Contextualizing Homeschooling Data: A Response to Rudner

Rudner (1999) presents the results of a survey and testing program, administered by Bob Jones University (BJU), for homeschooling students. In this response, we applaud Rudner's contribution to building a greater understanding of the homeschooling movement. However, we also voice a strong concern that what Rudner contributed with one hand, he took back with the other. We contend that Rudner's analysis of the BJU data fails to offer a straightforward explanation of important and striking limitations. The unfortunate result is an inaccurate portrayal of homeschoolers as a white, Christian, monolithic population. Although the results of Rudner's analyses are likely valid for the particular population he studied, his insufficient attention to the data's bias has led to an erroneous picture of homeschooling. At a time when most educational researchers pay little or no attention to the homeschooling movement, we appreciate Dr. Rudner's (1999) work as well as the

decision of EPAA to give that work a platform.Although Rudner's article does not vary far (either in methodology or in findings) from much of the body of the homeschooling research already in print (e.g., Gustavsen, 1981;Mayberry, 1987;Ray, 1990Ray, , 1997;;Wartes, 1988), keeping this research current and expanding the sample size contributes greatly to the limited assortment of currently-published studies.We believe that the questions asked of the data in this study are important; these issues need to be further explored.However, the data that Rudner analyzed are derived from only one section of the homeschooling population.And here lies the article's weakness: it fails to explain this limitation in a way that adequately alerts readers.
While we do not disagree with Rudner's tentative conclusions concerning homeschoolers' performance on standardized tests, we do think there is a need to offer several cautions to the readers of this study.The data employed in this study were taken from parents who used the Bob Jones University (BJU) standardized testing program.Rudner's article only briefly, and inadequately, addresses the fact that this may not be a cross-section of the homeschooling population: "...it should be noted that it was not possible within the parameters of this study to evaluate whether this sample is truly representative of the entire population of home school students" (this quotation is from the article's "Discussion" section).Rudner does not explain the relevance of this potential limitation as regards the demographic and achievement information that constitute the heart of these analyses.Nor does he offer the obvious reason why the BJU data may not be representative of the larger population: The University's image, at least partially deserved, is of racial intolerance and religious orthodoxy.Accordingly, some of Rudner's conclusions (e.g., that homeschoolers are overwhelmingly white and Christian) should instead be read as limitations on some of his other conclusions (concerning, e.g., median income, marital status, and achievement levels on standardized tests).
A related caution, which we would have liked Dr. Rudner to have offered, is that the data base drew a non-random, two-percent sample (even by the most conservative estimates) of the homeschooling population.(Note 1) Given that Rudner's sample involves 20,760 students, this sample then constitutes anywhere from 1.28% to 2.08% of the homeschooling population.Because the sample was biased in favor of a population associated with BJU, extrapolations from that data are very unreliable.
Yet Dr. Rudner, in his abstract, states that his article seeks to answer the following questions: Does home schooling tend to work for those who chose to make such a commitment?That is, are the achievement levels of home school students comparable to those of public school students?Who is engaged in home schooling?That is, how does the home school population differ from the general United States population?Notwithstanding these broadly worded questions, we note that Rudner acknowledges the fact that his data are not derived from "a controlled experiment" and must be understood within that context.Yet his acknowledgment fails to detail the broader context-namely those issues associated with BJU.He also fails, when setting forth these questions that he seeks to answer with this data, to heed his own warning about its limitations.Contrary to his stated aim, this data simply cannot be used to reliably compare homeschoolers' achievement levels with those of the general population or to describe the demographics of homeschoolers.
In addition to our concerns about the generalizability of Rudner's conclusions, we are concerned that Rudner's relative neglect of issues surrounding the selected sample will serve to perpetuate the common view of homeschoolers as a narrow and easily-defined section of the American population.Because of his omission, only those readers who already know about the full scope of homeschoolers or about BJU's image could raise this red flag.Importantly, the media's coverage of the release of Rudner's study portrayed his conclusions as descriptive of the broader homeschooling population (see Archer, 1999;Cook, 1999;Mathews, 1999;Schnaiberg, 1999;Toomer-Cook, 1999).We found only one article that even mentioned, albeit briefly, the ill fit between the homeschooling population described by Rudner and the homeschooling population in the area (south Florida) served by the newpaper (Nazareno, 1999).

Homeschoolers' Diversity
Notwithstanding the picture painted by the data presented in Rudner's article, today's homeschooling families represent a diverse sampling of the American population.Although many homeschoolers remain white and middle-class, the recent upsurge in homeschooling has drawn people from all ethnic and class groups (Knowles, 1988;Nazareno, 1999;Wahisi, 1995).Ideologically, parents who homeschool represent a similarly broad cross-section of American society (Knowles, 1988).While in its recent resurgence homeschooling began as a trend among fundamentalist Christians with primarily religious motivations, homeschoolers now represent a wide array of values and political mores (Bolick, 1987;Mayberry, 1987;Van Galen, 1988).As such, homeschoolers are no longer an easily defined segment of the population.(Note 2) Rudner's article, therefore, would have been more comprehensive and accurate had he acknowledged existing research demonstrating that his sample was not representative of the broader population.

Not All Homeschoolers Give Tests
In contrast to what is presented in Rudner's data, there exists a large and growing portion of the homeschooling population that does not administer standardized tests to its children.(Note 3) While some homeschoolers employ the "school at home" methodology that Rudner's questions allude to, replete with curricula and testing, other homeschoolers avoid these practices.These parents, often referring to themselves as "unschoolers," follow the philosophy of the late John Holt (see Holt 1981;1983;1989).They often choose to homeschool in order to avoid what they view as the restrictiveness of set curricula and testing (Franzosa, 1991).They believe in allowing a child's natural curiosities to set the scope and pace of education, even if it means waiting a long time before the child expresses interest in a particular topic (Wartes, 1988a).Feeling that the manner in which schools teach is not the way that children learn, they often view standardized testing as a part of the misguided system that they have left behind, and they put great effort into avoiding such testing in their children's education (Common & MacMullen, 1986;Gibbs, 1994).Notwithstanding the important role that this segment of the homeschooling population plays within the larger movement, its existence is not noted in Rudner's article.

Bob Jones University
The data Rudner analyzed was derived exclusively from parents who used the testing services of Bob Jones University, a fundamentalist Christian institution located in Greenville, South Carolina.The university prohibits interracial dating and marriage between its African-American and white students.Prior to 1971, African Americans were banned outright from attending the university (White, 1982).These racial policies were the subject of highly-publicized litigation before the United States Supreme Court in 1983, concerning the question of whether the University, given its explicit racially discriminatory policies, could maintain its "501(c)(3)" tax-exempt status (Bob Jones University v. United States, 1983;White, 1983).
Importantly, the University's racial views are anathema to many Americans, whether they be Christian or non-Christian, fundamentalist or non-fundamentalist.While many parents using the BJU testing service may not share the University's convictions, many other families have no doubt chosen not to employ the services of BJU precisely because of BJU's racial stance.The racial distribution of Rudner's homeschooled students showed 0.8% African American and 0.2% Hispanic.These statistics become much more meaningful when grounded in an understanding of the data source.
BJU's religious orientation may also have prompted many homeschoolers to shun a relationship with the University's testing service.The BJU web pages trumpet its Biblical grounding, noting that the University is "both orthodox and fervent in its evangelistic spirit" (see the Bob Jones University Website, http://www.bju.edu/aboutbju/history/).The religious distribution of Rudner's homeschooled students showed almost 58% Independent Fundamentalist, Baptist, or Independent Charismatic and only 6% non-Christian.These statistics, too, become much more meaningful when grounded in an understanding of the data source.

The Article's Perspective
Dr. Rudner is an accomplished scholar in the field of assessment.Using this expertise, he has testified on behalf of homeschoolers represented by the funder of this study, the Home School Legal Defense Association (HSLDA).We wish to acknowledge the contributions of the HSLDA in blazing a trail for the legal rights of homeschoolers.However, it is also important to note that the HSLDA is an advocacy organization.Moreover, while the HSLDA does not exclude from membership those who hold non-Christian beliefs, it is an overtly Christian establishment with an agenda dedicated to supporting the rights and duties of families as commanded by Biblical mandate.
We recognize that Dr. Rudner may have little or no experience with the homeschooling population outside of this Christian context and therefore might not be aware that many homeschoolers do not fit within the segment of the population who would consider employing the services of BJU.Like many Americans, Dr. Rudner may have simply taken for granted that the homeschooling population remains very narrow.When an analysis is inaccurately premised on the assumption that homeschooling is a phenomenon that is almost exclusively limited to conservative Christian parents, there is less reason to question the representativeness of a sample drawn from BJU.
In understanding this article's perspective, we also note that the HSLDA funded this study, at least in part, as a vehicle for gaining political support on behalf of homeschoolers.Toward this end, the organization distributed copies of the study to members of Congress.Michael Farris, the president of HSLDA, explained that he "hope[s] that [the study] will help judges and public policy makers make better decisions about our freedom" (Billups, 1999).We share this goal as well the belief that legislation concerning homeschooling is best based on a complete understanding of the homeschooling population.Accordingly, the following section presents a reconsideration of some of Dr. Rudner's findings in light of the source of his database.

