HOME VS. PUBLIC SCHOOLERS: DIFFERING SOCIAL OPPORTUNITIES

April Chatham-Carpenter
University of Northern Iowa
Department of Communication Studies
Cedar Falls, Iowa 50614 0357

Keywords: Homeschooling, home schooling, home education, socialization.
It has been assumed by proponents and opponents of home schooling alike that the home schooled child, who spends little time in institutional school‑related activities with peers, encounters different types of opportunities for interaction with adults, peers, and other children than does the traditionally schooled child, who spends up to eight hours a day at school with peers (cf. Maarse‑Delahooke, 1986). Recently, researchers have begun investigating the effects of home schooling on the personal and social lives of home schooled children, focusing on issues such as self‑concept (Hedin, 1991; Kelley, 1991; Kitchen, 1991; Stough, 1992; Taylor, 1986), emotional adjustment and social maturity issues (Maarse‑Delahooke, 1986; Shyers, 1992; Smedley, 1992), family atmosphere (Allie‑Carson, 1990), employment opportunities (Webb, 1989), and leadership skills (Montgomery, 1989). The few significant differences found between home schoolers and their traditionally‑schooled peers in these studies indicate that home schoolers have less behavior problems (Shyers, 1992), enact more adaptive behavior (Smedley, 1992), and have comparable or higher than the norm self‑concept scores (cf. Kelley, 1991) than do conventionally schooled children.
However, we still do not know much about the opportunities for social interaction of home schoolers (vs. traditionally schooled children), especially as they move into the crucial age period of adolescence, which is often associated with an increased incidence of peer interaction for the typical adolescent. The evidence which has been reported is primarily anecdotal in nature (e.g., Welch, 1984), or based on parental estimates of the children’s interaction (e.g., Bliss, 1989; Gustafson, 1988; Gladin, 1987; Schemmer, 1985).
The few studies of social opportunities of home schoolers illustrate that some, but not all, home schooling families do participate in interactions and activities with peers outside of the home. For example, respondents in Wartes’ (1987) study of 219 Washington State home schoolers indicated a median of 20‑29 hours/month for (1) contact with age peers, (2) contact with non‑age peers outside the immediate family, and (3) participation in organized community activities. In Groover and Endsley’s (1988) survey of 70 home schooling and 20 public schooling families with children in 1st to 6th grades, parents from both schooling groups reported their children had between 6‑10 friends outside of the home, with whom they spent time 1‑2 times a week. The parents also reported the children belonged to 3‑4 non‑family groups, with whom they participated one or more times a week. Bliss (1989) asked 70 home schooling families in Southwestern Michigan the question “do/does your child(ren) participate in any group activities with others their age.”  She found that 93% (n = 65) said “yes,” and only 7% (n = 5) said “no.”  However, Greene (1984), in surveying 88 home schooling families from Alaska, found that only 51% of the families answered “yes” to the question “do you participate in community activities such as clubs, sports, or large groups with others in your age group.”  The remaining 41 participants were asked “do you have friends nearby with whom you spend time,” and only 63% of these participants said “yes.”
There have been studies of the amount of peer interaction for traditionally schooled adolescents (e.g., Blyth, Hill, & Thiel, 1982; Crockett, Losoff, & Petersen, 1984; Montemayor & Van Komen, 1980). For example, Bo (1990) found that the majority of the contacts listed by her 15‑16 year old subjects were between the ages of 13 and 17 (cf. Cochran & Bo, 1989). Montemayor and Van Koman (1980) found, in their naturalistic study of the age‑segregation of 403 adolescents’ interactions, that adolescents most often were seen with peer friends, next with adults, and least of all with children. Bo (1990) found that approximately one‑half of each subject’s (174 Norwegian 15- and 16-year-olds) network contacts were interacted with by the subjects at least 4 (or more) days per week.
However, in contrast to traditionally‑schooled adolescents, there have been no empirical studies of home schoolers’ number of peer friends, amount of contact with different age groups (peers vs. adults vs. younger children), or the closeness of relationships. Therefore, this study poses three research questions. The first research question asks “is there a significant difference in the size of home and public schooled subjects’ social networks.”  In particular, “are there significant differences between home and public schooled subjects in the number of contacts they have that are younger than, the same‑age as, and/or older than themselves?” To investigate the amount of social interaction of home (vs. public) schoolers, the second research question asks, “are there significant differences between home and public schooled subjects for frequency of interaction with contacts?” To begin investigating the quality of home (vs. public) schoolers’ relationships, the final research question asks, “are there significant differences between home and public subjects on reported closeness of relationships overall?”

