Social Skills: A Comparison Study

Social Skills of Home Schooled and Conventionally

Schooled Children: A Comparison Study

 David J. Francis, Psy.D.

School Psychologist

 Saranac Lake Central School District

Saranac Lake, New York

Timothy Z. Keith, Ph.D.

Professor of Educational Psychology

University of Texas, Austin

Despite opposition from many local school boards, the National Education Association (NEA), and school administrators, home schooling continues to be a growing trend across the nation (Holtrop, 1996; Mayberry, 1989; Ray, 1996, Ray, 2002; Ritter, 1997). Recent estimates suggest that 1.5 to 2.0 million children may be currently home schooled in grades K through 12 nationwide (Ray, 2002). Concerns regarding the lack of religious teachings in public schools, negative influences of their children’s peer culture, and questionable quality of the education received in public schools are some reasons why parents have chosen to educate their children at home (Dahm, 1996; Mayberry, 1989; Murray, 1996). However, as support for home education increases, so does the skepticism from public school administrators, who are now forced to take a critical look at the reasons why students are leaving their school districts.
Public educators have questioned the socialization opportunities home schooled children receive if they are denied the traditional exposure to social interaction provided in conventional schools. Moreover, educators are concerned about the level of instructional quality these children will receive, as many home schooling parents are not certified to teach (Ray, 1996; Taylor, 1986). The media have begun to present the home school debate focusing on the social and academic concerns expressed by public educators (Knowles, 1988, Ray, 1992). However, most people know little about home schoolers: their backgrounds, their activities, or their achievements (Ray, 1997). Researchers have begun to address these concerns within the small body of literature that has steadily grown with the increasing number of home schooling families over the past two decades (Knowles, 1988). The majority of these studies have focused on the academic achievement of home schooled children; the social skills of home schooled children have received less attention (Ray, 1997).
Academically, the evidence suggests that home schooled children perform at or above the national average on standardized achievement tests when compared to conventionally schooled peers, (Frost, 1988; Murray, 1996; Ray, 1997; Wartes, 1987) and this performance resembles that of children in private schools (Lines, 1995). Moreover, home school advocates point to the increased interest from colleges and universities in actively recruiting home schooled children (Ray, 1999) and home schooled children’s success in academic competitions as further evidence of their academic success. Regardless of these academic accomplishments, detractors of home education continue to doubt that home schooling has positive effects on children’s overall social skills or competence.

Keywords: homeschooled, home-school,  home-based, education, comparison, children, social skills

Indirect Measurement of Social Skills: The Problem

Only recently have researchers examined the effects of home schooling on the social skills of home schooled children. Therefore, few data exist from which to draw conclusions about the social behavior of these children (Ray, 1997). Furthermore, many such studies have attempted to measure social skills by examining the self-concept and self-esteem of home schooled and traditionally schooled children, rather than social skills or social competence (Hedin, 1991; Kelley, 1991; Kitchen, 1991; Stough, 1992; Taylor, 1986; Tillman, 1995). These studies suggest that home-educated children have equivalent or higher self-concepts than the norm. Although such findings are of interest, self-concept and self-esteem are not equivalent to social skills or behavior (Eder, 1997).
Social skills are a difficult and complex variable to measure and can be approached in a variety of different ways (Gresham & Elliott, 1984). Within the home school literature, a variety of different methods and perspectives have been used to examine social skills. For example, social differences of home educated children have been investigated via self-esteem and self-concept (Hedin, 1991; Kelley, 1991; Kitchen, 1991; Stough, 1992; Taylor, 1986; Tillman, 1995), social adjustment and social maturity (Delahooke, 1986; Shyers, 1992; Smedley, 1992), and leadership skills (Montgomery, 1989). Additionally, family interaction patterns (Carson, 1990) and differences in social opportunities available to home educated children have also been examined (Chatham-Carpenter, 1994; Tillman,1995; Wartes, 1987). Yet few of these studies have attempted to address the socialization question with the use of more direct measures of social skills.
Although attempts have been made to address this problem empirically by examining social adjustment (Shyers, 1992) and social maturity (Smedley, 1992), the data in this area continue to be sparse and inconclusive (Aiex, 1994; Mayberry, 1989; Ray, 1997). Conducting additional studies with more appropriate and direct measures of social skills, therefore, would be an asset to the current home school research. A suggested method is to measure the social skills that are typically acquired during childhood (Gresham & Elliot, 1990).

