THE EFFECTS OF AN OBJECTIVE- CENTERED, SEQUENTIAL. PROGRAM OF PHYSICAL EDUCATION ON. THE ACADEMIC ACHIEVEMENT AND INTELLIGENCE: 0F ELEMENTARY SCHOOL-AGED CHILDREN Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY RICHARD EDWIN MOSHER 1 972 JIIIIIIIIIZII‘IIIIIIIIIIIII IIIIIII‘ H 3 Michiganb‘ . . IJEUanxficyr ‘r—iff' 5- ! This is to certify that the thesis entitled The Effects of an Objective-Centered, Sequential Program of Physical Education on the Academic Achievement and Intelligence of Elementary School-Aged Children presented by Richard Edwin Mosher has been accepted towards fulfillment of the requirements for Ph.D. degree in Physical Education fiefm Major professor Date__v~lU.D§ 30 .' 1972 0-7 639 BINDING BY M13“? I DUCK BINDERY ”II‘ ‘ LIBRARY BINIDIEI ”aileron. .I‘I. a IHHEI .. [IQ-Til ...I ABSTRACT THE EFFECTS OF AN OBJECTIVE-CENTERED, SEQUENTIAL PROGRAM OF PHYSICAL EDUCATION ON THE ACADEMIC ACHIEVEMENT AND INTELLIGENCE 0F ELEMENTARY SCHOOLeAGED CHILDREN By Richard Edwin.Mbsher The purpose of this study was to investigate the effects of a one- year, objective-centered, sequential program of physical education upon the academic achievement and mental ability of male and female children enrolled in kindergarten.through grade five. Two hypotheses were proposed: (1) a one-year, objective-centered, sequential program of physical education will not significantly affect the academic achievement and mental ability of elementary school children, at any level, kindergarten through grade five; (2) a one-year, objective- centered, sequential program of physical education will not affect males and females in a significantly different manner, at any level, kinder- garten through grade five, with respect to academic achievement and mental ability. Two schools, one experimental and one control, were selected from the Waverly School District. All students in kindergarten through grade * five, at both schools, were tested (N - 607). * Total N - 597 after deletions due to insufficient data. Richard Edwin Mosher Three tests, the Stanford Early School Achievement Test, the Stanford Achievement Test and the Otis-Lennon Mental Ability Test, formed the test battery administered to the children in grades K-5. Different forms of the same test were administered for pre-test and post-test data collection. The children in grades K-5 attending the experimental school received a one-year physical education program based on the curriculum developed for the Battle Creek Public School System. The program was a research-based, objective-centered, sequential program of physical educa- tion, covering a wide range of skills and activities. Stress was placed on the total development of the child and particular activities were presented to enhance social, emotional, cognitive and physical functioning. A series of multivariate covariance analyses were employed to determine if significant differences existed between the control and experimental groups with respect to the dependent variables, the post- test academic achievement and mental ability scores. Pre-test scores 'were used as the independent variables (co-variates) in the analyses. The results of the present study indicate that the one-year, objective-centered, sequential program of physical education did not facilitate the learning of academic concepts of children in grades kindergarten through five. Although isolated effects were noted through- out the grade levels, no consistent trends were evident. Further analysis indicates that the one-year, objective-centered, sequential program of physical education had no differential effect on the males and females in the experimental group. Although isolated sex differences were noted, no-consistent trends appeared throughout the study. THE EFFECTS OF AN OBJECTIVE-CENTERED, SEQUENTIAL PROGRAM OF PHYSICAL EDUCATION ON THE ACADEMIC ACHIEVEMENT AND INTELLIGENCE OF ELEMENTARY SCHOOL-AGED CHILDREN BY Richard Edwin Masher A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Health, Physical Education and Recreation 1972 Dedication To my wife Jean for her unfailing support and encouragement, without which the task would have been impossible 11 ACKNOWLEDGEMENTS The author wishes to acknowledge the invaluable assistance of Dr. Philip Reuschlein, major advisor for this dissertation, for his guidance and support throughout the author's term of study at Michigan State University. In addition, the author would like to thank Dr. Vern Seefeldt, Dr. Louise Sense and Dr. Gale Mikles for their assistance with the many facets of a Ph.D. program. Special thanks is also expressed to colleagues John Haubenstricker and Conrad Milne for their assistance in the data collection phase, Dr. H. Teitlebaum and Tom Gilliam for their help with the analysis of the data, and Dr. W. van Hues for his encouragement throughout the study. 111 Chapter I II III IV- TABLE OF CONTENTS INTRODUCT ION O O O O O O O O O O O O O O O O I O O O 0 Introduction . . . . . . . . . . . . . . . . . Need for the Study . . . . . . . . . . . . . Purpose of the Study . . . . . . . . . . . . Hypotheses . . . . . . . . . . . . . . . . . Scope of the Study . . . . . . . . . . . . . . Definition of Terminology. . . . . . . . . . .' Limitations of the Study . . . . . . . . . . . RELATED LITERATURE. . . . . . . . . . . . . . . Relationship Between Intellectual Functioning and Motor Performance . . . . . . . . . . . The Effect of Perceptual-Motor Programs on- Academic Achievement and Intelligence . . . Summary of Related Literature. . . . . . . . . “SEARCH “THODS O O O O O O O O O O O O 0 O O O O 0 Experimental Design. . . . . . . . . . . . . . The Subjects. . . . . . . . . . . . . . Collection of Data. . . . . . . . . . . The Tests . . . . . . . . . . . . . . . The Program . . . . . . . . . . . . . . Statistical Treatment of the Data. . . . . . . RESULTS 0 O I O O O O O O O O O O O O O O i O I O O O 0 Chi Square Tests of Hypotheses of No Associa- tion Between Dependent and Independent Variables Step-Wise Regression Analyses. . . . . . . . . . . Multivariate and Univariate Analyses of Covari- ance for School, Sex and Interaction Effects. . Kindergarten . . . . . . . . . . . . . . . . . . . Grade One. . . . . . . . . . . . . . .‘. . . . . . Grade Two. . . . . . . . . . . . . . . . . . . . . Grade Three. . . . . . . . . . . . . . . . . . . . Grade Four . . . . . . . . . . . . . . . . . . . . Grade Five . . . . . . . . . . . . . . . . . . . . DISCUSSION. 0 O C I O O C O O 0‘ O O O O O O O O O O I O 0 Summary and Implications of the Discussion . . . . iv Page O‘U‘lUIJ-‘J-‘UJH oo 16 26 28 28 28 29 30 34 35 39 40 4O 40 41 46 52 59 64' 71 81 87 Chapter Page VI SUMMARY, CONCLUSIONS AND RECOMMENDATIONS. . . . . . . . . 90 Summary. . . . . . . . . . . . . . . . . . . . . Purpose and Hypotheses. . . . . . . . . . Procedures. . . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . 91 Conclusions. . . . . . . . . . . . . . . . . . . 92' Recommendations. . . . . . . . . . . . . . . . . . 92 9O 9O 90 REFERENCES CITED 0 O O O O O O O O O O O O O O O O O O O I O O O O O 94 Table 10 11 12 l3 14 LIST OF TABLES Cell sizes by grade and sex for experimental and con- trol Beh0018 (N - 597) O O O O O O O O O O O O O O O O O 0 Academic achievement and mental ability tests (pre and post) administered to each grade. . . . . . . . . . . . . Academic achievement sub-tests administered by grade. . . Reliability coefficients for each sub-test of SESAT: L evel I O O O O O O O O O O O O O O O O O I O O O O O O O Split-half and Kuder-Richardson reliability coefficients by grade (Form J) 0 O O O O O O O O O O O O O O O O O O 0 Least square estimates adjusted for covariat es: re- ordered variables for kindergarten. . . . . . . . . . . . Cell means for experimental and control schools: re- ordered variables for kindergarten. . . . . . . . . . . . Univariate analysis of covariance test for school effects: kindergarten academic achievement total . . . . Multivariate and univariate analysis of covariance tests for school effects: kindergarten academic aChievemnt BUb-teets o o o o o o o o o o o o lo 0 I o o o Univariate analysis of covariance test for Sex effects: kindergarten academic achievement total . . . . . . . . . Multivariate and univariate analysis of covariance tests for Sex effects: kindergarten academic achieve- ment 8Ub-t88t8 o o o o o o o o o o o o o o o o o o o o o o Univariate analysis of covariance test for School x Sex interaction effects:‘ kindergarten academic achieve— ment , total 0 O O I O O O O O O O O O O O O O O O O O O I O O O Multivariate and univariate analysis of covariance tests for School x Sex interaction: kindergarten academic achievement sub-tests. . . . . . . . . .‘. . . . Least square estimates adjusted for covariates:I re- ordered variables for grade 1 . . . . . . . . . . . . . . vi Page 29 31 32 33 33 42 42 43 43 44 45 46 46 47 Table 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Page Cell means for experimental and control schools: re- ordered variables for grade 1 . . . . . . . . . . . . . . 48 Multivariate and univariate analysis of covariance. tests for school effects: grade 1 academic achievement total and mental ability. . . . . . . . . . . . . . . . . 48 Multivariate and univariate analysis of covariance tests for school effects: grade 1 academic achievement Bub-4268158 o o o o o o o o o o o o .o o o o o o o o o o o o 49 Multivariate and univariate analysis of covariance tests for Sex effects: grade 1 academic achievement total and mental ability. 0 O O O O O O O O O O O O O O O O O O O O 50 Multivariate and univariate analysis of covariance tests for Sex effects: grade 1 academic achievement sub-tests. 51 Multivariate and univariate analysis of covariance tests for School x Sex interaction: grade 1 academic achieve- ment total-and mental ability . . . . . . . . . . . . . . Sl_ Multivariate and univariate analysis of covariance tests for School x Sex interaction: grade 1 academic achieve- ment SHED-tests. o. o o o o o o o o ,' o o o o o o o o o o o 52 I Least square estimates adjusted for covariates: re- ordered variables for grade 2 . . . . . . . . . . . . . . 53 Cell means for experimental and control schools: re- ordered variables for grade 2 . . . . . . . . . . . . . . 53 Multivariate and univariate analysis of covariance tests for school effects: grade 2 academic achievement total and nut“ ability 0 O O I O O O O O O O O O O O O O I O O 54 Univariate analysis of covariance test for school effects: grade 2 academic achievement sub-tests. . . . . 55 Multivariate and univariate analysis of covariance test for Sex effects: grade 2 academic achievement total* and mntal ability 0 O O I O O O O O O O O O O O O I I I O 55 Univariate analysis of covariance test for Sex effects: grade 2 AA sub-tests. . . . . . . . .'. . . . . . . . . . 56 Multivariate and univariate analysis of covariance test for School x.Sex interaction: grade 2 AA TOT and MQAB. . 56 Univariate analysis of covariance test for School x Sex interaction:' grade 2 AA sub-tests. . .>. . . .'. . . 58 Least square estimates adjusted for covariates: re- ordered variables for grade 3 . . . . . . . . . . . . . . 59 vii Table 31 32 33 34 35 36 37 38 39 4O 41 42 43 44 45 46 47 48 Cell means for re-ordered variables for experimental and control schools: grade 3 . . . . . . . .-. . . . Multivariate and univariate analysis of covariance tests for school effects: grade 3 AA TOT and M.AB. . Multivariate and univariate analysis of covariance tests for school effects: grade 3 AA sub-tests . . Multivariate and univariate analysis of covariance test for Sex effects: grade 3 AA TOT and M.AB. . . Multivariate and univariate analysis of covariance test for Sex effects:' grade 3 AA sub-tests . . . . . Multivariate and univariate analysis of covariance test for School x Sex interaction: grade 3 AA TOT and M.AB. O O O O O O O O O O O O O I O O O O O I. O O I O Multivariate and univariate analysis of covariance test for School x Sex interaction: grade 3 AA sub-tests. Least square estimates adjusted for covariates: re- ordered variables for grade 4 . . . . . . . . . . . Cell means for experimental and control schools: re- ordered variables for grade 4 . . . . . . . . . . . . Multivariate and univariate analysis of covariance tests for school effects: grade 4 AA TOT and M.AB. Multivariate and univariate analysis of covariance tests for school effects: grade 4 AA sub-tests . . . Multivariate and univariate analysis of covariance tests for Sex effects: grade 4 AA TOT and M.AB.. Multivariate and univariate analysis of covariance tests for Sex effects: grade 4 AA sub-tests. . . . . Multivariate and univariate analysis of covariance tests for School x Sex interaction: grade 4 AA TOT and M.AB. O O O O O O O O O O I O O O O O O I O O O O Multivariate and univariate analysis of covariance tests for School x Sex interaction: grade 4 AA sub-tests Least square estimates adjusted for covariates:. re- ordered variables for grade 5 . . . . . . . . . . . . Cell means for experimental and control treatment groups: re-ordered variables for grade 5 . . . .‘. . . .‘ Multivariate and univariate analysis of covariance tests for school effects: grade 5 AA TOT and M.AB. . . . viii Page 60 6O 61 62 62 63 64 65 65 66 67 68 68 69 7O 71 72 73 Table 49 50 51 52 53 54 Page Multivariate and univariate analysis of covariance tests for school effects: grade 5 AA sub-tests . . . . . 74- Multivariate and univariate analysis of covariance tests for Sex effects: grade 5 AA TOT and MMAB.. . . . . 75 Multivariate and univariate analysis of covariance tests for Sex effects: grade 5 AA sub-tests. . . . . . . 76 Multivariate and univariate analysis of covariance tests for School x Sex interaction: grade 5 AA TOT and M.AB. O O O O O O O O O I O O O O O O O O .1 O O I O I O O 77 Multivariate and univariate analysis of covariance tests for School x Sex interaction: grade 5 AA sub- teats O O O O O C O O O O O O O O I O O O O O O O O O O O 78 Summary of significant main effects and interactions: grades K-5 0 O O O I O O O O O O O O O O O I O O O O O O O 79 ix Figure LIST OF FIGURES Page School x Sex interaction for grade 2 subjects: mental ability 0 O O O O O O I O O I O O O O O O O O O O O 5 7 School x Sex interaction for grade 2 subjects: academic achievement total. . . . . . . . . . . . . . . . 58 School x Sex interaction for grade 4 subjects: Academic Achievement sub-test 7 . . . . . . . . . . . . . 7O CHAPTER I INTRODUCTION Introduction The belief that the mind and body are closely interrelated is certainly not a new concept. The Gestaltist philosOphy that the indi- vidual functions as a "whole" has been accepted by many psychologists and educators and is expressed aptly_by Hall and Lindsey (l957) in the following quotation: "The organism.always behaves as a unified whole and not as a series of differential parts. Mind and body are not separate entities, nor does the mind consist of independent faculties or elements and the body‘ of independent organs and processes." (p. 297) Olson's "Organismic Age" (1949), through which the author professes a belief in the interrelatedness of such aspects of growth and develOpment as carpal age, height, mental age and school achievement, has also been. ‘well documented for more than 20 years. Assuming this "wholeness" of child development to be tenable, many investigators have devised motor-oriented programs which have as one of their primary goals the enhancement of academic and intellectual function- ing. One of these programs demanding considerable attention has been devised by Delacato (1963). Stressing the concept of "neurological organisation", the prOponents of this program allege that omissions in" the sequence of early motor patterns will result in subsequent behavior malfunction of one form or another. Thus a rigorous program.of motor 2 activities is prescribed that will purportedly stimulate the brain to its full functional capacity. Kephart's Perceptual Motor_Program (1960), while emphasizing the child's orientation to his environment to a greater extent than Delacato, also stresses the importance of motor experiences in the total education of the child. Similarly, the Visual-Perceptual Program devised by Frostig (1964) is based on the assumption that reading ability depends largely upon the acquisition of adequate visual—perceptual skills. Activity programs have been devised by Frostig to enhance specific visual-perceptual deficiencies (i.e., deficiencies in identifying mirror images from exact likenesses). Reading achievement is also thought to increase after exposure to a motor-oriented program emphasizing self- control (MbCormick, 1968). McCormick maintains that an internalization of self-control, gained through a sequential series of physical activities, will lead to an improvement in reading ability. An improvement in academic and intellectual functioning, through participation in the Dayton Program for develOping sensory and motor_ skills, is also suggested by Braley and his cadworkers (1968). Finally, Barsch's "Movigenic Theory" (1967) epitomizes the recent attempts at not only.linking the mind and body, but also in presenting a "theory ofv movement" as it relates to total learning. Movigenics is "the study of the origin and development of patterns of movement in man and the relationship of those movements to his learning efficiency." (p. 33) Barsch also views movigenics as "an effort to view men as a totality in everything he does and to account for all components of that totality in any of his performances." (p. 33) 3 Although there has been very little scientific data gathered under controlled conditions, the number of programs emphasizing the development of intellectual capacity through physical activity is increasing. Need for the Study Society has reached that point in time where experimental evidence is required in order to substantiate the heretofore subjective claims of various programs, not only in physical education, but in general education as well. School authorities, the government, and the community require that more than speculative data be produced to substantiate requests for financial and moral support of these educational programs. Many physical educators are convinced that physical education programs-can exist on the basis of psychomotor objectives alone. Pro- motion of the physical welfare of today's youth and adult populations alike is certainly foremost in terms of the priorities set by physical education specialists. However, since physical educators invariably include additional objectives in the cognitive and affective domains when-discussing the relative merits of the physical education profession, one of two tasks must be accomplished; either the claims must be sub- stantiated with scientific research, or those objectives which cannot be realistically and scientifically supported must be deleted. The first alternative is obviously the preferred approach in that a much stronger justification for physical education could be developed with a scientificallyébased, multi-objective program than from a pragram with a more.limited scOpe. To date, the scientific evidence supporting the theories dealing with the effects of motor activity on consequent academic and intellect- ual functioning is extremely contradictory. Conclusive evidence has not 4 been produced either favoring or disputing the effects of physical.activity programs on academic achievement. The educational programs which have been constructed, and the cons“ sequent evaluations of the programs, have been based primarily on atypical activities; atypical in the sense that special activities have been devised to meet particular needs.’ Many of the studies conducted have been clinical studies dealing with small numbers of subjects and no con- trol groups. Little evaluation has been conducted with respect to the effects of a physical education program on academic achievement. It should be noted as well that the total elementary school age range has seldom been considered. The overwhelming majority of studies have been restricted to one, or possibly two, age groups. Of interest, too, is the fact that scientific evaluation of sex differences, in relation to motor activity programs designed to facilitate classroom learning, is sparse and conflicting. Suggestions that the male may decrease the "academic gap" between he and his female counterpart, through participa- tion in a more male-oriented curriculum, are prevalent in the literature (Kettels, 1967; Bentzen, 1966;_Minuchin, 1966). Purpose of the Study The purpose of this study is.to investigate the effects of a one- year, objective-centered, sequential program of physical education upon the academic achievement and mental ability of male and female children enrolled in kindergarten through grade five. Hypotheses l. A onedyear, objective-centered, sequential program of physical education will not significantly affect the academic achievement and 5 mental ability of elementary school children, at any level, kindergarten through grade five. 2. A.one-year, objective-centered, sequential program of physical education will not affect males and females in-a significantly different manner, at any level, kindergarten through grade five, with respect to academic achievement and mental ability. Scope of the Study Two schools, one experimental and one control, were selected from the Waverly School District. All students in kindergarten through grade five, at both schools, were evaluated with respect to the variables of total academic achievement, academic achievement sub-tests, and intelli- gence (N - 607).* The Waverly School District offered an excellent opportunity to test the effects of a physical education program on academic variables in that no regularly scheduled physical education program was in Opera- tion in the elementary schools at that time. Thus, the control school was a."control" in the true sense, in that no regularly scheduled physical activity was provided for the pupils in the control school. Definition of TermdnoIggy_ For the purpose of this study, the following definitions are presented: 1. Academic Achievement: knowledges, skills and understandings commonly accepted as desirable outcomes of the major branches of the elementary curriculum (Kelly at al., 1964). *Total N - 597 after deletions due to lack of sufficient data. 6 2. Mental Ability: the pupil's facility in reasoning, and in dealing abstractly with verbal, symbolic, and figural test content sampling a broad range of cognitive abilities; measurement of the "g" or general intellective ability factor (Otis and Lennon, 1967). 3. Objective-centered: based on,a wide range of objectives con- structed for each grade level by the members of the Battle Creek Physical Education Curriculum Project (Vogel, 1969). 4. Perceptual-motor Programs: programs emphasizing a sequence of events which include sensory stimulation, cortical processing and overt behavior (Seefeldt, 1970). 5. Experimental Physical Education Program: a special program of physical activities constructed by the Battle Creek Physical Education Curriculum.Project to satisfy the physical, social, emotional and intel— lectual needs of the elementary school child. 6. Sequential:r following a lessonéby-lesson sequence of motor activities, as established by the Battle Creek Physical Education Curricu- lum Project, and based upon current knowledge regarding the growth and development and principles of learning of elementary sChool-aged children- (ngel, 1969). Limitations of the Study 1. The subjects were drawn entirely from the Waverly School System and thus, the generalizations may be limited by some rather specific characteristics of this community. 2. The degree of motivation throughout the testing period could not be rigidly controlled. 3. A-number of testers were employed which may have produced some inconsistencies in testing procedures. 7 4. The effect of the tests being administered by the testing team, as opposed to the regular classroom teacher, may have increased the. variability of the scores and therefore decreased the reliability of the tests. 5. Although the total sample size was 597 students, individual cell sizes were occasionally low. For example, the cell size for the experimental school, grade four females, numbered 15. A similar number was encountered in the category of control school, grade four males. CHAPTER II RELATED LITERATURE The literature pertinent to the present investigation has been divided into the following two subdivisions for review: 1. Those studies investigating the relationship between intellectual functioning and motor performance. 2. Those studies investigating the effects of perceptual-motor and physical education programs upon academic achievement and intelligence. Relationship Between Intellectual Functioning;and Motor Performance Much of the impetus for the construction of motorébased programs, aimed at the enhancement of intellectual functioning, originated as a result of the high positive correlations found between motor and mental ability in the retarded population. Francis and Rarick (1959) and.Howe (1959) demonstrated clearly that intelligence, as measured by standardized intelligence tests, was positively correlated with motor items. A secondary finding of the latter author, that mentally retarded subjects appeared extremely weak in balance activities, is interesting in the light of the findings of Ismail et al. (1965; 1969), to be presented later in the review. Sloan (1951) produced similar data to that reported by Francis and Rarick, and Howe, upon measuring the relationship between mental deficiency and motor proficiency in 10 year old males and females. This author reported that mental defectives were significantly inferior in motor proficiency to children of average intelligence. Sloan thus 8 9 concluded that motor proficiency was related to intelligence, and that motor functioning was.an aspect of the total functioning of the organism. IMalpass (1960) produced similar results when he compared 52 institution- alized retardates, 56 retarded children in public schools and 71 normal children on the Lincoln revision of the Oseretsky Test. No significant differences were found between the two retarded groups, but highly sig- nificant differences were noted when the scores of either retarded group ‘were compared with the normal children. All differences were in favor of the latter group. Similarly, in an attempt to ascertain the rela- tionship between raildwalking ability and the factors of mental age and etiological type among subnormal children, Heath (1942) tested 170 male' subjects. Heath's data supported the hypothesis that there is a positive relationship between motor performance and "quality of thought." A number of the early studies investigating the mindAbody relations ship in normal populations supported the assumption that a highly positive relationship did exist (Rarick and McKee, 1949; Kulcinski, 1945; Hackensmith and Miller, 1938; Jones, 1935; Rudisill, 1923). However, a greater number of researchers (Burley and Anderson, 1955; Brace, 1948; JOhnson, 1942; Ray, 1940; Halsey, 1938; DiGiovanna, 1936; Seegers and Postpichal, 1936; Landis et al., 1923) have found very slight, though invariably positive, correlations between the two abilities. Negative relationships are few (Slusher, 1964; Reels and Reese, 1939), but do exist. In 1963, Ryan utilized the concept of motivation in an attempt to understand the relationship between the academic and motor domains. This author investigated the relationship between performance on a selected motor skill and (1) intellectual capacity; (2) intellectual achievement; and (3) motivation, as measured by relative intellectual achievement. 10 When Ryan's group (N - 80) was dichotomized on the basis of college grade point averages and entrance examinations, no differences were' found between the sub-groups with respect to motor performance. How— ever, upon classifying his subjects as to "underachievers" and "over- achievers" (comparing college grades with entrance examination scores), the data indicated that the motor performance of the overachievers was significantly better than that of the underachievers. Ryan suggested that the motive to succeed in academic situations may be a more general characteristic than either intellectual or motor ability. As a result of these earlier studies, researchers conducted further investigations, many of which attempted to explore the relation- ship between the intellectual domain and perceptual-motor functioning. These studies (Skubic and Anderson, 1970; Singer, 1968; Singer and Brunk, 1967; Chang and Chang, 1967) were designed to include motor items which required a greater degree of cognitive functioning than was previously expected of the subjects. Research by Skubic and Anderson (1970) typifies the particular approach under consideration. Eighty—six fourth grade boys were studied- to investigate the relationship between perceptual-motor achievement, academic achievement and intelligence. 0n the basis of performance on. the Stanford Achievement Test, 41 and 45 pupils were categorized as low and high achievers respectively. The California Test of Mental Maturity and an 11-item-perceptual—motor battery were also administered. A positive relationship was found between.intelligence and perceptual-motor ability (ri- .491) and between academic achievement and perceptual-motor ability (r - .500). In discussing these relatively high correlations, the authors attributed them to their choice of perceptual-motor items: 11 "...in the battery of tests used in the present study, an attempt was made to eliminate problems encountered in previous research. Tests were constructed for the specific age level involved and were designed to present a challenge to the subject.both mentally and motorically." (p. 418) Singer and Brunk (1967) investigated the relationship between intelligence and the performance of a highly perceptual task, the Figure Reproduction Test. The authors found that although the majority of the correlations between the academic variables and the Figure Reproduction Test were significant, they accounted for very little of the total‘ variance. In fact, the authors noted that their highest correlation (r - .29) accounted for only .08 of the total variance. They thus con— cluded that the relationship, though significant, was highly suspect. Some researchers (Chang and Chang, 1967; Singer, 1968) have investigated the hypothesis that higher correlations may be found for younger subjects, as opposed to older children, with respect to the relationship between perceptual-motor and academic skills. Chang and Chang (1967) found that although higher relationships between visua- motor and reading skills were recorded for second graders, these coef- ficients were not significantly different from those obtained by those in grade three. In a similar study, Singer (1968) investigated the relationship between academic achievement, physical characteristics and perceptual-motor abilities within and between selected groups of third and sixth grade children (N - 42 and 30 respectively). The author found that even though a relatively low coefficient of .35 was required for significance, few coefficients for either grade exceeded this figure. In addition, no trends were apparent even where statistical significance was reached. Thus, in rejecting the hypothesis that the interrelation- ships between perceptual-motor, physical and cognitive variables would E. It 12 be greater_at the lower level, Singer concluded that individual abilities are fairly well task-specific, even with youngsters in the third grade. A second hypothesis, that those perceptual tasks with high intellectual components would correlate highest with academic achievement, was also rejected. The last decade has seen a shift from-the consideration of intelli- gence, perceptual-motor performance and academic achievement as total entities, to a consideration of the sub-components of these general classifications. A number of studies have been reported concerning auditory-motor and visual rhythm, and their relation to reading achieve- ment. Birch and Belmont (1965) produced data indicating a strong. relationship between perceptual ability, as measured by auditory-visual integration and reading achievement. More specifically, the perceptual test required the subject to choose one of three printed dot patterns that corresponded to a stimulus pattern tapped out with a pencil. Utilizing the same perceptual task to measure auditory-temporal rhythm perception, Sterritt and Rudnick (1966) found this aspect of perception related to reading achievement in fourth grade boys. Further study, by Werner et al. (1967), did little to clarify the issues at hand. These authors investigated the effectiveness of the Bender-Gestalt (B-G) test in identifying 10 and 11 year old children ‘with reading problems (N'- 750). The investigators found that although the incidence of reading problems increased with error scores on the perceptual task (B-G), the majority of children with reading problems had adequate B-G_reproductions. However, most‘poor B-G reproductions were found among children of below average intelligence. Other investigations focused upon spelling and language problems, as well as those problems experienced in reading, and the relationships '4. \ .o- o I 1 u: n: 'IJ 13 of these problems to perceptual-motor functioning. Coleman (1968) studied 87 children in grades one through six, all of whom possessed severe language and reading deficits. Upon evaluating for visual and visual-perceptual development, it was found that almost 50% of the sample had visual-perceptual dysfunction judged severe enough to significantly impede learning. Similarly, Bannatyne and Wichiarajote (1969) found significant relationships between unlearned ambidexterity and spelling (r - .42) and unlearned