The Social and Emotional Health of Homeschooled Students in the United States: A Population-Based Comparison with Publicly Schooled Students Based on the National Survey of Children’s Health, 2007

Guillermo Montes

Associate Professor, Ralph C. Wilson, Jr. School of Education, St. John Fisher College,

Rochester, New York



Although there is substantial research on homeschooling and its correlates, much of the research has used samples that are likely to suffer from self-selection biases (Kunzman & Gaither, 2013). In this article, we use the National Survey of Children’s Health (NSCH) 2007, a large survey collected by the US government to monitor the health of American children. Because the primary intent of the NSCH was not to study homeschooling but the health of American children it is likely that the degree of self-selection into the survey because of homeschooling is negligible. The purpose of this study was to determine if homeschooled students differed in social competencies and behavior problems or in their likelihood to participate in community activities compared to publicly schooled students on a U.S. population basis. Generally, this study failed to detect statistically significant differences between both groups in the NSCH Social Competencies scales (both social skills and behavior problems subscales), in items measuring social competencies, behavior problems, or depressed mood, or in overall participation in community activities, including paid work. Thus, homeschooled students are not at risk for socialization problems in the U.S. There were three minor exceptions: homeschooled children were less likely to participate in sports, and were reported as less likely to “argue too much” or “care about doing well in school” by their parent. We note that the “care for school” item was worded in a culturally insensitive manner for the homeschool population and could be interpreted ambiguously by families who homeschool as either caring for formal schooling or caring for educational work in the homeschooling context. Thus, differences in this item must be interpreted with caution.

Keywords: Behavior problems, social skills, socialization, participation in community, health survey, homeschooling, social competence, emotional intelligence, National Survey of Children’s Health, United States


The National Center for Education Statistics (NCES) estimated that 1.5 million children were homeschooled in the US in 2007 (Planty et al., 2009).  Although there is substantial research on homeschooling and its correlates, much of the research is predicated on samples that are likely to suffer from self-selection biases (Kunzman & Gaither, 2013). In this article, we use the National Survey of Children’s Health 2007, a large survey collected by the US government to monitor the health of American children. Because the primary intent of the NSCH was not to study homeschooling, but the health of American children, it is likely that the degree of self-selection into the survey because of homeschooling is negligible. Thus, this source of information, widely used in academic and medical research to determine population estimates of children with various health conditions—including social and emotional competencies and problems (Belanoff, McManus, Carle, McCormick, & Subramanian, 2012; Bell, Huebner, & Reed, 2012; Bethell et al., 2011; Close, Lee, Kaufmann, & Zimmerman, 2012; Guerrero, Garro, Chang, & Kuo, 2010; Jones, Kogan, Singh, Dee, & Grummer-Strawn, 2011; Kogan et al., 2009; Larson, Russ, Kahn, & Halfon, 2011; Lu et al., 2012; Singh, Siahpush, & Kogan, 2010a; Strickland, Jones, Ghandour, Kogan, & Newacheck, 2011)—can be used to provide nationally representative estimates of the condition of homeschooling children and their families in the United States.


Brief Literature Review


The question of socialization, “are children well socialized in homeschooling?,” is a persistent concern of people worried that homeschooling may somehow undermine children’s social and emotional development. Because there are extensive reviews of the question available (Kunzman & Gaither, 2013; Medlin, 2013; Ray, 2005), we briefly discuss the main issues in this section. As other researchers have noted (Medlin, 2000), socialization researchers, courts, family members, and concerned citizens may mean at least three different and inter-related constructs: (a) do homeschooled children participate in the wider social life of their community? (b) do homeschooling children have more behavior problems and/or worse social competencies than publicly schooled children? and (c) are homeschooled children exposed to a diverse group of people and points of view which is believed to be necessary to function effectively in a democratic, pluralistic society? (Medlin, 2000). Implicit in these questions is the unstated assumption that age-segregated, group education by certified teachers is the normative way to learn to function socially and emotionally in modern U.S. society (Ray, 2005).

Regarding the first question, researchers have shown that homeschooled students participated in more activities than children attending conventional school (Delahooke, 1986), and that homeschooled children actively participated in myriad forms of community activities from organized sports, scouts, volunteer and paid work, church activities and a whole variety of community offerings from dance and visual arts to hobby-related groups (Ray, 1997; Rudner, 1999; Tillman, 1995). Furthermore, as adults, homeschooled children are active participants in civic life, being more likely to work for political candidates or causes, attend public meetings, participate in protests or boycotts, vote in elections, and contribute money to political candidates than the general population (Ray, 2004; Van Pelt, 2003).