The Article's Conclusions
As a general matter, we suggest that the data from Bob Jones University add to our understanding of the particular homeschooling population served by BJU and raise interesting questions about the broader homeschooling population.But we would stop far short of drawing the more universal conclusions trumpeted by the HSLDA.
Consider, for example, Dr. Rudner's analysis of demographic information, concerning the high levels of formal education obtained by homeschooling parents and the high median income of homeschooling families.We caution against drawing firm conclusions from this data, particularly the analysis concerning median family income.Not all states require children to take standardized tests (and those that do generally provide a way for parents to take the tests without cost to the family).Consequently, those families who elect to pay a testing service may be in a higher income bracket than those who do not.Moreover, a 1990 survey of Maine homeschoolers revealed that 70% of respondents had an annual pretax income of less than $35,000 (Lyman, 1993).While this Maine study also had many limitations, it nonetheless raises the question of the generalizability of Rudner's findings.The analyses concerning homeschooling parents' level of formal education, computer use by students, and the amount homeschooling parents spend on school supplies could be tied to this income data and influenced by these same factors.Also, with regard to expenditure on school supplies, additional issues are raised (e.g., whether the parents have supplies from an earlier child and whether they can borrow supplies from a friend or family member).
Rudner also found that 98% of homeschooling students in the BJU data base live within married couple families.This statistic should, we contend, have been presented within a context explaining the conservative nature of BJU and its view of divorce as unbiblical.Further, given the view held by most conservative Christians that a woman's primary commitment is to her husband and her children, Rudner conclusion that 77% of homeschool mothers in this data base do not participate in the labor force is, we believe, also better understood within the BJU context.(Note 4) Rudner notes that homeschooled students watch much less television than do most students nationwide (with 65% of homeschooled students in the BJU data base watching one hour or less per day compared to 25% nationally).But again, conservative Christians tend to have strong moral objections to the quality of television programming-much more so than the general population.Accordingly, it may be a certain set of moral standards rather than homeschooling that drives this result.
The article further states that the "primary focus of many home schools is on religious and moral values."But, while many homeschoolers do see this as their primary focus and purpose, many do not-recall the "unschoolers" described above.Van Galen (1988) describes a group of parents, whom she labels as "Pedagogues," whose motivation is decidedly secular.These parents: ... teach their children at home primarily for pedagogical reasons.Their criticisms of the schools are not so much that the schools teach heresy, but that the schools teach whatever they teach ineptly....While diverse in other aspects of their lives, they share a respect for their children's intellect and creativity and a belief that children learn best when pedagogy taps into the child's innate desire to learn (p.55).
While these parents may have religious beliefs, the reason they chose to homeschool was not religiously motivated, and the focus of these homeschools is not the teaching of "religious and moral values." Other demographic characteristics that Rudner ascribes to homeschoolers could also be a result of the population sampled, a possible conflation of almost his exclusively Christian population with the trait of homeschooling.For example, it may be simply that Christian families are larger than those in the general population, not necessarily that most homeschooling parents have more children.Likewise, the fact that almost one in four homeschooled students in this study have a certified teacher as a parent may also be tied to the overwhelmingly Christian population, since many young Christian women see teaching as one of the few appropriate areas of employment and as a field where they can develop skills useful in both family and church life.Finally, as mentioned previously, the fact that the demographics reported by Rudner showed an almost exclusively white, Christian population could also be an artifact of the data source.
As Dr. Rudner noted, family income is strongly correlated with children's test scores.However, we do not know whether the general population of homeschoolers has the same high level of income as the families in the BJU data base.Further research is needed to demonstrate whether or not this difference in test scores would hold up if a lower income sample of homeschoolers were tested.That said, we do believe that homeschooled students can attain a significant benefit from the one-on-one learning experience, and this could be a powerful factor in driving higher test scores.(Note 5) Rudner concluded by stating that "these comparisons between home school students and students nationwide must be interpreted with a great deal of caution," and that "the reported achievement differences between groups do not control for background differences in the home school and general United States population and, more importantly, cannot be attributed to the type of school a child attends."Some researchers, in fact, would say that the test scores have nothing to do with how the children were schooled and simply show the results expected for children that come from this demographic group-households that are overwhelmingly white, well educated, two-parent, and middle class (see Coleman et al., 1966;Ogbu, 1987).This is not to say that these parents did not do a good job teaching their children, it is only to say that a comparable sample within the public or private schools may have scored just as well.

Our Conclusions
The actual analyses conducted by Dr. Rudner are important.Our critique is offered as a cautionary supplement, rather than as an objection, to his contribution.We feel that a more thorough explanation of the data's source and context helps us to build a better understanding of America's homeschooling population.

Notes
Patricia Lines conservatively estimates the number of homeschooled children at approximately 1 million (Lines, 1998).Less conservative appraisals among homeschooling associations and researchers place the number of homeschooled students at more than 1.2 million students (Hawkins, 1996;Kennedy, 1995;Ray, 1997).Newsweek recently estimated that number at 1.5 million (Kantrowitz & Wingert, 1998), a figure that the Home School Legal Defense Association -the sponsor of Rudner's study-also circulates in its literature.Others estimate that 1.
the number is as high as 1.6 million (see Yarnall, 1998).There are also many resources directed at homeschoolers that are not characterized by the demographics provided by Rudner (see the following web pages: "Bnos Henya Project: Jewish Orthodox Homeschooling"; "Al-Madrasah Al-Ula: The Magazine for Muslim Home Schoolers"; "Native American Homeschool Association Web Site"; and "Pagan Homeschool Page").

2.
Of course, if they are required to do so by law, then they comply with the state's requirements (these requirements for homeschoolers vary from state to state).However, these parent engage in the testing merely to satisfy their legal obligation, not because they believe testing to be an educationally worthwhile practice.

3.
It is true that generally, in order to homeschool, one of the parents must possess the ability to remain at home throughout the day, thus allowing that parent to teach and supervise the children.However, homeschooling parents may have jobs permitting them to also supervise their children -either through a flexible schedule, a home-based business, or a job allowing for on-site supervision of their children-with the result that both parents become part of the labor force.

4.
Further, many important forms of knowledge, which homeschooling parents may emphasize in their children's education, may not be assessable by standardized tests.

5.
decision of EPAA to give that work a platform.Although Rudner's article does not vary far (either in methodology or in findings) from much of the body of the homeschooling research already in print (e.g., Gustavsen, 1981;Mayberry, 1987;Ray, 1990Ray, , 1997;;Wartes, 1988), keeping this research current and expanding the sample size contributes greatly to the limited assortment of currently-published studies.We believe that the questions asked of the data in this study are important; these issues need to be further explored.However, the data that Rudner analyzed are derived from only one section of the homeschooling population.And here lies the article's weakness: it fails to explain this limitation in a way that adequately alerts readers.
While we do not disagree with Rudner's tentative conclusions concerning homeschoolers' performance on standardized tests, we do think there is a need to offer several cautions to the readers of this study.The data employed in this study were taken from parents who used the Bob Jones University (BJU) standardized testing program.Rudner's article only briefly, and inadequately, addresses the fact that this may not be a cross-section of the homeschooling population: "...it should be noted that it was not possible within the parameters of this study to evaluate whether this sample is truly representative of the entire population of home school students" (this quotation is from the article's "Discussion" section).Rudner does not explain the relevance of this potential limitation as regards the demographic and achievement information that constitute the heart of these analyses.Nor does he offer the obvious reason why the BJU data may not be representative of the larger population: The University's image, at least partially deserved, is of racial intolerance and religious orthodoxy.Accordingly, some of Rudner's conclusions (e.g., that homeschoolers are overwhelmingly white and Christian) should instead be read as limitations on some of his other conclusions (concerning, e.g., median income, marital status, and achievement levels on standardized tests).
A related caution, which we would have liked Dr. Rudner to have offered, is that the data base drew a non-random, two-percent sample (even by the most conservative estimates) of the homeschooling population.(Note 1) Given that Rudner's sample involves 20,760 students, this sample then constitutes anywhere from 1.28% to 2.08% of the homeschooling population.Because the sample was biased in favor of a population associated with BJU, extrapolations from that data are very unreliable.
Yet Dr. Rudner, in his abstract, states that his article seeks to answer the following questions: Does home schooling tend to work for those who chose to make such a commitment?That is, are the achievement levels of home school students comparable to those of public school students?Who is engaged in home schooling?That is, how does the home school population differ from the general United States population?Notwithstanding these broadly worded questions, we note that Rudner acknowledges the fact that his data are not derived from "a controlled experiment" and must be understood within that context.Yet his acknowledgment fails to detail the broader context-namely those issues associated with BJU.He also fails, when setting forth these questions that he seeks to answer with this data, to heed his own warning about its limitations.Contrary to his stated aim, this data simply cannot be used to reliably compare homeschoolers' achievement levels with those of the general population or to describe the demographics of homeschoolers.
In addition to our concerns about the generalizability of Rudner's conclusions, we are concerned that Rudner's relative neglect of issues surrounding the selected sample will serve to perpetuate the common view of homeschoolers as a narrow and easily-defined section of the American population.Because of his omission, only those readers who already know about the full scope of homeschoolers or about BJU's image could raise this red flag.Importantly, the media's coverage of the release of Rudner's study portrayed his conclusions as descriptive of the broader homeschooling population (see Archer, 1999;Cook, 1999;Mathews, 1999;Schnaiberg, 1999;Toomer-Cook, 1999).We found only one article that even mentioned, albeit briefly, the ill fit between the homeschooling population described by Rudner and the homeschooling population in the area (south Florida) served by the newpaper (Nazareno, 1999).

