|
Empirical Research Introduction
The body of research on home-based education has expanded dramatically since the first studies and academic articles of the late 1970s that dealt with the modern homeschool movement. Numerous researchers have examined the academic achievement of home-educated children and youth, their social, emotional, and psychological development, and their success into adulthood, and various aspects of homeschool families in general. Researchers have also explored myriad other aspects and issues related to home education in disciplines such as philosophy, sociology, and law. Only a handful of studies, however, have looked closely at a large nationwide sample of home educators and their children in the United States, and the last one of this nature was conducted about a decade ago.
Review of Literature and Conceptual Framework
Homeschooling grew from nearly nonexistent in the 1970s to roughly two million students in grades K to 12 by 2009 (Bielick, 2008; Ray, 2009a). Much of public opinion is positive toward this private educational practice. On the other hand, genuinely curious people and ideological skeptics continue to ask questions about home-based education. Research answers some of these key questions.
Concerns of Various Groups
Most children of about ages 6 through 17 have been placed in institutional schools with formally trained teachers and administrators for the past several generations. Homeschool parents, on the other hand, provide the majority of their children’s academic and social and emotional instruction and training in and based out of their homes without sending their children away to a place called school. Therefore, policymakers, educators, school administrators, judges, and parents often wonder whether ordinary mothers and fathers, who are not government-certified teachers, are capable of effectively teaching and rearing their children after age five.
Academic Achievement
Is it possible for adults without specialized, university-level training in teaching to help their children learn what they need to learn? Numerous studies by dozens of researchers have been completed during the past 25 years that examine the academic achievement of the home-educated (see reviews, e.g., Ray, 2000, 2005; 2009b). Examples of these studies range from a multi-year study in Washington State (Wartes, 1991), to other state-specific studies, to three nationwide studies across the United States (Ray, 1990, 1997, 2000; Rudner 1999), to two nationwide studies in Canada (Ray, 1994; Van Pelt, 2003). In most studies, the homeschooled have scored, on average, at the 65th to 80th percentile on standardized academic achievement tests, compared to the national school average of the 50th percentile (which is largely based on public schools). A few studies have found the home educated to be scoring about the same or a little better than public school students.
Researchers have examined relationships between several variables and homeschool students’ achievement (e.g., Ray, 2000; Ray & Eagleson, 2008; Rudner, 1999). Examples are parent educational attainment, family income, race or ethnicity, number of years the child had been home educated, time spent in formal instruction, and degree of regulation of homeschooling by the state. A few of these variables (e.g., parent education level) are consistently associated with homeschool students’ achievement, although the relationships are often relatively weak. Several variables studied to date show no or very little relationship to these students’ achievement; examples of such variables are the degree of regulation (control) of homeschooling by the state and whether the parents have ever been state-certified teachers.
Social, Emotional, and Psychological Development
“What about socialization?” Homeschool parents call it the “S question.” Socialization questions are asked of nearly every homeschool parent, homeschool teenager, and adult who was home educated. One part of the “S question” asks whether homeschool children interact with other people outside their nuclear family members. Research shows that the large majority of home-educated students consistently interact with children of various ages and parents outside their immediate family (see, e.g., Medlin, 2000; Ray, 1997, 2009b).
The second part of the socialization question asks whether home-educated children will experience healthy social, emotional, and psychological development. Numerous studies, employing various psychological constructs and measures, show the home-educated are developing at least as well, and often better than, those who attend institutional schools (Medlin, 2000; Ray, 2009b). No research to date contravenes this general conclusion. In a few studies, on some of the sub-measures within a study, the home educated have scored slightly lower (i.e., “worse,” according to the conceptual paradigm the researcher was using) than those in institutional schools.
The “Real World” of Adulthood
A corollary of the socialization question deals with whether the home-educated child will eventually function well in the world of adulthood, in which one is responsible for getting along with others on one’s own by not violating others’ inalienable rights, obtaining one’s own food, shelter, and clothing, and living a life that is “self-actualized,” noticeably autonomous in terms of critical thinking, or marked by some other worldview’s preferred traits. Various studies have addressed this issue in multiple ways. It appears that the home educated are engaged, at least as much as are others, in activities that predict leadership in adulthood (Montgomery, 1989), doing well on their college/university SAT tests (Barber, 2001, personal communication) and ACT tests (ACT, 2005), matriculating in college at a rate that is comparable or a bit higher than for the general public (Ray, 2004; Van Pelt 2003), performing well in college (Gray, 1998; Galloway & Sutton, 1995; Jenkins, 1998; Jones & Gloeckner, 2004; Mexcur, 1993; Oliveira, Watson, & Sutton, 1994), satisfied that they were home educated (Knowles & Muchmore, 1995; Ray, 2004; Van Pelt, Neven, & Allison, 2009), involved in community service at least as much as others (Ray, 2004; Van Pelt, Neven, & Allison, 2009), and more civically engaged than the general public (Ray, 2004; Van Pelt, Neven, & Allison, 2009). There is no research evidence that having been home educated is associated with negative behaviors or ineptitudes in adulthood.
Summary of Research Findings to Date on the Attributes of the Home Educated
More than two decades of research have shown that homeschooling – otherwise known as home-based education or home education – is associated with relatively high academic achievement, healthy social, psychological, and emotional development, and success into adulthood for those who were home educated (Galloway & Sutton, 1999; Ray, 2005). Conservatively speaking, one might say research “… simply shows that those parents choosing to make a commitment to home schooling are able to provide a very successful academic environment” (Rudner, 1999). Critics of homeschooling who emphasize the limitations of homeschool research claim, however, that research on the academic achievement of homeschool students can be used to reach “… very limited conclusions” (Reich, 2005, p. 115; see also, West, 2009). In another vein, Ray (2005, p. 11) stated it thus: In other words, the design of most research to date does not allow for the conclusion that homeschooling necessarily causes higher academic achievement than does public (or private) institutional schooling. On the other hand, research designs and findings to date do not refute the hypothesis that homeschooling causes more positive effects than does institutional public (or private) schooling. Along these lines, Ray (2000b), after reviewing many studies on homeschooling and conducting several himself, gingerly wrote: “Assuming, for the sake of discussion and based on a multitude of studies, that home schooling is associated with high academic achievement (and possibly causes it), one could ask whether there is any link between the preceding list of positive factors and the nature of the educational ‘treatment’ known as home schooling” (p. 92). In other words, research to date may not establish a cause-and-effect relationship between the practice of homeschooling and many positive attributes associated with homeschooling, but the research does allow for the possibility that home education causes desirable effects. Furthermore, this study should provide current information and analysis, and some of this may additionally plumb that potential causal connection.
Purpose of Study
The purpose of this nationwide cross-sectional, descriptive study (Johnson, 2001) is to examine the educational history, demographic features, and academic achievement of home-educated students and the basic demographics of their families, and to assess the relationships between the students’ academic achievement and selected student and family variables.