To answer the second question researchers have employed a variety of strategies. A commonly used approach is to rate homeschooled students using standardized social and emotional development research instruments. Studies have compared self-ratings, parental ratings and teacher ratings and have shown either no differences or better scores for the homeschool student group in comparisons with conventionally schooled students or when compared against national norms. Researchers have used well-known measures such as the Vineland Adaptive Behavior Scales, the family and community subscales of the Adaptive Behavior Inventory for Children, the Revised Rutter Scale, the Social Skills Rating System (SSRS) and the Roberts Apperception test for Children (Lee, 1994; Medlin, 2007; Rothermel, 2012; Smedley, 1992; Stough, 1992). Some studies have shown that homeschooled students are more likely than conventionally schooled peers to be less assertive, using the Children’s Assertiveness Behavior Scale, but still within the normal range (Rothermel, 2012; Shyers, 1992). In general, these results are remarkable because the established scales were developed and normed for conventionally schooled children and implicitly contain assumptions of what appropriate social behavior is that may very well apply only to school settings (e.g. high levels of disclosure to and compliance with unrelated adults). Indeed, some of those behaviors may be inappropriate outside of the conventional school context. Using both the Revised Rutter Scale and the Goodman Strengths and Difficulties Questionnaire, Rothermel established that behavioral scales may incorrectly screen homeschooled children as abnormal. She recommends “extreme caution” when health professionals use behavioral scales that were normed on school samples to screen homeschooled students (Rothermel, 2012).

Another strategy used by researchers is direct observation of social behavior rated by the research team using an established observational measure such as the Direct Observation Form of the Child Behavior Checklist (CBCL). Shyers found that homeschooled students were much less likely to exhibit behavior problems on the direct observation form of the CBCL (Shyers, 1992).

A third strategy has been to study students’ self-concept using established measures and comparing them against publicly schooled comparison groups or national norms. Again, results showed no evidence of worse self-esteem, with studies demonstrating either no difference between the groups or better scores for the homeschool group. In comparisons with national norms the homeschool group is typically above average (Lee, 1994; Medlin, 2000; Shyers, 1992; Stough, 1992; Taylor, 1986; Tillman, 1995)

Finally, regarding exposure to diverse group of people and points of views, research has shown that homeschooled students are more likely than conventionally schooled students to have frequent contacts with other adults and younger children (Chatham-Carpenter, 1994). Obviously, active participation in community activities and employment are clear indicators that homeschooled students are well integrated into American society.

In sum, research to date shows that children who are homeschooled are not at risk of socialization problems, a finding validated by domestic and international studies across different cultures and school systems (Medlin, 2013; Rothermel, 2012).


Purpose of the Present Study


A weakness of the current literature is that studies often suffer from self-selection bias and are not designed to be nationally representative of the entire non-institutionalized child population of the country. Using a nationally representative US data source, the National Survey of Children’s Health (NSCH) 2007, collected for purposes other than homeschooling research, is likely to provide us a source of information with little bias due to homeschooling status. Because the NSCH 2007 included items that measure the social and emotional health of children it offered an opportunity to answer the socialization question with a nationally representative data source with little homeschooling self-selection bias.

The purpose of the study was to determine if homeschooled students had better or worse social or emotional health using the NSCH 2003 social competencies and behavior problem scales, and to conduct an item analysis to determine if there were any differences at the individual item level.