Social Skills

Social skills are those skills that are “socially acceptable learned behaviors that enable a person to interact effectively with others and to avoid socially unacceptable responses” (Gresham & Elliott, 1990, p. 1). Behaviors such as sharing, helping, giving compliments, and having good manners are all examples of social skills that enable successful relationships throughout the life span (Gresham & Elliott, 1990). Consequently, each child’s mastery of these skills will enable or impair future relationships with both adults and peers (Gresham & Elliott, 1990). To date no study has approached the socialization debate specifically by measuring social skill differences of home schooled and publicly schooled children. A direct measure of social skills may provide some insight into social differences, if any, exhibited by home educated and traditionally educated children.
More conclusive data are needed so that the effect home education has on the socialization of children is further clarified for home and public educators alike. Such a study may contribute to resolution of the current debate between public educators and home school advocates and would meet the need for more causal-comparative studies (Ray, 1992). Past studies have examined “the social skills debate” by looking for differences between home schooled and conventionally schooled children while primarily focusing on the constructs of self-esteem and self-concept. These constructs, however, do not adequately explain the numerous behavioral differences found in children’s social behavior (Eder, 1997). As a result, this study addressed the socialization issue through the use of a social skill measure that examined parents’ perceptions of their children’s social skills. The purpose of this research was to determine whether home-educated children’s social skills differ from those of a paired comparison group of conventionally schooled children.




Participants included 34 pairs of children between the ages of 5 and 18 and their parents; one child from each pair was home schooled, and the other was conventionally schooled. All but one family resided in communities located in rural Western New York and the majority (98%), consisted of two parent families. Due to the nonequivalent nature of the comparison groups (i.e., home educated vs. conventionally educated) the pairing procedure was used to improve equivalency across the two groups. Briefly, home schooling parents nominated conventionally schooled friends to whom their children were compared (via a repeated measures ANOVA). The solicitation of home and conventionally school participants, along with the matching procedure used, is described in more detail in the Procedure section.
Data collection began in October 1998 and was completed by June 1999. Of the 48 home schooling families who were given the initial mailings, 39 (81%) participated. Nine families (19%) declined to participate in the study. Of the 39 who completed the survey, five families (13%) were dropped because they did not satisfy the researcher’s requirements for a home schooling family. Ninety-four percent of the conventional educating families completed the initial mailings. The two conventional educating families who did not participate were replaced.


Home schooled. Home-educating families were recruited by networking with home school organizations and individuals who subscribed to a home school newsletter within a rural area of western New York. For the purpose of this study, home schooled children were defined as those children between the ages of 5 and 18 who were currently being educated at home, and had been so educated for at least two consecutive years. Initially, the first author contacted home educating families to assess their interest for participating in the study. Parents were asked to complete a consent form and a demographic survey if they were willing to participate in the study.
Once a list was gathered of interested home schooling participants, we selected one child from each home schooling family to be assessed. For families with more than one home schooled child, the middle home schooled child, or the child closest to the fifth grade level (10.5 years of age) was selected, as this is the average age of home schooled children in the U.S. (Ray, 1997). In the event that only one child was being home schooled within a family, that child was used as a participant. Researcher selection of the home schooling children to be assessed was done to maintain control over possible selection bias of the home schooling parents, and to increase the response rate from each family by lessening the work involved for the respondents. Moreover, only one child was used from each home schooling family to avoid non-independence of observations. Home schooling parents were sent a Social Skills Rating System (SSRS) parent form (Gresham & Elliott, 1990) for the child selected by the researchers.
Conventionally schooled. For the purposes of this study, conventionally schooled children were defined as those children who were currently being educated within conventional education programs (public or private) and had been so educated for at least two consecutive years. The home schooling parents nominated the conventionally educated participants. Specifically, home schooling parents were asked to name up to three same sex conventionally schooled friends of their home educated children who they believed were similar to their own children. The home schooling parents who completed the SSRS form were encouraged to discuss with their children which friends should be included as potential comparison participants.
Conventionally educating families were contacted via phone and sent a letter briefly explaining the study along with a consent form and demographic survey. Families indicated their interest in participating in the study on the demographic survey and returned this form with the consent form. A second mailing was then sent to participating families including a letter and the SSRS parent form. Parents were instructed to complete the SSRS parent form for the child who was identified by the home educating family as a friend of the home schooled child. In the event that a conventional educating family did not wish to complete the SSRS form, another friend’s family was contacted.
We believe that generating a comparison group from a pool of home schooled children’s friends is particularly valuable because children make friends on the basis of common interests and activities (Hartup, 1996), and because interpersonal attraction among children is guided by similarities in social participation (Rubin, Lynch, Coplan, Rose-Krasnor, & Booth, 1994), beliefs, attitudes, and superficial markers such as age, race, and sex (Hartup, 1996). “Common ground is a sine qua non in friendship relations throughout childhood and adolescence, suggesting that friends ought to be similar to one another in abilities and outlook” (Hartup, 1996, p.5). Overall, the research evidence suggests that “birds of a feather flock together” (Aristotle, Rhetoric, Book 11 as cited in Hartup, 1996, p.5). Thus, we believed that this approach should yield more similar comparison children than other possible approaches. On the other hand, given likely similarities in friends’ social behaviors, this approach may bias our findings against finding differences in social skills across the two groups.
Many home schooling parents educate at home because they want to instill certain humanitarian or religious values into their children (Mayberry, 1989). Therefore, maintaining parental perceptions when choosing a comparison child was very important because parents were expected to select children who they believed shared similar values as their children. Input from each participating home schooled child was also encouraged. Moreover, parental perceptions of their children’s social skills were the criteria measured in this study; therefore, maintaining parental perceptions was the most logical process to use when generating the comparison group.
As an incentive for participation in the study, a drawing for four $25.00 gift certificates from local department stores was offered to all participants who returned a completed SSRS form. These gift certificates were to be spent on educational supplies. Additionally, a summary of results from this study was offered to all families who participated.