ambidexterity and balance (r - .28). The authors inferred a superior motor coordination between the two sides of the body and between these two variables and the eyes. In other words, the inference is that the superior speller possesses superior visuo-motor coordination. The authors concluded by suggesting that "spelling as an encoding function is heavily depending on automized motor/kinesthetic/ praxic processes" (p. 9). Physical education specialists have also been active in the assess- ment of relationships between perceptual-motor and academic functioning. Testing 172 male and female subjects, Plack (1968) found high positive relationships between achievement in reading and selected motor skills for children in grades one, three and five. This author also noted that mean motor skill scores were significantly different among high, middle and low reading achievement groups. More specifically, differences between the high and low reading achievement levels were generally sig- nificant, those between the high and middle were generally significant, and those between the middle and low were generally not significant. The work of Trussel (1969), however, yielded much lower correlations, with respect to the relationships between reading achievement, visual perception, motor development and eye-hand dominance. Seventy-five first and second grade children were tested on sub-tests of the MetrOpolitan ‘0: ‘1 l4 Achievement Test, reading, the Lincoln-Oseretsky Motor Development Scale, the Frostig_Developmenta1 Test of Visual Perception and two perceptual- motor tasks, the pursuit rotor and the stabilometer. Although correla- tions among many variables were significant, none were high enough for reliable prediction of one variable from another. Factor analysis failed to provide factors indicative of interrelationships among elements of visual, motor and reading functions. Attempts at predicting scores on one variable from results obtained on one or more other variables have been alluded to throughout this review of literature. At this point, the author would like to discuss a number of research findings which place a greater, though certainly not total, emphasis on the prediction of academic variables from information gathered with respect to motor variables. Ismail et al. (1963) initiated a group of "prediction" studies, in investigating the relationships between academic achievement, intel- ligence and motor aptitude. One hundred twenty 10, 11 and 12 year old males and females comprised the sample for this study. This investiga- tion indicated that motor aptitude can in fact predict, with some measure of success, achievement scores in academic variables. The authors noted (a) that intelligence scores could be predicted more accurately from motor aptitude test items than could academic achievement scores, and (b) both intelligence and academic achievement scores could be predicted more accurately by motor aptitude test items in high and low achievers than in medium achievers. No sex differences were noted with respect to prediction of academic variables from motor variables. Ismail and Gruber (1965) conducted a follow-up study which not only substantiated the above findings, but added a new dimension. The authors reported that M c b o . c l u ‘1 El 15 coordination and balance items, in that order, were the most beneficial motor performance items in predicting academic achievement. Three subsequent studies (Ismail et al., 1969; Yoder, 1968; Bengston, 1966) tended to support the results just presented. Ismail et al. (1969) provided a cross-validation of their previous work, using as subjects 94 primary school children from.London, England. This research also pointed to a substantial and positive association between motor aptitude, especially with respect to coordination and balance, and well established measures of intelligence and scholastic performance. Yoder (1968), utilizing a sample of 240 fourth and fifth grade boys and girls, found the same type of factors and relationships as those reported by Ismail. Bengston (1966) also confirmed the consistently positive rela- tionships found by the other investigators. The latter author examined the interrelationships among perceptual-motor development, motor per- formance, school achievement, and intelligence, using slightly younger children (9 years), fewer subjects (N - 55), and limiting the study to males only. Although the relationships between school achievement sub- scores and motor performance were generally low, the relationships between the Purdue Perceptual-Motor Survey sub-tests and the total school achievement were relatively high and positive. Not all studies tend to support the hypothesis that it is possible to predict academic scores from scores obtained on motor tasks. Research by Trussell (1969) and Bihlmeyer (1965) suggests the predictability of academic variables from motor performance measures to be virtually zero. The former author (Trussel, 1969) stated that: "It would be equally effective and far more exped- ient, to predict the performance of the child on the basis of age, rather than by employing any of the test batteries considered here." (p. 386) 16 The test batteries referred to included the Metropolitan Reading Test, LincolneOseretsky Motor Development Scale, Frostig Developmental Test of Visual Perception, and pursuit rotor and stabilometer tasks. Bihlmeyer (1965) produced data which cast doubt on the reliability of predicting academic scores from.motor variables. This author utilized the Lincoln-Oseretsky Motor Development Scale in an attempt to predict the academic achievement of fourth grade boys. He concluded that it did not possess a sufficient degree of relationship to reading, writing, drawing, and general educational achievement to be of value as a predictor of performance. It is of interest to note that both studies failing to substantiate the existence of a relationship between motor and academic variables, utilized the Lincoln-Oseretsky Motor Development Scale as the criterion measure with respect to motor performance. The Effect of Perceptual-Motor Programs on Academic Achievement and Intelliggnce Many researchers have devised perceptual-motor training programs which have as their ultimate goal the enhancement of academic functioning (McCormick, 1968; Barsch, 1965; Frostig, 1964; Delacato, 1963; Kephart, 1960). Evaluation of these programs has produced a number of studies, some supporting and others refuting the claims of the originator of the program cited. One of the earliest programs to be developed was that proposed by Delacato (1963), based on the_concept of "neurological organization." Delacato (1966) states that the basic difference between the nervous system.of man and that of lower animals lies not in the number of cells, but in the differentiation and organization of those cells. Thus, the concept of neurological organization, in addition to neurological develOp- ment, has to be recognized. Further, neurological organization is a 17 "whole." If any sepect is not complete, further development will be adversely affected. Delacato also states that: "Deprivation, trauma or enrichment all affect the development and organization of the nervous system. Children who have problems with reading have either been traumatized or have been deprived environmentally, resulting in a lack of complete Neurological Organization which, in turn, creates the reading problem." (p. 5) A number of studies have attempted to evaluate the validity of Delacato's theory of neurological organization. Robbins (1966), in the first of two investigations, selected one second grade class from each of three Chicago elementary schools. The first class (N - 43) served as a control group and carried on its normal curriculum. The second class (N - 38) was designated as the experimental group and, in addition to its regular curriculum, was subjected to a program consistent with the theory of Delacato. In order to control for experimental effect, the third class (N - 45) undertook a general program of activities not thought to_be correlated with reading achievement. The results of the three-month program suggested that creeping was not related to reading beyond chance.expectancy, nor did mean reading differences between sub- jects who were lateralized and those who were not lateralized exceed chance expectations. Further, mean post-test.differences in reading between the group exposed to the experimental program and the other two groups were not significant. If the reader is willing to accept the limitations of non-random selection of subjects and differential teaching skills among the three teachers involved, it must be concluded that the addition of the experimental program to the regular curriculum in no way enhanced the reading or lateral development of these second grade children. Similar findings resulted from a follow-up study (Robbins, 1967) employing retarded readers (N - 150) ranging in grade level from 18 three to nine. Experimental evidence did not support the postulated relationship between neurological organization (as measured by creeping and laterality) and reading achievement. The data of O'Donnell and Eisenson (1969) tended to support the findings of Robbins, although some measure of success was noted in.terms of producing changes in reading ability through neurological training. Utilizing a sample of 678 dis- abled readers (C.A. - 84-120 months), the authors noted that neither reading ability nor visual-motor integration were significantly higher in the experimental groups than in-the controls. However, in both cases the experimental group did score higher on the post-tests for both reading and visual-motor integration. The Perceptual-Motor Program designed by Kephart (1960), while being more difficult to investigate, has nevertheless been scientifically evaluated by some researchers (Brown, 1968; Painter, 1966; Falik, 1969) in an attempt to determine the relative effectiveness of perceptual- motor activity toward enhancing academic potential. Difficulties have arisen, as noted, "...due to the inability of investigators to pro- vide a rationale for a quasi-theoretical base. The lack of a well defined.methodology, with the impli- cation that a number of skills may be utilized to enhance various perceptual abilities, may also have discouraged scientific evaluation of the program." (Seefeldt, p. 84) Brown (1968) used 56 below-average readers as his experimental group and gave these subjects special perceptual-motor training, largely based on the activities suggested by Kephart (1960). Ninety-eight con- trol subjects.were given a physical education program which emphasized development of basic skills, with little emphasis on the development of perceptual-motor skills. The author's t-test analyses indicated that the experimental group did improve significantly in perceptual-motor 19 skills and, although greater gains in reading achievement were recorded by the experimental pupils than by the control group, significant dif- ferences were not detected. Two studies (Painter, 1966; Falik, 1969) on kindergarten children provide the reader with contradictory evidence concerning the effects of perceptual-motor programs. Painter, while not directly evaluating the effects of Kephart's Perceptual-Motor Program, investigated the effects of a program based on nine movement areas of Barsch's Moviegenic theory (1965) and on suggestions from Kephart (1960). Although signifi- cant differences were found in all areas, specifically the ability to draw a human figure, distortions of body-image concept, visual-motor integrity, sensory-motor spatial skills and psycholinguistic abilities, caution should be exercised in view of the small sample size (10 experi- mental, 10 control) and the lack of control for experimental effect. Falik's kindergarten study (1969) failed to substantiate the findings of Painter. The former researcher's results indicated that the experi- mental group, participating in a perceptual-motor program fashioned after that of Kephart, was not significantly superior in reading ability to the control group, either at the end of kindergarten or upon retesting in the middle of grade_two. Perhaps more investigators have concerned themselves with Frostig's VisualePerceptual Training Program (Frostig, 1964) than with any other single motor-oriented program. Jacobs (1968a) conducted one such evalua- tion on kindergarten, pre—kindergarten and first grade children (N - 500). Co-variance analyses indicated that children taking the Frostig program did significantly better on the Frostig Visual-Perceptual Test than did controls, with largest gains being made by those in first grade. However, no differences in reading readiness were found, although only kindergarten 20 pupils were evaluated on this variable. A follow-up study (1968b) was undertaken in an attempt to determine the predictive validity of the Frostig tests.upon.future reading achievement. The findings indicated that not only did pupils who took the Frostig program encounter little advantage in terms of future reading achievement, but the predictive ability of both the Frostig and the Metropolitan Reading Readiness Test was very low. However, as was shown in the previous work, pupils who~ took the Frostig program in pre-kindergarten, kindergarten and first grade tended to show higher levels of visual-perceptual performance on the Frostig test. Thus, the results of the two studies lead the reader to believe that if one's goal is simply to increase the visual- perceptual quotient of children, the Frostig program is beneficial. However, if the goal is to increase reading ability beyond that normally expected, the evidence provided by Jacobs is far from supportive. Identical conclusions can be drawn from the work of Rosen (1966). Studying a group.of 637 first grade boys and girls, Rosen found sig- nificant differences between treatment groups (Frostig and controls) in certain post—training perceptual abilities without concomitant.effects on reading achievement. Rosen concluded that additional time devoted to reading instruction was more important for improvement in reading than time devoted to the types of perceptual training utilized in-the investigation. WOrking with second graders (N - 71), Olson (1966) found that although the Frostig test had some value as a predictor of general achievement, reading achievement was predicted to a greater extent by more traditional diagnostic tests. Additional findings indicated, how- I ever, that four of the five Frostig sub-tests showed significant relation- ships with specific reading abilities. These statistically "significant" 21 relationships, however, often accounted for less than fifteen.percent of the variance. For example, many of the statistically significant correlations were in the neighborhood of .31 to .38. It should be noted also that the correlations obtained in Olson's study, although often significant, did not approach those presented by Bryan (1969). The- latter author presented some evidence concerning the effectiveness of the Frostig test in predicting reading readiness in kindergarten (.70) and grade one (.46), as well as significant relationships in grade one between Frostig scores and California Achievement Vocabulary (.50) and Comprehension (.51). Utilizing the factor analysis technique, Boyd and Randle (1970) found that the Frostig Test of Visual Perception did not appear to reflect essentially different and independent perceptual abilities. Thus, the content validity of the Developmental Test of Visual Percep- tion was seriously questioned. However, further analysis of both the Frostig Test and Visual Perceptual Program continued. In 1968, Lewis conducted a pilot study to determine if the Frostig program could be instrumental in improving perceptual ability as reflected in reading achievement gain. Five second grade youngsters, all diagnosed as having reading disability, were given a teneweek program consisting of three hourly sessions each week. Although improvement was noted in each area of visual perception, the small N prohibited the attaining of statistical significance. Nevertheless, the findings were similar to those reported earlier by Jacobs (1968a, 1968b) and Rosen (1966). More interesting perhaps were those gains in reading ability reported by Lewis. A mean percentile gain of 40.2 for the 5 subjects was significant at the .01 level. However, the small N tends to cast considerable doubt on this unusual finding . 