Steven D. Smith Ph.D.
Dow Center
Hope College
Holland, Michigan 49423

Jilanne Bannink-Misiewicz
Hope College
Holland, Michigan 49423

Shelly Bareman
Hope College
Holland, Michigan 49423

Keywords: Homeschooling, home schooling, home education, comparison, motor skill.

A topic of growing interest within the educational research field concerns the outcomes of home schooling. A review of the literature examining home school education reveals intense interest in academic outcomes and social development with no emphasis on psychomotor development.
The research regarding cognitive development in home schoolers has revealed interesting results. Children who are home schooled have been reported to perform very well in academic settings. Researchers have reported that standardized test scores for home schoolers are above that of conventional school children (Ray, 1990). Knowles (1991) investigated adults that were previously home schooled and found that 43 percent had attended college. Additionally, Colfax and Colfax (1988) cite many examples of former home schoolers who attended and excelled in university settings. Although the empirical data does not clearly indicate that the home school environment is superior in academic development, the research clearly does indicate that many home school children do not lag behind their conventional school peers.
Opponents of home school education often cite lack of normal socialization as a concern (Harris & Fields, 1982). This criticism has led to a variety of investigations into the socialization of home school children. Moore and Moore (1990) concluded that children in their experience were “far better socialized” than children in conventional education settings. Taylor (1986) reported significantly higher self-concept scores in home schoolers when compared to conventional education children using the Piers-Harris Children’s Self-Concept Scale. However, Knowles (1991) suggests that with the difficulties associated with investigating social development and socialization, there is a paucity of research investigating differences in home school and conventional school children in this domain.
            An examination of the literature concerning the psychomotor domain of education in home schoolers reveals a complete lack of research on this topic. An updated and comprehensive annotated bibliography by Ray (1991) and intensive investigation (Ray, 1990) reveal no empirical investigations of the psychomotor domain in home school settings. This indicates that comparisons of home schoolers and conventionally educated children in the area of physical education does not exist.
            The purpose of this study was to compare the motor performance and perceived competence scores of home school and conventional school children in Michigan. A secondary purpose of this research was to compare the motor performance scores with national norms.


Subjects for this study included 56 home school children and 44 conventional school children from Central and Western Michigan. Home school children were recruited by distribution of flyers at the State of Michigan home school convention. Respondents were then contacted and scheduled for testing. Conventional school children were recruited from a summer physical activity session and were tested prior to any instruction. The home school subjects included 34 boys and 22 girls ranging in age from 6 to 10 years with a mean age of 7.96   years (SD 1.22 years). Conventional school children included 27 boys and 17 girls ranging in age from 6 to 9 years with a mean age of 7.50 years (SD 0.63 years).
Although an available sample was used, the groups were considered comparable based upon age and activity characteristics. It should be noted that though the standard deviation differs between groups, the differences do not represent skewness in the population.
Concerns of nonequivalence could arise over the conventional school group being selected from a group already enrolled in a physical activity program. This concern was addressed by testing the children prior to receiving instruction. Additionally, the home school group responded to a questionnaire regarding physical activity patterns. Many of the home school children were also enrolled in physical activity programs.

The Test of Gross Motor Development (TGMD) by Ulrich (1985) was used to assess qualitative aspects of motor performance. The TGMD yielded three scores relevant to this research: (a) a locomotor skills subtest score based on running, hopping, skipping, jumping, sliding, leaping, and galloping performances; (b) an object-control skills subtest score based on throwing, catching, kicking, striking, and dribbling performances; and (c) a standard score derived from combining the locomotor and object-control scores. The maximum score on the locomotor subscale is 26 points and on the object-control subscale 19 points, with a higher score representing a more mature level of skill performance. The maximum standard score is 150 with a mean score of 100 (SD 15.0).
The psychometric properties of the TGMD have been examined and found acceptable (Ulrich, 1985). A stratified quota sample of 909 subjects was used to represent the population of the United States for the characteristics of gender, race, community size, and geographic region. Stability reliability of the TGMD was evaluated by Ulrich and Wise (1984). The estimated variance components for the locomotor and object-control subtests were .54 and .49, respectively. The magnitude of the variance measures, expressed in percent of total variance, were 1% and 2%. The generalizability coefficients of this analysis for the locomotor subtest and the object-control subtest, were .96 and .97 respectively (Ulrich, 1985).
Inter-scorer reliability also has been established (Ulrich, 1985). The estimated variance components for scorers in the locomotor and object-control subtests were .53 and .81 respectively. The reliability of mastery decisions with the total composite score was examined and reported in terms of proportion of agreement at .89.
Content validity for the TGMD was established by having three content experts judge whether the skills selected represented the skills that are frequently taught to children in the early elementary grades. The results of the independent ratings were unanimous as representative (Ulrich, 1985).
Construct validity was established for the TGMD by testing the hypothesis that gross motor development would improve across age levels. The validity of the test was reported for the locomotor subtest at .81, object-control subtest at .84 and gross motor composite at .86 (Ulrich, 1985).
The Self-Perception Profile for Children (Verbal Scale) (Harter, 1982) and the Pictorial Scale of Perceived Competence and Social Acceptance for Young Children (Harter, Pike, Efron, Chao, & Bierer, 1983) were used to measure perceived competence. The Verbal Scale was administered to the eight, nine, and ten year old children. The Pictorial Scale was administered to the six and seven year old children. The use of two tests was necessary due to the age limitations of each test.
The Verbal Scale includes 36 items which measure global self-worth and 5 specific domains: (a) scholastic competence, (b) social acceptance, (c) athletic competence, (d) physical appearance, and (e) behavioral conduct. The format of the test requires the subjects to compare two statements and select the statement which is most like him or her. Further discrimination as to whether the statement is “really true” or “just sort of true” for the subjects must also be selected.
The Pictorial Scale includes 24 items designed to measure 4 domains: (a) cognitive competence, (b) peer acceptance, (c) physical competence, and (d) maternal acceptance. The subject is instructed to compare two pictures depicting children involved in various activities and indicate which child is most like him or her. As in the scales for older children, further discrimination as to whether this is “really true” of “just sort of true” for him or her is required.
The validity of the Verbal Scale is based on a sample of over 2400 third through ninth grade children from four states. The test-retest reliability for the instrument following a three-month interval ranged from .87 for the physical competence subscale to .70 for the general self-worth subscale (Harter, 1982).
Validity of the Pictorial Scale was established using a sample of over 250 children aged four through seven years. Total scale reliability was reported at .87. Internal consistency of each subscale combined with each of their respective factors ranged from .75 to .89 (Harter, Pike, Efron, Chao, & Bierer, 1983).

This research was a static group comparison design (Campbell & Stanley, 1963). Data were collected on the experimental group of home school children over a three-week period in selected central sites across Michigan. The control group data were collected during a two day testing period.
The tests were administered by the project director and two undergraduate research assistants specifically trained in test procedures. Upon arrival to the test site the subjects were administered the appropriate perceived competence instruments individually. The children were then grouped (three to four subjects per group) to complete the TGMD. The performances on the TGMD were videotaped for later evaluation. During the testing, parents responded to a short survey regarding physical activity patterns of their children.


Because of the lack of existing research regarding the motor performance ability of home school children, no previous hypotheses were solidly stated. The amount of time or emphasis in physical activity and physical education in home school settings has not been reported in the literature. Therefore, this study regarded the motor ability section of this project as purely investigative.
Similarly, the perceived competence of home school children has not been investigated. Though some research has been accomplished in self-concept (Taylor, 1986), results of using a multi-domain instrument measuring perceived competence have not been reported. For this reason the perceived competence portion of this project was also considered investigative.
It should also be noted that because of the investigative nature of this study, significant difference are reported using the alpha = .10 level. This deviation from the commonly accepted alpha = .05 level in social science research was done in order to be sensitive to any differences which may be present. The actual p values are reported so that the reader can gain more information.

Motor Performance
Results of the testing using the TGMD revealed significant differences between the home school group and conventional school  group. MANOVA analysis of motor performance scores revealed a significantly lower score for the home school group on the locomotor skill subtest when compared to the conventional school group, F (1,98) = 3.66 p < .059. No significant differences between groups were observed for the object-control subscale F (1,98) = 2.14 p < .147. Significant differences between groups also were discovered on the sum of standard scores which combined the two subscales. The home school group scored significantly lower on the sum of standards when compared to the conventional school group using ANOVA procedures, F (1,98) = 3.80, p < .054. Mean scores are reported in Table 1.

–This table is missing in this web page presentation; you may find the PDF format elsewhere on this web site with the table included—

Table 1. Standard scores on the Test of Gross Motor Development (TGMD).

Comparisons of the home school group to national norms reported in the test manual of the TGMD were only possible using a converted score (from the sum of standards) known as a quotient score. The Student’s t-test results indicated that the home school group scored significantly lower than national norms, t (963) = 10.60 p < .001. Comparisons with national norms also can be understood in terms of percentile rankings. The home school group performed at the 26th percentile on the locomotor subscale. Similar results were discovered on the object-control subscale with the home school group performing at the 25th percentile. This indicated that 74-75 percent of the children tested nationally scored at or above the home school group level of motor performance.

Perceived Competence
MANOVA results of the perceived competence testing revealed a variety of significant differences between home school and conventional school groups. The Pictorial Scale revealed lower perceived competence scores for the home school group when compared to the conventional school group in the cognitive domain, F (1,43) = 9.02 p < .004; and the maternal acceptance domain, F (1,43) = 6.91, p < .012. No significant differences between groups were revealed in the peer acceptance domain, F (1,43) = 2.31, p < .136; and the physical ability domain, F (1,43) = 2.21, p < .144. Means for the Pictorial Scale are reported in Table 2.

–This table is missing in this web page presentation; you may find the PDF format elsewhere on this web site with the table included—

Table 2. Pictorial Scale scores.

The Verbal Scale also revealed significantly lower scores in perceived competence for the home school group in a variety of domains when compared to the conventional school children. The home school group scored significantly lower in perceived scholastic competence, F (1,53) = 4.33, p < .042; social acceptance, F (1,53) = 5.34, p < .025; athletic competence, F (1,53) = 3.15, p < .082; and behavioral condition, F (1,53) = 11.13, p < .002. No significant differences were revealed between home school and conventional school groups in the physical appearance domain, F (1,53) = 1.79, p > .10; and general self worth domain, F (1,53) = 1.67, p > .10. Means for the Verbal Scale are reported in Table 3.


Fundamental Motor Skill Ability of Home Schooled and Conventional Schooled Children
The results clearly revealed that the home school group performed at a significantly lower

–This table is missing in this web page presentation; you may find the PDF format elsewhere on this web site with the table included—

Table 3. Verbal Scale scores.

level than the conventional school group in fundamental motor skill ability. The data revealed that the differences between groups were more marked in the area of locomotor skills than in object-control skills. The mean standard scores reported in Table 1 indicated that both the home school and conventional school groups scored higher in locomotor skills when compared to object control skills. This may indicate that both groups have experienced more instruction or practice in locomotor skills. It also should be noted that the locomotor activities tend to emerge more naturally without specific instruction when compared to the skills tested in the object-control subscale. This may explain the lower scores on the object-control subscale for both groups.
Results of parent surveys suggest that lower scores for the home school group in motor performance may be explained by the absence of formal instruction in physical education. None of the reports indicated specific instruction in fundamental motor skills. It appears that low intensity exercises were the main focus of any physical instruction or experience that may have occurred in the home school setting. Community sport experiences such as swimming lessons, baseball, and soccer also were reported by some of the parents. Such experiences should have impact on skill development. However, many youth sport programs emphasize actual game play over skill instruction. It is probable that the home school children tested revealed the ramifications of lack of instruction in fundamental motor skills during the early elementary years.
The results of performances of the Michigan conventional school group also merits consideration. Though this group performed at a higher level than the home school group on fundamental motor skills, the scores indicate a low level of performance when compared to national norms. The conventional school group performed at the 36th percentile. This indicates that 64 percent of the national population scored at or above this level of fundamental motor skill performance. This is consistent with other findings regarding the motor skill ability of elementary aged children in Michigan (Reuschlein & Haubenstricker, 1985). It is possible that the lack of required physical education in Michigan may explain this considerable low level of ability as measured by the TGMD.