Significance of Study
The home-education population continues to grow (Bielick, 2008; Ray, 2009a) and the general public, parents, educators, policymakers, legal scholars, sociologists, and others are interested in those who comprise the population and the relative benefits or disadvantages, to children and society, of parent-led home-based education. The last nationwide study of significant size was done about a decade ago and people are curious about the current state of certain aspects of homeschooling. Further, this study is designed to improve on some of the limitations of past studies (see, e.g., Welner & Welner, 1999) by attempting to include a broader sampling of families and students.
Methods
This is a nationwide cross-sectional, descriptive study (Johnson, 2001). This section provides descriptions of the survey instrument, achievement measures used, the testing services that provided data, and the procedures used to develop the dataset.
Definitions
Students were included in the study if a parent affirmed that his or her student was “… taught at home within the past twelve months by his/her parent for at least 51% of the time in the grade level now being tested.” Definitions of terms that are not self-explanatory (e.g., degree of structure, structured learning time, and formal instruction) are provided in the “findings” section.
Background Survey
The background survey (questionnaire instrument) was designed by the researcher. The questions were determined by reviewing previous surveys designed and successfully used by Ray (1990, 1994, 1997, 2000) and then by Rudner (1999), prioritizing them, and selecting those that were most relevant to the objectives of the study. Current literature on home education and the objectives of this study were also considered in the development of the survey. Where possible, questions and responses were constructed to match those used by the U.S. Census, U.S. Department of Labor, and the National Assessment of Educational Progress (NAEP) to facilitate present or future comparisons of homeschool students with students nationwide.
The questionnaires used by Ray (1990, 1994, 1997, 2000) were designed by a cooperative effort of the researcher and others who had expertise in home education and survey research in the United States. The guidelines for conducting survey research delineated by Borg and Gall (1989) were followed for Ray’s studies and for the present study. The present instrument was designed to answer the research questions for this study (and research questions for other studies related to home education that may be pursued at a later date). This instrument was reviewed and revised by persons who are familiar with home education (e.g., homeschool leaders and researchers) and consensus was reached on the validity of the items and their wording.
The survey was designed to be shorter than those in some previous survey instruments. An effort was made to pose all questions in an objective format, and most items were very similar or identical to those used by Rudner (1999).
The instrument was comprised of five parts that were (a) qualifiers, (b) student demographics, (c) parent and family demographics, (d) scholastic information, and (e) other information. The items were either select-type (or forced-choice) format or involved the respondent simply marking one of two or more categories. The instrument resulted in 85 variables per child/student being available for analysis. The online survey system also recorded the date and time of submission of the surveys. Recoding and the creation of additional variables (e.g., collapsing many item response categories into a few) for analysis was done later.
The instrument was electronically posted online for parents of the students to input responses and thus data were automatically input to a database for analysis by the investigator. Parents also had the option of requesting a paper copy (hard copy) of the questionnaire and for those who did, one was sent them via the U.S. Postal Service and then the parent mailed the survey to the researcher, after which the responses were input to the online survey venue.
A unique ten-digit Student ID (identification) Number was given to each student and this number was used to merge the student’s background information with test-score data. A person could not complete and submit a survey more than once and have it be included in the study since the unique ID number would prevent this.
Measures of Academic Achievement
Academic achievement is considered to be the formal demonstration of learning (including knowledge, understanding, and thinking skills) attained by a student as measured by standardized academic achievement tests. For example, knowledge and ability in the areas of reading, language, and mathematics are included.
The author recognizes that there has been widespread debate for decades about the best way to measure academic achievement and how to use scores on tests (e.g., Cizek, 1988, 1993, 2001; Hardenbergh, 2008). Further, the author recognizes that within the homeschool community certain subpopulations might be less inclined to use standardized achievement tests than others. At the same time, standardized academic achievement tests have been widely used and well-regarded by the education profession and society-at-large for many decades. Standardized academic achievement tests were used to measure learning in this study.
The standardized academic achievement tests most used in this study were the Iowa Tests of Basic Skills (ITBS, Form A) and California Achievement Tests (CAT). The ITBS is published by Riverside Publishing Company. The tests were designed and developed by University of Iowa professors to measure skills and standards important to growth across the curriculum in the nation’s public and private schools. The ITBS reflects many years of test development experience and research on measuring achievement and critical thinking skills in reading, language arts, mathematics, science, social studies, and information sources. The testing services (see below) using this test used the 2005 norms. Some of the homeschool students took California Achievement Tests (CAT), Fifth Edition (CAT/5), published by CTB/McGraw-Hill. The CAT likewise has a long history of development and use in the United States for use by both public and private schools. Both the ITBS and CAT are considered valid and reliable instruments, as are the other tests (e.g., Stanford Achievement Test) used in this study.
Testing Services
Several organizations in the United States provide assessment (testing) services to homeschool families and their students on a fee-for-service basis. Several of these cooperated with the researcher in the present study to gather achievement test and demographic data on the students. Rudner (1999) worked with one such large testing service, BJU Press. For the present study, four major testing services plus several smaller ones were contacted to help with the study. It was theorized that using several testing services would provide a wider variety of homeschool students and therefore data from a more robust and representative sample of the homeschool population for the purposes of analysis (see, e.g., Welner & Welner, 1999).
In the Spring of 2008, homeschool students who were contracted to take tests via these testing services were given an achievement test and their parents were asked to complete a questionnaire, either online or a paper copy (that they would then mail to the researcher; further explained below).
Four major testing services and several smaller ones assisted in the study. BJU Press (Greenville, South Carolina, www.bjupress.com) began offering achievement testing in 1984. Both homeschool families and private schools use their services. Second, Family Learning Organization (Mead, Washington, www.familylearning.org) has provided testing services for many years. Third, Piedmont Education Services (Pfafftown, North Carolina, www.pesdirect.com, 2009) was founded in 1987 and offers many services to homeschool families. The fourth is Seton Testing Services, which is a function of Seton Home School Study that was founded in 1980 and grew out of parent-operated Catholic schools (in Manassas, Virginia, www.setonhome.org; Seton Home Study School, 2008).
Smaller testing services that were used for the study were Basic Skills Assessment and Educational Services (Oregon City, Oregon), Circle Christian School (Orlando, Florida), Covenant Home School Resource Center (Phoenix, Arizona), Idaho Coalition of Home Educators (Eagle, Idaho), and Whatcom Home School Association (Bellingham, Washington). Additionally, the several nationwide and statewide homeschool organizations that also worked closely with the researcher to contact homeschool families were Education Network of Christian Homeschoolers of New Jersey (Atlantic Highlands, New Jersey), Home School Legal Defense Association (Purcellville, Virginia), Homeschoolers of Maine (Camden, Maine), Massachusetts Homeschool Organization of Parent Educators (Holden, Massachusetts), NYS Loving Education at Home (Fayetteville, New York), and Oregon Christian Home Education Association Network (Portland, Oregon).
Data-Generation Procedures
The following steps were followed with testing services to produce data:
Parents who did not have access to the Internet or preferred to complete a print (hard) copy of the survey instrument requested one and it was mailed to them. They then mailed in the completed survey to the researcher for online data entry.
Reminders to participate in the study were e-mailed to the large majority of parents whose children were tested. For some of the testing services, postcards were sent via the U.S. postal service. Most parents received a second reminder. One of the four largest testing services did not send out reminders to parents.
The following steps were followed by the statewide and nationwide homeschool organization to produce data:
Several of the test administrators to whom tests were sent administered the tests to groups of students. It was estimated that about 90% of these large-group testers did not respond to the invitation to participate in the study and did not invite parents to participate in the study; there was no sure way to validate how many students were represented by the group testers who did not participate in the study.
Population and Sample
The target population was all families in the United States who were educating their school-age children at home and having standardized achievement tests administered to their children. An attempt was made to utilize a sample that was more representative than those studied by Ray (1990, 1997, 2000) and Rudner (1999). Ray’s (1990) sample was drawn only from the membership of one large nationwide home education organization. Ray’s (1997, 2000) sample drew from the same large membership organization and from many other organizations and those on other lists were asked to participate. Rudner’s (1999) sample was large and well-controlled, but it was from only one testing service and possessed what some thought to be notable limitations (c.f., Welner & Welner, 1999). The present study attempted to improve on the samples used in the three preceding mentioned nationwide studies. This was done by using numerous testing services to provide contacts with and test-score data from potential homeschool family and student participants. The researcher began with four notably large testing services that work with families nationwide, and then included a few more smaller testing services in the study. The expectation was this approach would provide a more robust sampling by utilizing several testing services from across the nation.
A total of 11,739 students provided useable questionnaires with corresponding achievement tests. The achievement test and questionnaire results were combined to form the dataset used in this analysis.
It was very challenging to calculate the response rate. One of the main problems was that, well into the study, it was discovered that many of the large-group test administrators were not communicating to their constituent homeschool families that they had been invited to participate in the study. Based on the best evidence available, the response rate was a minimum of 19% for the four main testing services with whom the study was originally planned, who worked fairly hard to get a good response from the homeschool families, and whose students accounted for 71.5% (n = 8,397) of the participants in the study. That is, of the students who were tested and whose parents were invited to participate in the study, both test scores and survey responses were received for this group. It is possible that the response rate was higher, perhaps as much as 25% for these four testing services. For the other testing services and sources of data, the response rate was notably lower, at an estimated 11.0%. These testing services and other sources of test data used a less-concentrated approach to soliciting participation and following-up with reminders to secure participation. The response rate for this study comparable to what many experience in this type of social science research (Fowler, 1988). On the other hand, the response rate in this study is lower than in many social science studies.
Pilot Study
No pilot study was done in the present study for the following reasons: (a) a pilot study was done in Ray's (1990, 1994) nationwide United States and Canada studies which were very similar to this study, (b) Ray’s (1997, 2000) and Rudner’s (1999) studies practically served as pilot studies to the present study, (c) similar instruments and research designs had been used by the researcher and others in previous studies on home education (Ray, 1990, 1994, 1995, 1997, 2000; Rudner, 1999), and (d) there was no reason to predict that this study would pose significant differences in terms of how home educators would respond to the requests made of them in this study.
Data Analysis and Statistical Hypotheses
The statistical software SPSS (SPSS, 2007) was used for data analysis. Students’ scores on tests were handled in the following manner. Percentile equivalents were converted to z-scores (Hopkins, Glass, & Hopkins, 1987, Appendix Table A; see also the Appendix in this report). Means were calculated and statistical tests were performed using z-scores.
Converting percentile equivalents to a standard score (e.g., normal curve equivalent, NCE; z-score) before statistical analysis is the proper method (Loveless, 2002; Tallmadge & Wood, 1978; Yin, Schmidt, & Besag, 2006). Using students’ scores on a variety of nationally normed standardized achievement tests in one study is a robust approach to measuring academic achievement of groups of students (Tallmadge & Wood, 1978; Tallmadge & Wood, 1978, e.g., p. 16, 19-21; Wechsler, 1991, e.g., p. 208-209), although some suggest it would be ideal to use only one test for a project such as this. Using a variety of standardized tests has been done before in research on home-educated students (e.g., Ray, 1990, 1994, 1997, 2000; Van Pelt, 2003) and thus this approach allows for a reasonable level of comparability to findings in previous studies. Further, student performance on major commercial standardized achievement tests is likely highly correlated (e.g., Horst, Tallmadge, & Wood, 1974, p. 29-30). In this study, z-scores were used because they provided the most reasonable way to aggregate scores from many students using a variety of tests, and to analyze how those scores compared to standardized test norms and to each other.
It is not assumed in this study that scores on different tests mean, necessarily, the same thing about the students who took them (Gronlund & Linn, 1990), nor is it assumed that students in this study are perfectly analogous to those students represented by norms for the standardized tests that these students took. It is assumed, however, that the use of aggregated scores from a variety of standardized achievement tests is an acceptable practice and provides valuable information (Frisbie, 1992; Hunter & Schmidt, 1990, p. 516-518; and previous references).
In many cases, simple descriptive statistics and frequencies were appropriate and reported. A number of hypotheses related to the research questions were tested. The hypothesis tested in all cases was the null hypothesis. For example, in testing correlations, the hypothesis was that there was no relationship between the variables. In comparing groups, the hypothesis was that there was no difference between the groups.
Alpha was set at 0.01 for statistical tests in this study and for the sake of determining when to call a relationship statistically “significant” in the narrative. However, p-values are reported so that the reader can make his or her own determination of what is significant. Alpha is the “. . . level of significance used to decide whether to accept or reject a [statistical] null hypothesis...” (Borg & Gall, 1989, p. 352). Alpha was set at 0.01 for several reasons. First, this level of alpha (rather than .05 or .10, for example) helps to take into account multiple error rate (Good, 1984). Second, this approach was consistent with prior research (Ray, 1990, 1994, 1997, 2000). Finally, this level of alpha helps reduce the probability of Type I error in this situation where the rejection of a true null hypothesis might involve potential harm to people like those involved in the study (Shavelson, 1988, p. 286). For example, concluding that there is a statistically significant (or practically significant) relationship between parents’ teacher certification status and children’s academic achievement, when in fact there is none, could do harm to family integrity and children’s learning in terms of subsequent policy decisions related to home education.
Findings
Several selected findings from the data are presented here. Although 11,739 homeschool students were included in the analysis, the total in any given table may differ from the 11,739 students involved in the study because of missing data (e.g., no response) about the students.
Characteristics of Homeschool Students and Families
The homeschool students in the study lived in all 50 states and 2 territories (Guam, Puerto Rico); none were from the District of Columbia.
Regarding gender, 50.3% (5,872) of the students were male and 49.7% (5,809) were female.
Age
Table 1 shows the ages of the home-educated students in the study at the time of achievement testing. About 56% were ages 9 through 13.
Table 1 Participating Homeschool Students Classified by Age
Note 1. The total in any given table may differ from the 11,739 students involved in the study because of missing data (e.g., no response) about the students.
Grade Level
Table 2 shows the grade levels of the homeschool students. Compared to students in public schools nationwide, the distribution of homeschool students in this study is somewhat positively skewed; a disproportionately high portion of them are in grades 3 through 8.
Table 2 Grade Level Completed by Spring 2008
a. National data for public schools in 2005: The 2009 Statistical Abstract, Table 233, retrieved 3/31/09 from http://www.census.gov/compendia/statab/cats/education/elementary_and_secondary_education_schools_and_enrollment.html, data from http://www.census.gov/compendia/statab/tables/09s0233.xls b. Total not 100 due to rounding.
Race/Ethnicity
Table 3 displays the racial/ethnic distribution of the home-educated students.
Table 3 Racial/Ethnic Distribution of Homeschool Students, in Percents
a. National data for race/ethnicity in 2007: The 2009 Statistical Abstract, Table 8, retrieved 3/31/09 from http://www.census.gov/compendia/statab/cats/population/estimates_and_projections_by_age_sex_raceethnicity.html; data downloaded 3-31-09 from http://www.census.gov/compendia/statab/tables/09s0008.xls b. Total not 100 due to rounding.
Marital Status of Parents
Table 4 shows the marital status of the students’ parents; 97.9% of the students’ homes are headed by a married couple. Nationwide, in families with children under age 18, 71.2% are headed by a married couple (U.S. Census Bureau, 2009, Table 63).
Table 4 Homeschool Students Classified by Parents’ Marital Status
Children at Home
Table 5 shows how many children aged 21 and under live in each student’s home. Some 6.6% of the households have one child, 25.3% have 2 children, and 68.1% have three or more children. These homeschool families had an average of 3.5 children age 21 or under. In all families across the United States with children under age 18, 42.6% have one child, 35.6% have two children, and 19.8% have three or more children, with an average of about 2.0 (U.S. Census Bureau, 2009, Table 63).
Table 5 Homeschool Families Classified by Family Size
a. Number of children age 21 or under in home.
Father’s and Mother’s Religion
Table 6 displays the religion of the students’ fathers and mothers.
Table 6 Homeschool Students Classified by Father’s and Mother’s Religion
a. Fathers, n = 11,498; mothers, n = 11,680.
Parents’ Academic Attainment
Table 7 shows the formal academic or educational attainment of the students’ fathers and mothers. Some 66.3% of the students’ fathers had attained a bachelor’s degree (i.e., a four-year degree) or higher. In 2007, 85.0% of all males aged 25 years and older nationwide had finished high school (ergo, 15.0% had not) and 29.5% had finished college (a bachelor’s degree or higher) (U.S. Census Bureau, 2007, Table 1). Of the homeschool students’ mothers, 62.5% had finished a bachelor’s degree or higher. In 2007, 86.4% of all females nationwide ages 25 years nationwide and over had finished high school or higher (so 13.6% had not) and 28.0% of females had finished a bachelor’s degree or higher (U.S. Census Bureau, 2007, Table 1).
Table 7 Educational Attainment of Fathers and Mothers
a. Fathers, n = 11,565; mothers, n = 11,715.
Family Income
The median family income was $75,000 to $79,999, while 4.8% of the families had an income of $29,999 or less. Table 8 shows incomes of the students’ families. For comparison, in 2006 in married-couple families with one or more related children under age 18, the median nationwide income was $74,049 (in 2006 dollars; or roughly $79,015 in 2008 dollars, Westegg.com, 2009) (U.S. Census Bureau, 2009, Table 678).
Table 8 Distribution of Family Income for Homeschool Families
Computer Use
A large majority of the homeschool students, 98.3% (n = 11,499; and 1.7% do not), have a computer at home that they use. By comparison, 82% of school children nationwide whose parents held a bachelor’s degree were using a computer at home and 84% of the same were using a computer at school in 2003 (i.e., children in nursery school and students in grades K-12 whose parents’ educational attainment was a bachelor’s degree; DeBell & Chapman, 2006).
Money Spent on Educational Materials
Table 9 shows the amount of money spent on homeschool education per student for one year. The median amount spent per this one year on the student’s education for textbooks, lesson materials, tutoring, enrichment services, testing, counseling, evaluation, and so forth is $400 to $599.
Table 9 Homeschool Students Classified by Money Spent On Homeschool Education in 2007-2008
Other Demographic Characteristics
Some 80.6% of homeschool mothers do not work for pay. Of the 19.4% who do work for pay, about 84.8% do so part time. Of the homeschool fathers, 97.6% work for pay, and 98.3% of these work full-time.
Of the students’ 23,182 parents, 89.4% had never been certified teachers. Some 15.8% of the homeschool mothers (n = 11,668) had ever been certified teachers; 5.3% of fathers (n = 11,514) had ever been certified.
Only 10.2% of the respondents were enrolled in a full-service curriculum program (i.e., a program that serves students and their parents as a “one-stop” primary source for textbooks, materials, lesson plans, tests, counseling, evaluations, record-keeping, and the like for the year’s core required subjects such as language, social studies, mathematics, and science).
Achievement Tests Used
Table 10 shows that the Iowa Test of Basic Skills (for 45.5% of students) and California Achievement Test (44.2%) accounted for a total of 89.7% of the scores used in the study.
Table 10 Academic Achievement Tests Used by Homeschool Students
Academic Achievement
Following are descriptive statistics about the homeschool students’ academic achievement. Then the relationship between several selected variables and achievement are explored in this section.
Overall Achievement
Table 11 shows the mean z-scores for home-educated students on the Reading Total, Language Total, Mathematics Total (with computation), Science, Social Studies, Core (with computation), and Composite (with computation) subtest scores. Core is comprised of combination of a student’s Reading, Language, and Mathematics scores. Composite is a combination of all subtests that the student took on the test. The corresponding percentiles shown in the table are the within-grade percentile scores for the nation that correspond to the given z-scores. By definition, the 50th percentile is the mean for all students nationwide (last column).
Table 11 Mean z-Scores and Corresponding National Percentile by Subtest for Homeschool Students
a. Following are a few z-score/percentile equivalents: -0.67 = 25th percentile; 0.00 = 50th percentile; 0.67 = 75th percentile; 1.00 = 84th percentile.
Whether Parents Knew Scores Before Participating in Study
An analysis of variance (ANOVA) shows that students whose parents knew their scores before participating in the study (27.7% of students) performed no better on the tests than students whose parents did not know their scores before participating (F=.06; df=1,10367; p=.812, n.s.; scores known, mean=1.1588; scores not known, mean 1.1630).
It should be noted that all F ratios reported here are from an ANOVA with core z-scores as the dependent measure, and typically two independent variables (including grade level). To assure adequate cell sizes, the analyses were restricted to Grades 1 through 11. A statistically significant difference only means that there is evidence of a difference in population values. The difference may be small and not meaningful in a practical sense. Also, “n.s.” is used to indicate not significant.
Years of Homeschooling
Over half of the respondents (59%) indicated that they had been home educated for every grade prior to their current grade, that is, their entire academic life. Table 12 shows that students who are homeschooled for their entire academic life perform the same as those students who have been home educated for only some of their academic years (F=.78; df=1,10405; p=.378, n.s.). There is a statistically significant interaction between grade and years homeschooled (F =2.56; df=10,10405, p=.004), indicating that the effectiveness of homeschooling varies with the student’s grade, but this interaction is minimal and explains only two-tenths of 1% of the variance in scores.
Enrolled in a Full-Service Curriculum
There is no significant difference in the mean core z-scores of home-educated students enrolled in a full-service curriculum and home-educated students not so enrolled (F=1.11.; df=1,10399; p=.292, n.s.) (Table 13). The pairs of means are quite similar at all grade levels; there is no significant interaction between enrollment in a full-service curriculum and grade level (F =1.56, df=10,10399; p=.111, n.s.).
Student Gender
There are statistically significant differences in the achievement levels of male versus female homeschool students (F gender=13.88; df=1,10352; p<.001), but the difference (i.e., females outperforming males) is very slight with gender explaining only one-tenth of 1% of the variance in scores (i.e., see the partial eta squared). The means are very similar at all grade levels; there is no significant interaction between gender and grade level (F =1.23, df=10,10352; p=.264, n.s.) (Table 14).
Money Spent on Educational Materials
There is a statistically significant difference in the achievement levels of homeschool students depending on the amount of money spent per child on educational materials including textbooks, lesson materials, tutoring, enrichment services, testing, counseling, and evaluation (see Table 15). At nearly every grade level, if the family spends $600 or more on the student, the student outperforms students for whom less is spent (F money spent=18.17; df=3, 10345; p <.001). It should be noted, however, that the amount spent is related to a very small effect size and explains only one-half of 1% of the variance in scores. There is no significant interaction between amount spent and grade level (F =1.413, df=30,10345 ; p=.067, n.s.).
Family Income
There is a statistically significant difference in the achievement of homeschool students based on family income. As shown in Table 16 students in higher income families consistently have higher mean core z-scores (F=25.15; df=3,10011; p<.001). The effect is small, practically speaking, with income explaining only about one-half of 1% of the variance in test scores. There is also a significant interaction of income and grade (F =1.50; df=30,10011; p=.038), but this interaction is minimal and explains just under one-half of 1% of the variance in scores.
Parent Certification as a Teacher
To determine whether there is a difference in achievement for students in households where at least one parent had ever held a state-issued teaching certificate, parent education level was controlled. As shown in Table 17, the achievement levels across groups are remarkably similar. Controlling for parent education level, there is a significant difference in the achievement levels of homeschool students whose parents are certified and those that are not (F=11.155; df=11,10141; p=.001); the students having neither parent ever certified performed slightly better. Although statistically significant, whether either parent has ever been a certified teacher explains less than one-tenth of 1% of the variance in test scores. There is no significant interaction of parent certification status and grade (F =.274; df=10,10141; p=.987, n.s.).
Parent Education Levels
Research consistently shows obvious differences in the performance levels of public school students nationwide as a function of parent’s educational level (Rudner, 1999). Differences by parent education level also appear in the performance levels of home-educated students (Table 18). At every grade level, homeschool children whose parents are both college graduates outperform children whose parents both do not have a college degree (F=132.36; df=2,10215; p<.001). There is also a significant interaction between grade and parent education (F=1.995; df=20, 10215; p<.005), indicating that the effect of parent education is more pronounced in some grades. It is notable, however, that parents’ education level explains only 2.5% of the variance in the scores and, at every grade level, the mean performance of homeschool students whose parents (both, if two) do not have a college degree is higher than the mean performance of students in public schools; their percentiles are at the 80th percentile or above.
Degree of State Regulation
Students’ scores were examined according to the degree of state regulation (definitions below) in their states at one point in time, the Spring of 2008, close to when most of the data were collected. Table 19 shows that there is no significant relationship between degree of state regulation of homeschooling and test scores (F=3.113; df=2,10410; p=.045, n.s.).
The fifty states and some of the territories were classified according to their degree of regulation of or control over homeschooling. Definitions that were essentially the same as those used by Ray (1997) were used for the present study. Experts on homeschool law at the Home School Legal Defense Association provided the ratings of the states, with review and oversight by the researcher. The states were classified according to the following definitions:
Table 19 Degree of State Regulation of Homeschooling in Spring 2008 and Test Scores
Low control states – AK, GU, ID, IL, IN, MI, MO, NJ, OK, PR, TX Medium control states – AL, AZ, AR, CA, CO, CT, DE, FL, GA, HI, IA, KS, KY, LA, ME, MD, MN, MS, MT, NE, NV, NH, NM, NC, OH, OR, SC, SD, TN, UT, VA, WV, WI, WY. High control states – MA, NY, ND, PA, RI, VT, WA.
The effect of the degree of state regulation over home-based education was also examined for only students who had been homeschooled for at least 5 years in a state in which the degree of regulation had not changed for at least 5 years (Table 20). In this analysis, also, there is no significant relationship between degree of state regulation of homeschooling and test scores (F=3.190; df=2,5602; p=.041, n.s.).
Table 20 Degree of State Regulation of Homeschooling and Test Scores, for students being home educated for 5 years or more in a state with same degree of regulation for those 5 years (Spring 2003 through Spring 2008)
Low control states – AK, GU, ID, IL, IN, MI, MO, NJ, OK, PR, and TX. Medium control states – AL, AS, AZ, AR, CA, CO, CT, DE, DC, FL, GA, HI, IA, KS, KY, LA, ME, MD, MN, MS, MT, NE, NV, NH, NM, NC, NMI, OH, OR, SC, SD, TN, VI, VA, WV, WI, and WY. High control states – MA, NY, ND, PA, RI, VT, and WA.
Finally, the effect of the degree of state regulation over homeschooling was examined for only students who had been homeschooled for at least 12 years in a state in which the degree of regulation had not changed for at least 12 years (Table 21). In this analysis, also, there is no significant relationship between degree of state regulation of homeschooling and test scores (F=2.499; df=2,243; p=.084, n.s.).