Demographic survey. A demographic survey was used to gather information about ages of children participating, the amount of time each child was educated in a conventional or home school environment, residential location, parental occupation and educational levels, religious affiliation of the family, and church attendance. One additional question was asked on the surveys sent to the home schooling families, asking them to indicate their primary reason for home schooling.
Social Skills Rating System. The Social Skills Rating System (SSRS) parent form was used to assess parents’ perceptions of their children’s social skills (Gresham & Elliott, 1990). The SSRS was standardized in 1988 on a national sample of 4,170 self-ratings of children, 1,027 ratings of children by their parents, and 259 ratings of children by their teachers. The SSRS was designed to provide assessment of children’s social competence in both educational and family settings. Moreover, the SSRS emphasizes positive behaviors or prosocial skills, and measures problem behaviors that may interfere with social skill performance (Benes, 1995; Gresham & Elliott, 1990).
The SSRS parent forms consist of an Elementary form (grades K-6) and a Secondary form (grades 7-12). These forms provide four social skill subscale scores (i.e., Cooperation, Assertiveness, Responsibility, and Self-control) in addition to Social Skills Scale Total Scores. The parent forms also surveys parents for Problem Behaviors, and yields a Total Problem Behaviors raw score that is generated from two subscales (i.e., Externalizing and Internalizing) on the Secondary level form, and from three subscales (i.e., Externalizing, Internalizing, and Hyperactivity) on the Elementary form. Items on both the parent forms ask the parent or guardian to rate the frequency of a behavior, (i.e., 0, Never; 1, Sometimes; 2, Very Often) as well as the estimated level of importance of each behavior (i.e., 0, Not Important; 1, Important; 2, Critical). The Problem Behavior Scale, however, only requires parents to indicate the frequency of a behavior. The data from each parent Social Skill subscale (i.e., Cooperation, Assertiveness, Responsibility, and Self-control) and the Problem Behavior scales, (i.e., Externalizing, Internalizing, and Hyperactivity) can be compared to the SSRS standardization norm group on a restricted scale (Fewer, Average, or More than average). The SSRS Parent form used here yielded Total Social Skills Scale Scores and Total Problem Behavior Scores that were converted into age-based standard scores with a mean of 100 and a standard deviation of 15.
The psychometric properties of the SSRS are well addressed in its manual (Gresham & Elliott, 1990). Specifically, in the normative sample the coefficient alpha measuring internal consistency of the parent form ranged from .87 to .90 for the Total Social Skills Scales, and from .81 to .87 for the Total Problem Behavior Scales. Test-retest reliability estimates were .87 for Social Skills and .65 for Problem Behaviors. Support for content and social validity has been provided based on extensive empirical research from the child development, clinical psychology, and special education literatures, and has been supported by the standardization analyses and item development of the forms (Benes, 1995; Gresham & Elliott, 1990). The SSRS has been correlated with a variety of measures including the Social Behavior Assessment, the Harter Teacher Rating Scale, the Piers-Harris Children’s Self-Concept Scale, and various forms of the Child Behavior Checklist (Benes, 1995). The results of these comparisons and evidence from additional studies provide strong evidence supporting the construct validity of the SSRS (Gresham & Elliot, 1990).


A one-way repeated measures analysis of variance (RANOVA) was used to examine whether differences existed between the comparison groups. The matched-pairs within-subject design utilized for this study combined with the above statistical procedure yielded a more sensitive test, as this procedure reduced the error variance by controlling for similarities within paired groups (Schweigert, 1994).