                               Methods

Subjects
The subject pool consisted of 21 home schooled children and 20 public schooled children, ages 12‑18, from central Oklahoma.¹ Twenty‑seven home schooling families (of 43 families contacted) agreed to participate in the study; a total of 19 families completed the study. (Two of these 19 families decided to have both a son and a daughter participate in the study; therefore, there was a total of 21 home schooling subjects.)  The mean age at which home schooling began for the subjects in the study was 10.48, with a mode of 10. The home subjects were home schooled for a mean of 4.10 years, and the modal number of years home schooled was 5 years.
Public schooling subjects, with demographic characteristics (e.g., socio‑economic status, residential location, church affiliation) similar to home schooling subjects, were obtained through names solicited from the home schooling families, as well as through a high school public schooling teacher in the central Oklahoma area. A deliberate attempt was made to obtain names of two‑parent public schooling families who regularly attended church, since all of the home schooling families in the study had two parents present in the home and were actively involved in church or in another religious body. Twenty‑eight families (of 55 families contacted) agreed to participate in the study, and 20 families completed the study.
The mean age for the children in the home sample was 14.48, and the mean age for the public sample was 14.85. There were 9 male subjects in the home sample, and there were 8 males in the public sample. There were 12 female subjects in the home sample and 12 in the public sample.
All of the families participating in the study were Caucasian in race, except for one Afro‑American home schooling family. Two subjects (one Home and one Public) were living with a step‑dad and their mother. (For more information regarding the subject pool, see Chatham, 1991.)

Procedures for Collecting Data
To gather information about the subjects’ social contacts, each subject was asked to keep a record of his/her interactions (interactions lasting at least 2 minutes or more) over a month’s period of time (i.e., who s/he talked to and what s/he talked about to each person recorded).² Data collection started in April and May for the subjects. After the month was completed, each subject was asked to have his/her parents and siblings review the list and add any additional person with whom the subject usually interacted at least once a month. Phone calls to the subjects served as a reminder for the project.
After the network list was constructed, follow‑up survey forms (see shortened form in the Appendix; cf. Chatham, 1991, for complete form) were given to the subject to gain information about the subject’s social contacts on the network list. These forms were used to gather information
regarding the subject’s relationships.
The score or value for the perceived level of closeness with a given person is the mean of the items B ‑ E (cf. Appendix) divided by four. Blyth (1984, as cited by Blyth & Traeger, 1988) found that across all types of relationships, the reliability coefficient alpha for the four‑item scale was .82. The scale has been found to discriminate equally well for both males and females and for a wide variety of types of relationships, with all reliability coefficient alphas over .77. Blyth and Traeger (1988) report that the scale has validly discriminated between relationships which are assumed to be at different intimacy levels (e.g., mother vs. extended family member).

Results

Research Question One: Size and Ages of Network
Is there a significant difference in the size of home and public schooled subjects’ social networks? The 21 Home schooled subjects made a total of 1035 contacts, with a mean of 49.29 contacts per subject; the range in contacts was from 7 to 150. The Public schooled sample of 20 subjects made 1125 contacts, with a mean of 56.25 contacts/subject; the range in contacts was from 18 to 154. A t‑test (unequal variances solution) found the difference to not be statistically significant, t (38.8) = ‑0.6319, p < .5312.
Are there significant differences between home and public schooled subjects in the number of contacts they have who are younger than, the same‑age as, and/or older than themselves? Table 1 reports the number of contacts at each age group. The categories in Table 1 were defined as follows: (1) Younger = 2+ years younger than the subject, (2) Peer = within 2 years of the subject’s age, (3) Older = 2+ years older than the subject’s age.
The repeated measures ANOVA indicated a significant difference between the different levels of age overall, F (2, 78) = 32.96, p < .0001 (Greenhouse‑Geisser³), and a significant interaction effect between age category and schooling type, F (2, 78) = 22.71, p < .0001 (Greenhouse‑Geisser). The schooling type main effect was not significant, F (1, 39) = 0.33, p < .5671.
Two‑correlated sample t‑tests were used for follow‑up multiple comparisons (cf. Maxwell, 1980). For home schooled subjects, there was a significant difference between numbers of older and peer contacts reported, t (20) = 5.06, p < .0001, as well as between numbers of older and younger contacts, t (20) = 4.28, p < .0004; however, for public schooled subjects, there was a significant difference between numbers of older and peer contacts, t (19) = ‑3.96, p < .0008, a significant difference between numbers of older and younger contacts, t (19) = 7.15, p < .0001, and a significant difference between the numbers of peer and younger contacts, t (19) = 7.79, p < .0001.