The National Survey of Children’s Health (NSCH), 2007 is a module of the National Center for Health Statistics and Centers for Disease Control and Prevention (CDC)’s State and Local Area Integrated Telephone Survey (SLAITS). The survey “was designed to produce national and state-specific prevalence estimates for a variety of physical, emotional, and behavioral health indicators and measures of children’s experiences with the health care system” (Blumberg, Foster, Skalland, Chowdhury, & O’Conner, 2012). The survey interviews the parent or guardian who knows best about the child’s health (typically the mother) using a random-digit-dialed sample of household with children less than 18 years old from all 50 states and the District of Columbia. In households with multiple children the child about whom the interview asks questions was sampled randomly. The questionnaire was based on the general framework and questionnaire used for the NSCH 2003 modified based on input from the SLAITS email listserv, CDC, Child Trends and the Advisory Committee for the Data Resource Center on Child and Adolescent Health. The survey resulted in 91,642 interviews. Interviews were conducted in English, Spanish, Mandarin, Cantonese, Vietnamese and Korean. The public data file includes weights that allow generalization to the non-institutionalized population of children in each state and nationwide. The overall response rate was 46.7%; the survey completion rate was 66.0%. The survey was funded by the Maternal and Child Health Bureau, Health Resources and Services Administration. All respondents consented to the study. There is a comprehensive report that details additional information on consent procedures and design, administration, and quality control of the survey (Blumberg et al., 2012). There is also an additional report on the size and direction nonresponse bias in the NSCH 2007 that found small biases (≤+/-3%) could be present in the NSCH 2007 estimates even when using weights that correct for nonresponse biases. Yet the presence of such biases and their direction varied depending on the method used to identify them and also on the question studied (Skalland & Blumberg, 2012).

The questionnaire section primarily used in this article was section 7, middle childhood and adolescence (ages 6 to 17), which contained the question on whether the child was homeschooled. This section also contained the NSCH Social Competence Scale described below.


Analytic Sample


The analytical sample was restricted to households with a child 6 to 17 years of age who were either homeschooled or publicly schooled. Private schooled students and the very few cases that reported not being in any of the three schooling types were excluded.




Homeschooling. Parents of children aged 6 to 17 were asked “during the last school year, what kind of school was (child) enrolled in? Is it a public school, private school or home-school?”


NSCH Social Competence Scale. The National Survey of Children’s Health Social Competence Scale (NSCH-SCS) was developed using a Rasch-family Graded Response Model on data from 67,405 households with children aged 6-17 from the NSCH 2003 (Blumberg, Carle, O’Connor, Moore, & Lippman, 2008). The measure has two four-item factors: Social Skills and Behavior Problems. Originally, the items were measured in a 4 point scale, but in the NSCH 2007, each item is measured on a five point scale (never, rarely, sometimes, usually, and always). In our sample, both factors had reliabilities comparable or superior to those reported in the measure development article (Cronbach’s alpha 0.65 for social skills and 0.71 for behavior problems) (Blumberg et al., 2008).


Behavior problem items. Parents were asked to rate the frequency of the following common childhood behavior problems on a five point scale (never, rarely, sometimes, usually and always): (a) (he/she) argues too much*, (b) (he/she) bullies or is cruel or mean to others*, (c) (he/she) is disobedient*, (d) (he/she) is stubborn, sullen or irritable*, (e) (he/she) feels worthless or inferior, (f) (he/she) is unhappy, sad or depressed, (g) (he/she) is withdrawn, and does not get involved with others. Starred items constitute the behavior problems subscale of the NSCH-SCS. The three items without stars measure depressed mood in various ways. For item analyses purposes these items were coded as either the behavior is observed (sometimes, usually or always) versus it is observed rarely or never.


Social competence items. Parents were asked to rate the selected child on the frequency of the following behaviors  using the five point scale (never, rarely, sometimes, usually and always): (a) (he/she) shows respect for teachers and neighbors*, (b) (he/she) gets along well with other children*, (c) (he/she) tries to understand other people’s feelings*, (d) (he/she) tries to resolve conflicts with classmates, family and friends*, (e) (he/she) cares about doing well in school, and (f) (he/she) does all required homework. Starred items form the social skill scale of the NSCH-SCS. For item analyses purposes these items were coded as either the behavior was frequently observed (usually or always) versus it was observed sporadically (never, rarely or sometimes).


Participation in activities. Parents were asked if during the past 12 months the child was on a sports team or took sports lessons afterschool or on weekends using a yes/no question. Similarly, parents were asked if during the past 12 months child participated in any clubs or organizations after school or on weekends. If the parents responded no to either or both questions then they were asked if during the past 12 months, the child participated in any other organized events or activities. A separate question asked if during the last 12 months the child had been involved in community service or volunteer work at school, church or in the community using a 4 point metric (never, a few times a year, a few times a month, once a week or more). We coded this variable as 1 if there was any volunteer work and zero otherwise. A total participation variable was constructed was equal to 1 if the child was said to participate in any of the activities described above.