Table 21 Degree of State Regulation of Homeschooling and Test Scores, for students being home educated for 12 years or more in a state with same degree of regulation for those 12 years (Spring 1996 through Spring 2008)
Low control states – GU, ID, IL, IN, MI, MO, NJ, OK, PR, and TX. Medium control states – AL, AS, AZ, CA, CO, CT, DE, DC, FL, GA, HI, IA, KS, KY, LA, MD, MN, MS, MT, NE, NH, NM, NC, NMI, OH, OR, SC, SD, TN, VI, VA, WI, and WY. High control states – MA, NY, ND, PA, RI, VT, and WA.
Number of Children in the Family
There is a significant inverse relationship between the number of children (aged 21 and under) living in the home and core z-scores (r=-.081, n=10348, p<.001). That relationship, however, is very small and explains only about one-half of 1% of the variance in scores.
Degree of Structure in the Home Education
Parents were told, in the survey instrument, the following: “The ‘degree of structure’ in the practice of home education varies greatly. It ranges from a very unstructured learning approach, (e.g., centered upon the child's interests or the eclectic nature of the teaching parent) to the use of a preplanned, structured, and highly prescribed curriculum.” Parents then rated their own practice, with their student, on a 7-point scale from “very unstructured” (with a value of 1) to very structured (with a value of 7). There is a significant relationship between degree of structure and core z-scores (r Spearman=.058, n=10417, p<.001). That relationship, however, is very small and explains only less than one-half of 1% of the variance in scores.
Structured Learning Time
Parents were told, in the survey instrument, the following: “’Structured learning’ is time during which the child is engaged in learning activities planned by the parent; it is a time during which the child is not free to do whatever he or she chooses.” Then parents indicated how many hours per day, on average, this child has been engaged in structured learning. There is a statistically significant relationship between amount of structured learning time in which a homeschool student is engaged and core z-scores (r=.083, n=10405, p<.001). That relationship, however, is very small and explains only about one-half of 1% of the variance in scores.
Age at Which Formal Instruction Began
Parents were instructed, in the survey instrument, as follows: “Consider ‘formal instruction’ to be planned or intentional instruction in areas such as reading, writing, spelling, or mathematics; it is done to meet a learning objective.” Parents then indicated the age at which they began formal instruction of this child. For analysis, students were selected whose parents reported they began formal instruction of the student some time from age 1 through 13. There is a significant inverse relationship between age at which formal instruction began and core z-scores (r=-.080, n=10652, p<.001). That relationship, however, is very small and explains only about one-half of 1% of the variance in scores.
Overall Adjustment to Test Scores?
Since the test scores of the homeschool students are so high, on average, an efforts was made to find whether there might be some clear ways to account for the relatively high scores.
Parents Knew Scores Before Participating Versus Did Not Know
As presented earlier, students whose parents knew their scores (mean of 88th percentile) before participating in the study performed no better on the tests than students whose parents did not know their scores (mean of 88th percentile) before participating.
“All” Scores Provided Versus Full Participants’ Scores
The scores of all students tested by three of the four major testing services were sent to the researcher. The scores of these students, a total of 22,584, nearly all of whom were home educated, are presented in Table 22. (That is, the testing services reported that a tiny minority might have been taught in small private schools.) These comprise the scores of both those who participated and those who did not participate in the present study.
Table 22 Mean z-Scores and Corresponding National Percentile by Subtest for All Students From Three Major Testing Services (i.e., Participants and Non-Participants)
Table 23 presents the subtest mean z-score and national percentiles of all students tested by the three testing services and only the participants in this study. The means of the all-student group were lower than the means of the participant-only-student group, with z-scores differences ranging from 0.10 to 0.16, and percentile differences ranging from 2 to 4. These are very small differences, in a practical sense.
Table 23 Comparison of Mean z-Score and National Percentiles of All Students Tested by Three Major Testing Services and Participants in the Study
a. Following are a few z-score/percentile equivalents: -0.67 = 25th percentile; 0.00 = 50th percentile; 0.67 = 75th percentile; 1.00 = 84th percentile. b. Sample sizes and standard deviations are presented in preceding tables.
Higher-Response-Rate Versus Lower-Response-Rate Student Groups
Another approach to accounting for the relatively high scores was to compare the scores of the participants from the four major testing services for whom the response rate was more readily calculated and was higher to the participants via the other eleven relatively minor test-score sources for which it appears the response rate was lower (see preceding section on response rates). Table 24 presents the Core subtest scores and shows that they are identical (88th percentile).
Table 24 Test Scores of Major Testing Services and Minor Testing Services
Summary of Major Findings in This Study
Major findings: Demographics
Major findings: Achievement
Discussion
The findings of the present study, in the context of all preceding research on the scholastic achievement of home-educated students, bring three main things to mind. First, the overall scores are well above public school student averages. This is consistent with the body of extant research. It is not a surprise. At the same time, however, some research and many anecdotes within the homeschool community and news stories suggest that the demographic, philosophical, and pedagogical variety within the homeschool movement has been broadening over the past decade and one might have expected homeschool achievement scores to lower or regress toward the mean somewhat. There is no evidence of such a decrease in academic performance in the present findings.
Second, the test scores seem, in some ways, notably too high. The researcher was expecting, if anything, a lowering of the scores compared to what was found in previous large-scale nationwide studies (Ray, 1990, 1997, 2000; Rudner, 1999). Reasons for this might have been a demographic broadening of the homeschool population (e.g., perhaps less focused on academic achievement), regression toward the mean, and with nationwide focus on efforts such as No Child Left Behind that emphasize academic achievement in government schools the relative performance of the home educated might be lower than before (Hong & Youngs, 2008, e.g., p. 4). But applying simple controls within the data set (e.g., such as checking whether there was any difference in scores based on whether the parents knew the student’s scores before participating in the study) and finding no major differences in scores suggest that the scores are solid and rather dependable. Furthermore, comparing all students’ scores from the major testing services and comparing them to only participants’ scores revealed rather small differences between the two groups. Perhaps homeschool students are simply doing as well or better than in the past on achievement tests due to their and their parents’ educational efforts. Perhaps the difficulty level of standardized tests has been reduced somewhat over the past decade. Or perhaps the ability of the norm group (public and private school students) has declined somewhat over the past decade. Finally, perhaps the present study captured a sample of the homeschool population at large that is over-represented by high achievers even though it was designed to sample from a broad and robust sampling of the homeschool community that uses standardized tests.
The third notable finding is the lack of statistically significant relationships and effect sizes of practical importance between academic achievement and the several factors examined. It was intriguing to find that the students of parents who had never been state-certified teachers did better than those whose parents were certified, although the effect size appears basically trivial. Only parent education level explained a noticeable or practically significant amount of variance in student scores, and that was 2.5%. One explanation for why so few practically significant effect sizes were found might be that since the scores are relatively high overall that there is attendant range restriction. That is, when the range of scores is relatively narrow, practically significant associations are less likely to be found.
Developing a sample from the widest source ever of homeschool student test scores, this study offers plentiful information concerning the students’ demographics and achievement. It shows, as did Rudner’s (1999) study, that homeschool students and their families are in some ways a select population: although the income for homeschool families with school-age children in this present study is at about the national median, “… education levels are well above national averages. The family structure is traditional with married couples as parents, several children, father as bread winner, and a stay-at-home mother. ….. Home school families do not spend a great deal of money on educational materials and tend not to subscribe to pre-packaged full-service curriculum programs.”
It must also be kept in mind that there are notable limitations to this study. First, homeschool families and their students do not appear to be a completely representative cross-section of all families in the United States. And it was not possible within the constraints of this study to confirm whether this sample is representative of the population of home-educated students.
The content of the standardized tests used is another major limitation of this study. Again, as Rudner (1999) aptly noted: “While home schools teach the basic skill areas of reading, mathematics, social studies, and science, they do not necessarily follow the same scope, sequence, or emphasis as traditional public and private schools. The primary focus of many home schools is on religious and moral values. ….. Public and private schools usually select [a standardized test] … due to its close alignment with their curriculum; home schools select the test primarily out of convenience.”
As previously noted, the results of the present study are consistent with preceding studies of the academic achievement of homeschool students (Ray, 1990, 1994, 1997, 2000; Rudner, 1999; van Pelt, 2003). The above-average nature of these achievement test scores is also consistent with state-provided data in states that have mandated or used testing of the home educated (e.g., Alaska Department of Education, 1993; Arkansas Department of Education, 1998; Oregon Department of Education, 1999). Comparisons between home-educated students and institutional school students nationwide should, however, be interpreted with thoughtfulness and care. As stated at the beginning of this report, this is a nationwide cross-sectional, descriptive study (Johnson, 2001). It is not an experiment and readers should be careful about assigning causation to anything.
One could say, as Rudner (1999) wrote: “This study simply shows that those parents choosing to make a commitment to home schooling are able to provide a very successful academic environment.” On the other hand, it may be that something about the typical nature and practice of home-based education causes higher academic achievement, on average, than does institutional state-run schooling (Ray, 1997; 2000, p. 91-100; 2005). Similar to what Holt (1983) suggested nearly three decades ago, academic leaders could entertain this possibility and consider what those ingredients might be, and how the theoreticians and practitioners involved in conventional institutional schools might be informed by their counterparts in the parent-led home-based education community.
References
ACT. (2005). ACT high school profile report, h s graduating class 2005, hs graduating class, national report, Code 990‑000 . Retrieved 11/28/05 online http://www.act.org/news/data/05/pdf/data.pdf. Alaska Department of Education. (1993). Centralized Correspondence School: Summary of the school district report card to the public. Juneau, AK: Author. Arkansas Department of Education, Student Assessment. (1998, February 17). Personal communication with Yvette Dillingham, program support manager. Barber, G. (2001, February 20). Barber of Educational Testing Service to Brian D. Ray, February 20, 2001, personal communication. Bielick, S. (2008, December). 1.5 Million Homeschooled Students in the United States in 2007. Washington, DC: U.S. Department of Education (National Center for Education Statistics). Retrieved December 23, 2008 from http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2009030. Borg, W. R., & Gall, M. D. (1989). Educational research: An introduction (5th ed.). New York, NY: Longman. Cizek, G. J. (1988). Applying standardized testing to home-based education programs: Reasonable or customary? Educational Measurement Issues and Practice, 7(3), 12-19. Cizek, G. J. (1993). The mismeasure of home schooling effectiveness: A commentary. Home School Researcher, 9(3), 1-4. Cizek, Gregory J. (Ed.) (2001). Setting performance standards: Concepts, methods, and perspectives. Mahwah, NJ: Lawrence Erlbaum Associates, Inc., Publishers. DeBell, M, & Chapman, C. (2006, September). Computer and Internet use by students in 2003 (NCES 2006–065) [Figure 2]. Washington, DC: U.S. Department of Education, National Center for Education Statistics. Retrieved 4/8/09 from http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2006065. Fowler, F. J., Jr. (1988). Survey research methods (revised edition). Newbury Park, CA: SAGE Publications. Frisbie, D. (1992). Book review. Journal of Educational Measurement, 29(2), 275. Galloway, R. A., & Sutton, J. P. (1995). Home schooled and conventionally schooled high school graduates: A comparison of aptitude for and achievement in college English. Home School Researcher, 11(1), 1-9. Galloway, R. S., & Sutton, J. P. (1999). Are adults who were home educated experiencing success in their adulthood? Private School Monitor, 20 (3), 4-6. Good, R. (1984). A problem of multiple significance tests. Journal of Research in Science Teaching, 21(1), 105‑106. Gray, D. W. (1998). A study of the academic achievements of homeschooled students who have matriculated into post-secondary institutions. Dissertation Abstracts International, 59, 021. Doctoral dissertation, University of Sarasota, Sarasota, Florida, 1998. Gronlund, N. E., & Linn, R. L. (1990). Measurement and evaluation in teaching, 6th edition. New York, NY: Macmillan Publishing Company. Hardenbergh, Nicky. (2008, March 30). Validity of high stakes standardized test requirements for homeschoolers: a psychometric analysis. Otherwise Instructed: Issues In Education. Retrieved December 21, 2009 from http://otherwiseinstructed.com/docs/validityofhighstakeshomeschooltesting.pdf. Holt, J. C. (1983). Schools and home schoolers: A fruitful partnership. Phi Delta Kappan, 64, 391-394. Hong, W-P., & Youngs, P. (2008). Does high-stakes testing increase cultural capital among low-income and racial minority students?. Education Policy Analysis Archives, 16(6). Retrieved April 3, 2009 from http://epaa.asu.edu/epaa/v16n6/. Hopkins, K. D., Glass, G. V., & Hopkins, B. R. (1987). Basic statistics for the behavioral sciences (2nd ed.). Englewood Cliffs, NJ: Prentice-Hall, Inc. Horst, D. P., Tallmadge, G. K., & Wood, C. T. (1974, October). Measuring achievement gains in educational projects. Los Altos, CA: RMC Research Corporation. Also ERIC number ED114406, retrieved January 15, 2009 from http://www.eric.ed.gov/ERICWebPortal/contentdelivery/servlet/ERICServlet?accno=ED114406. Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis: Correcting error and bias in research findings. Newbury Park, CA: Sage Publications. Jenkins, T. P. (1998). The performance of home schooled students in community colleges. Doctoral dissertation, Texas A & M University-Commerce. Johnson, B. (2001). Toward a new classification of nonexperimental quantitative research. Educational Researcher, 30(2), 3-13. Retrieved 5/30/06 online http://www.aera.net/uploadedFiles/Journals_and_Publications/Journals/Educational_Researcher/3002/AERA3002_Johnson.pdf. Jones, P., & Gloeckner, G. (2004, Fall). A study of admission officers’ perceptions of and attitudes toward homeschool students. Journal of College Admission, No. 185, 12-21. Knowles, J. G., & Muchmore, J. A. (1995). Yep! We're grown-up home-school kids C and we're doing just fine, thank you. Journal of Research on Christian Education, 4(1), 35-56. Loveless, T. (2002). How well are American students learning? Washington, DC: Brookings Institute. Retrieved January 16, 2009 from http://www.brookings.edu/~/media/Files/rc/reports/2002/09education/09education.pdf. Medlin, R. G. (2000). Home schooling and the question of socialization. Peabody Journal of Education, 75(1 & 2), 107-123. Mexcur, D. (1993). A comparison of academic performance among public school graduates, conventional Christian school graduates, accelerated Christian school graduates, and home school graduates in three Christian colleges. Doctoral dissertation, Bob Jones University, Greenville, SC. Montgomery, L. R. (1989). The effect of home schooling on the leadership skills of home schooled students. Home School Researcher, 5(1), 1-10. Oliveira (de Oliveira), P. C. M., Watson, T. G., & Sutton, J. P. (1994). Differences in critical thinking skills among students educated in public schools, Christian schools, and home schools, by. Home School Researcher, 10(4), 1-8. Oregon Department of Education [Office of Student Services]. (1999, May 20). Annual report of home school statistics [1998-99]. Salem, OR: Author. Piedmont Education Services. (2009). Retrieved 3/12/09 from http://www.pesdirect.com/. Ray, B. D. (1990). A nationwide study of home education: Family characteristics, legal matters, and student achievement. Salem, OR: National Home Education Research Institute. Ray, B. D. (1994). A nationwide study of home education in Canada: Family characteristics, student achievement, and other topics. Salem, OR: National Home Education Research Institute. Ray, B. D. (1995). Learning at home in Montana: Student achievement and family characteristics. Salem, OR: National Home Education Research Institute. Ray, B. D. (1997). Strengths of their own—Home schoolers across America: Academic achievement, family characteristics, and longitudinal traits. Salem, OR: National Home Education Research Institute. Ray, B. D. (2000). Home schooling: The ameliorator of negative influences on learning? Peabody Journal of Education, 75(1 & 2), 71-106. Ray, B. D. (2004). Home educated and now adults: Their community and civic involvement, views about homeschooling, and other traits. Salem, OR: National Home Education Research Institute. Ray, B. D. (2005). A homeschool research story. In Bruce S. Cooper (Ed.), Home schooling in full view: A reader, p. 1-19. Greenwich, CT: Information Age Publishing. Ray, B. D. (2009a, August 10). Research facts on homeschooling. Retrieved December 16, 2009 from http://www.nheri.org/Research-Facts-on-Homeschooling.html. Ray, B. D. (2009b). Home education reason and research: Common questions and research-based answers. Salem, OR: National Home Education Research Institute. Ray, B. D., & Eagleson, B. K. (2008, August 14). State regulation of homeschooling and homeschoolers’ SAT scores. Journal of Academic Leadership, 6(3). Retrieved January 23, 2009 from http://www.academicleadership.org/emprical_research/State_Regulation_of_Homeschooling_and_Homeschoolers_SAT_Scores.shtml. Reich, R. (2005). Why home schooling should be regulated. In Bruce S. Cooper (Ed.), (2005), Homeschooling in full view: A reader. Greenwich, CT: Information Age Publishing. Rudner, L. M. (1999). Scholastic achievement and demographic characteristics of home school students in 1998. Retrieved 5/16/08 from http://epaa.asu.edu/epaa/v7n8/. Seton Home Study School. (2008). Retrieved 7/23/08 from http://www.setonhome.org/history.php. Shavelson, R. J. (1988). Statistical reasoning for the behavioral sciences (2nd ed.). Boston, MA: Allyn and Bacon. SPSS, Inc. (2007). SPSS 16.0 for Windows (release 16.0.1, November 15, 2007). Chicago IL: Author. Tallmadge, G. K., & Wood, C. T. (1978, January). ESEA Title I evaluation and reporting system: User's guide (revised, January 1978). Mountain View, CA: RMC Research Corporation. United States Census Bureau. (2007). Educational Attainment in the United States: 2007; Table 1. Educational Attainment of the Population 18 Years and Over, by Age, Sex, Race, and Hispanic Origin: 2007. Retrieved April 6, 2009 from http://www.census.gov/population/www/socdemo/education/cps2007.html. United States Census Bureau. (2009). The 2009 Statistical Abstract. Retrieved tables on or about 3-31-09 from http://www.census.gov/compendia/statab/. Van Pelt, D. (2003). Home education in Canada: A report on the pan-Canadian study on home education 2003. Medicine Hat, Alberta: Canadian Centre for Home Education. Van Pelt, D. A. N., Allison, P. A., & Allison, D. J. (2009). Fifteen years later: Home-Educated Canadian Adults. London, Ontario, Canada: Canadian Centre for Home Education. Retrieved 12/2/09 from http://www.hslda.ca/cche_research/2009Study.pdf. Wartes, J. (1991, December). Five years of homeschool testing within Washington State. (Available from the Washington Homeschool Research Project at 16109 N. E. 169 Pl., Woodinville, WA, 98072). Wechsler, D. (1991). Wechsler Intelligence Scale for Children, third edition, manual. San Antonio, TX: The Psychological Corporation, Harcourt Brace Jovanovich, Inc. Assessment. Welner, K. M., & Welner, K. G. (1999, April 11). Contextualizing homeschooling data: A response to Rudner [Lawrence M. Rudner. 1999. Scholastic achievement and demographic characteristics of home school students in 1998. Education Policy Analysis Archives]. Education Policy Analysis Archives, 7(13). Retrieved 4/13/00 from http://epaa.asu.edu/epaa/v7n13.html. West, R. L. (2009, Summer/Fall). The harms of homeschooling. Philosophy and Public Policy Quarterly, 29(3/4), 7-12. Retrieved December 17, 2009 from http://www.puaf.umd.edu/files.php/ippp/vol29summerfall09.pdf. Westegg.com. (2009). The inflation calculator. Retrieved 8/5/09 from http://www.westegg.com/inflation/. Yin, R. K., Schmidt, R. J., & Besag, F. (2006). Aggregating student achievement trends across states with different tests: Using standardized slopes as effect sizes. Peabody Journal of Education, 81 (2), 47-61. © Copyright 2010 by Academic Leadership | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