Data were obtained on 34 pairs of children (i.e., 34 home educated and 34 conventionally educated children). Of this group, four conventionally educating families were referred more than once by separate home schooling families. As a result, data were collected twice from four of the conventionally educating families; however, a different child from each conventionally educating family was used. One family served as both a home and conventionally educating family because the children assessed satisfied this study’s methodological requirements for both groups. Because this overlap violates the assumption of independence of observations, analyses were conducted both with and without these families. Findings were the same with and without inclusion of the overlapping cases, therefore results are reported using the entire sample.

Demographic Information
Data were collected across seven counties in rural western New York State (one conventionally schooled family had moved to Ohio). This section will describe the demographic characteristics of the sample, as well as compare those characteristics for home schooled versus conventionally schooled participants. Except where noted, the two groups showed no statistically significant differences on these demographic characteristics, a finding which suggests the equivalence of the home schooled and conventionally schooled groups.
Table 1 compares home schooled and conventionally schooled children on several of the categorical demographic variables. Participants were primarily from rural areas across both groups, and sex of the participants was well distributed between the groups and consisted of 36 males (52.9%) and 32 (47.1%) females. All but one home educating family and one conventionally educating family reported being married. In one of the few differences across groups, statistically significantly more of the mothers in home schooled families reported being homemakers rather than working outside the home (c2 [1, N = 68] = 11.90, p<. 01). This makes sense; if children are home schooled, an adult needs to stay home. In this sample, that adult was generally the mother.
Religious affiliation is a distinguishing variable when discussing home educating families (Ray, 1997). In this sample, 70.6% of the home schooling families were fundamentally, biblically-based, or “born again” Christians, 5.9% were Catholic, 8.8% were Protestant, 5.9% were Baptist, and 8.8% were Mennonite. Conventionally educating families also reported a majority (55.9%) to have a fundamentally, biblically-based Christian religious background, whereas approximately 24% percent were Catholic, 14.7% were Protestant, 2.9% were Lutheran, and 2.9% were Baptist. A chi-square analysis comparing fundamental Christian versus other religious affiliations showed no statistically significant difference between home and conventionally schooled families (χ² [1, N = 68] = 1.58, p>. 10). The groups were similar in terms of their predominately fundamental Christian religious affiliation.



Home Schooled (%)

Conventionally Schooled (%)
Rural 85.3 82.4
Female 47.1 47.1
Two-Parent Families 97.1 97.1
Fundamentalist Christian 70.6 55.9
Mothers Who Are Homemakers 79.4 38.2

Table 1. Characteristics of home schooled and conventionally schooled children and families.

Table 2 compares the means of home schooled and conventionally schooled children and families on a number of continuous variables collected on the demographic survey. Each of these sample characteristics was compared using a repeated measures ANOVA; none were statistically significant. As shown in the Table, the average age of student participants was approximately 10 years. Home schooled children averaged 4.67 consecutive years within their home educational environments, whereas conventionally educated children averaged 5.15 years within their current educational environments. Most (85.3%) conventionally educated students were enrolled within public schools, the remainder, (14.7%) were educated within private/religious schools. Fifty-three percent of the home schooled respondents reported that they had not attended another type of school other than the home education environment.
Three or more children is quite typical of home educating families (Ray, 1992); for this sample, the average number of children per home educating family was 3.59 with approximately 56% of the home schooling families reporting more than three children in their homes. Conventionally schooled families were quite similar, and reported an average of 3.1 children at home, with 32% of these families reporting three or more children in their homes.
As shown in Table 2, the occupational status of fathers and educational attainment of mothers and fathers were quite consistent across groups. Although a high percentage of home schooling mothers stayed at home, they had educational attainments in line with conventionally schooling mothers; indeed, 67.7% had a college or higher degree. Ninety-one percent of home schooling families reported attending church once or more per week; 97% of conventionally schooled families reported this level of church attendance.

Home Schooled
Conventionally Schooled
Age of Participant 10.44 10.47
Years in Current Educational Setting 4.67 5.15
Number of Children in Family 3.59 3.09
Father Occupational Status (1=laborer to 14=professional) 9.69 8.67
Father Educational Attainment (1=some high school to 6=doctoral degree) 3.79 3.76
Mother Educational Attainment 3.71 3.53
Church Attendance (0=do not attend to 5=more than once a week) 4.47 4.15

Table 2. Means of Home schooled and conventionally schooled children and families on continuous demographic variables.