Finally, the survey also asked if the child was employed, defined as earning money from work, during the last week “including regular jobs as babysitting, cutting grass or other occasional work” in a yes/no question. Because the temporal frame of this question was only one week it was analyzed separately.


Demographic information. The NSCH 2007 public data file comes with a number of derived variables that are quite useful to characterize the family demographically. Age of the child in years was measured using the derived variable ageyr_child. Education was coded as either high school graduate or less or more than a high school graduate based on the derived variables Educ_momr, educ-dadr and educ_resr. The maximum level of education for the household was obtained by comparing these variables. Hispanic or Latino ethnicity of the child was measured using the derived variable Hispanic. Two poverty level variables were constructed based on the derived variable poverty_levelr, POV200, a measure of poverty, that measured if the household was at or below 200% the Federal poverty line (FPL) and POV400, a measure of affluence, that measured if the household was at or above 400% FPL. The FPL depends on annual household income and number of dependents in the household. Minority ethnicity of the child was coded as child is minority if child was Black, multiracial or of other ethnicity versus child is not minority if child was identified as white.



We used Stata (Stata Corp, College Station, TX) to adjust all estimates for complex sample design using the sampling weights that permit generalization to the noninstitutionalized child population in the US.




Analysis proceeded in four steps. First, demographic information about the child and household was compared for publicly schooled children versus homeschooled children using cross-tabulations and design-adjusted F tests. Second, information on the key variables of the study (social skill items, behavior problem items, participation in activities) was compared for both groups using cross tabulations and design-adjusted F tests. Third, the NSCH-SCS subscales’ averages were compared for both groups using design-adjusted linear regression (i.e. ANOVA). Finally, statistically significant differences in item analysis and NSCH-SCS subscales were studied further using a multivariate logistic or linear regression that included sex, age, household education, minority status, Hispanic ethnicity, and both the poverty and affluence variables as covariates. A p-value ≤.05 was considered statistically significant.




Prevalence of Homeschooling


Based on the full NSCH 2007 sample ages 6-17, the proportion of students homeschooled in the United States is 2.4% (95% Confidence Interval 2.09%-2.68%), thus the US homeschooled population ages 6-17 is between 1.02 and 1.31 million children. The sample had 1530 homeschooled students  and 54,320 publicly schooled students. Because the focus of this article was comparisons between publicly schooled students and homeschooled students all subsequent analyses excluded privately schooled students.


Demographic Differences Among Groups


Table 1 shows the demographic characteristics of the sample by schooling status. Both groups of students were comparable by age of the child (F=1.17, ns), and sex (F=1.63, ns). Minority students (15.6% vs. 28.6%, p≤.01) and Hispanic students (6.6% vs. 20.4%, p≤.01) were less likely to be homeschooled. Population estimates were that more than 150,000 minority and 66,000 Hispanic students in the 6-17 age range were homeschooled in the US in 2007. Homeschooling households were more likely to have a parent who had education beyond high school (77.7% vs. 64.8%, p≤.01). Marginally significant, households that homeschooled were more likely to be at or below 200% the federal poverty line (65.1% vs. 58.2%, p=.05), and were less likely to be affluent households, defined as at or above 400% FPL, compared to publicly schooled households (18.8% vs. 27.9%, p≤.01).  In sum, homeschooled students were likely to live in less affluent white non-Hispanic households than their publicly schooled counterparts.

These variables – gender, age, education, minority status, Hispanic ethnicity, at or below 200% FPL , and at or above 400% FPL – became the standard covariates used in multivariate analyses.


Behavior Problems


Table 2 provides item-level comparison for behavior problems, including depressed mood, between both groups. Parents of homeschooled students reported their child argued too much less frequently than parents of publicly schooled students (51.0% vs. 62.5%, p≤.01). This result persisted after controlling for the covariates of table 1 in multivariate logistic regression (Adjusted Odds Ratio (AOR) 0.61, p≤.01) There were no statistically significant differences in any of the other problem behavior items indicating that the likelihood of parents classifying their child as bullies, disobedient, stubborn, worthless or depressed was unrelated to type of schooling.


Social Competencies


Similarly, Table 2 shows the percentage of students rated as having social competencies by their parents in each group. Both groups were statistically indistinguishable in five out of the six items (respect for adults, getting along with other children, trying to understand other people’s feelings, trying to resolve conflicts, and completing homework). Homeschooled students were less likely to be rated as caring about doing well in school than publicly schooled students (80.8% vs. 86.8%, p≤.05), a result that persisted in multivariate analyses (AOR=0.57, P≤.05).


NSCH SCS Comparisons


There were no statistically significant differences in the NSCH SCS behavior problem subscale and indeed both groups obtained almost identically scores (M=8.98 vs. M=8.99, ns). This result persisted in multivariate analyses after controlling for the covariates in Table 1 (b=-.31,ns). Similarly, there were no statistically differences in the NSCH SCS social skills subscale scale (M=16.7 vs. M=17.1, ns), a result confirmed by multivariate analysis (b=.01,ns).


Participation in Activities


Table 3 shows the percentage of students by various types of community participation activities. Homeschooled students were less likely to have participated in sports teams or taken sport classes in the last 12 months than publicly schooled students (45.9% vs. 56.9%, p≤.01) but were more likely to participate in afterschool clubs and organizations (62.3% vs. 55.5%, p≤.05). Among those students who did not participate in either sports or clubs, homeschooled students were more likely to have attended in organized events or activities (36.4% vs. 23.9%, p≤.05).




Table 1.

Demographic Characteristics of Homeschooled Students and Students Attending Public Schools



Table 2.

Behavior Problems and Social Competencies by Schooling Status


Table 3.

Participation in Activities by Schooling Status






Homeschooled students were equally likely to participate in volunteer work or community service in school, church or community than publicly schooled students (83.0% vs. 76.3%, ns). The overall participation of both groups was quite similar (87.8% vs. 84.0%, ns). It is notable that more than 8 in 10 homeschooled students were engaged in outside activities. Multivariate analyses confirmed the results on sports (AOR = .57, p≤.01), but not those in afterschool activities (AOR=1.19, ns), or events (AOR= 1.57, ns).


Finally, both groups were equally likely to be employed during the week prior to the survey (42.7% vs. 40.3%, ns).




The strength of this study is that it uses a nationally representative data set that was collected for purposes other than homeschooling research, thus self-selection biases into the study based on homeschooling status are likely to be very small. Because this study uses a nationally representative data source widely used in medicine to calculate prevalence of conditions affecting the health of American children (Belanoff et al., 2012; Bell et al., 2012; Bethell et al., 2011; Close et al., 2012; Guarnizo-Herreño & Wehby, 2012; Guerrero et al., 2010; Jones et al., 2011; Kogan et al., 2009; Larson et al., 2011; Lu et al., 2012; Ness, Barradas, Irving, & Manning, 2012; Singh et al., 2010a; Singh, Siahpush, & Kogan, 2010b; Sisson, Broyles, Newton Jr, Baker, & Chernausek, 2011; Strickland, Jones, Ghandour, Kogan, & Newacheck,  2011; Visser, Bitsko, Danielson, Perou, & Blumberg, 2010), this study also provides information on the prevalence of homeschooling and the demographic characteristics of students who are homeschooled. First, we estimate the prevalence of homeschooling among children ages 6-17 at 2.6% in 2007. Second, on a population basis, white children are more likely to be homeschooled than minority or Hispanic children. Yet, more than 66,000 Hispanic and 150,000 minority students were homeschooled in the U.S. Families who homeschooled were less affluent than families who sent their children to public schools in the United States. Other studies have found that homeschooling households have similar or slightly below annual incomes and have a greater number of children; since the federal poverty line is calculated based on annual income and number of dependents living in the households it makes sense that homeschooling households are more likely to be at lower multiples of the federal poverty line. The fact that homeschooling families are, on average, less affluent than families that send their children to public schools nationally is important contextual information for policymakers that intend to impose onerous and often ineffective regulations on homeschooling families. Homeschooling families not only have lower income than their publicly schooled counterparts, but face additional costs of educating as well as the school taxes. Therefore, many of these families may be economically vulnerable.

The main purpose of this study was to investigate the social and emotional health of homeschooled students in the United States using a population-based governmental data source routinely used by both the government and academic researchers to monitor the health of American children. The study was based on more than 55,000 students, of whom about 1,500 were homeschooled. Therefore the study was sufficiently powered to detect small differences between the groups. It was surprising that this study revealed almost no systematic differences in the social and emotional health of homeschooled students in the United States.  There were no detectable differences in the NSCH 2003 social competencies and behavior problem scales and indeed both groups had remarkably similar overall scores. Similarly, item analysis discovered that there were no detectable differences in the majority of the social competencies and behavior problem items between both groups. With regards to participation in activities, homeschooled students had similar overall participation as publicly schooled students.

Given that statistical inference cannot ever rule out differences in the population (because, theoretically, even the smallest of differences would become statistically significant at sufficiently large sample sizes), one cannot state that there are really no differences in the US population between these groups of students. We can only state that we failed to detect any.  However, given the sample size of the study and its concomitant statistical power we can assert that whatever true population differences we did not detect are likely to be as small in size as to have negligible public policy and practical significance. Thus, the main conclusion of this study is that homeschooled students in the US have comparable mental and emotional health and comparable community participation to their publicly schooled counterparts, even though their households are less affluent.  There is no evidence that homeschooled students are at greater risk of socialization problems than publicly schooled students in the US.

We found some minor exceptions to this conclusion at the item level. First, homeschooled students were substantially less likely than publicly schooled students to be rated as arguing too much.  This is somewhat remarkable because homeschooled students spend substantially more time with their parents than publicly schooled students. Homeschooled students must negotiate with their parents an additional set of complex educational tasks (e.g. curriculum, instructional practice, homework, balance between educational work and community activities, etc.). Thus, it was a positive finding to find out that they are able to accomplish these additional complex educational tasks and additional interaction time and yet argue less.

The second item where significant differences were detected had to do with whether the student “cares about doing well in school.” Unfortunately, the wording of the item prevents us from reaching a definite conclusion because it is insensitive to the homeschool culture. In particular, some homeschooling mothers undoubtedly interpreted the item as it was intended, namely that the child cares about educational work; but others mothers, particularly those who homeschooled as a response to some recent real or perceived failure of the public school, may have interpreted the item as asking if the child cares about doing well in the conventional school.  The issue is that most screening and assessment items are developed with the tacit assumption that every child is in a conventional school setting and thus may be inappropriate to students who are homeschooled. Researchers have pointed out that such items must be used with great caution (Rothermel, 2012). Thus, the only definite conclusion is that designers of national survey questionnaires must be more culturally sensitive in the wording of their items when using the word “school” to mean broader concepts such as educational work.

Third, homeschooled students participated less in organized sports than publicly schooled students. This is not entirely surprising because in the US sports are organized mostly by schools. There are well-documented legal as well as  informal barriers to the full participation of homeschool students in these sports activities (Batista & Hatfield, 2005; Gardner & McFarland, 2001). It seems, however, that homeschooled students are able to compensate and participate in other clubs, and activities so that their overall participation rate is comparable to the publicly schooled rate.

The lack of detectable differences in the other items is remarkable for two reasons: first, homeschooled households were on average less affluent than publicly schooled households and lower income is associated with increased behavior problems and lower social competencies. Second, compared with parents of publicly schooled students, parents of homeschooled children spend more time with their children and thus have greater opportunity to observe many of the behaviors and social competencies asked in the items. So one could expect a greater frequency of behavior problems simply because of greater opportunity to both exhibit the behavior and observe it. Similarly, in some items related to educational work one would expect homeschooling parents would be “tougher” simply because they have better information than parents of publicly schooled students. For example, not many parents of a 16 year old student in public school have complete knowledge of the homework assigned and thus answer the question about completed homework with significantly less information than a homeschool parent who likely is assigning and correcting the homework the homeschooled student is completing.

In sum, this study corroborates previous research summarized by Medlin (2013) that found no reason to be concerned about the socialization of homeschooled children in the United States. We concur with Medlin’s suggestion that future research should focus on the process of socialization in the homeschooling context instead of outcomes. In the larger context of arguments against homeschooling both from educators and policy makers, this research suggests that opponents need not be concerned about the socialization of homeschooled children and are either unfamiliar with or choose to ignore the cumulative empirical evidence (Ray, 2013).


Limitations of the Study


This study has a number of potential limitations. Causality cannot be inferred from a cross-sectional study. First, by including all homeschooling students in a particular year, this study includes transitory homeschooled students, those who are just beginning to homeschool or who will homeschool only for a brief period and these students may have worse social and emotional health than students who are homeschooled in a stable manner.  Second, some of the items in the measures used assume the child is in conventional school, as has been previously discussed. Third, the study relies on parental report but these specific questions were both validated and widely used in academic research.




Using a widely used population-based health governmental data source, we established that social and emotional health and community participation of US homeschooled students, ages 6 to 17, is comparable to that of publicly schooled students. Homeschooled students are not at risk of socialization problems in the U.S.




Batista, Paul J., & Hatfield, Lance C. (2005). Learn at Home, Play at School: A State-by-State Examination of Legislation, Litigation and Athletic Association Rules Governing Public School Athletic Participation by Homeschool Students. J. Legal Aspects Sport, 15, 213.

Belanoff, Candice M., Mcmanus, B. M., Carle, A. C., McCormick, M. C., & Subramanian, S. (2012). Racial/ethnic variation in breastfeeding across the US: A multilevel analysis from the National Survey of Children’s Health, 2007. Maternal and child health journal, 16(1), 14-26.

Bell, Janice F., Huebner, C. E., & Reed, S. C. (2012). Oral health need and access to dental services: Evidence from the National Survey of Children’s Health, 2007. Maternal and child health journal, 16(1), 27-34.

Bethell, Christina D., Kogan, M. D., Strickland, B. B., Schor, E. L., Robertson, J., & Newacheck, P. W. (2011). A national and state profile of leading health problems and health care quality for US children: key insurance disparities and across-state variations. Academic Pediatrics, 11(3), S22-S33.

Blumberg, Stephen J., Carle A. C., O’Connor, K. S., Moore, K. A., & Lippman, L. H. (2008). Social competence: Development of an indicator for children and adolescents. Child Indicators Research, 1(2), 176-197.

Blumberg, Stephen J., Foster, E., Skalland, B., Chowdhury, S., & O’Conner, K. (2012). Design and operation of the National Survey of Children’s Health, 2007. Vital and Health Statistics. Series 1: Program and Collection Procedures. Department of Health and Human Services; 2009.

Chatham-Carpenter, April (1994). Home versus public schoolers: Differing social opportunities. Home School Researcher, 10(1), 15-24.

Close, Heather A., Lee, L.-C., Kaufmann, C. N., & Zimmerman, A. W. (2012). Co-occurring conditions and change in diagnosis in autism spectrum disorders. Pediatrics, 129(2), e305-e316.

Delahooke, Mona M. (1986). Home educated children’s social/emotional adjustment and academic achievement:a comparative study. (Doctoral Dissertation), California School of Professional Psychology, Los Angeles, CA.

Gardner, Kathryn, & McFarland, A. J. (2001). Legal precedents and strategies shaping home schooled students’ participation in public school sports. J. Legal Aspects Sport, 11, 25.

Guarnizo-Herreno, Carol C., & Wehby, G. L. (2012). Explaining racial/ethnic disparities in children’s dental health: a decomposition analysis. Journal Information, 102(5).

Guerrero, Alma D., Garro, N., Chang, J. T., & Kuo, A. A. (2010). An update on assessing development in the pediatric office: has anything changed after two policy statements? Academic Pediatrics, 10(6), 400-404.

Jones, Jessica. R., Kogan, M. D., Singh, G. K., Dee, D. L., & Grummer-Strawn, L. M. (2011). Factors associated with exclusive breastfeeding in the United States. Pediatrics, 128(6), 1117-1125.

Kogan, Michael D., Blumberg, S. J., Schieve, L. A., Boyle, C. A., Perrin, J. M., Ghandour, R. M., van Dyck, P. C. (2009). Prevalence of parent-reported diagnosis of autism spectrum disorder among children in the US, 2007. Pediatrics, 124(5), 1395-1403.

Kunzman, Robert, & Gaither, M. (2013). Homeschooling: A comprehensive survey of the research. Other Education: The Journal of Educational Alternatives, 2(1), 4-59.

Larson, Kandyce, Russ, S. A., Kahn, R. S., & Halfon, N. (2011). Patterns of comorbidity, functioning, and service use for US children with ADHD, 2007. Pediatrics, 127(3), 462-470.

Lee, Walter J. (1994). The socialization of home-schooled and public-schooled children. (Doctoral Dissertation), University of La Verne, La Verne, CA.