Perceived Competence of Home School and Conventional School Children
The home school group perceived competence scores were significantly lower in many of the domains tested when compared to the conventional school group. The most interesting results relevant to the psychomotor domain occurred in the areas pertaining to physical abilities using the Pictorial Scale. The mean scores reveal that the home school group actually perceived themselves to be quite high in physical ability. This contrasts sharply to the results of the motor performance testing. This arouses questions regarding who they are comparing themselves with in terms of performance.
It is also interesting to note that the above findings did not occur for the older children on the Verbal Scale. Perceived athletic competence scores on the Verbal Scale were significantly lower for the home school group when compared to the conventional school group. This perceived lower ability on the part of home school children may indicate that at older ages, these children are perceiving their abilities with greater accuracy.
The perceived cognitive abilities of home school children also produced interesting results. Home school children perceived themselves to be lower in cognitive abilities (Pictorial Scale) and scholastic abilities (Verbal Scale) when compared to conventional school children. This contrasts sharply with the results of Taylor (1986) which indicated higher scores attributed to home school children. These differences may be explained by the use of different test instruments. Additionally, the children in Taylor’s (1986) study were older than the population in this study. Further study is warranted based on these opposite findings.
The lower perceived competence scores for home schoolers in peer acceptance (Pictorial Scale) and social acceptance (Verbal Scale) also contrast with Taylor’s (1986) findings which attributed higher scores to home schoolers in the category of popularity. Examinations of the two instruments reveal similar test question items. Mean scores between test instruments are not comparable. Explanation for these contrasting results may again be related to age differences or slight differences in instrumentation.
Delahooke (1986) also may offer insight into lower perceived peer acceptance scores for home school children. She found that home school children appeared to be less peer oriented than those in conventional school settings. If so, it is possible that the lower perceived scores for this population may be a reflection of less concern for peer acceptance. Further analysis with larger populations would be beneficial in understanding these scores.
The lower perceived maternal acceptance scores for home schoolers was unexpected. One could assume that in a home environment where the mother is often the main educator, perceived maternal acceptance would be quite high. Perhaps the double role of the mother as a formal educator and the necessary demands of this role has an impact on perceived maternal acceptance. This area of affective development would profit from further investigation with instruments more specific to this domain. Other explanations are tenuous in the absence of research to support or contrast with these findings.
The nonsignificant differences between home school and conventional school scores on the physical appearance and general self worth domains deserve discussion. These scores indicate that both groups perceived themselves similarly and are quite satisfied with their physical appearance. The results of the general self worth subscale reinforces the need for domain specific instruments in measuring perceived competence. That is, when perceived competence is measured as a single general domain, the domain specific differences described above may be masked. Therefore, it is imperative that research examining self-concept and perceived competence use domain specific measures as well as general measures.