Reasons for Home Schooling
Historically, ideological concerns about the climate of the public school system have been indicated as a primary driving force behind the home schooling movement (Guterson, 1990; Mayberry, 1989). Ideological concerns continue to be important, along with a rise in pedagogical reasons (Cizek, 1993; Common & MacMullen, 1986; Mayberry, 1989). The majority of home schooling families in this study (67.6%), indicated their primary motivation for home schooling was ideological, 17.6% indicated pedagogical reasons, and 11.8% indicated a blend of both pedagogical and ideological reasons for educating at home.

Social Skills
A one-way repeated measures analysis of variance (RANOVA) was used to determine if social skill differences existed between matched pairs of home and conventionally educated children. Differences in social skills were examined using the SSRS parent forms, using the total social skills standard score. The results of this analysis showed a statistically significant difference between the overall social skill standard scores of home and conventional educated groups [F (1, 33) = 4.49, p = .042] (see Table 3). Specifically, home-educated children were found to have higher total social skill standard scores than the matched group of conventionally educated students (Means 113.12 and 107.15, respectively). The type of schooling (home vs. conventional) accounted for 12% of the variance explained in social skills (h² = .120), a moderate effect size using conventional rules of thumb.
RANOVA was also used to test for differences between home and conventionally schooled children on the three SSRS Social Skills subscales. As noted in the Method, the SSRS subscales provide a score on a restricted scale only (fewer, average, or more than average, or 0, 1, and 2 for data analysis). This means that the dependent variable in these analyses had a restricted range, and thus these analyses had considerably lower power than did the analysis of overall social skills. Nevertheless, one of the components of social skills, self-control, was still statistically significant in these comparisons [F (1, 33) = 6.45, p = .016,]. Home educated students showed statistically significantly higher self-control scores than did paired conventionally schooled children (M=2.29 versus 2.00). Although home schooled children scored higher on each of the other subscales (Cooperation, Assertiveness, and Responsibility), the effects were small (h² = .066, .010, .014) and statistically nonsignificant.

Problem Behaviors
A RANOVA was again used to examine the total problem behavior standard scores of the home and conventionally educated subjects (see Table 4). This analysis showed no statistically significant differences between the home and conventional students in terms of their problem behavior standard scores [F (1, 33) = .514, p = .478], and a small effect size (h2 = .015). The standard score means for the Problem Behaviors scale were 94.50 and 96.38 for home schooled and conventionally schooled children, respectively (higher scores represent more problem behavior). None of problem behaviors subscales (internalizing, externalizing, and hyperactivity, using a restricted scoring range) showed statistically significant differences for the home schooled versus conventionally schooled students.






Type of Schooling 1 606.015 606.015 4.490 .042 .120
Between pairs 33 6573.309
Error 33 4454.485 134.984
Total 67 11633.809
Table 3. Repeated measures ANOVA of overall social skills for home schooled versus conventionally schooled children.






 p h2
Type of Schooling 1 60.235 60.235 .514 .478 .015
Between pairs 33 5511.765
Error 33 3864.765 117.114
Total 67 9436.765

Table 4. Repeated measures ANOVA of overall problem behaviors for home schooled versus conventionally schooled children.


Heated debates of concerned parents, educators, and administrators about the social implications home schooling may have on children’s ability to adapt, cope, and maintain themselves within the mainstream social environment was a primary motivating factor for conducting this study. This study assisted in addressing the paucity of research within the home schooling socialization domain. With few studies measuring social skills within the literature to reference, the aforementioned debates may be based primarily on speculation and the “conventional wisdom” of the public at large. To address these concerns, this study investigated whether home educated children’s social skills differed from the social skills of a matched comparison group of similar, but conventionally educated children.
The results from this study indicate that the home schooled children earned higher social skill standard scores than their conventionally educated peers. Although both groups earned social skills standard scores above the average in relation to the standardization sample, the home schooled children earned scores higher above the average in comparison to conventionally schooled children, means 113.1 and 107.1 respectively. These differences occurred despite the high degree of similarity between the groups and strongly suggest that home schooling had a statistically significant positive effect on the home schooled children’s social skills. Home schooled children also earned statistically significantly higher scores on the self-control component of overall social skills.
An additional finding was that no differences occurred between the groups when measuring problem behaviors. The home schooled group did not differ from the conventionally educated group in their total problem behavior standard scores (M 94.50 and 96.38, respectively) or in any of the components of problem behaviors (internalizing, externalizing, or hyperactivity). Both groups were considered to have problem behaviors typical for their age, and were slightly better than average in relation to the standardization sample of the SSRS.
The findings of this research suggest that home schooling does not appear to have any negative effects on the development of proper social skills. To the contrary, the results to this study suggest that the children benefited from an exposure to an education at home as their social skills appear to have been enhanced when compared to their conventionally educated counterparts. In addition, home schooled children show, on average, neither more nor fewer social skill problems than do conventionally schooled children.

A potential limitation of this study is related to the convenience sample utilized by the researchers, which may affect the level of generalizability of these results. Although this study sampled subjects across seven counties in rural Western New York, it is unclear if this sample is representative of the entire state or home schooling families across the nation. However, in comparison to a recent nationwide study by Ray (1997) some similarities in demographics were identified. In particular, similarities were noted in the areas of family size, age of home schooled children, percentage of single parents who home school, and number of years in home education.
A second limitation of this study concerns the potential of sampling bias. Specifically, all participants were volunteers. Therefore the participants may have naturally been more satisfied with home schooling than families who refused to participate. To increase generalizability and variability of this sample, home schooling subjects were recruited using a variety of tactics by the researchers, including networking via word-of-mouth, contacting independent home school support church groups, and by chapter representatives for Loving Education At Home (LEAH) support organizations. Use of a variety of methods to recruit subjects was believed to be the best way of generating the most varied group of home schooling families, and has been often practiced in previous research (e.g., Knowles, 1988; Mayberry, 1989; Ray, 1990).
Third, the use of parents as evaluators of their children’s social skills allows the possibility of bias to be introduced when completing the rating forms. However, parents were thought to provide the most valid assessment of their children’s social skills. This approach was the most logical as parents are likely to spend more time in a day with their children than either teachers (Conners, 1990) or independent observers; therefore, parents should have the greatest knowledge of their children’s social behavior across time and varying situational contexts.
Moreover, a high degree of consistency typically has been found between observers who play similar roles when rating children’s behavioral/emotional problems (i.e., parents vs. parents and teachers vs. teachers) across different situations (Achenbach, McConaughy, & Howell, 1987). Most importantly, it is likely that whatever bias might have been introduced by having parents rate their children’s social skills should have been distributed equally to home schooled and conventionally schooled children.
Fourth, one measure was utilized when examining the socialization domain. This measure used parent’s perceptions of their children’s social skills while measuring social skill differences. This approach may seem to have some limiting qualities in relation to more “objective” measures such as direct observation. However, ratings based on observers’ impressions have been found to equal the predictive power of direct observations (Weinrott, Reid, Bauske, & Brummett, 1981; as cited in Achenbach, McConaughy, & Howell, 1987), and have been used to refine direct observations of children’s behavior (Elliot, Busse, & Gresham, 1993). Moreover, although a multiple assessment method is typically suggested to take into account situational-specific conduct while measuring social behavior (Elliot, Busse, & Gresham, 1993; Gresham & Elliott, 1990) we sought to gain the most valuable information from the participating families while remaining minimally intrusive. As a result, parents were the only evaluators of their children’s social skills.

Advocates of home schooling have conducted much of the previous home schooling research. The researchers in this study were employees of a public school and a University and were neither home school advocates nor fundamental Christians (a common characteristic of home educators; Ray, 1997). The interpretive implications of this study were, therefore, thought to be minimized as the researchers had no incentive to endorse this non-conventional teaching method.
This study is apparently the first to explore the socialization domain of home schooled children using a more direct measure of the socialization domain (e.g. social skills). By using a measure that was specifically designed for parents to complete for the purpose of evaluating their children’s social skills, interpretation of results was assisted, and speculation minimized.
Finally, the matching procedure usedhaving home schooling families nominate a similar conventionally schooled childsucceeded in creating a demographically similar comparison group for the home schooled children studied. This similarity, in turn, bolsters confidence that the social skill differences found are in fact a result of some characteristic related to the home schooling experience. Although the home schooling sample collected is unusual in some ways (e.g., their level of religious commitment), they are representative of home schoolers, in general (Ray, 1997). Furthermore, the use of repeated measures ANOVA for analysis capitalized on the matching procedure to reduce error variance and allowed a more sensitive statistical test.


The current study explored the question: Do home schooled children differ in social skills in relation to a matched group of conventionally educated students? The results of this study indicated that differences did occur between the groups in terms of Total Social Skills Standard Scores. Home schooled children achieved higher scores on this scale than their conventionally educated counterparts. In addition, home schooled children scored higher on the four component scores of overall social skills, and statistically significantly higher on one of these (Self Control). The higher rating scores on these scales strongly suggest that home schooled children in this study were not harmed socially due to their home education. On the contrary, these data suggest home schooled children may have benefited as a result of their education at home in terms of displaying higher overall social skills and self control than the conventionally educated group.