22 A different approach was taken by Fretz et a1. (1969) in a study which investigated the effects of therapeutic play on the intellectual and perceptual-motor development of maladjusted children. The findings indicated that while the changes in perceptual-motor performance, as measured by the Frostig and Bender-Gestalt, were significant in favor of the experimental group, changes in intelligence were not nearly so evident. Performance IQ's increased significantly for the controls. In fact, the sub-test changes in both the Performance and Verbal.TQ's revealed a mixed pattern of significant changes. The Frostig program.has also been evaluated with respect to its effect on educably mentally retarded children (Alley, 1968). The perceptual-motor training programuwas conducted daily for two months, during which time the controls spent a comparable amount of time in regular special education activities. Covariance analysis suggested that there were no advantages to be derived by EMR children from the Frostig visual-perceptual program, with regard to subsequent sensori- motor, visual-motor, perception and concept formation performance, over general special education classroom activities. To date, little scientific evidence has been produced to substanti- ate the success of McCormick's "attention-orienting" program (1968), from which improvement in reading skills is alleged to be concomitant with the development of self-discipline in the presence of surrounding stimuli. To evaluate their own.program, McCormick et al. (1968) matched fortyetwo underachieving grade one children for age, sex, intelligence and Lee-Clark reading grade level. Group one received the perceptual- motor program in 45-minute periods, twice a week for seven weeks. The exercises began with cross-lateral crawling and proceeded through walking, balancing, jumping rape and other similar exercises. Group two, formed 23 to control for "extra-activity" effects, received standard physical education training. A third group formed the control group proper and thus received no extra training or activity. The data were analyzed for significance by the Wilcoxen Matched-Pairs Signed Rank Test and the results indicated statistically significant gains in reading achievement for the group which received the perceptual-motor training but not for either of the two control groups. A number of studies have attempted to explore the effectivness of perceptual-motor training procedures other than those specifically set down by Delacato, Kephart, Barsch, Frostig and McCormick. Ines 1963 study, Chansky (1963) investigated whether children trained to perceive with greater precision would subsequently become better achievers. Underachieving children received training in reproducing designs with blocks differing in size, shape and color. Upon analysis, it was found that trained children, in contrast to the controls, improved in both reading and spelling. Also noted was the fact that many of the experimental subjects increased as much as two standard errors of measurement in intelligence scores. A follow-up study (Chansky and Taylor, 1964) of a similar nature, conducted on educably mentally retarded children, produced similar results. An additional finding of note was the fact that group training appeared to be as effective as individualized training in producing achievement and.intelligence test improvements. Changes in intelligence scores as a result of visual-perceptual training were also the concern of Boger (1952). working-with elementary pupils, the author's experimental group participated in a visual- perceptual program.consisting of work with such materials as hidden picture puzzles, directional drawings, visual crossword puzzles, jig-saw . ' 24 puzzles and other similar tasks. In general, the exercises were devised to provide practice in following directions, noting details, perceiving spatial relationships and developing increased eye-hand coordination. Covariance analysis indicated that the exercises did cause an increase in intelligence scores of rural elementary school pupils. In addition, retests five months later showed that the experimental group maintained their superiority with respect to intelligence scores. From the literature cited, it would appear that most children who are given extensive training in.perceptual and motor tasks are able to: (a) show an improvement in tasks involving identical cognitive processes, e.g., left-right discriminations in perceptual forms such as the Frostig training procedures, and (b) show no significant increase in their school achievement. Bibace and Hancock (1969) conducted a study to examine the above possibilities. More specifically, the authors conducted a pilot study to test the theoretical assumption of the mastery.of lower (perceptual- motor) processes as necessarily preceding the mastery of higher (cognitive) processes, and hence academic achievement. It should be noted at this point that this theoretical assumption is extremely important in that it has shaped both the methodological assumptions guiding research in the field and the clinicalhpedagogical efforts of various specialists who attempt to offer remedial programs for children with learning disa- bilities. Examples of both cases, many of which have formed the basis of the review to this point, are prevalent. The experimental design employed by the authors consisted of three variables: age, level of schOlastic achievement, and level of perceptual- motor achievement. Boys aged 7-8 and 12-13 were selected in order that the assumption could be investigated for both those who had only 25 recently acquired the higher cognitive processes and those in whom such processes should have been welléestablished. Since this study was a pilot investigation, only 8 boys were tested. In addition to the Kephart Perceptual—Motor Survey and school grades, data were collected on learning tasks, each of which was developmentally organized such that lower (perceptual-motor) or higher (conceptual) means of functioning might be recognized in the performance. The authors found that those who were high in scholastic achievement, irrespective of whether they were high or low in perceptual-motor functioning, utilized conceptual means rather than perceptual means on the three experimental tasks. Conversely, subjects low in scholastic achievement relied primarily on perceptual means whcn faced with the experimental tasks. The authors concluded that the the theoretical assumption of the mastery of lower processes as necessarily preceding the mastery of higher processes must be more fully researched. While the effects of various types of perceptual-motor programs, many employing widely varying techniques and equipment, have been extensively investigated, the effects of regular physical education programs upon the academic functioning of the child are still unknown. In fact, to this author's knowledge, the only scientific attempt to evaluate the relative effectiveness of a.well-organized regular physical education program on intelligence and academic achievement was that conducted by Ismail (1967). The sample consisted of ten, eleven and twelve year old subjects (N - 142). Experimental and control groups were formed by matching according to intelligence. Three sub-groups were formed within both the experimental and control groups and labelled "high achievers" (N - 18), "medium achievers" (N - 44), and "low achievers" (N - 9). Analysis of variance produced only two significant F values vi—I 26 with respect to intelligence and one of these, "levels of achievement", 'was obviously expected. The second significant P value, "treatments by. sex interaction", indicated that higher mean intelligence scores in boys were associated with the experimental group. Opposite findings were obtained for girls. The results obtained from investigating the relative effectiveness of the physical education program on academic achievement presented a somewhat different picture. Together with the expected significant.F value for "levels of achievement", the authors also found significance for the "treatments" main effect, in favor of the experi- mental group. This indicated that the physical education program had' a significantly favorable effect on the academic achievement of children. Summary of Related Literature The relationships noted between academic and motor domains are generally positive, though rarely high enough for predictive purposes. The highest relationships occurred when extreme population groups were compared. For example, relationships of a high magnitude are often noted when comparing retarded to normal papulations in motor ability. 1 Similarly, studies dichotomized as to "high" and "low" achievers often produce high positive relationships with respect to academic and motor, functioning. Lastly, higher positive relationships are invariably found in studies which have investigated a retarded population alone, rather than those studies conducted on normal papulations. In reviewing the large number of studies investigating the effects of perceptual-motor programs upon academic achievement and intelligence, it appears that the principle of specificity must be upheld. For the most part, children receiving perceptual—motor training improve most in perceptual—motor skills. Those children receiving extra reading practice 27 appear to improve most in reading skills. Exceptions are noted, how- ever, when_investigating retarded readers, in which case significant gains in reading, or other academic skills, are often in evidence. However, in general, the more closely related the learning experience is to the ultimate task, the more significant are the gains noted. In a great number of the studies reviewed, one or more of the following limitations were evident: 8) b) e) d) lack of a control group very small sample size lack of control for the "Hawthorne Effect lack of a well-defined methodology (i.e., specific tasks, methods, etc.) with which to conduct an experimental program. CHAPTER III RESEARCH METHODS The purpose of this study was to determdne the effects of a one‘ year, objective-centered, sequential program of physical education on. the academic achievement and mental ability of male and female children enrolled in kindergarten through grade five. Experimental Design Theisubjects The children involved in the study were obtained from the Waverly School District, to the west of Lansing, Michigan. The elementary schools of the district were studied to determine if it was possible to locate two schools equated with respect to facilities, qualifications of teaching staff, socio-economic status, and caliber of student. Two such schools, namely Elmwood and Colt Elementary Schools, were identi- fied by the research team from Michigan State University and the super- intendant, principals and teachers of the Waverly School System. Elwood Elementary School was identified as the experimental school by random selection. Complete data were obtained for 597 children. This figure includes all subjects, grades K—5, at both schools, with the exception of those' 28 29 * for whom insufficient data were available. The 597 children were dis- tributed as indicated in Table 1. Table 1. Cell sizes by grade and sex for experimental and control schools (N - 597) v:— -_— :— ‘1 II Experimental School Control School Male Female Male Female Grade K 33 30 19 29 1 37 25 27 30 2 16 21 25 30 3 27 21, 26 25 4 21 15 15 20 5 26 31 18 20 Collection of Data The testing teamwwas comprised of two professors, three dOctoral candidates and four master's students from the Department of Physical Education, Michigan State University. The master's candidates were trained by the five senior investigators prior to the testing period. Parents of the children enrolled in both schools served as recorders. The initial round of testing began in early September and continued for one week at each school. The tests were administered in the regular classrooms by a member of the testing team.** When possible, only one testing session was administered to each class per day. When the number of sub-tests required that more than one.sub-test per day be administered, IA variety of reasons resulted in ten subjects being drapped from the initial population of 607 students. By far the most common was the fact that some subjects moved from the area between testing sessions. Thus, the total sample (N - 597) included kindergarten (111), grade one (119), grade two (92), grade three (99), grade four (71), grade five (95). **The pre-tests for grades 2 and 4 were administered by the Waverly School System as these two grades were scheduled by the School Board to be tested at that time (September 1969). 30 one session was scheduled in the morning and one in the afternoon. Make-up testing of-absentees was conducted on an individual basis byra member of the testing team. Procedures for administration were as outlined in the annuals provided by the publishers, Harcourt, Brace and World, Inc. The post-program.testing period began in early June and continued_for one' week at each school. The Tests Three tests, the Stanford Early School Achievement Test, the Stanford Achievement Test and the Otis-Lennon Mental Ability Test, formed the test battery administered to the children of grades K-5. Different forms of the same test were administered for pre-test and post- test data collection. Table 2 lists the tests administered to each grade at each testing session and Table 3 provides a summary of the sub-tests for academic achievement administered at each grade level. The Otis-Lennon Mental Ability Test was utilized in grades 1-5 and for the post-testing of the kindergarten children (no pre-test for mental ability was available for this group). Although many of the mental ability tests were administered in a number of parts, the total score was used to provide a comprehensive assessment of general mental ability or scholastic aptitude. Two of the most‘fundamental requirements of any measurement are that it be valid and reliable. Reliability coefficients and means of attaining validity for the Stanford Early School Achievement Test, the Stanford Achievement Test and the Otis-Lennon Mental Ability Test are presented below: a) Stanford Early School Achievement Test: Presented in Table 4 are the split-half (odd-even) reliability coefficients corrected by the 31 .9009 UGOBU>UH£U¢ UHOMfldumfl .umma unmao>mfino< Hoonum edema vuomnmum « a» M HH hHfluflQEflHm .udOfldfldlmHuo 3 HH QUNHVQEHUudH ”H4m m M HH hufludmfiflflm udOddfldlmHuO N H UuwflvflfihfludH. uHuH uHoq uamwnu¢ momma Awmmeuumomv usHHmm h HH MHGUGQEQHM udOfidmglmHuO 3 H fludfivflfihmudH uHmH 1Hmg suaoanu< . mouse HommHIouniHHuH momuw some ou monoumwoflavm Aaron use many mumou huaaaom Houses one udosm>ownum afiamomum .N manna V I" \l 32 Table 3. Academic achievement sub-tests administered by grade W I_(g Grade 1 1. The Environment 1. Word Reading 2. Mathematics 2. Paragraph Meaning 3. Letters and Sound 3. Vecabulary 4. Oral Comprehension 4. Spelling 5. Word Study Skills 6. Arithmetic Grade 2 Grade 3 1. Word Meaning 1. Word Meaning 2. Paragraph Meaning 2. Paragraph Meaning 3. Spelling 3. Science and Social 4. WOrd Study Skills Studies Concepts 5. Arithmetic 4. Spelling 5. Word Study Skills 6. Language 7. Arithmetic Comprehension 8. Arithmetic Concepts Grade 4 Grade 5 l. Werd Meaning 1. WOrd Meaning 2. Paragraph Meaning 2. Paragraph Meaning 3. Spelling 3. Spelling 4. Wbrd Study Skills 4. Language 5. Language 5. Arithmetic Comprehension 6. Arithmetic Comprehension 6. Arithmetic Concepts 7. Arithmetic Concepts 7. Arithmetic Application 8. Arithmetic Application 8. Social Studies 9. Science Spearman-Brown Prophecy Formula. The coefficients obtained are of the magnitude expected since each sub-test is intentionally short, and relia- bility is expected to increase with test length. No mention is made of validity with respect to the SESAT battery, although content or curricular validity is undoubtedly assumed. b) Stanford Achievement Test: The medians of the split-half relia- bilities obtained on the sub-tests for each of grades 1-5 respectively were .89, .89, .91, .91, .89. For complete reliability data, by sub-test 33 Table 4. Reliability coefficients for each sub-test of SESAT: Level 1* Subtest Kg Grade 1 1. The Environment .85 .82 2. Mathematics .79 .82 3. Letters and Sounds .79 .89 4. Oral Comprehension .76 .77 * Reliability coefficients for Level II are presently being con— structed and thus are not available at this time. and by grade, the reader is referred to Kelley et al. (1964), Stanford Achievement Test: Technical Supplement (Harcourt, Brace and Werld, Inc., 1966). Content or "curricular" validity is assumed by the authors to have been met. c) Otis-Lennon Mental Ability Test: Presented in Table 5 are the split-half (corrected by the Spearman-Brown Prophecy Formula) and the Table 5. Split-half and Kuder-Richardson reliability coefficients by grade (Form J) Reliability Coefficients Grade* Split-Half K—R #20» 1 .90 .90 2 .89 .88 3 .92 .91 4 .94 .93 5 .95 .95 * No coefficients available for kindergarten. 