Table 1, 2.6″ tall

Table 1. Contact ages in relation to subject ages: frequencies, percentages, and mean number of contacts per subject.

To investigate further where significant differences between home and public schooled subjects were in the different age groups, I used the Games‑Howell solution for unequal n’s (cf. Games & Howell, 1976; Toothaker, 1991) to compute the cell mean multiple comparisons. The mean numbers of peer contacts for home and public schooled subjects was significantly different, t (29.7) = ‑4.30, p < .01, and the mean numbers of younger contacts for home and public schooled subjects were significantly different, t (31.4) = 2.13, p < .05. However, the two schooling groups were not significantly different on the mean number of older contacts.
Research Question Two: Frequency of Interaction
Are there significant differences between home and public schooled subjects for frequency of interaction with contacts? The frequencies and percentages of responses at each level of interaction frequency (cf. Appendix, Item A) are listed in Table 2.

Table 2, 4.0″ tall

Table 2. Frequency of interaction individual categories: frequencies and percentages of responses.

Overall, the mode and median of frequency of interaction for the Home sample was 4 (i.e., 1‑2 times a week). For the Public sample, the mode frequency of interaction was also 4 (i.e., 1‑2 times a week), but the median was 5 (3‑6 times a week).
Differences between home and public schooled subjects on interaction frequency with contacts can be clearly seen in the analysis done with frequency of interaction using a grouping of categories, as seen in Table 3 (i.e., less than once per month, monthly, weekly, and daily).
The repeated measures ANOVA on the above grouping indicated a significant main effect for frequency of interaction, F (3, 117) = 22.57, p < .0001 (Greenhouse‑Geisser), and a significant interaction effect between schooling type and interaction frequency, F (3, 117) = 11.45, p < .0001 (Greenhouse‑Geisser).
Follow‑up multiple comparisons on the cell means indicated that there were significant differences between home and public schooled subjects on the following levels of frequency of interaction (using the Games‑Howell procedure for unequal n’s): (a) less than once a month, t (24.3) = 3.239, p < .01, with home schoolers reporting significantly more contacts than public schoolers; (b) monthly, t (25.0) = 2.21, p < .05, with home schoolers reporting significantly more contacts than public schoolers; and (c) daily, t (20.7) = ‑6.39, p < .01, with public schoolers reporting significantly more contacts than home schoolers.

Table 3, 2.5″ tall

Table 3. Frequency of interaction grouped categories: frequencies, percentages, and mean number of contacts per subject at different levels.

Research Question Three: Closeness‑Intimacy with Contacts
Are there significant differences between home and public schooled subjects for perceived closeness of relationships? Items B through E in the Appendix were used to gain information about a subject’s level of closeness or intimacy with each contact person. There were three possible answers for subjects to use, i.e., (1) “not at all,” (2) “some,” and (3) “a lot”; however, in a few cases, subjects responded with a “0” or “not applicable” response. (For Home, there were 31 out of a possible 4,140 “0” responses; for Public, there were 30 out of a possible 4,500 “0” responses.)  Two runs of the data were done ‑‑ one in which the “0” responses were counted as “0’s,” and one in which the “0’s” were counted as “1’s.”
Overall means for the four items summed together were as follows (“0’s” as “1’s” in parentheses): Home = 1.90 (1.91) and Public = 2.12 (2.13). Each item and the mean response for the item, as well as t‑test results for each item, are reported in Table 4 (with “0’s” as “1’s” means in parentheses).
As seen in Table 4, on the first two items the Public sample indicated they experienced significantly more closeness with their relational network contacts than the Home sample did, with specific results (0’s as 0’s analysis) as follows: (a) go to person for advice, t (33.9) = ‑3.24, p < .0027 and (b) share inner feelings with person, t (37.0) = ‑2.22, p < .0330. T‑test results also indicated an overall significant difference on all four closeness items averaged together (overall closeness measure) on both runs of the data, again with the Public sample scoring significantly higher on closeness than the Home sample, t (35.5) = ‑2.19, p < .0350 (0’s as 0’s analysis).