The results of this study question a conventional approach toward education and a frequently referenced secondary goal of conventional school systems. Socialization of children is often believed to be most successfully addressed through immersion into mainstream culture via the public school system (Klicka, 1993; Murray, 1996). The results from this study question this “conventional wisdom.”  In fact, this study may ease concerns of educators (e.g., parents and professionals) who share anxiety over the possible social implications of home education, as the data suggests concern of social ineptness are most likely unwarranted.

This study also demonstrates the importance of using appropriate measures when exploring the socialization domain of home schooled children. Measures should be selected that are designed in a manner in which they can be appropriately applied to a home schooling population. Moreover, they should be sensitive enough to identify and evaluate the specific behavioral differences that may exist between home and conventionally educated children. Identifying the specific social and behavioral differences between home and conventional students is a crucial area for future research.

Reasons for higher social skill scores among the home educating group are unclear. This study allows for speculation as to the primary variables that resulted in the social score differences between the groups. Variables such as church attendance, and lower teacher-child ratio may have been primary factors leading to higher social skills. However, an examination of the demographic data suggests a high level of similarity between the groups. Moreover, home education allows for a consistent and high amount of parent-to-child contact. An increased opportunity for parental feedback for appropriate and inappropriate social behavior may be a primary variable influencing these results. Research in this area, however, continues to be infrequent and varied and leaves a high level of inference for why differences occur between home and conventional educating groups. More research is needed to further explore the socialization domain as well as the specific variables involved that affect the development of home schooled children’s social skills.
Overall, the data continue to support the notion that home schooled children are not socially deprived, as the evidence indicates the development of appropriate social skills, and suggests home schooled children are reaping the benefits socially from an education at home.

