Kramer, and our own exploratory study—do not suggest that homeschooled students are taking this route in significant numbers, adding to existing questions about what low SAT-taking may mean for homeschool college attendance.
What about data that suggest high rates of college attendance?
Homeschool advocates often claim that homeschool graduates have higher college attendance rates than other students. To make this argument they tend to cite Brian Ray’s 2003 study, Homeschooling Grows Up, which found that “[o]ver 74% of home-educated adults ages 18–24 have taken college-level courses, compared to 46% of the general United States population.” In a study of homeschool alumni conducted by Homeschool Alumni Reaching Out (HARO) in 2014, 87% of respondents reported that they had completed at least some college (CRHE provided the data analysis for this study). However, both studies relied on volunteer samples rather than using nationally representative randomly selected samples; while these studies’ results can be considered descriptive of their respondents, they cannot be assumed to be representative of the population of all homeschool graduates.
In a study of unschoolers (a term typically used for secular homeschoolers who take a child-led, hands-on approach to education) published in 2014, researcher Peter Gray found that 44% of respondents had completed a bachelor’s degree or were currently enrolled in a bachelor’s degree program. 83% of respondents had pursued at least some higher education. However, Gray’s study suffers from the same weakness that besets both Ray’s study and the HARO study—all three drew on volunteer samples rather than drawing on a representative cross-section of homeschoolers. Of the three, only the HARO survey reported the level of education obtained by respondents’ parents. The level of parental education reported by respondents was substantially higher than the average level of education held by homeschool parents as reported by the NCES, suggesting that the survey over-sampled homeschool graduates from college-educated homes.
The CARDUS Education Survey, conducted in 2011, is the only study of homeschool graduates that we are aware of to use a random sample. This survey, which looked only at individuals whose mothers attended religious services weekly, found that homeschooled students raised in religious homes had a lower level of educational attainment than non-homeschooled students raised in religious homes. Researchers found that homeschool graduates had lower SAT scores, completed fewer years of higher education, and were less likely to receive a college degree. In other words, the findings of the only study of homeschool graduates to use a random sample reinforce concerns raised by low homeschool SAT-taking.
A Call for Research
We could answer questions about low homeschool SAT-taking by looking directly at the rate of homeschooled students attending four-year colleges and universities, but such data does not currently exist. We do have data on the rate of homeschooled students attending a smattering of individual colleges. 7% of freshmen at Messiah College were homeschooled during their high school years, along with 12% of students at Grove City College; both are Christian colleges popular with evangelical homeschoolers. Other private colleges that report this data have far lower homeschool attendance rates. Elite private colleges and universities tend to report extremely low homeschool enrollment—Harvard, Princeton, and MIT report homeschool enrollment between 0.1% and 0.5%. The few data points we have on homeschool attendance at four-year state colleges or universities, institutions that educate far more students overall than private colleges, are troubling, ranging from 0.2% to 1%. A comprehensive study of homeschool college attendance is needed.
Homeschool advocates often point to positive studies of homeschooled students’ performance in college as proof that homeschooling not only works but actually has better results than other educational methods (there are also studies that find more mediocre performance). However, if only a narrow slice of homeschooled students take the SAT and attend college, it is to be expected that these students—the best and brightest—would both score better on the SAT and perform better in college than their traditionally schooled peers, not because homeschooling is a superior method of education but rather because lower-performing homeschooled students have been removed from the sample. CRHE believes that homeschooling should offer the same opportunities for success as other educational methods. If homeschoolers have a lower rate of college attendance than other students, this would indicate that they are not receiving these opportunities.
Given how little research has been conducted on homeschool graduates who do not attend college, there is not much we can say definitively about this group. Proponents of homeschooling argue that independent learning encourages entrepreneurship; perhaps these students have started their own businesses or found other ways to create careers without attending college. On the other hand, advocates for homeschooled alumni, such as HARO and CRHE, spotlight the stories of alumni who struggle, held back by their lack of education. We know very little about what goes into homeschool graduates’ decisions about college. Are they motivated by a desire to avoid debt, or by a distaste for traditional classroom learning? Are they held back by a lack of math attainment, or by a lack of knowledge about the admissions process in the absence of high school guidance counselors? We do not currently have the data we need to answer these critical questions. More research is needed on the full range of outcomes for homeschooled children—not just for the best of the best.
 We submitted a research request to the College Board in September 2016, hoping to gain access to this and additional data, but were informed that the College Board no longer releases data on homeschooling.
 Some states have begun requiring all students to take the SAT (or ACT) before graduating, but students from these states made up only around 6% of 2014 graduating seniors who had taken the SAT. Even removing these states from the count, a full 50% of 2014 high school seniors in the remaining states took the SAT at some point before graduating.
 The National Center for Education Statistics (NCES) has estimated that 1.77 million students were being homeschooled in 2012; of these students, 514,000 were in grades 9-12. Assuming that the homeschooling continued to grow at the same 3% annual rate at which it grew between 2007 and 2012, we estimate that there were 545,000 homeschooled students in 2014. For a number of reasons, we have chosen to assume that these students were evenly distributed between grades 9-12. While some state enrollment data shows a decline in homeschool enrollment in the later grades of high school, this is likely due to these students passing the age of compulsory attendance.
 Belfield found that 6,033 SAT-takers indicated that they were homeschooled; the NCES data suggests that there were 275,000 homeschooled students in grades 9-12 at that time, roughly 68,750 of them seniors. Like the NHERI release, Belfield puts homeschool SAT-taking rates at below 10%.
 The ACT has increased in popularity in the past decade, overtaking the SAT in 2012. While more students nationwide took the ACT than the SAT in 2014, homeschooled students were still equally likely to take the SAT as the ACT. It is possible that homeschool parents have yet to catch up with new trends in college admissions test-taking. It is also possible that higher ACT-taking nationwide is the result of more states using the ACT as a high school exit examination.
 Whether students take the SAT or ACT has traditionally varied widely by region. A study out of California, where students traditionally take the SAT, found that 39% of SAT-takers also took the ACT and 86% of ACT-takers also took the SAT. In other words, 37% of all California students who took the SAT or ACT took both tests. Whether the same percentage of test-takers take both tests nationwide is unclear.
 Using the data provided in footnote 8, we can estimate that roughly one-third of the estimated 50% of homeschooled students without access to a non-homeschool SAT code take at least one of the tests. However, we have no way of knowing whether California’s pattern of test-taking is reflected nationwide, or whether homeschooled students who the SAT or ACT take both at the same rate as other test-takers.
 These states are AL, AK, CA, CO, FL, LA, ME, MD, PA, SC, TN, and WA, home to approximately one-third of the nation’s school-age population, and roughly one-third of the nation’s SAT-takers. Alabama and South Carolina are the only states on this list that require all students to be homeschooled through umbrella schools or homeschool associations. We estimate that at least half of the students in these states are homeschooled independently, and not through alternate programs.
 68,000 is half of the estimated 136,000 homeschooled high school seniors in 2014. 13,549, the number of SAT-takers indicating that year that they were homeschooled, is 20% of 68,000.
 Villanueva, Brian D. “An Investigation of the Admissions Standards of United States College and Universities for Home-Schooled Students.” Home School Researcher, 13, No. 2 (1999), 1-6.
 Barno, Richard Joseph. “The Selection Process and Performance of Former Home-Schooled Students at Pennsylvania’s Four-Year Colleges and Universities.” Lehigh University, Ed.D. (2003), 140.
 Barno, ibid., 141.
 When we compare 2012 student demographic data with demographic data on 2012 SAT-takers, we find that not having a parent with a bachelor’s degree has a negative effect on student SAT-taking, while having a parent with a bachelor’s degree has a positive effect and having a parent with a graduate degree has an even stronger positive effect. According to the NCES, students being homeschooled in 2012 were slightly more likely than other students to have a parent with a bachelor’s degree but slightly less likely than other students to have a parent with a graduate degree. When looking at parental education alone, this generation of homeschooled students should be taking the SAT at the same rate as, or a slightly higher rate than, students overall.
 We do not have detailed household income data for homeschooled students more recent than 2007. However, by comparing that data with 2007 student demographic data and demographic data on 2007 SAT-takers, we can ascertain how this measure might have affected homeschool SAT-taking at that time. In 2007, living in a household with an income of less than $20,000 had a negative effect on student SAT-taking while living in a household making over $100,000 had a positive effect on student SAT-taking. Students being homeschooled in 2007 were slightly less likely than other students to live in households with incomes under $20,000 and slightly less likely to live in households with incomes over $100,000. As with parental education, when considering household income alone, this generation of homeschooled students should be taking the SAT at the same rate as, or a slightly higher rate than, students overall.
 By comparing 2012 student demographic data with demographic data on 2012 SAT-takers, we find that Hispanic students are less likely than other students to take the SAT and that Asian students are substantially more likely than other students to take the SAT. In 2012, homeschooled students were more likely to be white and less likely to be black, Hispanic, or (especially) Asian. As with the previous two factors, when looking at student race alone, this generation of homeschooled students should be taking the SAT at the same rate as, or a slightly higher rate than, students overall.
 Because the NCES has only released the percentage of full-time homeschooled students who are poor or non-poor, we cannot make a direct comparison between this data and the College Board data, which is released by income level rather than by poverty status. As for race, the College Board data suggests that white students take the SAT in relative proportion to their share of the student population, even as Hispanic students take the test at a lower rate and Asian students take it at a higher rate. Our analysis of full-time homeschooled students’ level of parental education, as compared with data released by the College Board, suggests that this population should take the SAT at a rate nearly 10% higher than the nationwide average.
 Bagwell reported that 169 students enrolled at York Technical College during the 2007-2008 school year listed homeschooling as the school from which they matriculated. York Technical College reports having 4,731 students enrolled in fall 2007. Of Bagwell’s full sample (he looked at all homeschooled students enrolled in the college from spring 2001 through fall 2007), one-third were under 18. Bagwell notes that in many cases these students dual-enrolled, taking community college courses while completing other credits at home. Whether these students would be considered transfer students when enrolling in a four-year college or university would depend on how many credits they completed at York.
 Homeschooled students accounted for 20 out of 14,628 incoming students.
 Between 16 and 35 new homeschooled students were enrolled each year from 2014 through 2011, compared to a total enrollment of between 6,041 and 10,512 students. Assuming that newly enrolled homeschooled students accounted for about one third of total enrolled homeschooled students (based on Bagwell’s findings), this amounted to a total homeschool enrollment of between 0.7% and 1.2% each year.