34 Kuder-Richardson #20 reliability coefficients. The validity of the Otis- Lennon tests is based on.content or "curricular" validity. With respect to the scoring of the tests, the pre-tests for grades 2 and 4 were scored by the classroom teachers in the two schools. The pre-tests for grade one were machine-scored by the publishing company as these data were to be incorporated in national norms. The remainder of the pre-tests and all the post-tests were scored by the research team from Michigan State University. The Program The children in grades K-5 attending the experimental school received a one-year physical education program based on the curriculum developed for the Battle Creek Public School System.* This curriculum was constructed jointly, over a faur-year period, by physica1.education specialists at Michigan State University and the members of the Battle Creek Physical Education Curriculum Project from the Battle Creek Public Schools. The program was a research-based, objective-centered, sequen- tially developed program of physical education, covering a wide range of skills and activities. Although a great deal of emphasis was placed on the development of perceptual-motor skills, the program can in no way be compared to those programs specifically constructed to increase perceptual-motor functioning (McCormick, 1968; Delacato, 1966; Barsch, 1965; Frostig, 1964; Kephart, 1960). Stress was placed on the total development of the child and particular activities were presented to enhance social, emotional, cognitive and physical functioning. Theo activities presented were consistent with the objectives and sequences * Unpublished material, Curriculum Study Center, Room 40, women's Intramural Building, Michigan State University, East Lansing, Michigan. 35 previously determined by the Battle Creek Curriculum Project. Thus, the present age, grade level and state of the learner dictated the particular activities that he or she would undertake at each stage of his or her development. The experimental physical education program was conducted from September 1969, immediately following the pre-testing period, to May-1970, immediately preceding the post-testing period. The number and duration of classes was administered according to the following schedule: Grade K: 3 classes/week - 20 minutes/class l: 3 classes/week - 30 minutes/class 2: 3 classes/week - 30 minutes/class 3° 2 classes/week - 35 minutes/class 4: 2 classes/week - 35 minutes/class 5: 2 classes/week - 45 minutes/class Two teachers, both qualified female physical education instructors, were employed on a half-time basis by the Waverly School District. The teachers were rotated such that each instructor was exposed to all grade levels, thus helping to eliminate the teacher bias. An established elementary school teacher at the experimental school was assigned as a coordinator for the program. The control school received no formal physical education classes throughout the entire duration of the study. Statistical Treatment of the Data Chi square tests of hypotheses of no association between dependent and independent variables were calculated for each grade level. These tests were computed in order to provide statistical evidence as to the apprOpriateness of covariance analyses in this study. Stepdwise regression analyses, indicating the contribution of each independent variable, were calculated for each co-variable used in the analysis, i.e., a separate step-wise regression analysis was computed 36 for each sub-test at every grade level. These analyses were conducted in order to obtain an estimate as to the necessity, or contribution, of each co-variable used in the study. A series of multivariate covariance analyses were employed to determine if significant differences existed between the control and experimental groups with respect to the dependent variables, the post- test academic achievement and mental ability scores. Pre-test scores were used as the independent variables (co-variates) in the analyses. Each grade was treated separately in a multivariate analysis, and within each grade, males and females were analyzed individually. The academic. achievement sub-test scores were converted to standard scores (T-scores) in order to obtain a measure of total academic achievement. However, the multivariate covariance analysis was carried out on the raw scores obtained for mental ability. The nature of the dependent variables necessitated that two separate multivariate covariance analyses be conducted for each hypothesis, at all grade levels. Total academic achievement and mental ability were investigated in one analysis and the academic achievement sub-tests comprised the second analysis. The linear relationship between the academic achievement sub-tests and total academic achievement precluded the use of a single multivariate covariance analysis. The multivariate covariance program yielded three separate F values: a) multivariate; b) univariate; and c) step-down F values. The step-down F values were included in order to obtain an estimate of the relative contribution of each variable involved in the analysis. It is possible to enter the variables to be analyzed in any order. For example, if 37 AAS3* was entered as the fifth variable in the analysis, the experimenter would be able to investigate the relative contribution of AAS3 after accounting for the variability due to the first four variables. Upon completion of the initial covariance analyses, all programs were re-ordered on the basis of the step-down F values, in a further attempt to obtain reliable estimates of the contribution of each variable involved. The variables were re-ordered on the basis of their contribu- tion to the overall multivariate F value. Those sub-tests which appeared to contribute the most to the variability noted in the multivariate F value were placed first in the subsequent analysis. Thus, the variables were ranked in order, from highest to lowest, with respect to their overall importance in accounting for the multivariate F value obtained. Only those sub-tests which obtained a step-down value of p<.10, for at least one of the hypotheses under consideration (school, sex, interaction), were included in the re-ordered program. When interpreting the F values obtained in the re-ordered co- variance analyses, emphasis was placed firstly on the multivariate F values, secondly on the univariate values, and finally on the step-down values. In fact, due to the susceptibility of the step-down F values to the-order in which the variables were placed in the analysis, little attention was placed on these values once the re-ordering had been com- pleted. In particular instances in which the univariate values were very close to the level of significance required for rejection of the null hypothesis, step-down values were consulted in making the decision with respect to rejecting or accepting the null hypothesis in question. * Academic achievement sub-test three. 38 The multivariate alpha level necessary for rejection of the null hypothesis was set at o = .05. Similarly, the step-down values (when"‘ consulted) were also set at a - .05. The univariate values were set at H', where o = .05, and n = the number of variables (tests) involved in the analysis. For example, if the re-ordered covariance program contained five sub-tests, the alpha value required for rejection of the null hypothesis would be .01. This procedure is in accord with that set forth by the Bonferroni Equality (1966), in that the Bonferroni effect states that the division of the alpha level (i.e., H') does not need to be made in equal amounts. In other words, where the univariate alpha is equal to .01, as noted above, this value of .01 can be adjusted slightly to allow univariate F values that come very close to .01 to be rejected. Thus, in this paper, other criteria are utilized (step— down F values) in making the decision of whether or not to reject the null hypothesis in particular cases where the univariate F value approaches a that value required for rejection (i.e., n or, in this case, .01). All computations were carried out on the CDC 3600 computer, with the exception of the T-score conversions, which were computed on the CDC 6500 computer, at Michigan State University. CHAPTER IV RESULTS The results of this study will be presented in the following order: 1. The results of Chi Square tests of hypotheses of no association. between independent and dependent variables. 2. The results of stepdwise regression analyses indicating the contribution of each independent variable. 3. Separate analyses of covariance, by grade, presented in the following order: a) b) C) d) e) Least square estimates, adjusted for covariates; Pre-test and post—test cell means for experimental and control schools; Multivariate* and univariate covariance tests for School effects: Academic Achievement Total and Mental Ability;g Multivariate and univariate covariance tests for School effects: Academic Achievement sub-tests; Multivariate and univariate covariance tests for Sex effects: Academic Achievement Total and Mental Ability; * . In the case where only one variable was analyzed, a univariate analysis is presented alone. 39 40 f) Multivariate and univariate covariance tests for Sex effects: Academic Achievement sub-tests; g) Multivariate and univariate covariance tests for School x Sex interaction: Academic Achievement Total and Mental Ability; h) Multivariate and univariate covariance tests for School x Sex interaction: Academic Achievement sub-tests. 4. Summary Table of Results. Chi Square Tests of Hypotheses of No Association Between Dependent and Independent Variables The results of Chi Square tests of hypotheses of no association between dependent and independent variables indicated P values less than .0001, for all grades., It_was thus concluded that the covariance analyses were appropriate for use in this study. Step4Wise Regression Analyses The results of stepdwise regression analyses depicting the contri- bution of each independent variable (co-variable) indicated P values less than .01 for all but two co-variables used across all six grades. The majority of the P values were less than .0001. It was thus concluded that all the cavvariables used, with the exception of two, were necessary for the analysis of the data in this study. Multivariate and Univariate Analyses of Covariance for School, Sex and Interaction Effects Least square estimate tables will be presented for each grade level. With respect to interpretation, the sign (minus or plus) indicates the direction of the difference between a) experimental and control schools and b) males and females. For the main effect of Schools, a minus sign 41 indicates that any difference, whether significant or not, is in favor of the control school. For the main effect of Sex, a minus sign indicates that the difference is in favor of the females. Plus values indicate differences in favor of the experimental school, or in the case of the Sex main effect, males. The numerical values indicate the magnitude of the differences. However, significant differences are noted in the tables that follow (univariate and multivariate analysis of covariance tables). The reader, finding a significant difference, should then consult the least square estimates table in order to locate the direction (experimental-control,male-female) and magnitude of the difference. Thus, the least square estimates tables, presented at the beginning of- the analyses for each grade level, are for further reference, once sig- nificant differences have been located in the covariance analyses. Tables of cell means for experimental and control schools, males and females, are also presented for each grade level. This allows the reader to refer to the actual cell means for pre—test and post-test scores. Thus, further evidence is provided for use in interpreting any significant-F values noted in the following covariance analyses. Kindergarten A summary of the least square estimates adjusted for covariates for the re-ordered variables is presented in Table 6. The direction of change for the main effects of School and Sex are indicated as positive and negative values. Table 7 provides a summary of the pre-test and post—test means for the re-ordered variables. Cell means are presented for both schools, experimental and control, and both sexes, male and female. 42 * Table 6. Least square estimates adjusted for covariates: re-ordered variables for kindergarten Variables AA83** AA81 AA TOT Hyp. 1: School -8.066 -0.950 -2.l4l Hyp. 3: Interaction -0.l49 -2.768 0.267 * See Chapter III, p. 37, for a discussion on the re-ordering procedures. **Throughout the presentation of the results of this study, aca- demic achievement sub-tests will be represented in this form, i.e., AAS3 denotes academic achievement sub-test 3, AASl denotes academic achieve- ment sub-test 1. Similarly, AA TOT will be used to denote academic achievement total and M.AB. to denote mental ability (I.Q.). Table 7. Cell means for experimental and control schools: re-ordered variables for kindergarten Variables AAS3 AASl AA TOT Pre Post Pre Post Pre Post. Exp. Males 45.409 44.327 52.182 50.921 48.877 48.992 Exp. Females 49.247 47.457 50.263 48.813 51.003 49.133 Cont. Males 52.300 54.721 51.595 54.016 52.201 53.475 Cont. Females 54.486 56.003 46.200 47.524 48.797 49.766 A summary of the results of the univariate analyses of covariance test for School effects (AA TOT) is presented in Table 8. 43 * Table 8. Univariate analysis of covariance test for school effects: kindergarten academic achievement total II D.F. - 1 and 106 Between Univariate» P Less variable Eben Square F Than AA TOT 107.2693 4.5501 .0353 * No mental ability test for kindergarten. The univariate analysis of covariance test for School effects (AA TOT) was significant (P<.0353) in favor of the control group, thus allowing for the rejection of the null hypothesis of no difference between schools at the kindergarten level. A summary of the results of the multivariate and univariate tests of the mean vectors for Schools (AA sub-tests) is presented in Table 9. Table 9. Multivariate and univariate analysis of covariance tests for school effects: kindergarten academic achievement sub-tests D.F. - 2 and 104 F - Ratio - 16.3254 P<0.0001 Between Univariate P Less Step- P Less Variable Mean Square F Than Down F Than AAS3* 1587.0239 31.6761, .0001 31.6761 .0001 AASl 3.9949 0.0506 .8226 0.9806 .3244 D.F. for Hypothesis - 1 D.F. for Error - 105 * , Academic Achievement sub-test 3. Multivariate and step-down.alpha level required for rejection = .05. univariate alpha level required for rejection - .025. 44 The multivariate main.effect for Schools (AA Sub-tests) was.ata- tistically significant (P<.0001) in favor.of the control school. Of interest is the fact that AASB, Letters and Sounds, accounted almost” entirely (P<.0001) for the significant difference noted between schools. A.summary of the_results of the univariate analysis of covariance test for Sex effects (AA TOT) is presented in Table 10. Table 10. Univariate analysis of covariance test for Sex effects: kindergarten academic achievement total D.F. I 1 and 106 Between Univariate P Less' Variable Mean Square F Than AA TOT 48.0541 2.0383 .1564 Univariate alpha level required for rejection - .05. The univariate main effect of Sex was not statistically significant (P<.1564). A summary of the results of the multivariate and univariate tests of the mean vectors for Sex (AA sub-tests) is presented in Table.ll. The multivariate main effect for sex was statistically significant (P<.0290). Of interest, however, is the fact that neither AAS3 or AASl are significant by themselves when considered in the univariate analysis. This indicates that the two sub-tests do not contribute significantly, when taken individually, to the overall significant difference for sex effects. However, considered together, the two sub-tests account for a significant amount of the variability between.sexes at the kindergarten 45 level. Of note also is the fact that AASl, the Environment, indicated a greater increase by the males, while AASB, Letters and Sounds, showed superior performance by the females. Table 11. Multivariate and univariate analysis of covariance tests for Sex effects: kindergarten academic achievement sub-tests D.F. = 2 and 104 F - Ratio = 3.6674 P<0.0290 Between Univariate P Less Step- P Less Variable. Mean Square F Than Down P Then AASB 134.4902 2.6844 .1044 2.6844 .1044 AASl 235.0893 2.9752 .0875 4.5595 .0351 D.F. for Hypothesis - 1 D.F. for Error - 105 Multivariate and step-down alpha levels required for rejection - .05. Univariate alpha level required for rejection - .025. A summary of the results of the univariate analysis of covariance test for School x Sex interaction (AA TOT) is presented in Table 12. The interaction of School x Sex (AA TOT) was not statistically significant (P<.889l). A summary of the results of the multivariate and univariate covari— ance tests for the School x Sex Interaction (AA Sub-test) is presented in Table 13. The interaction of School x Sex (AA Sub-tests was not statistically significant (P<.7228). 46 Table 12. Univariate analysis of covariance test for School x Sex interaction effects: kindergarten academic achievement total D.F. = l and 106 Between Univariate P Less variable Mean Square F Than AA TOT 0.4613 0.0196 .8891 Univariate alpha level required for rejection 8 .05. Table 13. Multivariate and univariate analysis of covariance tests for School x Sex interaction: kindergarten academic achievement sub-tests D.F. = 2 and 104 F - Ratio - 0.3257 Between Univariate P Less Step- P Less Variable Mean Square F Than Down F Than AAS3 0.1469 0.0029 .9570 0.0029 .9570 AASl 50.4483 0.6384 .4261 0.6484 .4226 D.F. for Hypothesis - l D.F. for Error 8 105 Multivariate and step-down alpha levels required for rejection - .05. Univariate alpha level required for rejection = .025. Grade One A summary of the least square estimates adjusted for covariates for the re-ordered variables is presented in Table 14. The direction of change for the_main effects of School and Sex are indicated as positive and negative values. 47 Table 14. Least square estimates adjusted for covariates: re—ordered variables for grade 1 Variables AASl AAS4 AASS AAS6 AA TOT M.AB. Hyp. 1: School -5.203 -3.494 —7.000 -2.425 -3.073 -3.269 Hyp. 2: Sex -0.523 -0.061 -O.374 -3.351 -O.996 0.487 Hyp.-3: Interaction 6.119 -7.483 -3.338 0.381 -0.814 -0.529 Table 15 provides a summary of the pre-test and post-test means for the re-ordered variables. Cell means are presented for both schools, experimental and control, and both sexes, male and female. A summary of the results of the multivariate and univariate tests of the mean vectors for Schools (AA TOT and M.AB.) is presented in: Table 16. The multivariate main effect for Schools was statistically signifi- cant (P<.0002) in favor of the control school. Thus, the null hypothesis of no difference between schools was rejected at the Grade 1 level. It is noted that AA TOT alone (P<.0002) is significant in the univariate and step-down.analyses, thus indicating that the greater proportion of the overall variability between schools is accounted for by AA TOT. A summary of the results of the multivariate and univariate tests of the mean vectors for Schools (AA sub-tests) is presented in Table 17. The multivariate main effect for Schools (AA Sub-tests) was sta- tistically significant (P<.0001) in favor of the control school. Thus, the null hypothesis of no difference between schools was rejected at the 48 Table 15. Cell means for experimental and control schools: re-ordered variables for grade 1 1| Variables AASl AAS4 AASS Pre Post Pre Post Pre Post Exp. Males 51.068 50.224 50.532 47.108 47.114 45.530 Exp. Females 49.116 47.568 53.520 52.216 48.412 49.152 Cont. Males 54.482 49.822 49.389 52.159 51.370 51.937 Cont. Females 47.173 51.600 46.940 47.943 53.650 52.620 AAS6 AA TOT M.AB. Pre Post Pre Post Pre Post Exp. Males 52.681 45.130 49.757 48.137 43.000 39.054 Exp. Females 55.496 51.132 51.531 50.669 43.040 39.840 Cont. Males 44.474 50.815 49.476 50.508 37.963 40.518 Cont. Females 47.107 72.056 49.495 51.282 38.900 40.133 Table 16. Multivariate and univariate analysis of covariance tests for school effects: grade 1 academic achievement total and mental ability D.F. = 2 and 112 F - Ratio = 9.3328 P<.0002 Between Univariate P Less Step- P Less Variable Mean Square. F Than Down F Than AA TOT 265.1438 15.7939 .0002 15.7939 .0002 M.AB. 265.6993 4.6708 .0328 2,6423 .1069 D.F. for Hypothesis - l D.F. for Error 3 113 4 Multivariate and step-down.a1pha levels required for rejection - .05. Univariate alpha level required for rejection - .025. 49 Table 17. Multivariate and univariate analysis of covariance tests for school effects: grade 1 academic achievement sub-tests D.F. a 4 and 108 F - Ratio - 9.0915 P<.0001 Between Univariate P Less Step- P Less Variable Mean Square' F. Than Down P Then AASl‘ 505.1982 9.3047 .0029. 9.3047 .0029 AAS4 280.7770 3.0501 .0835 5.3397 .0028 AASS 1016.3902 11.3220 .0011 9.8161 .0023 AAS6. 738.7188 0.9811 .3241 8.5389 .0043 D.F. for Hypothesis - l D.F. for Error - 111 Multivariate and step-down alpha levels required for rejection - .05. Univariate alpha level required for rejection - .0125. Grade 1 level. Of interest is the fact that AASl and AASS appeared sig- nificant in both the univariate and step-down analyses and are thus con- sidered to be the greatest contributors to the overall rejection of the School hypothesis. It is also interesting to note that the step-down value for AAS6 (P<.0043) is also significant. This means that after, taking out the variance due to academic achievement sub-tests one, four and five (AASl, AAS4, AASS), academic achievement sub-test six (AAS6) is still significant when considered together with the other three sub-tests. Similar findings are in evidence for the fourth re-ordered variable, AAS4: Spelling. However, the reader will note the criteria for sig- nificance as stated earlier (i.e., a sub-test must be significant, or extremely close to significant in the univariate analysis before consider- ation would be given to the step-down values). Thus, in this analysis, 50 the only variables considered significant are AASl: word Reading, and AASS: word Study Skills. A summary of the results of the multivariate and univariate tests of the mean vectors for Sex (AA TOT and M.AB.) is presented in Table 18. Table 18. Multivariate and univariate analysis of covariance tests for Sex effects: grade 1 academic achievement total and mental ability D.F. = 2 and 112 F - Ratio - 0.9846 P<.3769 Between Univariate P Less Step- P Less Variable Mean Square P Then Down F Than AA TOT 29.3502 1.7483 .1888 1.7483 .1888 M.AB. 6.6926 0.1176 .7323 0.2327 .6305 D.F. for Hypothesis - l D.F. for Error - 113 Multivariate and step-down alpha levels required for rejection - .05. Univariate alpha level required for rejection = .025. The main effect of Sex (AA TOT and MJAB.) was not significant (P<.3769). A summary of the results of the multivariate and univariate tests of the mean vectors for Sex (AA sub-tests) is.presented in Table 19. The.main effect of Sex (AA sub-tests) was not significant (P<.7748). A summary of the results of the multivariate and univariate co- variance tests for the School x Sex interaction (AA TOT and M.AB.) is presented in Table 20. 51 Table 19. Multivariate and univariate analysis of covariance tests for Sex effects: grade 1 academic achievement sub-tests D.F. ‘ 4 and 108 F — Ratio - 0.4464 P<.7748 Between. Univariate P Less Step- P Less variable Mean Square F Than Down F Than AASl‘ 5.2913 0.0975 .7555 0.0975 .7555 AAS4 0.6108 0.0066 .9353 0.0217 .8833 AASS 4.5098 0.0502 .8231 0.0613 .8050 AAS6 5.3125 0.7183 .3986 1.6061 .2078 D.F. for Hypothesis - 1 D.F. for Error - 111 Multivariate and step-down alpha levels required for rejection - .05. Univariate alpha level required for rejection - .0125. Table 20. Multivariate and univariate analysis of covariance tests for School x Sex interaction: grade 1 academic achievement total and mental ability D.F. - 2-and 112 F - Ratio = 0.1487 P usuaunonou "unsouw unoauoouu Honuaoo use Huunoaauonxo sou ensue maoo .nq manna 73 A summary of the results of the multivariate and univariate tests of the mean vectors for Schools (AA TOT and M.AB.) is presented in Table 48. Table 48. Multivariate and univariate analysis of covariance tests for school effects: grade 5 AA TOT and MQAB. Ti ‘—- —— D.F. I 2 and 88 F - Ratio I 0.5910 P<0.5560 Between Univariate P Less Step- P Less Variable Mean Square F Than Down F Than AA TOT 6.9887 0.7312 .3948 0.7312 .3948 M.AB. 30.2392 0.7463 .3900 .4554 .5016 D.F. for Hypothesis I l D.F. for Error I 89 .44 Multivariate and step-down alpha levels required for rejection I .05. Univariate alpha level required for rejection I .025. The main effect of School (AA TOT and M.AB.) was not statistically significant (P<.5560). A summary of the results of the multivariate and univariate tests of the mean vectors for Schools (AA Sub-tests) is presented in Table 49. The main effect of Schools (AA sub-tests) was significant (P<.0005). The following sub-tests were considered significant after referring to the univariate F values: a) AAS9: Science (P<.009l)* * c Although the required value for significance was .0083, i.e., -, where n I the number of sub-tests and a I .05, it was considered realistic to include AAS9 (P<.0091) as significant. 74 Table 49. Multivariate and univariate analysis of covariance tests for school effects: grade 5 AA sub-tests D.F. I 6 and 80 F - Ratio I 4.6387 P<.0005 Between Univariate P Less Step- P Less Variable Mean Square F Than Down F Than . AASZ 53.2863 3.3650 .0701 3.3650 .0701 AAS9 162.9940 7.1451 .0091 4.6067 .0348 AAS3 34.2647 1.1705 .2824 3.9526 .0501 AASS 367.2261 6.6768 .0115 5.8167 .0182~ AAS6 3.1207 0.0755 .7842 0.1946 .6603 AAS8 148.4657 4.7965 .0313 7.4357 .0079 D.F. for Hypothesis Isl D.F. for Error I 85 Multivariate and step-down alpha levels required for rejection I .05. Univariate alpha level required for rejection I .0083. b) AASS: Arithmetic Comprehension (P<.0115)* Consideration of the_step-down F values necessitated the inclusion. of AAS8: Social Studies (P<.0079), together with AAS9 and AASS, in the list of sub—tests contributing significantly to the difference noted between schools in.academic achievement sub-tests. Of particular interest is the fact that the sub-tests noted as sig- nificant did not favor the same school. Whereas AAS9 favored the experi- mental school, AASS and AAS8 were found to be in favor of the control school. * Included as significant for the same reason as cited above'for the inclusion of AAS9. 75- A summary of the results of the multivariate and univariate tests of the mean vectors for Sex (AA TOT and M.AB.) is presented in Table 50. Table 50. Multivariate and univariate analysis of covariance tests for Sex effects: grade 5 AA TOT and M.AB. D.F. I 2 and 88 F - Ratio I 1.0825 P<.3433 Between univariate P Less Step- P Less Variable Mean Square F Than Down F Than AA TOT 0.1105 0.0116 .9147 0.0116~ .9147 M.AB. 80.4145 1.9847 .1624 2.1533 .1459 D.F. for Hypothesis I 1 D.F. for Error I 89 ._:__1 Multivariate and step-down alpha levels required for rejection I .05. Univariate alpha level required for rejection I .025. The main effect of Sex (AA TOT and M.AB.) was not statistically significant (P<.3433). A summary of the results of the multivariate and univariate tests of the mean vectors for Sex (AA.Sub-tests) is presented in Table 51. The main effect for Sex (AA sub-tests) was statistically signifi- cant (P<.0003). The following sub—tests were considered significant after referring to the univariate F values: a) AASZ: Paragraph Meaning (P<.0049) b) AAS9: Science (P<.0179)* *It was again considered realistic to include AAS9 as significant in that the univariate F value fell just short of significance. Further evidence for accepting AAS9 as significant was provided by an inspection of the step-down F value (P<.0013). 76 Table 51. Multivariate and univariate analysis of covariance tests for Sex effects: grade 5 AA sub-tests D.F. I 6 and 80 F - Ratio I 4.8582 P<.0003 Between Univariate P Less Step— P Less Variable Mean Square F Than Down F Than AASZ 132.3122 8.3556 .0049 8.3556 .0049 AAS9 133.1454 5.8366 .0179 11.1988 .0013 AAS3 4.4583 0.1523 .6974 0.5486 .4610 AASS 86.9934 1.5817 ' .2120 0.4310 .5134 AAS6 61.5680 1.4898 .2257 4.3474 .0403 AASS 70.1357 2.2659 .1360 2.2440 .1381 D.F. for Hypothesis I 1 D.F. for Error I 85 Multivariate and step-down alpha levels required for rejection I .05. Univariate alpha level required for rejection I .0083. Analysis of the step-down F values did not provide additional sig- nificant sub-tests. Of particular interest was the fact that AASZ was significant in favor of the females while AAS9 was in favor of the males. A summary of the results of the multivariate and univariate co- variance tests for the School x Sex interaction (AA TOT and M.AB.) is presented in Table 52. The interaction of School x Sex (AA TOT and MeAB.) was not sta— tistically significant (P<.2330). 77 Table 52. Multivariate and univariate analysis of covariance tests for School x Sex interaction: grade 5 AA TOT and M.AB. ‘— —-— D.F. I 2 and 88 F - Ratio I 1.4813 P<0.233O Between Univariate P Less Step- P Less Variable Mean Square F Than Down F Than AA TOT 5.0303 0.5263 .4701 0.5263 .4701 M.AB. 116.8763 2.8846 .0930 2.4279 .1228 D.F. for Hypothesis I 1 D.F. for Error I 89 Multivariate and step-down alpha levels required for rejection I .05. Univariate alpha level required for rejection I .025. A summary of the results of the multivariate and univariate covar- iance tests for the School x Sex interaction (AA Sub-tests) is presented in Table 53. The interaction of School x Sex (AA Sub-tests) was not statistically significant (P<.5048). Table 54 contains a summary of the significant main effects obtained in this study. The reader will note that these main effects are grouped by grade level. It should be remembered that two separate covariance analyses (a. AA TOT, M.AB.; b. AA sub-tests) were conducted for each hypothesis, at each grade level, and thus, in some cases, different multivariate significance levels are reported for the same main effect, i.e., Grade 1 school effects. 78 Table 53. Multivariate and univariate analysis of covariance tests for School x Sex interaction: grade 5 AA sub-tests D.F. I 6 and 80 F - Ratio I 0.8922 P<.5048 Between Univariate P Less Step- P Less Variable Mean Square F Than Down F Than AASZ 51.0994 3.2269 .0760 3.2269 .0760 AAS9 10.6358 0.4662 .4966 0.0122 .9123 AASB 79.3905 2.7120 .1033 1.2159 .2734 AASS 0.0624 0.0011 .9733 0.2574 .6133 AAS6 21.5890 0.5224 .4719 0.7809 .3795 AASS 11.5892 0.3744 .5423 0.0072 .9324 D.F. for Hypothesis I 1 D.F. for Error I 85 Multivariate and step-down alpha levels required for rejection I .05. Univariate alpha level required for rejection I .0083. 79 nowononounaoo Houusou naoov m 03353.3 ”34¢ 380 m Hooeom m. Hounuawuonxm goof m ooeuaom "mg mooo.v m Hooeom m mode: Houuaoo ._. + 1.33230 noauoouounH $336.1 . ea 83¢ a 636232 "$3. 239 a now a Hooeum n 3683636 28¢ m mango: 263 £92 380 .1 H6918 e. Houses—annexe” .88 .v m . Q .2 309v m Hoonom 1» 36836213 88 .v m 8.... 5. 38 .v .1 H628 s I I ionoz @506 m Now m +Houunou use sowuomnouaH Haunoaauuaxm I «eeodoz ago.v m Now a Hoonom N Houusoo Hoof m mafia—m 3.3m v.33 ”mm: 386m Hoonom H H6380 28 .v m 9:331 E 63 £3 88 .v .1 H6213. H Houusoo Nooo .v m you. << «000 .v m Hooeom H I I .5932 came .v M Show M muesom .. Houusoo Hooo.vm use 33qu ”mm: .88 .vm Hooeom M Houucou mmno.vm you. 3‘ «osoz Hooeom M mo H954 .wfiamwm 3.33.3.» 33.31, .325 . mesmem uoowwu can: mush—u uo>um 5 3336.8: Ian: . madmam ouuwuozud—z mIM seesaw “moowuocnounfl one 30030 new... ugowmenwwo mo hugs—5m .3 0.33. .Amaoo.vmv unuoamaswwm enema: use o=Hu> m naouInuum on» .Amnao.vmv uonuoawfiswwm mo uuosm kum usam> m uudauu>an= unu nanosuao unnu.nH uncoamwawwm mo .uonowom "mm<< uvnaonfl ou ofiumaaduu uuuuvwmnoo can uH mo .Aanoo.vmv unooawasmfim Manage was 05Hu> naovInuum on» .Amamo.vmv ooamoumwamwm no uuoem Hana u=Hc> m ouuwuc>fis= osu nwsonuau unnu sq unuoemeewwm we .muausum Hoaoom .munuonoo oauuanuau< Imam one mo nofiuuonoun owned a now wounoooom .wswsooz enouweuum vouuouunfi sofiuuunmnfi Hoeuunm .mm was mm momma new 80 0903 .usuanouwbnm use uwm<< mundane ou oaumaaouu conovwmsoo was uH o .AHwHo.vmV undefimanmam magmas was o=Hd> naovInoum oeu .Auomo. my oonooamwnwfim mo uuonm Adam unau> m uuuwuo>wns unu emsozuaw uonu sq unmoumwewam mm u~m<< you soauomuuunw Nam N Hooeom oeu ovaaonfi ou oeumaacou vuuuvaonoo was uH t; .uoomwo news ounwum>auana unmoaman nume< was .muauuaoo unuoaauaua "mm<< omen .unmowwfinwwm haamnva>wvnfi on? anon .moanmauu> vuuuvuolou snow on» «0 +1. .uonnuww mdoauomuousw Nam x Hoonom one use aw confluuno onus nuances uuamonn0+ .noauoeuuunw ouofiuw>wuana unmoamaamwm one now wouanooom menu .mwmhauno ousfiuu>fiuana onu ca nonuowou soMmu ..o. H .unmowmfiswfio maauovw>avsa uuoa ..m<. 2 no How << .moanafiuu> one onu mo Honuuuze "Hm¢< no .mvnsom use muuuuoq a5. .uoommo sans uuofiun>wuasa undo Iamaewwm one you vouenoooo menu .mammaweu ouuwuu>auana one ow nonuowou swamp ..o.H .uneoumanMHm haawovfi>wvefi "mm<¢ .muanmfluu> wouovuolou one mo Hoeuwoz «e .mammaosu menu ea manuwuo> Mano onu mos Hmuoa ususo>ofieo< ueaovoo< « «one: aaao.vm osmoamfiom “amee mooo.vm sum m mundane meoo.vm magnum: namummumm “Nmee mooo.vm sum m Houuaoo nnoo.vm mmofiesum Hmnuom "mm<< mooo.vm Hoonom m we Hm>ua .maamem moanmnua> unmana> Hu>oa .wfiamnm acumen new: means uo>mm nH ouawuo>finn Ian: .wfidmam ouafiun>fiuanz A.e.naoov en manna CHAPTER V DISCUSSION A discussion of the findings of this study with respect to the results reported in previous investigations is difficult, for a number of reasons. First, the scope of this study, in terms of the number of subjects and grade levels involved, is far greater than has been attempted in the majority of previous investigations. Secondly, previous studies have focused primarily on the effects of specific perceptual- motor programs on various academic parameters. Thus, the outcome of these studies is specific to these experimental programs, each with its own stated objectives, and thus are not easily related to the outcomes of this study, which attempts to evaluate the effects of a regular physical education program on normally functioning subjects. Thirdly, most of the perceptual-motor programs have been employed in an attempt to investigate the effects of such programs on reading achievement alone, rather than total academic achievement. Thus, it is within the above limitations that comparisons are drawn between the results of prior investigations and the present study. A number of conditions existed which must.be recognized before the results at individual grade levels are discussed. One of these condi- tions was the completeness of the program which was followed by the subjects in the experimental group. Although the program.was fully formulated (objectives, grade sequences, unit plans and so on) specific lesson plans were not fully developed for all grade levels. Therefore, 81 82 it is very difficult to evaluate the extent to which the guidelines of the physical education program were actually adhered. Another considera- tion, applicable at all six grade levels, is the possibility that a condition of "latent learning" may exist. In other words, it may take a period of two to three years, or longer, for the effects of the program to be fully identified. In the lower elementary grade levels, kindergarten and grade one, significant school effects were noted, all in favor of the control school. More specifically, in kindergarten, the control subjects exhibited significantly greater gains in total academic achievement as well as in one sub-test, Sounds and Letters. In grade one, greater increases were registered by the control subjects with respect to total academic achievement. Sub-test one, word Reading, and sub-test five, word Study Skills, also indicated significant gains in favor of the control school. No significant school effects were noted between experimental and control subjects in grade two. The results of the present investigation thus agree with those obtained by Brown (1968) and Rosen (1966), the former working with grade one subjects, the latter with pupils in grade two. Whereas Brown found greater gains in reading achievement by controls, Rosen noted greater increases in "idea comprehension", a selected measure of read- ing ability, by his control group. In fact, Rosen indicated that, based on his own study, additional time devoted to reading instruction was more important for reading achievement than time devoted to the types of perceptual training used in his study. The results obtained in the present study are thus in contradiction with those presented by Painter (1966), McCormick (1968), Boger (1952), and Lewis (1968), all of whom obtained significant increases in academic 83 functioning as a result of perceptual-motor training. Falik (1969), Jacobs (1968a), Jacobs (1968b), O'Donnell and Eisenson (1969) and Robbins (1967) noted no significant differences between experimentals and controls following perceptual-motor training programs introduced at the grade levels under consideration (K, 1, 2). As noted earlier, however, direct comparisons between the present study and those noted above are difficult, in that varying emphases were placed on perceptual— motor activities. A possible explanation for the highly significant school effect in favor of the control school at the kindergarten level is the differ- ing emphases of the two schools at this level. Whereas the control school is markedly structured and "academic-oriented", the experimental school adopted an approach stressing the "socialization" aspect to a much greater extent in its approach to teaching at the kindergarten level. Thus, it is possible that the highly academic approach apparent in the control school accounted for the differential increase in academic achievement. Similarly, the possibility exists that a carry- over of the results of this academic approach to learning at the kinder- garten level is experienced at the first grade level, thus helping to account for the highly significant school effects once again registered in favor of the control subjects. Nevertheless, the results presented for school effects at the kindergarten and first grade levels indicate that the objective-centered sequential program of physical education appeared to do little or nothing to counteract the academic approach of the control school. One might continue with this reasoning, in reference to the non-significant school main effects at the grade two level, sug- gesting that the waning influence of the highly academic program at the kindergarten level was finally counterbalanced by the positive effects 84 of the physical education program at the experimental school in grade two. To this writer, however, the foregoing explanation appears to be highly speculative and the non-significant results at the grade two level will not be subjected to further interpretation. Nevertheless, it is interesting to note that after reaching equality with respect to school main effects at the grade two level, the trend begins to swing toward the experimental school. Although non-significant multivariate values are evident in both analyses (AA TOT, MMAB., and AA Sub-tests) at the grade three level, the only variable which approaches significance is in favor of the experimental school. This variable, AASZ: Paragraph Meaning, approaches significance at the univariate level (P<.0311) and is in fact significant in the step-down analysis (P<.0309). Furthermore, four of the six academic variables, though not significant, favor the experimental school. At the grade four level, the shift with respect to the main effect of schools is complete. At this level, the experimental school exhibits significant superiority in terms of academic increases. Both total academic achievement and mental ability, as well as one sub-test, AASl: Word Reading, are statistically significant. These results concur with those presented by Ismail (1967), in the only investigation reported in the literature similar in nature to the present study.‘ While Ismail obtained significance in favor of the experimental group, in total academic achievement, non-significant results are reported with respect to mental ability. It is interesting to note that in the present investigation, both total academic achievement and mental ability were highly significant in favor of the experimental school at the fourth grade level. No reasons are readily apparent for the shift in experimental effects, first from superiority of the control school to apparent 85 non-significance, and finally to superiority, in terms of academic gains, to the experimental school. An obvious interpretation is to suggest that the physical education program had a beneficial effect, with respect to academic achievement, upon the upper elementary aged pupils, while exhibiting a negative effect on those enrolled in the lower grade levels. However, this conclusion is rendered untenable upon consulting the results presented for the grade five subjects. Two of three sig- nificant sub-tests indicate greater gains by the control school, thus contradicting the results obtained by Ismail as well as those presented in this study for grade four students. Also of interest is the fact that the three significant sub-tests noted at the grade five level are completely different than those noted as significantly different at earlier grade levels. Whereas previous significant differences involved primarily language arts variables (Sounds and Letters, word Meaning, word Reading and word Study Skills), the differences noted in grade five appeared in Arithmetic Comprehension, Social Studies and Science. At this time no logical explanation can be offered for such results. To this point, no mention has been made of the differential effects of the program on males and females. Although the null hypothesis of no sex differences was rejected at the kindergarten level, as well as in grades three and five, no significant trends are apparent. In two of the three significant multivariate sex main effects (level K and. grade 3), no individual univariate values reached significance. However, of those sub-tests appearing to account for the major portion of the variability between the sexes, the sub-tests differed as to favoring one sex or another, without exception. Stated otherwise, in each situa- tion where significant multivariate values were evident, one of the 86 major contributors to this variability would show larger increases by the males, and the other, the larger increase by the females. As indicated above, no single grade showed consistent increases by one sex or the other as a result of the physical education program. However, of interest is the fact that the females showed greater gains in the subject areas of Letters and Sounds (kindergarten controls), Paragraph Meaning (grade three experimentals and grade five experimentals). This indicates that the increases in the sub-test of Paragraph Meaning, in favor of the experimental school, were largely due to the gains made by the female pupils. Similar reasoning may be applied to the increases in the sub-test of Letters and Sounds by the control females. Conversely, the males accounted for the greater portion of the increases in sub- tests involving the Environment (kindergarten controls), Arithmetic Concepts (grade three experimentals), and Science (grade five experi- mentals). Thus, it can be seen that few trends have evolved with reference to the sex differences noted in the results of this study. Perhaps the most interesting point to emerge from a consideration of the sex main effect is the fact that the respective sexes showed greater gains in those subjects which have been traditionally considered "female- oriented" (language arts) and "male-oriented" (Arithmetic and Science). This is in opposition, however, to the results provided by Minuchin (1966), in which the author did not find the boys to be better "problems solvers" or the girls to excel in more "imaginative" and "introspective" subjects. The only significant school x sex interactions were noted for grade two and grade four subjects. Since contradictory results were obtained with respect to the performance of the males and females in the control and experimental schools, in terms of the respective increases in 87 academic achievement total and mental ability (grade two), little sig- nificance is attached to this statistically significant interaction. Of interest is the fact that Ismail (1967) reported a significant School x Sex interaction on the variable "mental ability" when studying the effects of.a physical education program on grade five and six pupils. Ismail found that higher mean intelligence scores in boys were associated with the experimental rather than the control groups, while in girls, Opposite results were obtained with higher mean intelli- gence scores being associated with the control rather than the experi- mental groups. In the present investigation, the significant interaction for grade four, AAS7, indicated opposite findings. Higher mean scores in boys were associated with the control school while, for girls, higher mean scores were found in the experimental school. Summary and Implications of the Discussion As indicated in the preceding paragraphs, the data obtained in this study indicate few consistent results when examined across the six grade levels, kindergarten through grade five. Trends are conspicuously absent with the exception of the main effects for School. As noted, the physical education program did not have a facilitating effect at the kindergarten or grade one levels. Following non-significant results in grade two, the experimental program appeared to slightly enhance the academic progress of those pupils enrolled in grade three at the experimental school. The shift continued and extremely significant increases in academic variables were noted by those pupils having under- gone the experimental program in grade four. Although the superiority of the experimental students was not established at the fifth grade level, nevertheless, a slight trend across the six levels was noticeable. 88 Investigating the relationships between perceptual and cognitive processes, Bibace and Hancock (1969) tested the oft-held theoretical assumption of the mastery of lower order (perceptual-motor) processes as necessarily prior to higher order (cognitive) processes, i.e., scholastic achievement. The results of their study led the authors to conclude that the theoretical assumption must be at least qualified and that the clinical-pedagogical practices based on this assumption needed to be re-examined. The present investigation supports the findings of these authors in that little evidence is produced to support the hypothesis that those individuals participating in a physical education program, consisting in part of perceptual training, achieve better than those not participating in such a program. A similar hypothesis (Myers and Hamill, 1969), in which the authors proposed that motor activity was a precursor to, or a useful adjunct to, skills which are essential for academic achievement, is another approach that requires further investigation in light of the findings of the present study. Seefeldt (1970) notes that: "...the results of perceptual-motor training pro- grams, with few exceptions, indicate that training is specific to the skill which has been practiced. Programs based on the hypothetical transfer of improvement in perceptual-motor skills to reading achievement or readiness for reading have generally met with failure." (p. 11) The above is not surprising in view of the fact that specificity of training, in both verbal and motor skills, has been well documented (Henry, 1958; Williams, 1969; Singer, 1966; Bachman, 1961; Henry and Rogers, 1960). Thus, it appears that "on the basis of present knowledge, we cannot assume that what is learned in one skill will transfer to another" (Seefeldt, p. 21). 89 In view of the results of this study, it is possible that the approach taken by Humphrey (1965) should be seriously considered. Humphrey is of the belief that direct teaching of academdc skills, through the medium of games and physical activity, is superior to those programs relying on transfer, with respect to producing changes in academic achievement. In this approach, the author devises active games in which academic concepts are directly involved. Thus the pupils, in order to participate maximally in the game situation, must understand the concepts necessary for participation in the activity. Thus, the possibility exists that academic concepts may be strengthened to a greater degree through the use of the physical education medium as employed by Humphrey, as opposed to the programs relying on transfer from one task to another. This is not to say, however, that the approach undertaken in this study is without value. Future investi- gations which avoid the limitations noted in this study may produce different results with respect to the effectiveness of motor-oriented programs on the academic functioning of the child. CHAPTER VI SUMMARY, CONCLUSIONS AND RECOMMENDATIONS Summary Purpose and Hypotheses The purpose of this study was to investigate the effects of a one- year, objective-centered, sequential program of physical education upon the academic achievement and mental ability of male and female children enrolled in kindergarten through grade five. Two hypotheses were thus proposed: (1) a one-year, objective-centered sequential program of physical education would not significantly affect the academic achieve- ment and mental ability of elementary school children, at any level, kindergarten through grade five; (2) a one-year, objective-centered sequential program of physical education would not affect males and females in a significantly different manner, at any level, kindergarten through grade five, with respect to academic achievement and mental ability. Procedures Two schools, one experimental and one control, were selected from the Waverly School District. The total sample included kindergarten (111), grade one (119), grade two (92), grade three (99), grade four (71), and grade five (95). Three tests, the Stanford Early School Achievement Test (SESAT), the Stanford Achievement Test (SAT), and the Otis-Lennon Mental Ability Test, formed the battery administered to the 90 91 children. The experimental program consisted of a physical education program stressing the total development of the child and thus, particular activities were presented to enhance his/her social, emotional, cognitive and physical functioning. The control school received no formal physical education classes throughout the entire duration of the study. A series of multivariate covariance analyses were employed to determine if sig- nificant differences existed between the control and experimental groups with respect to the dependent variables, the post-test academic achieve- ment and mental ability scores. Results 1. Statistically significant differences were noted, in favor of the experimental school, for academic variables in Grade 4 (Academic Achievement Total, Word Meaning, Mental Ability) and Grade 5 (Science). 2. Statistically significant differences were noted in favor of the control school for academic variables in kindergarten (Academic. Achievement Total, Letters and Sounds), Grade 1 (Academic Achievement Total, Word Reading, Werd Study Skills) and Grade 5 (Arithmetic Compre- hension, Social Studies)._ 3. Statistically significant differences were noted in favor of the females for academic variables in Grade 5 (Paragraph Meaning). 4. Statistically significant differences were noted in favor of the males for academic variables in Grade 5 (Science). 5. A statistically significant School x Sex interaction noted for Grade 2 (Mental Ability) indicated that within the experimental school, larger gains were registered by the females. Within the control school, larger increases were made by the males. Opposite results were obtained for Academic Achievement Total. 92 6. A statistically significant School x Sex interaction noted~for Grade 4 (Arithmetic Concepts) indicates that within the experimental school, the largest gains were registered by the females, while in the control school, opposite results were obtained. Conclusions Within the limitations of this study, the following conclusions are noted. 1. The one-year, objective-centered, sequential program of physical education did not facilitate the learning of academic concepts, as measured by standardized academic achievement and intelligence tests, of children in grades kindergarten through five. Although isolated effects were noted throughout the grade levels, no consistent trends were evident. 2. The one-year, objective-centered, sequential program of physical education had no differential effect on the males and females in the experimental group. Although isolated sex differences were in evidence, no consistent trends were noted throughout the study. Recommendations Further investigations concerned with the effects of physical edu- cation programs upon the academic functioning of elementary school children should consider the following points: 1. The emphases of participating schools or school systems should be more clearly evaluated, with respect to their approach to teaching, i.e., "academic" or "socialization" at the outset of the investigation. 2. Spurious results are often obtained when only one measure of a particular trait is taken. It is suggested that at least two measures of each academic ability, i.e., Arithmetic Concepts, should be included in future studies of this type. 93 3. More meaningful measures should be developed to evaluate the*‘ effects of perceptual-motor and physical education programs on academic~'* achievement and intelligence. It is very possible that many essentia1“'” knowledges and abilities are enhanced by motor-oriented programs, yet are not in evidence due to the nature of the evaluative instruments. Measures of perception, imagination and creativity are examples of variables which might be included in evaluative batteries for future program testing. 4. It was concluded by the author that many overriding variables (teacher effect, program orientation, completeness of material and instructions) resulted in the inability of this investigation to pro- duce clear and unbiased estimates of program effects. Thus, it is suggested that future studies attempt to control for those variables which may have obscured the findings of this investigation. 5. A closer examination should be taken of those programs (Humphrey, 1965), or modifications of such programs, emphasizing speci- ficity of training. To date, very little evidence is forthcoming to suggest that the learning of academic concepts will "transfer" from.a program of motor activities. However, if specific academic concepts are taught, through the medium of physical activity, more positive results may be in evidence. REFERENCES CITED REFERENCES CITED Alley, L. Perceptual-motor performances of mentally retarded children after systematic visual—perceptual training. American Journal of Mental Deficiency, 1968, 73, 247-250. Bachman, J. C. Specificity vs. generality in learning and performing two large muscle motor tasks. Research Quarterly, 1961, 32, 3-11. Bannatyne, A. D., and Wichiarajote, Penny. Relationships between written spelling, motor functioning, and sequencing skills. Jaurnal of'Learning Disabilities, 1969, 2(1), 4—16. Barsch, R. Basic Constructs oj’Mbviegenics. Seattle:- Special Child Publications, 1967. Barsch, R. A Moviegenic Curriculum, Bulletin NO. 25: State Department of Public Instruction, Madison, Wisconsin, 1965, Vol. 1. Bengston, C. M. The relationship between perceptual-motor development and motor performance of nine-year—old boys. Unpublished Master's Thesis, University of Colorado, 1966. Bentzen, Francis. Sex ratios in learning and behavior disorders. The National Elementary.Principal, 1966, 46, 13-17. Bibace, R., and Hancock, K. Relationships between perceptual and con- ceptual cognitive processes. Jaurnal of'Learning Disabilities, 1969, 2(1), 17-29. Biddulph, L. G. Athletic achievement and personal and social adjust- ment of high school boys. Research Quarterly, 1954, 25, 1-7. Bihlmeyer, E. W. A study of the relationship between motor ability and ‘the performance of male fourth grade pupils in selected school skills. Unpublished Ed.D. Dissertation, University of South Dakota, 1965. Birch, H. G., and Belmont, L. Auditory—visual integration, intelligence, and reading ability in school children. Perceptual and.Mbt0r Skills, 1965, 20, 295-305. Boger, J. H. An experimental study of the effects of perceptual train- ing on group intelligence test scores of elementary pupils in rural ungraded schools. Journal of’Eaucational Research, 1952, 46, 43-52. 94 95 Boyd, L., and Randle, K. Factor analysis of the Frostig Developmental Test of Visual Perception. Journal of’Learning Disabilities, 1970, 3, 253-255. Brace, D. K. Motor learning of feeble-minded girls. Research Quarterly, 1948, 19, 269-275. Braley, W. T., Konicki, G., Leedy, C. Daily Sensorimotor Training Activities. Freeport: Educational Activities, Inc., 1968. Bryan, Q. The relative importance of intelligence and visual perception in predicting reading achievement. California Jaurnal of’Educa- tional Research, 1964, 15, 44—48. Brown, R. C. The effect of a perceptual-motor education program on perceptual-motor skills and reading readiness. Presented at research section, AAHPER National Convention, St. Louis, Mo. Burley, L. R., and Anderson, R. L. Relation of jump and reach measures of power to intelligence scores and athletic performance. Research Quarterly, 1955, 26, 28-35. Carton, A. M. Relationship of auditory-motor rhythm to reading achieve- ment. Unpublished Master's Thesis, Purdue University, 1963. Chang, T. M. C., and Chang, v. A. C. Relation of visual-motor skills and reading achievement in primary grade pupils of superior- ability. Perceptual and Motor Skills, 1967, 24, 51-53. Chansky, N. M. A study of perceptual training with elementary school under-achievers. Jaurnal of’School Psychology, 1963, 1, 33—41. Chansky, N. M., and Taylor, M. Perceptual training with young mental retardates. American Jaurnal ofTMental Deficiency, 1964, 68, Coleman, H. M. Visual perception and reading dysfunction. Jaurnal of Learning Disabilities, 1968, 1(2), 26-33. Delacato, C. H. The Diagnosis and Treatment of Speech and Reading Problems. Springfield, Illinois: Thomas, 1963. Dalacata, C. H. Neurological Organization and.Reading. Springfield, Illinois: Thomas, 1966. DiGiovanna, V. G. A comparison of the intelligence and athletic ability of college men. Research Quarterly, 1937, 8, 96-106. Falik, L. H. The effects of special perceptual-motor training in kinder- garten on reading readiness and on second grade reading performance. Journal of'Learning Disabilities, 1969, 2(8), 395-402. 96 Francis, R. J., and Rarick, G. L., Motor characteristics of the mentally retarded. American Journal of'Mental Deficiency, 1959, 63, 792-811. Fretz, B. R., Johnson, W. R., and Johnson, J. A. Intellectual and per- ceptual-motor develOpment as a function of therapeutic play. Research Quarterly, 1969, 40, 687-691. Frostig, M., Maslow, P., LeFever, W., and Whittlesey, R. The Marianne Faustig Developmental Test of Visual Perception. Palo Alto, California: Consulting Psychologists Press, 1964. Hackensmith, C. W;, and Miller, L. A comparison of academic grades and intelligence scores of participants and non-participants in. intramural athletics at the University of Kentucky. Research Quarterly, 1938, 9, 55-56. Hall, C. 8., and Lindsey, G. Theories of'Personality. New York: Wiley & Sons, 1957. Halsey, E. The relation between various measures of mental and physical traits. Research Quarterly, 1938, 9, 55-56. Heath, 8. Roy. Rail-walking performance as related to mental age and etiological type among the mentally retarded. American Journal of'Psychology, 1942, 55, 240-247. Henry, F. M. Specificity vs. generality in learning motor skills. Proceedings College Physical Education Association, 1958, 61, Henry, F. M., and Rogers, D. Increased response latency for complicated movements and a "memory drum" theory of neuromotor reaction. Research Quarterly, 1960, 31, 448-458. Howe, R. A comparison of motor skills of mentally retarded and normal children. Eaceptional Children, 1959, 25, 352-354. Humphrey, J. H. Child Learning Through Elementary School Physical Education. Dubuque, Iowa: wm. C. Brown Compnay, 1965. Ismail, A. H., and Gruber, J. J. Predictive power of coordination and balance items in estimating intellectual achievement. Proceed- ings of 1st International Congress on Psychology of Sport, Rome, Italy, April, 1965. Ismail, A. H. The effect of a well organized physical education program on intellectual performance. Research in Physical Education, 1967, 1(2), Ismail, A. H., Kane, J., and Kirkendall, D. R. Relationship among intellectual and non-intellectual variables.. Research Quarterly, 1969, 40, 83I92. 97 Ismail, A. H., Kephart, N., and Cowell, C. C. Utilization of motor aptitude test batteries in predicting academic achievement. Technical Report No. 1, Purdue University Research Foundation, August, 1963. Jacobs, J. N., Wirthlin, L. D., and Miller, C. B. A follow-up evalua- tion of the Frostig Visual-Perceptual Training Program. .Educa- tional Leadership, 1968b, 26, 169-175. Jacobs, J. N. An evaluation of the Frostig Visual-Perceptual Program. Educational Leadership, 1968a, 25(4), 332-340. Johnson, G. B. A study of the relationship that exists between physical skill as measured and general intelligence of college students. Research Quarterly, 1942, 13, 57-59. Jones, R. H. A comparison of the intelligence of high school athletes with non-athletes. School andISociety, 1935, 42, 415-416. Kelly, T. L., Madden, R., Gardner, E. F., and Rudman, H. C. Stanford Achievement Test. New York: Harcourt, Brace and World, Inc., 1964. Kephart, N. C. The Slow Learner in the Classroom. Columbus, Ohio: Charles E. Merrill Publishing Company, 1960. Kettels, A. School is for girls. Michigan Education Journal, 1967, November, 45, 15-16. Kulcinski, L. E. The relation of intelligence to the learning of funda- mental.musculsr skills. Research Quarterly, 1945, 16, 266-276. Landis, M. H., Burtt, H. E., and Nichols, J. H. The relation between physical efficiency and intelligence. American Physical Educa- tion Review, 1923, 28, 220-221. Lewis, J. N. The improvement of reading ability through a developmental program in visual perception. Journal of'Learning Disabilities, 1968, 1, 23-25. McCormick, C. C., Schnobrick, J. N., Footlik, S. W., and Poetker, B. Improvement in reading achievement through perceptual-motor training. Research Quarterly, 1968, 39, 627-633. Malpass, L. F. Motor proficiency in institutionalized and non-institu- tionalized retarded children and normal children. American Jaurnal of'Mental Deficiency, 1960, 64, 1012-1015. Miller, Rupert. Simultaneous Statistical Inference. New York: McGraw, Hill, 1966. Minuchin, P. Sex differences in children: research findings in an education context. The National Elementary.Principal, 1966, 46, .4546. 98 Myers, P. 1., and Hamill, D. D. Methods for Learning Disorders. New York: Wiley & Sons, 1969.~ O'Donnell, P. A., and Eisenson, J. Delacato training for reading achieve- ment and visual-motor integration. Jaurnal of’Learning Disa- bilities, 1969, 2(9), 441-447. Olson, A. V. Relation of achievement test scores and specific reading abilities to the Frostig Developmental Test of Visual Perception. Perceptual and Motor Skills, 1966, 22, 179-184. Olson, W. C. Child Development. Boston: D. C. Heath, 1949. Otis, A. S., and Lennon, R. T. Otis-Lennon Mental Ability Test. New York: Harcourt, Brace and World, Inc., 1968. Painter, G. The effect of a rhythmic and sensory motor activity program on perceptual motor spatial abilities of kindergarten children. Exceptional Child, 1966, 33, 113-116. Plack, J. T. Relationship between achievement in reading and achievement in selected motor skills in elementary school children. Research Quarterly, 1968, 39, 1063-1068. Rarick, G. L., and McKee, R. A study of twenty third-grade children exhibiting extreme levels of achievement on tests of motor pro- ficiency. Research Quarterly, 1949, 20, 142-152. Ray, H. C. Interrelationships of physical and mental abilities and achievements of high school boys. Research Quarterly, 1940, 11, 129-141. Reels, W. H., and Rees, R. C. High school lettermen: their intelligence and scholarship. SChool Review, 1939, 47, 534-539. Robbins, M. P. A study of the validity of Delacato's Theory of Neuro- logical Organization. Eaceptional Children, 1966, 32, 517-523. Robbins, M. P. Test of-the Doman-Delacato rationale with retarded readers. Journal of’American Medical Association, 1967, 202, 87-91. Rosen, C. L. An experimental study of visual perceptual training and reading achievement in first grade. .Perceptual and.Mbtor Skills, 1966, 22, 979-986. Rudisill, E. S. Correlations between physical and motor capacity and intelligence. School andISociety, 1923, 18, 178-179. Ryan, E. D. Relative academic achievement.and stabilometer performance. Research Quarterly, 1963, 34, 185-190. Seefeldt, Vern. Perceptual-motor programs:- their implicatiOns for physical education. Paper presented at New Hampshire Association of Health, Physical Education and Recreation, Plymouth State College, Plymouth, New Hampshire, March, 1971. 99 Seefeldt, Vern. Perceptual-motor skills. Chapter 3 in Montoya: An Introduction to Measurement in Physical Education (vol. 2). Indianapolis, Indiana: Phi Epsilon Kappa Fraternity, 1970. Seegers, J. C., and Postpichel, 0. Relation between intelligence and certain aspects of physical activity. Journal ofTEducational Research, 1936, 30, 104-109. Singer, R. N. Interrelationship of physical, perceptual-motor, and academic achievement variables in elementary school children. Perceptual and Motor Skills, 1968, 27, 1323-1332. Singer, R. N., and'Brunk, J. W. Relation of perceptual-motor ability and intellectual ability in elementary school children. Per- ceptual Motor Skills, 1967, 24, 967-970. Singer, R. N. Interlimb skill ability in motor skill performance. Research Quarterly, 1966, 37, 406-410. Skubic, V., and Anderson, M. The interrelationship of perceptual-motor achievement, academic achievement and intelligence of fourth grade children. Journal of’Learning Disabilities, 1970, 3(8), 413-420. Sloan, W. Motor proficiency and intelligence. American Journal of’ .Mental Deficiency, 1951, 55, 394-406. Slusher, H. S. Personality and intelligence characteristics of selected high school athletes and non-athletes. Research Quarterly, 1964, 35, 539-545. Sterritt, C. M., and Rudnick, M. Auditory and visual rhythm perception in relation to reading ability in fourth grade boys. Perceptual and Motor Skills, 1966, 22, 859—864. Trussel, E. M. Relation of performance of selected physical skills to perceptual aspects of reading readiness in elementary school children. ,Research Quarterly, 1969, 40, 383-390. Vogel, P. Battle Creek Physical Education Curriculum Project. Jaurnal of'Realth, Physical Education and Recreation, 1969, 40, 25-29. Werner, E. E., Simonian, K., and Smith, R. S. Reading achievement, language functions and perceptual-motor development of ten and eleven-year-olds. .Perception and Motor Skills, 1967, 25,~ 409-420. Williams, Lois A. Specificity versus generality of motor response consistency. Journal of’Mbtor Behavior, 1969, 1(1), 45-52. Yoder, J. H. The relationship between intellectual and non-intellectual performance. Unpublished Ph.D. Thesis, Purdue University, 1968. nICHIoaN STRTE UNIV. LIBRARIES \lHIWW”W"I“Wll”WWWIWIIIWIWI 31293006431690