Recommendations for further research should consider the testing of home school children on a national scale. This research was limited to home schoolers in Michigan. Comparisons to national norms would become more meaningful with a national sample of home school children. It also would be profitable to examine and compare home school children and conventional school children in each state making comparisons within states.
It also should be noted that this research examined only a small portion of the total scope of physical education. It would be beneficial to investigate other aspects of physical education such as abilities in physical fitness, specific sport skills, and aquatics. Analyses of abilities in these areas of physical education would be useful for prioritizing curricular material needs of home schoolers.
Conventional educators should also be concerned about the results revealing low motor performance ability in the Michigan conventional population tested. It would be beneficial to investigate the impact of greater time and curricular commitment to outcome oriented physical education on motor skill development during these early elementary years.
Further inquiry should also investigate specifics regarding perceived abilities. It would be interesting to determine whether individuals who perceived their abilities to be high actually perform at higher levels and vice-versa.
The internal validity of results could be increased by more specific matching of the two populations. The reader is cautioned about drawing strict causal relationships between low motor performance and home schooling. It is possible that other external concurrent factors interacted to bring about significant results. This research matched populations by age. Matching based on other variables such as socio-economic levels, cognitive abilities, and parental education levels would be beneficial.

                         Final Comments

The results of this research clearly indicate that the fundamental motor skill ability of the home schoolers tested was deficient. These results indicate a need to expand this type of research to a national level. If this population is representative of home schoolers, there is a tremendous need for intentional programming and curriculum emphasizing motor skill development in the early elementary years. It is obvious that for this population tested community sports programs and occasional physical activity is not filling the need in psychomotor education.
Home school educators will find that a search for help in programming this area is fruitless at this time. The physical education literature available for conventional education settings often do not meet the needs of more individualized settings of home schools. While children have similar psychomotor developmental needs in both home school and conventional school settings, the methods of instructing this development will likely differ in both settings. Therefore, one of the great needs for home schoolers is quality materials that will assist them in instruction in physical education.


Campbell, D. T., & Stanley, J. C. (1963). Experimental and quasi-experimental designs for research. Boston, MA: Houghton Mifflin Co.
Colfax, D. & Colfax, M. (1988). Homeschooling for excellence. New York, NY: Warner Books.
Delahooke, M. M. (1986). Home educated children’s social/emotional adjustment and academic achievement: A comparative study. Unpublished doctoral dissertation, California School of Professional Psychology, Los Angeles.
Harris, J. J. & Fields, R. E. (1982). Outlaw generation: A legal analysis of the home instruction movement. Educational Horizons, 61(1), 26-31, 50, 51.
Harter, S. (1982). The perceived competence scale for children. Child Development, 53, 87-97.
Harter, S., Pike, R., Efron, C., Chao, C., & Bierer, B. (1983). Pictorial scale of perceived competence and social acceptance for young children, manual. Denver, CO: University of Denver.
Knowles, J. G. (1991). We’ve grown up and we’re OK: An exploration of adults who were home-educated as students. Paper presented at the 13th National Conference of the New Zealand Association for Research in Education, Dunedin, New Zealand.
Knowles, J. G., Marlow, S. E., & Muchmore, J. A. (1992). From pedagogy to ideology: Origins and phases of home schools in the United States, 1970-1990. American Journal of Education, 100(1).
Moore, R., & Moore, D. (1990). Home grown kids. In Pederson, A., & O’Mara, P., (Eds.) Schooling at home: Parents, kids, and learning. Santa Fe, NM: John Muir Publications.
Ray, B. D. (1991). Home centered learning annotated bibliography. Salem, OR: National Home Education Research Institute.
Ray, B. D. (1990). A review of home school research: Characteristics of the families and learner outcomes. Salem, OR: National Home Education Research Institute.
Reuschlein, P. S., & Haubenstricker, J. L. (1985). 1984-1985 Physical education interpretive report: Grades 4, 7, and 10. Michigan Education Assessment Program. Lansing, MI: Michigan Department of Education.
Taylor, J.W., 5th (1986). Self-concept in home school children. Home School Researcher, 2(2).
Ulrich, D. A. (1985). The test of gross motor development. Austin, TX: Pro-ed.
Ulrich, D. A., & Wise, S. L. (1984). The objectives-based motor skill assessment instrument: Validation of instructional sensitivity. Perceptual and Motor Skills, 59, 175-179.


1. This research was funded by the Hope College Collaborative Faculty-Student Development Grant and by the Frost Social Science Research Center Grant.

0 replies

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply