Homeschooling as “Educational Neglect” or Neglected Research Standards? 2
Brian D. Ray, Ph.D.
National Home Education Research Institute
PO Box 13939, Salem, Oregon 97309
Kelly, Barr, & Weatherby (2006), the authors of the report entitled Educational Neglect & Compulsory Schooling: A Status Report, claimed that school-age children who are privately home educated and not registered with state agencies – even if the law does not require such registration – are “missing” and “educationally neglected.” They garnered national attention with their assertions (Allen, 2006; Associated Press, 2006). The authors of the report, however, provided essentially no empirical data to support their most important claims, conclusions, and policy recommendations. In fact, it is possible that the data they reported could be used to argue just the opposite of their proposals. This is a brief review of their report.
Philosophical, Theoretical, and Legal Issues
Kelly, Barr, and Weatherby were correct that certain academics (e.g., themselves) and government officials think the state should be concerned about whether the parents of public-school children (and their children) obey compulsory school attendance laws and whether the state provides an adequate academic education for those enrolled in tax-funded, public schools. They were also correct in implying that for the sake of administering public schools and the law related thereto, it is helpful to have in place clear definitions of truancy and educational neglect.
On another and more substantive note, however, there were several severe weaknesses in this report. First, the authors presented a simplistic and imbalanced history of compulsory school laws across the United States (U.S.) and they misrepresented the Idaho constitution and law, particularly by being incomplete in their presentation of them. For example, the Idaho constitution (like many states’ constitutions), as the authors correctly quoted it, says that the state must provide (i.e., offer) tax-funded schools to the people; it nowhere, however, gives the state authority to control or regulate the education of all the children within the state.
Second, it should also be noted that the report’s authors appealed to the federal No Child Left Behind legislation (NCLB) to argue for control of privately-educated students. The authors are remiss, however, to not include in their report that the NCLB does not apply to homeschool students. i
Third, since the authors presented this as a research-based report, it is appropriate to consider several methodological errors they committed in developing their study. At the outset, they used essentially no logical rationale nor empirical foundation to equate a compulsory-school-age child not being registered with the state’s education agencies with “educational neglect” and being a “missing student” (e.g., p. 2 & p. 40). The authors did this despite the fact they found that “[s]tates do not have a common definition of educational neglect. Twenty-four (24) states have no statutory definition of educational neglect” (p. 1) and Idaho needs to clearly define educational neglect (p. 2). The implication they made was that any student not registered with a state education agency is a “missing” student (and an “educationally neglected” student) even though there may be no legal (or moral) requirement in the state for the child to be registered with state education agencies. In a number of states, neither private school nor homeschool students are required to be registered with state agencies. Idaho, for example, does not require registration or contact with state agencies by homeschool families (Klicka, 2004, p. ID-1, ID-2). Kelly, Barr, and Weatherby claimed “… Idaho’s educational institutions are ‘missing’ almost 14,000 children, or 6.5% of all children within the state’s compulsory education age range” (p. 44). Contrary to the implications of this claim, Idaho’s (tax-funded and state-controlled) educational institutions are not missing any students who are not required by law to be enrolled in them. (Further, the same applies to all states.) That is to say, those children and youth who are home educated are not legally required to be in state-run schools and the authors provided no evidence (e.g., law) that those students enrolled in private institutional school were legally required to be enrolled in or to register with state-run educational agencies. The authors confused presenting age-old philosophical and theoretical views and debates over the state’s responsibility for and authority over the education of children with empirical research methods.
Methods, Data, and Analyses
Kelly, Barr, and Weatherby claimed that a discrepancy between U.S. census data and state education agency records about the number of school-age children could be used to make dependable conclusions about how many children were “missing” in any given state (p. 51). For example, they assumed that if national census data showed more children existed in a state than state education agencies had registered with them, then the difference consisted of “missing” children. Further, the authors clearly implied that those who were “missing” were at high risk of being “educationally neglected” (if not “educationally neglected,” ipso facto).
One reason that this approach is problematic is that the methods used to gather U.S. census data may not result in estimates that are precise and dependable enough for the purposes of the present study. Strangely enough, along these lines, the authors found that in some states the education agencies reported far more students registered with them than the U.S. census data showed existed in those states. By their methods, for example, New Jersey had 9,051 more and Florida had 82,838 more compulsory-age students “enrolled in educational institutions” than existed in the individual states. This paradox alone should make the authors’ conclusions suspect.