                             Discussion

In describing the social opportunities of the 21 home and 20 public schooled adolescents in this study, the home schoolers did have differing opportunities for interaction with contacts than the public schoolers did in the areas of contact ages, frequency of interaction with contacts, and closeness.
The public schoolers had a slightly, although not significantly, larger network of contacts than the home schoolers. The public schoolers had significantly more peer (within 2 years of subject’s age) contacts than the home schoolers; however, the home schoolers had significantly more younger (2+ years younger than subject) contacts than public schoolers. Overall, the home schoolers had significantly more older contacts than either peer or younger contacts; however, the public schoolers had significantly more peer contacts than older or younger contacts, as well as significantly more older than younger contacts. The finding that home schoolers have more older contacts than peer contacts is not surprising, since they do not have the traditional schooling place to provide them with potential peer contacts. Neither is the finding surprising that public schoolers have more peer contacts than the home schoolers, as the public schoolers spend so much more time at a conventional school, with numerous potential peer contacts.
These results for public schoolers are consistent with studies by Bo (1990) and Montemayor and Van Komen (1980), who found that adolescents interacted with more peers than adults or children; however, the results for the home schoolers contradicted these

Table 4, 4.5″ tall

Note: Some items were shortened to fit the specifications of the table. Means in parentheses are from the 0’s as 1’s analysis. Scale values: 0 = not applicable; 1 = not at all; 2 = some; 3 = a lot.

Table 4. Closeness of relationships: mean closeness scores per subject for contacts, and t-test results.