Lu, Emily, Dayalu, R., Diop, H., Harvey, E. M., Manning, S. E., & Uzogara, S. G. (2012). Weight and mental health status in Massachusetts, National Survey of Children’s Health, 2007. Maternal and child health journal, 16(2), 278-286.

Medlin, Richard G. (2000). Home schooling and the question of socialization. Peabody Journal of Education, 75(1-2), 107-123.

Medlin, Richard G. (2007). Homeschooled children’s social skills. Home School Researcher,  17(1), 1-8.

Medlin, Richard G. (2013). Homeschooling and the question of socialization revisited. Peabody Journal of Education, 88(3), 284-297.

Ness, Maria, Barradas, D. T., Irving, J., & Manning, S. E. (2012). Correlates of overweight and obesity among American Indian/Alaska Native and non-Hispanic white children and adolescents: National Survey of Children’s Health, 2007. Maternal and child health journal, 16(2), 268-277.

Planty, Michael, Hussar, W., Snyder, T., Kena, G., KewalRamani, A., Kemp, J., . . . Dinkes, R. (2009). The condition of education 2009 (NCES 2009–081). Washington, DC: National Center for Education Statistics, Institute of Education Sciences, US Department of Education.

Ray, Brian D. (1997). Home education across the United States: family characteristics, student achievement and other topics. Purcellville, VA: HSLDA Publications.

Ray, Brian D. (2004). Home educated and now adults: Their community and civic involvement, views about homeschooling and other traits. Salem, OR: NHERI Publications.

Ray, Brian D. (2005). Worldwide guide to homeschooling: facts and stats on the benefits of homeschool. Nashville, TN: Broadman & holman Publishers.

Ray, Brian D. (2013).  Homeschooling associated with beneficial learner and societal outcomes but educators do not promote it. Peabody Journal of Education, 88, 324-341.

Rothermel, Paula J. (2012). Home educated children’s psychological well-being. Estudios sobre la educacion, 22, 13-36.

Rudner, Lawrence M. (1999). Scholastic achievement and demographic characteristics of home school students in 1998. Education Policy Analysis Archives, 7(8), 2-37.

Shyers, Larry E. (1992). A comparison of social adjustment between home and traditionally schooled students. Home School Researcher, 8(3), 1-8.

Singh, Gopal K., Siahpush, M., & Kogan, M. D. (2010a). Disparities in children’s exposure to environmental tobacco smoke in the United States, 2007. Pediatrics, 126(1), 4-13.

Singh, Gopal K., Siahpush, M., & Kogan, M. D. (2010b). Rising social inequalities in US childhood obesity, 2003–2007. Annals of epidemiology, 20(1), 40-52.

Sisson, Susan B., Broyles, S. T., Newton Jr, R. L., Baker, B. L., & Chernausek, S. D. (2011). TVs in the bedrooms of children: does it impact health and behavior? Preventive medicine, 52(2), 104-108.

Skalland, B., & Blumberg, S. (2012). Nonresponse in the National Survey of Children’s Health, 2007. Vital and health statistics. Series 2, Data evaluation and methods research(156), 1-22.

Smedley, Thomas C. (1992). Socialization of home schooled children. Home School Researcher, 8(3), 9-16.

Stough, L. (1992). Social and emotional status of home schooled children and conventionally schooled children in West Virginia. (Master’s Thesis), University of West Virginia, Morgantown.

Strickland, Bonnie B., Jones, J. R., Ghandour, R. M., Kogan, M. D., & Newacheck, P. W. (2011). The medical home: health care access and impact for children and youth in the United States. Pediatrics, 127(4), 604-611.

Taylor, John W. (1986). Self-concept in home schooling children. Home School Researcher, 2(2), 1-3.

Tillman, Vicki. D. (1995). Home schoolers, self-esteem, and socialization. Home School Researcher, 2(2), 1-3.

Van Pelt, Deani A. (2003). Home education in Canada. London, Ontario, Canada: Canadian Centre for Home Education.

Visser, S., Bitsko, R., Danielson, M., Perou, R., & Blumberg, S. (2010). Increasing prevalence of parent-reported attention-deficit/hyperactivity disorder among children—United States, 2003 and 2007. Morbidity and Mortality Weekly Report, 59(44), 1439-1443.