In many ways, in both their methods and narrative, the authors ignored the fact that some states’ laws do not require private institutional-school and private homeschool students to be registered with state education agencies. This clearly presents a methodological problem, and it is a philosophical problem in that they ignored the fact that this variety in law reflects the freedom of education that citizens still enjoy in many states of the U.S.
Perhaps most problematically, Kelly, Barr, and Weatherby asserted and concluded that more oversight and control of homeschooling by the state would ensure that fewer children in the a state would be “educationally neglected” despite the fact that they provided no data to support this claim. Even if one were to equate being “unaccounted for” with “educationally neglected” (which is not substantiated by the authors in any way), the authors do not support their conclusions with evidence. Promoting policy changes that are not factually founded is risky business in terms of violating individuals’ rights in the arenas of family life and children’s education.
Furthermore, using the data that the authors provided, the reader can do some simple statistical analyses that lead to serious doubts about the authors’ conclusions. For a first example, the authors wrote of “levels of oversight” by the state. They were not clear about this in their report but if they considered these levels to be ordinal in nature, that is, able to be ordered from less to more oversight, then one preliminary analysis by this author found no correlation between level of oversight and the percent of “missing” students. ii
As a second example, the authors were also not clear about whether they considered the “levels of oversight” to be categorical in nature but if the levels were to be considered as such, then a preliminary analysis by this author found no significant difference in percent of “unaccounted for” students according to level of oversight. iii
Kelly, Barry, and Weatherby did not provide any empirical evidence to suggest that if the state were to control homeschool families more then the education of homeschool children would improve. The preceding two examples of statistical analysis by this author make it even clearer that the authors of the report did not present a positive correlation between degree of state control and less “unaccounted for” (or “missing”) school-age children, let alone present data that would support a causal relationship between more state control of homeschooling and fewer children being educationally neglected. The report’s authors also failed to mention prior research findings that showed no correlation between the degree of state control of homeschooling and homeschool students’ academic achievement (Ray, 2000). The authors were apparently trying to solve a problem of which they had not shown the existence.
The report’s authors concluded, among other things related to education in general, that the law should be changed in Idaho such that all homeschool children shall be compelled to be registered with the state and be annually subjected to standardized achievement testing. They provided, however, no evidence that this would help homeschool children in particular or all school-age children in general. Along these lines, there was no inclusion in their report of the fact that the current body of research on homeschooling indicates that home-educated students generally perform above average in terms of academic achievement, do well with respect to their social, emotional, and psychological development, and are experiencing above-average success into adulthood (Medlin, 2000; Ray, 2004, 2005; Rudner, 1999).
Since the authors of the report proposed that all homeschool parents should be compelled to register their children with the state and have them tested, they begged the following salient question that they ignored: Should not the state compel all private institutional schools and the children in them (e.g., Catholic schools, Islamic schools, Jewish schools, Lutheran schools, Mormon schools, New Age, and “nonsectarian” schools) to be registered with the state, controlled by the state, and tested by the state even if they are not funded by a state’s citizens’ tax dollars? Why did the authors not address this question?
On another note related to the authors’ findings, if they believed their methods to be valid, why did they not ask more serious questions about the eight states in which more school-age students were registered with state education agencies than the federal census showed existed? For example, Is it possible that public schools are claiming more students on their rosters than are truly on them? If so, might this suggest fraud upon the states’ tax payers?
Finally, there was no indication in the report that its contents (e.g., operational definitions, analyses of data, conclusions, recommendations) were reviewed by a jury of peers before its publication. This does not necessarily diminish the quality of the report but the fact that it was not reviewed by scholarly peers before its publication should be noted by its readers and those who might attempt to base social policy or law on its contents.
In summary, while the authors of Educational Neglect & Compulsory Schooling: A Status Report (Kelly, Barry, & Weatherby, 2006) made a few observations worth considering and included some interesting data, the report was full of methodological research errors, incomplete philosophical and theoretical discussion, and conclusions that were not based on empirical evidence. Some of the most notable concerns of this author are the following:
1. The report’s authors presented an overly simplistic and imbalanced history of compulsory school laws across the U.S., misrepresented Idaho constitution and law, and misapplied the federal No Child Left Behind legislation.
2. With no logical rationale, no legal basis, nor any empirical research foundation, the report equated a compulsory-school-age child not being registered with the state’s education agencies with “educational neglect” and being a “missing student.”
3. The authors claimed that a discrepancy between U.S. census data and state education agency records about the number of school-age children in a state can be used to make dependable conclusions about how many children are “unaccounted for,” “missing,” or “educationally neglected” in any given state. This approach lead them, for example, to conclude that Florida had 82,838 more compulsory-age students “enrolled in educational institutions” than existed in Florida at the time. This foundational paradox alone should make the authors’ conclusions suspect.
4. Perhaps most problematically, both philosophically and empirically, the authors concluded that more oversight and control of homeschooling by a given state will ensure that fewer children in the state will be “educationally neglected” despite the fact that they provided no data to support this claim.
5. Two different preliminary statistical analyses by this author (using the report’s data) showed no relationship between “level of oversight” by the state of homeschooling and the number of students “unaccounted for” or “missing” in a state. Simply put, there was no correlation. One could argue, therefore, that reducing state control might lead to fewer “unaccounted for” children just as well as one might argue the reverse.
6. Kelly, Barr, and Weatherby (2006) provided no empirical evidence that their recommendations would help homeschool children in particular or all school-age children in general.
Allen, Anne Wallace [Associated Press Writer]. (2006, January 27). Report: Conn. has more school students than Census counted. Retrieved 1/28/06 online http://www.newsday.com/news/local/wire/connecticut/ny-bc-ct–educationreport0127jan27,0,5508559,print.story?coll=ny-region-apconnecticut.
Associated Press. (2006, January 27). Report cites ‘educational neglect’ in Idaho: Thousands of school-age children not accounted for. Retrieved 1/27/06 online http://www.columbiatribune.com/2006/Jan/20060127News014.asp.
Kelly, Philip, Barr, Robert, & Weatherby, James. (2006 or 2005 [no publication date is on the report]). Educational neglect & compulsory schooling: A status report. Center for School Improvement and Policy Studies Final Report, 2004 – 2005. Boise, ID: College of Education, Boise State University. Retrieved 1/26/06 online http://www.idaho-post.org/MDteams/documents/EducationalNeglectFinalReport.pdf.
Klicka, Christopher J. (2004, August). Home schooling in the United States: A legal analysis. Purcellville, VA: Home School Legal Defense Association.
Medlin, Richard G. (2000). Home schooling and the question of socialization. Peabody Journal of Education, 75(1 & 2), 107-123.
Ray, Brian D. (2000). Home schooling: The ameliorator of negative influences on learning? Peabody Journal of Education, 75(1 & 2), 71-106.
Ray, Brian D. (2004, Fall). Homeschoolers on to college: What research shows us. Journal of College Admission, no. 185, 5-11.
Ray, Brian D. (2005). Worldwide guide to homeschooling. Nashville, TN: Broadman and Holman, Publishers.
Rudner, Lawrence M. (1999). Scholastic achievement and demographic characteristics of home school students in 1998. Educational Policy Analysis Archives, 7(8). Retrieved 8/14/03 online http://epaa.asu.edu/epaa/v7n8/.
1 The “Perspectives – News and Comments” section consists of pieces that have not undergone peer review.
2 Neither this review nor the statistical analyses conducted by its author have been reviewed by research peers. This article is based on one entitled “A Brief Review of ‘Educational Neglect and Compulsory Schooling: A Status Report’” first published January 28, 2006 (revised January 30, 2006) online http://www.nheri.org/modules.php?name=Content&pa=showpage&pid=32.
i. Title 20, Section 7886(b) of United States Code.
ii. For the sake of this preliminary statistical analysis, it was assumed that the ordinal categories of less to more oversight should be arranged in the following order: no oversight, students register with local school/school district, multiple oversights, progress monitoring/assessment data, and curricular oversight (Appendix F, p. 58). Second, only five categories of oversight were used; the category of “other oversight” was not included because it did not appear to fit the ordinal model. Third, states with a negative percent of “unaccounted for” students (p. 13-14) were not included. Finally, the nonparametric Spearman correlation coefficient of .09 was not significant (p = .57).
iii. For the sake of this statistical preliminary analysis, all 6 categories of oversight were used. Second, states with a negative percent of “unaccounted for” students were not included. Finally, a one-way ANOVA showed no significant difference between the 6 groups (i.e., categories), F(5, 35) = .419, p = .832.