findings, as the home schoolers reported more older than peer contacts, demonstrating that the home schooling process does have the potential to restructure a child’s social world, in providing the home schooler more mixed‑age than same‑age interaction and socialization opportunities. We now need to gather more information on how these different social opportunities affect home and public schooled children in the future (e.g., in dealing with people from many different age groups; cf. Webb, 1989).
There were significant differences between levels of interaction frequency with all the contacts. Overall, the public schoolers interacted with more contacts more often. Specifically, the public schoolers talked to more contacts on a daily basis than the home schoolers, while the home schoolers reported talking to more contacts less than once a month and on a monthly basis than the public schoolers did. Home schoolers reported talking to more of their contacts weekly than they did daily, monthly, or less than once a month, while the public schoolers reported talking to their contacts weekly and daily at approximately the same level, both of which were greater than monthly or less than once a month. These results demonstrate that the schooling process does indeed make a difference in how often an adolescent interacts with his/her contacts.
More research and valid data are needed on interaction frequency in an attempt to distinguish between “surface” and “deep” interactions; the way this study measured interaction frequency is problematic in that it does not clearly answer the question of how much interaction is taking place within a relationship, nor does it consider the quality of the interaction that is taking place. For example, there is a qualitative difference between a 2‑minute and a 30‑minute interaction, but item A on the survey did not take this into account.
Regarding closeness to relationships with contacts, the public schoolers reported being more close to contacts than the home schoolers did overall. On specific closeness items, the public schoolers reported significantly higher scores than the home schoolers did for contacts on the items “how much do you go to this person for advice,” and “how much do you share your inner feelings with the person.”  As reflected by the above significant closeness items, the public schoolers report more opportunities to talk to and share with their contacts than the home schoolers do, perhaps as a function of their greater opportunities for more frequent interaction with their contacts. Perhaps the contacts just are not around when the home schoolers need them, as opposed to public schoolers who, when they have a problem during the school day, have available numerous peers with whom they can talk.
A question for future research is, “for home and public schoolers, does frequency of interaction and/or length of acquaintance in a relationship interact with schooling type on mean closeness scores?”  Perhaps one reason home schoolers do not report feeling as close to their contacts is because they do not spend as much time with them. Do home schoolers (vs. public schoolers) feel closer to the contacts with whom they spend more time? Is there a significant difference between home and public schoolers on mean closeness scores for the relationships with which they interact on a weekly and/or daily basis? Are different types of interactions present for home and public schoolers in their close vs. non‑close relationships?
Future research should also study longitudinally the relationships of the home schooler (vs. the public schooler), starting in childhood, when the home schooling process begins. Further, what happens to the social networks when home schooled children are placed into a traditional school or taken out of conventional school to be home schooled? Studies need to investigate the difference between the social networks of children who have been totally home schooled and children who have been both traditionally and home schooled.
I think demographic factors are important to consider as well in future studies (cf. Chatham, 1991), as they can play a role in a child being or not being “at risk” from the process of home schooling. For example, in this study there was a range of 7 to 150 contacts identified by the home schooling subjects. The home schooled female who only listed 7 contacts lived in a rural area, was an early adolescent (was not able to drive), and attended a smaller community church (the family was actually in transition between churches). The home schooled male who listed the most contacts of all home schoolers (i.e., 150) lived in a metropolitan city neighborhood across the street from a YMCA, attended a large traditional church (at which he worked part‑time), and was old enough to drive. What role did place of residence, church youth group size, and age of subject, for example, play in providing contacts for these youth? An interesting ethnographic project would be to follow potential “at risk” (in terms of low social opportunities) home and public schooling families over time, to understand more about the life of these families C their thinking and communication behaviors. Mapping their ecological “systems” (cf. Bronfenbrenner, 1979) could be enlightening.
Future research should aim for a more broad‑based sample of home schoolers from different backgrounds, reasons, and methods of home schooling, as well as a sampling of home and public school networks during different parts of the school year. This study only generalizes to the religious or ideological home schooling family, and does not tap into the pedagogical or New Age home schooler (cf. Mayberry & Knowles, 1989; Van Galen, 1986). In addition, perhaps there are seasonal variations in size and diversity of social networks for both home and public schoolers. What differences are there between the networks of public schoolers and home schoolers, for example, in the summer months? Further, the small overall sample size also creates a problem for generalizing this study past this particular sample of home and public schoolers.
However, this study did add to the present home schooling literature both methodologically and conceptually. Methodologically, it went beyond the use of the common parental (i.e., mother) self‑reports (cf. Groover & Endsley, 1988), using child self‑reports to explore the social networks of an age group (i.e., adolescence) that has not been investigated a great deal by home school researchers. It is the first study of its kind to use social networks to investigate the actual numbers of contacts that home schoolers have and to begin describing the characteristics of the relationships of home schoolers. This study used a similar comparison group of public schoolers against which to compare the home schoolers’ reports, making it possible to start drawing basic conclusions about the role of the schooling process in social networks of adolescents.
In conclusion, this study found that home schoolers are not “at risk” socially (as compared to a similar group of public schoolers), in terms of the total numbers of people with whom they interact, although this varies from family to family (e.g., the home schooled girl above who reported only seven total contacts), showing the importance of the family in providing opportunities for social interaction outside the home for home schoolers, especially for early adolescents. However, this study demonstrates that the home schooling process does indeed affect the nature of the relationships experienced in adolescence for home schoolers. Are the home schoolers “at risk” socially because of this? The home schoolers in this study reported a mean of 11 peer contacts each, while the public schoolers reported an average of 30 peer contacts. It is not that the home schoolers are having no peer contacts, but just less peer contacts than the public schoolers. With home schoolers reporting less peer contacts than public schoolers and less closeness to their contacts overall, perhaps some home schooling parents would be happy with these results, saying that their adolescent is not experiencing “peer dependency” (cf. Harris, 1984). But one must ask if these contacts are close enough for the home schoolers to be developing good friendships, as well as a sense of companionship with their peers and other contacts. There has been no research to date indicating that a certain amount of contact or number of friends is needed to learn communication skills or develop a good friendship, only that a child/adolescent needs to experience both same‑age and mixed‑age interactions (cf. Hartup, 1979; Hunter, 1984; Sebald, 1986; Valiant, 1983), which was found to be true of both the home and public schoolers in this study. What kind of interaction is necessary to learn these social skills is the crucial question which needs to be addressed in future research.

                             References

Allie‑Carson, J. (1990). Structure and interaction patterns of home school families. Home School Researcher, 6(3), 11‑18.
Bliss, B. A. (1989). Home education: A look at current practices. East Lansing, MI: Michigan State University. (ERIC Document Reproduction Service No. ED 304 233)
Blyth, D. A. (1984). The intimate worlds of early adolescents: Using social mapping techniques for exploration and prediction. Symposium presented at the biennial meeting of the Midwestern Society for Research in Life‑Span Development, Akron, OH.
Blyth, D. A., Hill, J. P., & Thiel, K. S. (1982). Early adolescents’ significant others: Grade and gender differences in perceived relationships with familial and non‑familial adults and young people. Journal of Youth and Adolescence, 11(6), 425‑450.
Blyth, D. A., & Traeger, C. (1988). Adolescent self‑esteem and perceived relationships with parents and peers. In S. Salzinger, J. Antrobus, & M. Hammer (Eds.), Social networks of children, adolescents, and college students (p. 171‑194). Hillsdale, NJ: Lawrence Erlbaum Associates.
Bo, I. (1990). The significant people in the social networks of adolescents. In K. Hurrelmann & U. Engel (Eds.), The social world of adolescents: International perspectives (p. 141‑165). New York: Walter de Gruyter Co.
Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press.
Cauce, A. M. (1986). Social networks and social competence: Exploring the effects of early adolescent friendships. American Journal of Community Psychology, 14, 607‑628.
Chatham, A. D. (1991). Home vs. public schooling: What about relationships in adolescence? (Doctoral dissertation, University of Oklahoma, Norman, 1991). Dissertation Abstracts International, 52(6), 2100A.
Cochran, M., & Bo, I. (1989). The social networks, family involvement, and pro‑ and antisocial behavior of adolescent males in Norway. Journal of Youth and Adolescence, 18(4), 377‑398.
Cochran, M. M., & Brassard, J. A. (1979). Child development and personal social networks. Child Development, 50, 601‑616.
Crockett, L., Losoff, M., & Petersen, A. C. (1984). Perceptions of the peer group and friendship in early adolescence. Journal of Early Adolescence, 4(2), 155‑181.
Games, P. A., & Howell, J. F. (1976). Pairwise multiple comparison procedures with unequal n’s and/or variances. Journal of Educational Statistics, 1, 113‑125.
Gladin, E. W. (1987). Home education: Characteristics of its families and schools. Unpublished doctoral dissertation, Bob Jones University.
Greene, S. S. (1984). Home study in Alaska: A profile of K‑12 students in the Alaska Centralized Correspondence Study program. (ERIC Document Reproduction Service No. ED 255 494)
Groover, S. V., and Endsley, R. C. (1988). Family environment and attitudes of home schoolers and non‑homeschoolers. Athens, GA: University of Georgia. (ERIC Document Reproduction Service No. ED 323 027)
Gustafson, S. K. (1988). A study of home schooling: Parental motivation and goals. Home School Researcher, 4(2), 4‑12.
Harris, G. (1984). Problem and cause of peer dependency revealed. In S. Welch (Ed.), Socialization (p. 8‑9). Portland, OR: Christian Home Schools.
Hartup, W. W. (1979). The social worlds of childhood. American Psychologist, 34, 944‑950.
Hedin, N. S. (1991). Self‑concept of Baptist children in three educational settings. Home School Researcher, 7(3), 1‑5.
Hunter, F. T. (1984). Socializing procedures in parent‑child and friendship relations during adolescence. Developmental Psychology, 20(6), 1092‑1099.
Kelley, S. W. (1991). Socialization of home schooled children: A self‑concept study. Home School Researcher, 7(4), 1‑12.
Kitchen, P. (1991). Socialization of home school children versus conventional school children. Home School Researcher, 7(3), 7‑13.
Maarse‑Delahooke, M. (1986). Home educated children’s social/emotional adjustment and academic achievement: A comparative study (Doctoral dissertation, California School of Professional Psychology, Los Angeles, 1986). Dissertation Abstracts International, 47, 475A.
Maxwell, S. E. (1980). Pairwise multiple comparisons in repeated measures designs. Journal of Educational Statistics, 5, 269‑287.
Mayberry, M., and Knowles, J. G. (1989). Family unit objectives of parents who teach their children: Ideological and pedagogical orientations to home schooling. The Urban Review, 21(4), 209‑225.
Montemayor, R., & Van Komen, R. (1980). Age segregation of adolescents in and out of school. Journal of Youth and Adolescence, 9(5), 371‑381.
Montgomery, L. (1989). The effect of home schooling on the leadership skills of home schooled students. Home School Researcher, 5(1), 1‑10.
Salzinger, S., Hammer, M., & Antrobus, J. (1988). From crib to college: An overview of studies of the social networks of children, adolescents, and college students. In S. Salzinger, J. Antrobus, & M. Hammer (Eds.), Social networks of children, adolescents, and college students (p. 1‑16). Hillsdale, NJ: Lawrence Erlbaum Associates.
SAS Institute Inc. (1985). SAS User’s Guide: Statistics. Version 5 Edition. Cary, NC: SAS Institute Inc.
Schemmer, B. A. S. (1985). Case studies of four families engaged in home education (Doctoral dissertation, Ball State University, 1985). Dissertation Abstracts International, 46, 2560‑265lA.
Sebald, H. (1986). Adolescents’ shifting orientation toward parents and peers: A curvilinear trend over recent decades. Journal of Marriage and the Family, 48(1), 5‑13.
Shyers, L. E. (1992). A comparison of social adjustment between home and traditionally schooled children. Home School Researcher, 8(3), 1‑8.
Smedley, T. C. (1992). Socialization of home school children. Home School Researcher, 8(3), 9‑16.
Stohl, C. (1989). Children’s social networks and the development of communicative competence. In J. Nussbaum (Ed.), Life‑span communication: Normative processes (p. 53‑77). Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.
Stough, L. (1992). Social and emotional states of home schooled children and conventionally schooled children in West Virginia. Unpublished Master’s thesis, University of West Virginia. (ERIC Document Reproduction Service No. ED 353 079)
Taylor, J. W. (1986). Self‑concept in home‑schooling children. Home School Researcher, 2(2), 1‑3.
Tietjen, A. M. (1982). The social networks of preadolescent children in Sweden. International Journal of Behavioral Development, 5, 111-130.
Tietjen, A. M. (1989). The ecology of children’s social support networks. In D. Belle (Ed.), Children’s social network and social supports (p. 37‑69). New York: John Wiley & Sons.
Toothaker, L. E. (1972). Adjustment of ANOVA F‑tests for violation of assumptions for repeated measures designs. Paper presented at the meeting of the Southwestern Psychological Association.
Toothaker, L. E. (1991). Multiple comparisons for researchers. Newbury Park, CA: SAGE Publications, Inc.
Valiant, G. L. (1983). Adolescents, parents, and peers: What is one with or without the other? Journal of Adolescence, 6(2), 131‑144.
Van Galen, J. (1986). Schooling in private: A study of home education (Doctoral dissertation, University of North Carolina, Chapel Hill, 1986). Dissertation Abstracts International, 47(5), 1683A.
Wartes, J. (1987). Report from the 1986 homeschool testing and other descriptive information about Washington’s homeschoolers: A summary. Home School Researcher, 3(1), 1‑4.
Webb, J. (1989). The outcomes of home‑based education: Employment and other issues. Educational Review, 41(2), 121‑133.
Welch, S. (1984). Socialization experience shared through survey. In S. Welch (Ed.), Socialization (p. 10‑11). Portland, OR: Christian Home Schools.

Appendix

Social Network Survey
NAME OF CONTACT
A. Frequency of interaction (at least two minutes or more)
1. less than once per month
2. 1 time per month
3. 2 ‑ 3 times/month
4. 1 ‑ 2 times/week
5. 3 ‑ 6 times/week
6. 1 time per day
7. 2 ‑ 4 times per day
8. 5+ times per day
For questions B ‑ E, please choose one of the following three answers, writing down the number of the answer you choose in the space provided underneath each question.
1. not at all    2. some    3. a lot
B. How much do you go to this person for advice?
C. How much do you share your inner feelings with this person?
D. How much does this person accept you no matter what you do?
E. How much does this person understand what you’re really like?

                              Endnotes

1. Adolescents were chosen for the study for both theoretical and pragmatic reasons C theoretically, since during adolescence, “peer dependency” especially becomes an issue of concern as there is a often a psychological “moving away” from home to establish one’s own identity, and pragmatically, because of adolescents’ ability to answer the questionnaires since I didn’t want to rely on parental self‑report data. Also, most socialization studies of home schoolers have been done with primary or middle‑school aged children, and I wanted to begin investigating another age group.
2. This study investigated the “social networks” of home vs. public schooled children, since a child’s social networks can be used to describe the relationships that a child (or adult) is experiencing (cf. Bo, 1990; Cauce, 1986; Cochran & Bo, 1989; Cochran & Brassard, 1979; Salzinger, Hammer, & Antrobus, 1988; Stohl, 1989; Tietjen, 1982, 1989). For this study, a child’s “social network” consisted of all of the people who interact on a regular basis with the child‑‑at least once a month.

3. Type III sums of squares in SAS (or “partial sums of squares,” SAS Institute Inc., 1985) were used throughout the study to deal with the problem of unequal n’s, and the degrees of freedom were corrected for violation of the sphericity assumption by using the Greenhouse‑Geisser epsilon (Toothaker, 1972).