Homeschoolers' Diversity
Notwithstanding the picture painted by the data presented in Rudner's article, today's homeschooling families represent a diverse sampling of the American population.Although many homeschoolers remain white and middle-class, the recent upsurge in homeschooling has drawn people from all ethnic and class groups (Knowles, 1988;Nazareno, 1999;Wahisi, 1995).Ideologically, parents who homeschool represent a similarly broad cross-section of American society (Knowles, 1988).While in its recent resurgence homeschooling began as a trend among fundamentalist Christians with primarily religious motivations, homeschoolers now represent a wide array of values and political mores (Bolick, 1987;Mayberry, 1987;Van Galen, 1988).As such, homeschoolers are no longer an easily defined segment of the population.(Note 2) Rudner's article, therefore, would have been more comprehensive and accurate had he acknowledged existing research demonstrating that his sample was not representative of the broader population.

Not All Homeschoolers Give Tests
In contrast to what is presented in Rudner's data, there exists a large and growing portion of the homeschooling population that does not administer standardized tests to its children.(Note 3) While some homeschoolers employ the "school at home" methodology that Rudner's questions allude to, replete with curricula and testing, other homeschoolers avoid these practices.These parents, often referring to themselves as "unschoolers," follow the philosophy of the late John Holt (see Holt 1981;1983;1989).They often choose to homeschool in order to avoid what they view as the restrictiveness of set curricula and testing (Franzosa, 1991).They believe in allowing a child's natural curiosities to set the scope and pace of education, even if it means waiting a long time before the child expresses interest in a particular topic (Wartes, 1988a).Feeling that the manner in which schools teach is not the way that children learn, they often view standardized testing as a part of the misguided system that they have left behind, and they put great effort into avoiding such testing in their children's education (Common & MacMullen, 1986;Gibbs, 1994).Notwithstanding the important role that this segment of the homeschooling population plays within the larger movement, its existence is not noted in Rudner's article.

Bob Jones University
The data Rudner analyzed was derived exclusively from parents who used the testing services of Bob Jones University, a fundamentalist Christian institution located in Greenville, South Carolina.The university prohibits interracial dating and marriage between its African-American and white students.Prior to 1971, African Americans were banned outright from attending the university (White, 1982).These racial policies were the subject of highly-publicized litigation before the United States Supreme Court in 1983, concerning the question of whether the University, given its explicit racially discriminatory policies, could maintain its "501(c)(3)" tax-exempt status (Bob Jones University v. United States, 1983;White, 1983).
Importantly, the University's racial views are anathema to many Americans, whether they be Christian or non-Christian, fundamentalist or non-fundamentalist.While many parents using the BJU testing service may not share the University's convictions, many other families have no doubt chosen not to employ the services of BJU precisely because of BJU's racial stance.The racial distribution of Rudner's homeschooled students showed 0.8% African American and 0.2% Hispanic.These statistics become much more meaningful when grounded in an understanding of the data source.
BJU's religious orientation may also have prompted many homeschoolers to shun a relationship with the University's testing service.The BJU web pages trumpet its Biblical grounding, noting that the University is "both orthodox and fervent in its evangelistic spirit" (see the Bob Jones University Website, http://www.bju.edu/aboutbju/history/).The religious distribution of Rudner's homeschooled students showed almost 58% Independent Fundamentalist, Baptist, or Independent Charismatic and only 6% non-Christian.These statistics, too, become much more meaningful when grounded in an understanding of the data source.

The Article's Perspective
Dr. Rudner is an accomplished scholar in the field of assessment.Using this expertise, he has testified on behalf of homeschoolers represented by the funder of this study, the Home School Legal Defense Association (HSLDA).We wish to acknowledge the contributions of the HSLDA in blazing a trail for the legal rights of homeschoolers.However, it is also important to note that the HSLDA is an advocacy organization.Moreover, while the HSLDA does not exclude from membership those who hold non-Christian beliefs, it is an overtly Christian establishment with an agenda dedicated to supporting the rights and duties of families as commanded by Biblical mandate.
We recognize that Dr. Rudner may have little or no experience with the homeschooling population outside of this Christian context and therefore might not be aware that many homeschoolers do not fit within the segment of the population who would consider employing the services of BJU.Like many Americans, Dr. Rudner may have simply taken for granted that the homeschooling population remains very narrow.When an analysis is inaccurately premised on the assumption that homeschooling is a phenomenon that is almost exclusively limited to conservative Christian parents, there is less reason to question the representativeness of a sample drawn from BJU.
In understanding this article's perspective, we also note that the HSLDA funded this study, at least in part, as a vehicle for gaining political support on behalf of homeschoolers.Toward this end, the organization distributed copies of the study to members of Congress.Michael Farris, the president of HSLDA, explained that he "hope[s] that [the study] will help judges and public policy makers make better decisions about our freedom" (Billups, 1999).We share this goal as well the belief that legislation concerning homeschooling is best based on a complete understanding of the homeschooling population.Accordingly, the following section presents a reconsideration of some of Dr. Rudner's findings in light of the source of his database.

The Article's Conclusions
As a general matter, we suggest that the data from Bob Jones University add to our understanding of the particular homeschooling population served by BJU and raise interesting questions about the broader homeschooling population.But we would stop far short of drawing the more universal conclusions trumpeted by the HSLDA.
Consider, for example, Dr. Rudner's analysis of demographic information, concerning the high levels of formal education obtained by homeschooling parents and the high median income of homeschooling families.We caution against drawing firm conclusions from this data, particularly the analysis concerning median family income.Not all states require children to take standardized tests (and those that do generally provide a way for parents to take the tests without cost to the family).Consequently, those families who elect to pay a testing service may be in a higher income bracket than those who do not.Moreover, a 1990 survey of Maine homeschoolers revealed that 70% of respondents had an annual pretax income of less than $35,000 (Lyman, 1993).While this Maine study also had many limitations, it nonetheless raises the question of the generalizability of Rudner's findings.The analyses concerning homeschooling parents' level of formal education, computer use by students, and the amount homeschooling parents spend on school supplies could be tied to this income data and influenced by these same factors.Also, with regard to expenditure on school supplies, additional issues are raised (e.g., whether the parents have supplies from an earlier child and whether they can borrow supplies from a friend or family member).
Rudner also found that 98% of homeschooling students in the BJU data base live within married couple families.This statistic should, we contend, have been presented within a context explaining the conservative nature of BJU and its view of divorce as unbiblical.Further, given the view held by most conservative Christians that a woman's primary commitment is to her husband and her children, Rudner conclusion that 77% of homeschool mothers in this data base do not participate in the labor force is, we believe, also better understood within the BJU context.(Note 4) Rudner notes that homeschooled students watch much less television than do most students nationwide (with 65% of homeschooled students in the BJU data base watching one hour or less per day compared to 25% nationally).But again, conservative Christians tend to have strong moral objections to the quality of television programming-much more so than the general population.Accordingly, it may be a certain set of moral standards rather than homeschooling that drives this result.
The article further states that the "primary focus of many home schools is on religious and moral values."But, while many homeschoolers do see this as their primary focus and purpose, many do not-recall the "unschoolers" described above.Van Galen (1988) describes a group of parents, whom she labels as "Pedagogues," whose motivation is decidedly secular.These parents: ... teach their children at home primarily for pedagogical reasons.Their criticisms of the schools are not so much that the schools teach heresy, but that the schools teach whatever they teach ineptly....While diverse in other aspects of their lives, they share a respect for their children's intellect and creativity and a belief that children learn best when pedagogy taps into the child's innate desire to learn (p.55).
While these parents may have religious beliefs, the reason they chose to homeschool was not religiously motivated, and the focus of these homeschools is not the teaching of "religious and moral values." Other demographic characteristics that Rudner ascribes to homeschoolers could also be a result of the population sampled, a possible conflation of almost his exclusively Christian population with the trait of homeschooling.For example, it may be simply that Christian families are larger than those in the general population, not necessarily that most homeschooling parents have more children.Likewise, the fact that almost one in four homeschooled students in this study have a certified teacher as a parent may also be tied to the overwhelmingly Christian population, since many young Christian women see teaching as one of the few appropriate areas of employment and as a field where they can develop skills useful in both family and church life.Finally, as mentioned previously, the fact that the demographics reported by Rudner showed an almost exclusively white, Christian population could also be an artifact of the data source.
As Dr. Rudner noted, family income is strongly correlated with children's test scores.However, we do not know whether the general population of homeschoolers has the same high level of income as the families in the BJU data base.Further research is needed to demonstrate whether or not this difference in test scores would hold up if a lower income sample of homeschoolers were tested.That said, we do believe that homeschooled students can attain a significant benefit from the one-on-one learning experience, and this could be a powerful factor in driving higher test scores.(Note 5) Rudner concluded by stating that "these comparisons between home school students and students nationwide must be interpreted with a great deal of caution," and that "the reported achievement differences between groups do not control for background differences in the home school and general United States population and, more importantly, cannot be attributed to the type of school a child attends."Some researchers, in fact, would say that the test scores have nothing to do with how the children were schooled and simply show the results expected for children that come from this demographic group-households that are overwhelmingly white, well educated, two-parent, and middle class (see Coleman et al., 1966;Ogbu, 1987).This is not to say that these parents did not do a good job teaching their children, it is only to say that a comparable sample within the public or private schools may have scored just as well.

Our Conclusions
The actual analyses conducted by Dr. Rudner are important.Our critique is offered as a cautionary supplement, rather than as an objection, to his contribution.We feel that a more thorough explanation of the data's source and context helps us to build a better understanding of America's homeschooling population.

Notes
Patricia Lines conservatively estimates the number of homeschooled children at approximately 1 million (Lines, 1998).Less conservative appraisals among homeschooling associations and researchers place the number of homeschooled students at more than 1.2 million students (Hawkins, 1996;Kennedy, 1995;Ray, 1997).Newsweek recently estimated that number at 1.5 million (Kantrowitz & Wingert, 1998), a figure that the Home School Legal Defense Association -the sponsor of Rudner's study-also circulates in its literature.Others estimate that 1.
the number is as high as 1.6 million (see Yarnall, 1998).There are also many resources directed at homeschoolers that are not characterized by the demographics provided by Rudner (see the following web pages: "Bnos Henya Project: Jewish Orthodox Homeschooling"; "Al-Madrasah Al-Ula: The Magazine for Muslim Home Schoolers"; "Native American Homeschool Association Web Site"; and "Pagan Homeschool Page").

2.
Of course, if they are required to do so by law, then they comply with the state's requirements (these requirements for homeschoolers vary from state to state).However, these parent engage in the testing merely to satisfy their legal obligation, not because they believe testing to be an educationally worthwhile practice.

3.
It is true that generally, in order to homeschool, one of the parents must possess the ability to remain at home throughout the day, thus allowing that parent to teach and supervise the children.However, homeschooling parents may have jobs permitting them to also supervise their children -either through a flexible schedule, a home-based business, or a job allowing for on-site supervision of their children-with the result that both parents become part of the labor force.

4.
Further, many important forms of knowledge, which homeschooling parents may emphasize in their children's education, may not be assessable by standardized tests. 5.