Aiex, Nola K. (1994). Home schooling and socialization of children (Report No.CS 508 666). ERIC Clearinghouse on Reading, English, and Communication, Bloomington, IN. (ERIC Document Reproduction Service No. ED 372 460)
Achenbach, Thomas, McConaughy, Stephanie, & Howell, Catherine (1987). Child/Adolescent behavioral and emotional problems: Implications of cross-informant correlations for situational specificity. Psychological Bulletin, 101, 213-232.
Benes, Kathryn M. (1995). Review of the Social Skills Rating System. In J. C. Conoley, J. C. Impara, & L. L. Murphy (Eds.), The twelfth mental measurements yearbook (pp. 965-967). Lincoln, NB: University of Nebraska Press.
Carson (Allie-Carson), Jayn. (1990). Structure and interaction patterns of home school families. Home School Researcher 6(3), 11-18.
Chatham-Carpenter, April. (1994). Home vs. public schoolers’ relationships: Differences in social networks. University of Northern Iowa, Department of Communication Studies. (ERIC Document Reproduction Service No. ED 361 784)
Cizek, Gregory J. (1993). The mismeasure of home schooling effectiveness: A commentary. Home School Researcher, 9(3), 1-4.
Common, R. W., & MacMullen, M. (1986). Home schooling: A growing movement. Education Canada , 4-7.
Conners, C. Keith (1990). Manual for Conners’ Rating Scales. North Tonawanda, NY: Multi-Health Systems.
Dahm, Leslie (1996, October). Education at home with help from school. Educational Leadership, 68-71.
Delahooke, Mona Maarse. (1986). Home educated children’s social/emotional adjustment and academic achievement: A comparative study. Unpublished doctoral dissertation, California School of Professional Psychology, Los Angeles, CA. Dissertation Abstracts International (University Microfilms No. 8608759).
Eder, Rebecca A. (1997). The overemphasis on self-esteem series: Week of the young child: [Five Star Lift Edition]. St. Louis Post-Dispatch [On-line], B, 7:2. Available: Proquest Direct.
Elliot, Stephen N., Busse, R.T., & Gresham, Frank M. (1993). Behavior rating scales: Issues of use and development. School Psychology Review, 22, 313-321.
Frost, Gene (1988). The academic success of students in home schooling. Illinois School Research and Development, 24 (3), 111-117.
Gresham, Frank M., & Elliott, Stephen N. (1984). Assessment and classification of children’s social skills: A review of methods and issues. School Psychology Review, 13, 292-301.
Gresham, Frank M., & Elliott, Stephen N. (1990). Social skills rating system: Manual. Circle Pines, MN: American Guidance Service.
Guterson, D. (1990). When schools fail children. In J. W. Noll (Ed.), Taking sides: Clashing views on controversial educational issues (6th ed., pp. 144-155). Guilford, CT: Dushkin.
Hartup, Willard W. (1996). The company they keep: Friendships and their developmental significance. Child Development 67, 1-13.
Hedin, Norma S. (1991). Self-concept of Baptist children in three educational settings. Home School Researcher, 7(3), 1-5.
Holtrop, Stephen D. (1996, October). Individualization starts at home. Educational Leadership, 74-76.
Kelly, Steven W. (1991). Socialization of home schooled children: A self-concept study. Home School Researcher, 7(4), 1-12.
Kitchen, Paul (1991). Socialization of home school children versus conventional school children. Home School Researcher, 7(3), 7-13.
Klicka, Christopher J. (1993). The right choice: The incredible failure of public education and the rising hope of home schooling: An academic, historical practical, and legal perspective (3rd ed.). Gresham, OR: Noble.
Knowles, Gary J. (1988). Introduction: The context of home schooling in the United States. Education And Urban Society, 21(1), 5-15.
Lines, Patricia M. (1995). Home Schooling. ERIC Digest, Number 95. Eugene, OR: ERIC Clearinghouse on Educational Management. (ERIC Document Reproduction Service No. ED381849).
Lines, Patricia M. (1996, October). Home schooling comes of age. Educational Leadership, 63-67.
Mayberry, Maralee (1989). Home-based education in the United States: Demographics, motivations and educational implications. Educational Review, 41(2), 171-180.
Montgomery, Linda (1989). The effect of home schooling on the leadership skills of home schooled students. Home School Researcher, 5(1), 1-10.
Murray, Bridget (1996, December). Home schools: How do they affect children? The APA Monitor, 27 (12), pp. 1, 43.
Ray, Brian D. (1990). A nationwide study of home education (Report No. CG 026 205). Home School Legal Defense Association, Paeonian Springs, VA. (ERIC Document Reproduction Service No. ED 381 725)
Ray, Brian D. (1992). Marching to the beat of their own drum! A profile of home education research. Paeonian Springs, VA. The Home School Legal Defense Association.
Ray, Brian D. (1996). Home education research fact sheet IIIb. (Available from the National Home Education Research Institute, P.O. Box 13939, Salem OR. 97309.)
Ray, Brian D. (1997). Strengths of their own: Home schoolers across America: Academic achievement, family characteristics, and longitudinal traits. Salem, OR. NHERI.
Ray, Brian D. (1999). Home schooling on the threshold. A survey of research at the dawn of the new millennium. Salem, OR: NHERI.
Ray, Brian D. (2002). A quick reference worldwide guide to homeschooling: Facts and stats on the benefits of home school, 2002, 2003. Nashville, TN: Broadman & Holman.
Ritter, B. (Reporter). (1997, October 8). Good Morning America. New York: American Broadcasting Company.
Rubin, Kenneth H., Lynch, D., Coplan, Robert, Rose-Krasnor, Linda, & Booth, Cathryn L. (1994). “Birds of a feather…”: Behavioral concordances and preferential personal attraction in children. Child Development, 65, 1778-1785.
Schweigert, W. A. (1994). Research methods & statistics for psychology. Pacific Grove, CA: Brooks/Cole.
Shyers, Larry E. (1992). A comparison of social adjustment between home and traditionally schooled students. Home School Researcher, 8(3), 1-8.
Smedley, Thomas C. (1992). Socialization of home school children. Home School Researcher, 8(3), 9-16.
Stough, Lee (1992). Social and emotional status of home schooled children and conventionally schooled children in West Virginia. Master’s thesis, University of West Virginia. (ERIC Document Reproduction Service No. ED 353 079)
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, 11(3), 1-6.
Wartes, Jon (1987). Report from the 1986 homeschool testing and other descriptive information about Washington’s homeschoolers: A summary. Home School Researcher, 3(1), 1-4.
Authors’ note: This manuscript is based on the first author’s doctoral dissertation at Alfred University, Alfred, New York. We are grateful to committee members, Drs. Eugene A. Lovelace, Mark Fugate, and James F. Curl for their assistance and thoughtful suggestions. Correspondence should be addressed to Timothy Z. Keith, Department of Educational Psychology, 504 SZB, 1 University Station, D5800, The University of Texas at Austin, Austin TX 78712, ph. 512-471-0287,

1 reply

Trackbacks & Pingbacks

  1. […] Another study applied a Social Skills Rating System to evaluate thirty-four pairs of homeschooled and traditionally schooled children between the ages of five and eighteen. The researchers found that homeschooled children achieved higher scores on this scale than conventionally educated students. […]

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply