AN ASSESSMENT OF THE ESEFULNEfl OF THE WARTEGG DRAWING CGMPLETQCJN TEST £35 A CRQSS-CULTURAL NON-LANGUAGE PREMCTOR 0F ACADEMEC ACHIEVEMENT AMQNG ELEMENTARY SCHGGL CMRDREN EN GUATEMALA Thesis Fm the bag?” as? Pit D. MECHEGAN HATE UNIV’ERSlTY Qfiee E. Giiberfi‘ W615 IHESIs J i LIBRARY Michigan State University This is to certify that the thesis entitled An Assessment of the Usefulness of the Wartagg Drawing Completion Test as a Cross-Cultural Non- Language Predictor of Academic Achievement Am0ng Elementary School Children in Guatemala presented by Otto Ernest Gilbert has been accepted towards fulfillment of the requirements for Ph.D. degree inGuidance and Counseling Date {flog/VJ/—/4é5p 0-169 ABSTRACT AN ASSESSMENT OF THE USEFULNESS OF THE WARTEGG DRAWING COMPLETION TEST AS A CROSS-CULTURAL NON-LANGUAGE PREDICTOR OF ACADEMIC ACHIEVEMENT AMONG ELEMENTARY SCHOOL CHILDREN IN GUATEMALA By Otto E. Gilbert The Problem The existing limited school facilities in developing countries could be used more efficiently by admitting only students that have high achieving potentiali Selection of such students could probably be accomplished by admin— istering a group, cross—cultural, non-language test. Several recent studies suggested that the Wartegg Drawing Completion Test (DCT) seemed to fulfill the criteria needed for such a test. The main purpose of this study, therefore, is to test the predictive validity and reliability of the DCT. The Sample A total sample of 283 second-grade subjects was obtained from 16 randomly selected public schools in four counties of different geographical areas of Guatemala, Central America. OTTO E, GILBERT MethodolOgy The data was collected by administering the Wartegg Drawing Completion Test under standard conditions, by classrooms, at randomly selected schools, to all second- grade students. The tests were scored individually by means of a new scoring scale. The validating criterion (mean mid— year grades) was obtained from the official school—records. The resulting data were coded and punched on I.B.M. cards and processed through the CORE programs at the Michigan State University Computer Laboratory. Re ult and Conclusions U) U) J l. The Wartegg total score, as tested by the Ebel method, is well within the expected levels of reliability (.99—-see Appendix F). However, correlations between Wartegg total scores and its sub—tests did not reach the established levels for intra-test reliabilityt D.) The predictive ability of the DCT for the selected validation criterion (mean—class—grades) is very low. However, results also indicate that the reliability of the criterion, as a measure of academic achievement, is also low, 3. Significant differences between urban and rural school children's scores on the Wartegg total and its parts OTTO E° GILBERT were found. Rural children score lower than urban children. Significant differences in mean class-grades between urban and rural children were found. Rural children grades are lower than those of urban children. Due to certain limitations in this study, further tests of validity and reliability of the Wartegg Test, incorporating several changes, are recommended. AN ASSESSMENT OF THE USEFULNESS OF THE WARTEGG DRAWING COMPLETION TEST AS A CROSS—CULTURAL NON—LANGUAGE PREDICTOR OF ACADEMIC ACHIEVEMENT AMONG ELEMENTARY SCHOOL CHILDREN IN GUATEMALA By P" 4* ( Otto E. Gilbert A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Counseling, Personnel Services and Educational Psychology College of Education 1965 ACKNOWLEDGMENTS The author wishes to express his deepest gratitude to Dr. John E. Jordan for his advice and encouragement while pursuing doctoral studies at Michigan State University, and for his assistance in the preparation of this thesis. Appreciation is also extended to Drs. Harold H. Anderson, William W. Earquahr and Carl ‘ 2) Gross -or their assistance throughout the doctoral program and as members of the thesis committee. The author would also like to thank Dr. Irwin Tanaka for his help in preparing the data for machine computation and Dr. John P. Keith for his suggestions for improving the formal presentation of this work. To the Pan American Union, the University of San Carlos and The American School of Guatemala, which sponsored his graduate studies at Michigan State University, the writer also wishes to express his sincere thanks. Last but not least, gratitude is also extended to the county technical supervisors, school directors, teachers, and pupils who wholeheartedly gave of their time and efforts to make this study possible. ii TABLE OF CONTENTS ACKNOWLEDGMENTS LIST OF TABLES LIST OF APPENDICES Chapter I. INTRODUCTION. . . . . . . . . . The Problem The Nature of the Problem The Instrument Reasons for the Study Limitations Organization of the Thesis II. RELATED RESEARCH . . . . . . . III. METHODOLOGY . . . . . . . . . . The Sample Hypotheses to be Studied The Scoring System Administration of the Test Statistical Procedures Validating Criteria IV. ANALYSIS OF THE DATA Results of the Tests of Reliability Results of the Test of Validity V. SUMMARY, CONCLUSIONS, AND IMPLICATIONS FOR FURTHER RESEARCH . . . . . . Summary Conclusions Implications for Further Research REFERENCES APPENDICES iii n, 2 ,.. \g , 1: "u LIST OF TABLES Table Page 1. Estimated school-age pepulation (7—14 years) in Guatemala and total number of children registered in schools . . . . . . . . 3 2. Distribution of school-age children (6-13 years) according to probable levels of . instruction. . . . . . ., . . . . . 4 3. Registration and mean attendance in rural and urban primary schools in Guatemala in I962 . 5 iv Appendix A LIST OF APPENDICES Population Density Map, Republic of Guatemala Wartegg Drawing Completion Test Blank. Academic Achievement Prediction Scale. Code Book . . . . . . . . . . . Data Tables 4. One—way analysis of variance of Wartegg total scores: geographical regions. 5. One-way analysis of variance of Wartegg total scores: classrooms 6. One-way analysis of variance of Wartegg total scores: urban, town, and rural schools . . . . . . . . . . . 7. One—way analysis of variance of mean— grades: geographical regions. . 8. One—way analysis of variance of mean— grades: classrooms 9. One-way analysis of variance of mean- grades: urban, town, and rural schools lO. Simple correlations of Wartegg total and its parts: total sample . . . ll Multiple correlation coefficients between Wartegg total and its parts: total sample . . . . 12 Highest order partial correlation coefficients between Wartegg total and independent variables including mean— grades: total sample . . . . . l3. Beta Weights of the Wartegg total and the independent variables: total sample Page “5 A? 49 60 Si Appendix IA. 15. l6. 17. 18. 23. 2A. 25. 26. Simple correlation coefficients between scores and re-scores of the Wartegg total test and its parts: total sample Simple correlations of mean- grades and the Wartegg total and its parts: total sample . . . Simple correlation coefficients between transformed mean—grades and Wartegg total and its parts: total sample . Multiple correlation coefficients between mean-grades and Wartegg total and its parts: total sample . . . . Highest order partial correlation coefficients between mean- grades and Wartegg total and its parts: total sample . . . . . . . . . . Beta weights of mean-grades and Wartegg total and its parts: total sample Means and standard deviations of the Wartegg total scores: total sample. Means and standard deviations of the Wartegg total scores: geographical regions . . . . . . . . Means and standard deviations of the Wartegg total scores: classroom groups Means and standard deviations of the Wartegg total scores: urban, town, and rural schools . . . . . Means and standard deviations of mean— grades: total sample and geographical regions . . . . . Means and standard deviations of mean- grades: classroom groups Means and standard deviations of mean— grades: urban, town, and rural schools F Computation of Reliability by Analysis of Variance . . . . . . . vi Page 62 63 6a 66 66 67 68 68 69 7O 71 CHAPTER I INTRODUCTION The Problem The present research is part of a larger project being conducted in different cultural regions of the world and is based on several previous studies which ap— peared to indicate that the Wartegg Drawing Completion Test (DCT), with some improvements in its scoring system, could be used as a cross-cultural, non—language measure- ment of academic achievement. If the test could be so used, it would be an important contribution toward the improvement of education in developing countries. A Unesco report, The World Survey of Education II (26:15), states that in the period from 1950 to 1954 there were a total number of 550 million children between the ages of 5 to 14 years in the world. Of these, only 300 million were enrolled in primary or secondary schools. This means that only slightly more than one—half of the world‘s children are receiving some kind of education in schools today. Of these 300 million, a large number drop out after one or two years of school. Therefore, a con- siderable proportion of those who do go to schools, partic— ularly in rural areas, do not attend long enough to insure permanent functional literacy. Of the total primary school- age population of over 260 million, only 70.9 (21%) reach secondary school. It seems, therefore, reasonable that the developing countries might make better use of their limited primary school facilities by admitting only those that obtain the highest scores on a test which predicts academic achievement. The Nature of the Problem In Guatemala, the present situation of education is even more serious than the average in the rest of the world. Seventy-five per cent of its total population above 14 years of age is illiterate and the population, as a whole, is increasing at an annual rate of 3% (9:35). The school—age population is increasing at a much higher rate. It has gone up almost 200% in the last twenty years: from 427,000 in 1940 to 779,000 in 1960 (10:63). In 1940, 296,000 children were unable to attend school because of lack of space. In 1960, there were 482,000 children in the same situation. The absolute numbers keep increasing as indicated in Table 1. Table 1 also includes data about the total number of children in—and-out of schools during a four year period, and the breakdown between urban and rural schools. In 1960 it was estimated that 11,800 more classrooms with furniture and equipment were needed to accommodate all TABLE l.--Estimated school—age pOpulation (7-14 years) and the total number registered in schools (10:64). Population 1957 1958 1959 1960 Total population 679,038 696,902 707,344 779,100 Total registered in day primary schools 249,832 259,890 281,950 297,009 Urban schools 153,788 161,168 174,659 185,323 Rural schools 96,044 98,722 107,291 111,686 Total number out | of schools 429,206 437,012 425,394 482,091 school—age children in the country at that time. Since every classroom needs a teacher, there was also a shortage of 11,800 teachers (3:154). The total cost to the nation to fill this need would amount to a 300% increase in its present annual budget. The country does not have the capital to finance such an educational program in the near future, since the popula— tion keeps increasing every day and the annual income (per- capita) is low: $174.89 in 1960 (3:10). Table 2 shows the efficiency of Guatemala's educational system during the last decade. During the 1950 to 1960 period, 525,500 children registered in first grade. Of these, only 93,500 finished sixth grade, which is the official completion of primary TABLE 2.——Distribution of school-age children (6—13 years) according to probable levels of instruction in 1960 (4:6). Relative Distribution Absolute Number Percentage of (in thousands) School Age Children Total school—age population 779.1 100 Absolute illiterates 504.9 65 Potential illiterates 168.8 22 In school 83.8 11 Out of school 85.0 11 Deficient literates 50.4 6 In school 25.0 3 Out of school 25.4 3 Satisfactory literates 55.0 7 school in Guatemala. This is oh1y 17.8% of the total number that started school (4:7). One of the main reasons for this loss is that many children register but never attend classes as is shown in Table 3. Of those that do attend, the highest percentage drop out in the first three years. In 1961—1962 the percentage of dropouts in first grade was 49.47% (10:64-75). According to the report of the Commission of the Economic Planning Council (3:61-72) a high percentage of those who drop-out are students who fail grades or have low academic achievement. TABLE 3.—-Registration and mean attendance in primary school in 1962 (10:65). Registration Mean Attendance Male Female Sum Male Female Sum Total 189,321 149,532 338,853 158,903 125,489 284.392 Rural 78,082 52,371 130,453 63,172 42,228 105,400 Urban 111,239 97,161 208,400 95,731 83,261 178,992 Of a total of 338,800 students who registered in 1962, only an average of 284,400 attended classes. Table 3 also shows that the average attendance is lower in rural than in urban schools. All evidence reviewed so far seems to indicate that in the foreseeable future only a portion of the total school population in Guatemala will find a place in school. There- fore, every effort should be made to increase the efficiency of the limited resources available. This could be accom— plished, at least in part, by admitting to school only those students who seem to have the potential to be good academic achievers. This selection could be accomplished through a combination of school—grades, teacher ratings, and test scores made at the beginning of the second school year. The screening test could be used as an entrance test to the second year, and its results could be used in con— junction with the child's grades and teacher ratings in the first grade. An efficient screening test for these purposes would have to meet the following criteria. The test should be: 1. U"! a group test that can be easily administered and scored, because the number of children to be tested is large; a non-language test, because Guatemalan children speak different languages and dialects, and at those ages they are still unable to read fluently; a "culture fair" test, because it has to include only those cultural elements common to the hetero- geneous cultural groups in the country; a low cost device, because the educational system operates on a very limited budget; a good discriminator of children's academic abilities, because its main purpose is to separate children according to their differing degrees of ability; and a good predictor of children's future academic achievement. A test, which at least "a priori” seems to meet most, if not all, of the above-mentioned criteria, is the Wartegg Drawing Completion Test. The Instrument The prototype of the Wartegg Drawing Completion Test, as a projective technique for personality assessment, was originally constructed by Sander at the University of Leipzig and was known as the Phantasie Test (14:3). The present form of the test, based on Sander's work, was con— structed by Wartegg (27, 28, 39, 30). In 1952, Kinget introduced the Wartegg test to the United States. She had previously conducted research, in- volving 383 adult subjects, using the Wartegg Test for her doctoral dissertation in Belgium. She developed a scoring system to aid in the interpretation of the drawings. Her system scores personality on four components: emotion, imagination, intellect, and activity (14:9—10). Based on the work of Kinget, Stark (24), Keith (12), and Matheney (19) developed, tested, and suggested a new scoring system in order to use the Wartegg Drawing Completion Test as a test of intelligence. The DCT blank, as used in the present study, is pre- sented in Appendix B. It consists of eight frames encased in a heavy black border arranged adjacent to each other on the upper half of the form. Within each frame there is a different stimulus of very small dimension. The character— istics of the stimuli that appear in the eight frames have been described by Kinget as follows: Stimulus l, the dot, represents smallness, light- ness, circularity and centrality. The stimulus itself is not imposing and could be omitted by the less sensitive or perceptive subject. Stimulus 2, the wavy line, suggests something alive, mobile, loose undulatory, fluid or growing. The qualities of this stimulus resist matter-of- fact treatment or technical use. It required integration into something organic or dynamic. Stimulus 3, the three vertical lines with propor— tional increments, express the qualities of rigidity, austerity, regularity, order, and progression. These qualities can be combined and can produce complex impressions of dynamic organization, gradual develop- ment, methodic construction, and similar concepts. Stimulus 4, the black square, looks heavy, solid, massive, angular, static, and evokes concrete materiality. While Stimulus 3, in spite of its mechanical nature, still shows some growth and dynam— ism, Stimulus 4 is completely inorganic and inert. Stimulus 51 two slanted lines in opposite directions, express the idea of conflict and dynamism. The position of the longer line evokes something directed decidedly upwards to which the shorter line shows frank Opposition. The rigidity of the lines and their perpendicular relation also suggests their technical or construction use. Stimulus 6, the horizontal and vertical lines, have a strictly matter—of-fact, sober, rigid, dull and uninSpiring aspect. At first sight they seem fit only for completion into simple geometric patterns or elementary objects. Experience shows, however, that this stimulus may be worked into a variety of interesting combinations. However, the off-center position of each of the lines makes their completion into a balanced whole, a tough task requiring con— siderable planning activity. Stimulus 7, the dotted half circle, suggests something very fine, delicate, round and supple, that is at the same time both appealing and a little puzzling because of its complex bead—like structure. This structure-like aspect of the stimulus, together with its somewhat awkward location within the square, forces the selective activity of the mind and resists casual or crude treatment. Stimulus 8, the broadly curved line, has the organic qualities of roundness and flexibility of Stimulus 7, but whereas Stimulus 7 has something irritating in its complexity and smallness, Stimulus 8 appears restful, large, fluent, and easy to deal with. Its smooth curve readily suggests completion into organic subject matter, animate or inanimate, while its downward bending movement and location connote the idea of cover, shelter, and protection. Its relatively large dimension also evokes eXpansion and vastness as proved by the frequent completion of 9 this stimulus into natural phenomena such as raih- bows or sunsets (14:35—37). Though these stimuli emphasize the measurement of personality, the DCT can also measure intelligence as will be described in Chapter II while the DCT’sscoringsystem, as used in this study, will be described in Chapter III. Reasons for the Study Some of the most serious educational problems being confronted by Guatemala at the present time have been pre- sented in previous pages of this thesis. The great disparity of cultural and sub—cultural groups in the country, the great variety of languages and dialects spoken by rural school children, the limited educational facilities, and the high number of school drop-outs among low academic achieving children, indicated the need for an instrument which can differentiate children of high and low academic ability. This could probably best be accomplished by a group test. The ideal tool to accomplish this task would be a culture-fair,non—verbal, economical, and simple test (to administer and score) that would make possible such differentiation among eight to ten year old children, an age at which children initiate their schooling in Guatemala. The DCT appears to be such an instrument. It is a relatively culture—free test because it does not include typical elements of a certain culture at the expense of children of other cultural groups. It is a non~verbal, 10 graphic test that circumvents the problem of different languages and dialects. It does not require that children be fluent in Spanish in order to perform well on it. Since the DCT is a graphic test it does not require that the subjects know how to read and write. This is very impor- tant because the children proposed for the Guatemalan sample have not yet learned how to read and write. The test is so simple to administer and score that primary school teachers, with some training, could admin— ister and score it. The test's directions are sufficiently simple that first-grade children can follow them. The DCT is also economical because it is a group test that can be given simultaneously to a whole class, consists of only a sheet of paper, is administered in approximately forty minutes, and its objective scoring system is such that it requires only a few minutes to score each test. A careful survey of all research publications both in Guatemala and the United States indicates that no previous research has been conducted with the DCT as a predictor of pupil academic potential in Guatemala. The purposes of the present study are: 1. To test the reliability of the DCT as a measuring instrument. 2. To determine if there is any significant relation- ship between the DCT test scores and the student‘s mean grades. ll 3. To determine the validity of the DCT in predicting future academic achievement of students starting their primary education. Limitations This study has been designed in such a way that the stratified random sample would include a representative number of cases in three areas, with different population densities, of second—grade children in official Guatemalan day schools during the 1963 school year. The three areas included in the sample are: 1. Rural area—~This includes public, primary, co- educational day schools found in villages with less than 1,000 habitants. 2. Semi-urban area or towns—~includes public, primary, coeducational day schools found in small towns with a population of 1,000 to 5,000 habitants. 3. Urban area—-includes public, primary, coeducational day schools found in towns with a population of 5,000 or over. Since all schools sampled are coeducational, there is almost an equal number of boys and girls included in this study: 157 boys and 126 girls. It is neither the in— tention nor the purpose of this study to standardize the DCT for Guatemalan public schools. Therefore, no general- izations, at a national level, should be made on the basis 12 of this sample. Although there is a considerable range in ages of children included in this sample, no age norms for children can be established from this data. These age distributions are not normal; in all probability ”brighter" than normal children would be found at lower ages and a greater number of "duller" than normal children would be found at older ages since the children sampled are all in the second grade. Guatemala's educational system does not place emphasis in promoting children on the basis of chrono- logical age. Children of low ability frequently repeat first and second grades and, therefore, ”duller” than normal 9, 10, and 11 year old children are found in the second grade. For the same reason, "brighter” than normal seven year olds are found in second grade because they are the ones that have started school early and have not failed the first grade. The generalizations that can be made, based on this study, are only with reference to the charac— teristics of second grade children found in rural, semi— urban, and urban public, primary, coeducational Guatemalan day schools in 1963. These generalizations can be made in regard to academic achievement, DCT scores, and distribution of ages and sexes at the second grade level. 13 Organization of the Thesis Chapter I, the introduction presents the educational problems faced by the developing countries of the world today, including those faced by Guatemala. The statement of the problem, reasons for the study and its limitations are included in this section. Chapter II, the related research is reviewed, placing special emphasis in reviewing the research that has been published on the Wartegg Drawing Completion Test. Chapter III, the method used for sample selection, data collection, hypotheses to be tested, the scoring system, test administration, and statistical analysis are described in this chapter. Chapter IV, an analysis of the data, using apprOpriate tables to aid the interpretation, is given here. Chapter 7, a summary of the obtained results, con— clusions that can be drawn, and recommendations for further research are presented in this section. CHAPTER II RELATED RESEARCH Interest in children's drawings as a means of measuring general ability can be traced to the year, 1885, when Cooke (5) published an article in which he described the different stages of intellectual development shown by the drawings of children. From 1900 to 1915, Lamprecht (17) of the University of Leipzig collected drawings of children from all parts of the world, Unfortunately, he never completed this research and, therefore, we do not have a summary of all the data that was gathered. Later, Kerschensteiner (l3) conducted one of the most extensive studies in drawings of children. He gathered more than 100,000 and classified them in three categories: 1. Purely schematic drawings. These correspond to the so—called ideoplastic stage in which the child draws what he knows and not what he sees; 2. Two dimensional, visual appearance drawings; and 3. Three dimensional drawings, in perSpective, in which the child tries to give the impression of three~dimensional Space. .1 4 15 Kerschensteiner devoted several pages in his book to reporting the differences he had noted between high- and low-achieving school children. He discovered that there were quantitative and qualitative differences between these and that low achieving children make more primitive drawings than the others. Rouma (22) established many of these differences between the drawings of retarded and normal children. Rouma also conducted other experiments in which he compared the drawings of European children with those of children in contemporary primitive societies. He found, among other things, that the order of development in children's drawings is astonishingly constant even among children of radically different social and cultural back~ grounds. In all cultures that were studied, Rouma found that the first drawings children attempt are a graphical enumeration of things. At a later stage, they develop proportionality among the different parts and Spatial re— lations among the different things. He also discovered that children of lower mental ability can imitate well but cannot produce good original drawings while mentally—gifted children show real creative ability in their drawings. More recently, Gestalt psychologists have emphasized this type of research by stating that the basic personality structure of the individual can be inferred from his drawings. The Machover Test of the human figure (18) and 16 the Buck House-Tree—Person Test (HTP) (2) are some of the tests designed for this purpose. Several other studies have been conducted with the purpose of using the drawings as the basis to measure intelligence. Of these, the most widely known is undoubt— ably that of Goodenough which resulted in the ”Draw—A-Person Test" (8). This test has demonstrated its usefulness as a nonverbal intelligence test. Bender has tried to use drawings as an index of visual motor coordination development (1). Koppitz designed an efficient scoring system for this test (16). This system makes it possible to determine a development age. The scores on the Bender correlate fairly well with scores ob— tained by means of the Wechsler Intelligence Scale for Children (WISC) between the ages of 5 and 10 years (16:413— 416). Another exponent of Gestalt psychology, Sander of the University of Leipzig, constructed a prototype of the present Drawing Completion Test which he entitled Phantasy Test (23). Sander's work furnished the basis for that of Wartegg, one of Sander's colleages at the University of Leipzig (29). Wartegg is the author of the present form of the Drawing Completion Test which is known as the Wartegg Test. Kinget (14) conducted a study in which she administered the Wartegg Test to 383 adult normal subjects and presented her results in the form of a doctoral dissertation at the 17 University of Iouvain in Belgium. Kinget developed a rela- tively elaborate scoring system designed to aid in the interpretation of the drawings. According to Kinget‘s scoring system the Wartegg Test can be analyzed through four factors: emotion, imagination, intelligence, and activity (14:9—10). This evaluation system, which uses the Wartegg Test as a projective technique to analyze person- ality for clinical purposes, was introduced by Kinget to the United States in 1952. Since that date, interest in this test has continued to grow in the United States. In a review of all publications appearing in the Psychological Abstracts the present writer was unable to find any study that used the DCT in an under-developed society. However, some studies have been conducted with the DCT with more than 2,000 children in developed countries. Two studies, one conducted by Erna Duhm (6) and the other by Hemmo Muller Suur (20) are listed in the Psychological Abstracts of 1953, These researchers found that certain characteristics differentiate the drawings of children of high- and low» academic achievement within the same culture. The charac— teristics are: a. Children of low academic achievement do not integrate the initial elements, given by the test form, in their drawings; 18 b. The same children show a marked repetition of simple graphic themes in each drawing; and c. They alwo show a tendency to disregard or ”burst" the spatial divisions of the tests. Another study was conducted by Stark (2A) at the University of Detroit in 1954. She suggested that the DCT could be scored objectively as an intelligence test. A Pearson product moment correlation coefficient of .79 was established between the DCT and the WISC scores, which seemed to confirm this fact. The evaluation system used by Stark was based to a great extent in the variables sug~ gested by Kinget but she also added some others taken from the work of Goodenough. Her scoring system included the following variables: (1) orientation, (2) detail, (3) organization, (A) proportion, (5) dimension, (6) symmetry, (7) symbolism, (8) movement, (9) originality, (10) variety, and (ll) time. Matheny (19) investigated the usefulness of the DCT as an instrument to measure the general ability of fourth graders at Waverly School District in Lansing, Michigan. His sample includes 176 students in the fourth grade of elementary school. He divided the children, according to sex into a validation and a cross-validation group. Several comparisons were made between the scores obtained by the students in the DCT and in the Primary Mental Ability Test (PMAT), grade point average (GPA), and their Stanford 19 Achievement Test scores in arithmetic and in reading. The results showed that the DCT scores have a statistically significant relationship with the P.M.A.T., I.Q. scores, the Stanford's arithmetic and reading standard scores, and with grade-point averages. The DCT scoring variables that correlated best with the validation criteria are: dimen- sionality, proportionality, and detail. Another study by Keith (12) presented the results of administering the DCT to school-age children of three sub— Saharan African tribes. He attempted to evaluate the aca— demic achievement of 98 eleven—year old children in rural schools using the three scoring variables suggested by Duhm and Muller Suur in order to differentiate between the different levels of intelligence. These variables are: (1) integration of the stimulus in the drawings, (2) repeti— tion of graphic themes, and (3) disregard for the spatial divisions of the test. Keith‘s sample was divided into high and low academic achieving students according to their grade point averages and teacher evaluations. He found that integration of the test—stimulus in the drawings con- tributed significantly to differences between the mean- scores of high and low academic achieving children. The two other variables did not contribute consistently to dif- ferenciate between the means of high and low achieving children of the African tribes that were studied. CHAPTER III METHODOLOGY The Sample A stratified random sample of coeducational public primary schools was used, and 283 second grade pupils were thus tested. In order to meet the necessary assumptions for a statistical interpretation of results, each school was assigned an identifying number, and a table of random numbers was run to select the schools in each stratified area. The sample is stratified according to the population density in the site where the school is located. The sample included: (1) schools located in rural areas, i.e., schools found in villages of less than one thousand inhabitants; (2) semi-urban schools located in towns of more than 1,000 but less than 5,000 inhabitants; and (3) urban schools in towns of 5,000 inhabitants or over. Although this study did not aim to gathera nation— wide sample from which norms for all Guatemalan second- grade children could be established, nevertheless in order to avoid the urban cultural influence of the capital city, it seemed preferable to have samples from different parts of the country. Therefore, the sample includes schools of the extremes east and west of the country as well as from 20 21 the south and center. (See map of population density and locations of counties included in this sample in Appendix A.) The selected schools from the stratified area are: Urban: 1. José J. Palma, Guatemala. 2. Alejandro Marure, Guatemala. Small city: 1. Santa Elena, Chiquimula. 2. José A. Palma, Chiquimula. 3. E. Palo Gordo, San Marcos. A. El Salto, Escuintla. Rural: 1. San Estan, Chiquimula. 2. El Ingeniero, Chiquimula. 3. Vado Hondo, Chiquimula. A. Raul Mejia, Chiquimula. 5, Chamac, San Marcos. 6. Champollap, San Marcos. 7. Federacion, San Marcos. 8. quuihuila, San Marcos. In several rural schools there were children from either first or third grade in the same room with the second grade pupils. 'With only two exceptions, the first or third grade pupils were asked to leave the room while the second grade subjects took the test. In these two instances, the test was also administered to the few (2 or 3) third 22 graders, but their test sheets were later excluded from the sample. In order to determine the location of the schools, to obtain the necessary data about them, and to insure the cooperation of principals and teachers for this study, it was necessary to visit the Technical Superintendent of Education of the county in which the schools to be included in this study were located. The Technical Supervisors agreed to sign a letter addressed to the primary school principals in their county, which asked them to cooperate in every possible way to make this study possible. In all cases, supervisors, directors, and teachers were very courteous and willing to help. They permitted the tests to be administered upon appearance of the tester and provided all the additional information needed to complete the study. They also prepared a list with the pupil's name, age, sex, grade placement, and midterm grades in the four main academic subjects. Hypotheses to be Studied There are four hypotheses which were to be tested in this study: Hypothesis 1: The Wartegg Drawing Completion Test is a reliable instrument. Therefore, the total score of the DCT is predicted by the independent variables (its sub-tests). The sub-hypotheses relative to this analysis are: a. The number of dimensions score contributes sig~ nificantly to the total score. 23 b. The number of objects in the drawings makes a significant contribution to the total score. 0. The number of drawings in which the stimulus was integrated contributes significantly to the total score. d. The number of meaningful objects or drawings with meaningful lines makes a significant con— tribution to the total score. e. The number of proportional drawings contributes significantly to the total score. Hypothesis 2: The Wartegg Drawing Completion Test is a valid instrument for predicting academic achievement of second grade children. Therefore, a significant relationship exists between the Wartegg total score and its part~scores with second grade student’s mean mid—year school grades. Hypothesis 3: There is a significant difference between urban, town and rural school children scores on the Wartegg Drawing Completion Test. (For the purpose of this study, urban children are those that attend public primary schools in Guatemala city (over 500,000 inhabitants); town children are those that attend public primary schools located in towns of 1,001 to 4,999 inhabitants; and rural children are those that attend public primary schools located in areas with less than 1,000 people in Guatemala.) Hypothesis 4: There are significant cultural differences among Guatemalan school population which affect pupil's mean—grades and total and partial scores on the Wartegg Drawing Completion Test. These differ- ences exist both between and within different geo— graphical regions of the country. (Cultural differences,for the purposes of this study, are those found between the ”Spanish” or ”Western” popula— tion which predominates in the urban sample of this study and the "lndian" population of Guatemala which prednminates in the town and rural samples of this study. These differ— ences include language, social, and religious customs, nutritional habits, living standards, and costumes.) The Scoring System The Academic Achievement Prediction Scale (AAPS) of the Drawing Completion Test is based on previous work done by Keith (12) and Matheney (19). Their recommenda— tions concerning the improvement of the validity and the reliability of the tests scoring system were incorporated. The scoring system appears in Appendix C. The eight frames of the test are scored separately on a g to Q9 point scale. The highest possible score for any subject is A80, since a raw score of fig can be obtained for each The scoring is performed according to the following directions: 1. Determine whether the drawings are one, two, or three dimensional in nature. 2. Determine the number of objects in each picture. 3. Select the appropriate column of dimensionality and correct row for none, one, two, or more objects, with or without background detail. U. Determine integration of stimulus or non- integration of same. 25 5. Determine whether the drawing is meaningful or has meaningful lines. 6. Determine for or against proportionality of drawing. Example: A 3-D object with background detail. Enter column 1, row 5 (counting upwards). If the object is integrated, scores can range from El to.flfl. If the object is not integrated, scores can range from _§Z to £9, If the object is integrated and meaningful, possible scores can be fl§_or £3; 3; if not proportional,_flfl if it is proportional. If the object is integrated but not meaningful, scores can be_fll or 5g, In Special cases additional scoring criteria is used: (1) two or more 2—D drawings which are simple stick drawings, score no more than single 2—D drawings; (2) minus one point for each repetitious theme drawn. The terms used in the scoring system are defined by Matheny (l9:41-42)as follows: 1. Dimensionality; Drawings may be classified as one, two, or three dimensional in nature, The proper— ties of dimensionality are sufficiently wellndefined as to make further definition unnecessary. 2. Integration: This variable is judged to be present when there is clear evidence that the subject 26 has taken cognizance of the stimulus in his drawing. The degree of integration is not considered at this point. The sole criterion is whether or not there is clear evidence that the subject has attempted to incorporate the stimulus into his drawing. 3. Meaningfulness: This variable refers to the ability of the drawing to convey something of a representational nature to the examiner. Since the child is not asked to verbally identify the drawing, meaningfulness must be inherent in the projected qualities of the drawing. 4. Proportionality: This variable refers to the relationship of the various parts of the picture to the whole. It depends exclusively upon the meaning- fulness of the picture. Consequently, if a drawing is not perceived as having meaningfulness, there is no way of rating the degree of proportionality offered by the drawing. 5. Detail: Drawings which add ornamentation beyond what is necessary for clear recognition of the item represented are given credit for detail. 6. Repetition: Drawings which appear to be replicating a previous theme suffer a penalty of one point. In a sense, this is a reverse procedure for scoring variety of content. It appears to lend itself to objective scoring more fully than does variety as a scoring variable. Administration of the Test The DCT was administered to the subjects by units of classrooms. The following procedures and directions were practiced in order to insure uniformity: l. Pupils were seated a suitable distance apart to render opportunities for "cheating” less likely. 2. The drawing blank was placed on a manila folder. 3. The subjects were furnished Number Two drawing pencils with uniformly sharpened points. 4. The following instructions, suggested by Kinget, were read to the subjects in Spanish: On this form you see eight squares. Each of these squares contains little signs. These signs have n~ \ special meaning; they are to be part of the drawings which I want you to make in each of the squares. You may draw whatever you like and you may start with the sign you like best. You may work as long as you wish, and you may use the eraser. Do not, however, turn the sheet. This must be the top (Examiner points) (45:28—29). Whenever necessary, the instructions were repeated to help the pupils understand what they were to do. The time required for administering the test to an entire class ranged from MD to 60 minutes. Statistical Procedures The reliability and validity of the DCT were first determined by computing multiple regression equations and analysis of variances for the entire sample of 283 subjects, using the CORE program of the Michigan State University Computer Laboratory. The program includes the following computations: 1. Multiple regression equations with the total Wartegg score as the dependent variable and (l) dimensionality, (2) number of object, (3) integration, (3) meaningfulness, and (5) proportionality as the independent variables. R.) a Multiple regression equations with the mean mid—year class grades as the dependent variable and the (1) total Wartegg score, {2) dimensionality, (3) number of objects, {‘) integration, (5) meaningfulness, and (6) preportionality as the independent variables. 3. Analysis of variances with the mean mid—year class grades as the dependent variable and the different regions as the independent variables. 28 A. Analysis of variance with the mean mid—year class grades as the dependent variable and the differ- ent class groups as the independent variables. 5. Analysis of variance with the total DCT score as the dependent variable and the different regions as the independent variables. 6. Analysis of variance with the total DCT score as the dependent variable and the different classroom groups as the independent variables. 7. The ability of the DCT was further tested by the method outlined in Appendix E, which is a computation of reliability by analysis of variance. Validating Criteria In first and second grades, academic achievement is officially determined by mid-year and final grades assigned by the grade teacher. No official written tests are admin- istered and there are no known standardized achievement tests in Spanish suitable for second grade Guatemalan pupils. Therefore, the validity of the Drawing Completion Test, as a predictor of academic achievement was tested against mean mid—year class grades in the second grade. The average mid-year grades in the main academic Subjects: Spanish, Arithmetic, Natural Sciences, and Social Sciences constituted the criterion variable of academic achievement. CHAPTER IV ANALYSIS OF THE DATA The statistical data for analysis are found in Tables A through 26. The data were coded according to a code book specifically devised for this purpose (see Code Book in Appendix D), and was key-punched in I.B.M. cards. The oom— putations and card punching were performed by the Michigan State University Computer Laboratory. Means and variances of Wartegg total scores were computed for (l) classroom groups, (2) urban, town and rural samples, and (3) geographical regions samples. These appear in Appendix E, Tables 21, 22, and 23. Analyses of variance were also computed for each one of these categories and the differences were found significant beyond the one per cent level. As shown in Tables 4, 5, and 6 (Appendix E), F values are 12.2 for classroom—groups, 24.2 for urban, town and rural groups, and 20.A for eographical regions. Means and variances of mean-grades were computed for (l) classroom—groups, (2) urban, town and rural, and {3) geographical region samples (Appendix B, see Tables 25, 25, and 26). Analyses of variance computed for each one of these groups were found significant beyond the one per 30 cent level. Tables 7, 8, and 9 (Appendix E) show their respective F values 5.6 for classroom—groups, 16.1 for urban, town and rural and 11.5 for geographical regions samples. Results of the Tests of Reliability Reliability is the consistency of measurement of a particular instrument. It can be tested statistically by means of several techniques. The techniques used in this study are (a) the part to whole correlation method and (b) the score—rescore correlation. Completion Test's total score was correlated with each one of its constituent sub— test scores. As shown in Table 11, a corrected multiple correlation coefficient of .5687 was obtained between the DCT's total score and its five sub—tests, with a .323A coefficient of multiple determination. Partial correlation coefficients, which appear in Table 12 reveal that Number of Objects and Integration make significant contributions to the total score but Proportionality, Dimensionality, and Meaningfulness contribute little to the total score. With the exception of Dimensionality, all simple cor~ relation coefficients between the Wartegg total score and its part scores were found to be statistically significant at the .05 level. These coefficients appear in Table 10, and emphasize that Number of Objects and Integration 31 contribute substantially to the total score with coeffi— cients of .AA and .32. Beta weights were also computed for each of the part scores on the total score. Their values appear in Table 13. Again, Number of Objects and Integration appear to be the best predictors of the total score. Meaningfulness predicts the total score to a lesser degree and Dimensionality and Proportionality do not predict it at all. To conclude, thistest of reliability indicates that the Drawing Completion Test is partially consistent as a measuring instrument. The inter-scores reliability of the DCT was also determined. Simple correlations between scores and re—scores cf the total and partial scores were computed for a randomly selected sub-sample of 50 subjects. All coefficients were found to be statistically significant beyond the .01 level (see Table 14, Appendix E). Results of the Test of Validity Multiple correlation coefficients were computed between the criterion variable of mean-grades and the pre- dictor variables of the Wartegg total score and its parts. As shown in Tables 15 and 17, these coefficients are low which means that the present Wartegg scoring scale is unable to predict grades. The proportion of explained variance is 6.8 per cent, leaving a large percentage of 32 unexplained variance that lowers the value of its predictive power. The multiple correlation coefficients between mean- grades and Wartegg total score is .26 as indicated in Table 17. With the exception of Dimensionality, all simple cor- relation coefficients between mean-grades and the Wartegg total score and its sub—tests, which appear on Table 15, were found to be statistically significant at the .05 level. However, the interrelationship of all independent variables to the criterion are low. The Number of Objects subetest score has the best predictive power with a .21 correlation coefficient, but even this relationship leaves an unexplained error variance of 95 per cent. Additional information on the directional effect of the independent variables is provided by beta weights which appear in Table 19. Here again, Number of Objects and Inte- gration have the best predictive relationship with the validating criteria with .17 and .18 beta weights. The pre- dictive relationship of all other independent variables with mean-grades is insignificant. As an additional test of validity, mean class grades "2" values and these were then correlated. were transformed to The transformation of the raw scores to ”z" values did not affect the correlation coefficients. The results are identical to those of Table 15 (see Table 16, Appendix E). CHAPTER V SUMMARY, CONCLUSIONS, AND IMPLICATIONS FOR FURTHER RESEARCH Summary The Problem.—-The main purpose of this research was to determine the usefulness of the Wartegg Drawing Com- pletion Test as a predictor of academic achievement (mean— grades) among public school children in Guatemala, Central America. The sample was limited to second grade population in different geographical regions of that country. The Wartegg total score and its parts constituted the independent variable. One of the hypotheses of this study dealt with the reliability of the DCT's scores; another with the validity of these to predict academic achievement; and two others were concerned with differences between the drawings in diverse areas of the country. The Sample.—-A stratified random sample of second grade students in rural, town or semi-rural, and urban public schools in different geographical regions of Guatemala was used. The total sample consisted of 283 second grade students from 16 randomly selected schools in four counties. There were 157 males and 126 females in this sample. A table of random numbers was used for school selection within each of the four counties. 33 3A Methodology.--The data was collected administering the Wartegg Drawing Completion Test to all second grade students, by classrooms, at the selected schools. The DCT was scored individually by means of a new scoring scale and its sub—tests were also scored. The validating criterion (mean mid-year grades in the main subjects) was Obtained from the schools' Official records. The resulting data was arranged according to a code book (see Appendix D); punched on I.B.M. cards; and processed through the CORE programs at the Michigan State University Computer Laboratory. The results were tabulated, analyzed, and interpreted. The Results.--Computation of reliability by the Analysis of variance method established that the total DCT score is well within the expected range of reliability (see Appendix F). Intra—Test correlations, although positive and statistically Significant, did not reach the expected levels. A multiple correlation coefficient of .57 was Obtained between the Wartegg total and its parts. The independent variables, Number Of Objects and Integration, contributed most to the total score. Three other indepen— dent variables contributed little tO the total. Score-rescore correlations between Wartegg total and its sub-test scores, for a partial subsample of 50, were 35 all found to be statistically significant beyond the .01 level, the coefficient for the total score being .84. The test of validity revealed that mean mid-year grades are not predicted by the total or partial Wartegg scores. A multiple correlation coefficient Of .26 was found between grades and Wartegg total and its parts. A simple correlation coefficient of .12 was established between the criterion and the predictor. Simple correla- tions between other predictors (sub—test scores) and the criterion show little relationships which range from .11 for Dimensionality to .21 for Number Of Objects. Partial correlation coefficients are also low which seems to con— firm that the criterion (mean-grades) and the predictors (Wartegg total and part scores) do not reach the necessary relationship for predictive validity. The computed beta weights indicate that the sub-tests Number Of Objects and Integration are the best predictors; but all, including these two, have Very low predictive power. Conclusions l.. The results of this study do not appear to support Hypothesis 1; i.e., that the total score Of the DCT is predicted by the independent variables. Although correlations between the Wartegg total score and its sub— tests are all positive, and well beyond the one per cent level, they do not reach the established levels for the intra-test reliability. 36 2. Hypothesis 2 suggesting that the Wartegg Drawing Completion Test predicts academic achievement was not supported by the resulting low correlations between mean school grades and Wartegg total and part scores. 3. Analyses of variance of Wartegg total and partial scores of urban, town, and rural school children were found statistically significant. Urban children's DCT total and partial scores are higher than those of rural children. These results support Hypothesis 3. 4. Cultural differences, which affect the pupil's total and partial scores on the DCT, were also found to affect pupil's mean—grades among Guatemalan school population. Means and variances Of both mean-grades and Wartegg total and partial scores differ to a greater extent between urban, town, and rural groups than within each of them. Analyses Of variance also indicate that these differences are statistically significant between these samples. Since, as stated before, the "Spanish" culture predominates in the urban samples and the "Indian" culture predominates in the town and rural samples, these results seem to support Hypothesis 4. Implications for Further Research 1. The findings reported here suggest that the reliability and validity of the DCT have to be improved if the test is to become a useful cross—cultural, non-language 37 measuring instrument to predict academic achievement among elementary school children. Therefore, the scoring scale should undergo further revisions and refinement. A gradual improvement Of the scale could be obtained through computations of successive inter-scorer correlations after each new change in the scales is introduced. 2. A pilot study should be conducted to determine if the reliability of the test is improved if larger squares are used covering a considerably larger area Of the sheet. Larger drawings would permit a better Observation Of test variables such as Proportionality, Dimensionality, and Meaningfulness. 3. The significant negative correlations between Integration and several other variables suggests that this variable does not measure the same psychological traits and could thus be excluded from the scoring scale. 4. Analyses of variance results indicate that mean class-grades differ significantly between classroom groups. If these variations had shown a high degree of correlation with the results on the Wartegg scores, it could have been inferred that these differences were due to actual differ- ences in achievement levels Of these groups. However, they do not coincide and thus seem to be due to the teachers' personal grading practices. Mean-class grades, therefore, cannot be considered to be a valid and reliable criterion to predict future academic achievement of pupils since 38 their future achievement grades would depend to a con— siderable extent on their future teachers' grading practices. 5. The best validation criterion to determine the validity of the Wartegg Test as a predictor of academic achievement would be pupils' success in passing grades. Use Of such criterion would require a longitudinal study which would follow pupil progress until the end of their primary education. 6. Due to the design of this study, the age factor was not controlled. All second grade children, regardless of age, were included in the sample. Heterogeneity of ages probably obscured the interpretation Of validity and reliability results, since raw scores on the Wartegg were correlated with the criterion variable. 7. Age norms for the Wartegg scores should be develOped and these should be correlated with the criterion variable. 8. Further tests of reliability and validity Of the Wartegg Test should be conducted incorporating some or all of the changes in research design, test form, and criterion variables suggested by the results of this study. REFERENCES Bender, Lauretta. A Visual Motor Gestalt Test and Its Clinical Use. New York: The American Orthopsychia— tric Association, 1938. Buck, John N. "The H-T-P, a Projective Device,” American Journal of Mental Deficiency, 51 (1947), 606-610. ComisiOn del Consejo Nacional de Planificacidn EconOmica. Situacidn Demogréfica, EconOmica, Social y Educative de Guatemala. Guatemala, Ministerio de Educacidn Publica. Tercera Edicidn, 1963. Consejo Superior Universitario Centroamericano. Eficiencia del Sistema Educativo Primario en Centroamerica y Escolaridad que Produce. Recursos Humanos en Centroamerica. San Jose, Costa Rica: CSUCA, 1963. Cooke, E. "Art Teaching and Child Nature,” Journal of Education, cited by Goodenough, F., Measurement_9f Intelligencekby Drawings. New York: World Book Company, 1926. Duhm, Erna. "The Meaning of the Starting Signs in the Wartegg Drawing Test,” Psychological Research, 1952, 242—248, reviewed in Psychological Abstracts, 27, no. 5867. _E§posici6n del Presidente de la Delegacidn de Guatemala, Lic. Juan Anchisi Caberes Ministro de Educacion Publica. (Conferencia sobre Educacidn y Desarrollo Economico y Social en América Latina, celebrada en Santiago de Chile del 5 a1 19 de Marzo de 1962.) Guatemala, MisiOn de Asistencia Tecnica de la Unesco 1962. Goodenough, Florence. Measurement of Intelligence by Drawings. Chicago, Illinois: World Book Company, 1926. Guatemala en Cifras 1961—1962; Anual Guatemala: Direccidn General de Estadistica, 1962. 39 IO. 11. 12. 13. 14. 15. 16. 17. 18. 19. 40 Informe de las Labores Realizadasypor 1a Seccién de Estadistica Escolar. Anual. Guatemala: Ministerio de Education. Seccién de Estadistica Escolar, 1962. Informe Nacional sobre el Desarrollo de la Educacién, 1961-1962. (Presentado a la III Reunion Inter— americana de Ministros de la Educacion, Bogota, 1963. Utilizado con el titulo de Educacién, Ciencia y Cultura en Guatemala.) Guatemala: Ministerio de Educacidh Publica, 1963. Keith, J. ”Assessing Academic Achievement with Specific Variables of the Drawing Completion Test in Certain Sub—Saharan Tribal Groups.” A Pilot Study. Unpublished doctoral thesis, Michigan State University, 1963. Kerschensteiner, D. S. Die Entwicklung der Zeichnerischen Begabung, cited by Goodenouch, F. Measurement of Intelligence by Drawings. New York: World Book Company, 1926. Kinget, G. M. The Drawingypompletion Test: A Projec- tive Technique for the Investigation of Personality. New York: Greene and Stratton, Inc., 1952. Koppitz, Elizabeth M. ”Relationship between the Bender—Gestalt Test and the Wechsler Intelligence Test for Children,” Journal of Clinical Psychology, (October, 1958), 413—E16 Koppitz, Elizabeth M. "The Bender—Gestalt Test for Children: A Normative Study," Journal of Clinical Psychology, 16 (1960), 432—435. Lamprecht, Karl. ”Les Dessins d'Enfants Comme Source Historique,” Bulletin de l' Academie Royale de Belgique, cited by Goodenough, F., Measurement of Intelligence by Drawings. New York: World Book Company, 1926 Machover, Karen. PersonaligyiProjection in the Drawing of the Human Figure. Springfield, Illinois: Charles C. Thomas, Publisher, 1949. Matheny, K. B. ”An Assessment of the Usefulness of the Wartegg Drawing Completion Test as a Measurement of Intelligence Among Children. Unpublished doctoral thesis, Michigan State University, 1963. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 41 Muller—Suur, Hemmo. ”Psychiatric Experiences in the Wartegg Drawing Test,” Nervenarzt, 1952, 446—450. Pintner, R. and Toops, H. A. "A Drawing Completion Test,” Journal Applied Psychology, 2 (1918), 164— 173. Rouma, G. Le Language Graphique de 1' Enfant. Misch et Thron, Paris, 1913, as cited by Goodenough, F., Measurement of Intelligence by Drawings. New York: World Book Company, 1926 Sander, F. ”Experimentelle Ergebnisse der Gestalt Psychologie,” as cited by Kinget, G. M., The Drawing Completion Test: A Projective TeERHique for the Investigation of Personality. New York: Greene and Stratton, Inc., 1952. Stark, Rosemary. ”A Comparison of Intelligence Test Scores on the Wechsler Intelligence Scale for Children, and the Wartegg Drawing Completion Test with School Achievement of Elementary School Children. Unpublished thesis submitted as partial fulfillment of the requirements for an M.A. degree, University of Detroit, 195A Tanaka, I. I. "The Development of the Drawing Com- pletion Test as a Cross— Cultural Non— —Language Measurement of Academic Achievementamong Elementary School Children in Hawaii. Unpublished doctoral thesis, Michigan State University, 1964 UNESCO. World Surveyiof Education II. Primary Edu— cation. Paris, 1958. Wartegg, E. Gefdhl. ”Neue Psychologische Studien," cited by Kinget, The Drawing Completion Test: A Projec— tive Technique for the Investigation of Personality. New York. Greene and Stratton, Inc. 1932. Wartegg, E. "Gefuhl und Phantasiebild," Bericht fiber den XV Kongress der Deutschen Gessellschaft fur Psychologie. Jena, 1937. Wartegg, E. ”Gestaltung und Character, ” Beiheft 84 der Zeitschrift fur Angewandte Psychologie. Leipzig, 1939 Wartegg, E. Schichtdiagpostik, Der Zeichentest (WOT). Gotinga, 1953. S E C I D m P P A 42 APPENDIX A POPULATION DENSITY MAP REPUBLIC OF GUATEMALA 43 44 REPUBLIC OF GUATEMALA POPULATION DENSITY REFERENCES Inhabitants por Km: Less than IO IO to less than 25 25 to less than 50 NM 50 to lessthan 75 - 75 to less than 100 100 and over Population census of april 18, I950 "Direccién General de Estadl'stica Oficina permanente’ del Censo." The counties included in this study are those marked with heavy borderlines. APPENDIX B WARTEGG DRAWING COMPLETION TEST BLANK 45 WARTEGG DRAWING COMPLETION TEST APPENDIX C ACADEMIC ACHIEVEMENT PREDICTION SCALE SCORING SYSTEM 47 I_‘“"‘“—‘5T’Afi: I : fi:”-’<’f“’: if I: :1, : 7,", 3 ;imension Drawings Two or more 3—U or" I fleets witnout bac-~ ground detail I I . I 1 Single 3—D objey without background detail. I I : I I I I I . I - M I I . .I I ’ +I I j v 1”: I :l _ l_I ." 1 ~A+:+_‘ IQ‘I I _I+ _I .7 I —I—I—IwI ‘I — LIIIAIJ +I + I+I 12 SPECIAL CASES: i Duplication of or +I + I—I III 1. Two or more 2—3 drawings voicn are completion Lf +I — I—I .9 simple stick drawings 5 . n: stimuli + — I-. 2‘ more than single 2—D drawings. in wholes--or -I + I+I S 1in-:s. - + — I 2. Minus 1 point for each repetition: l —I — I—I 5 theme drawn. I A~iJr_l ,QJ I‘I*I "I I ’fi’d I I I I A ACADEMIC ACHIEVEMENT PREDICTION SCALE SCORING SYSTEM APPENDIX D CODE BOOK 49 50 CODE BOOK The Development of the Drawing Completion Test as a Cross—Cultural Non—Language Measurement of Academic Achievement Among Elementary School Children. John E. Jordan Otto E. Gilbert Instructions for the use of this CODE BOOK. U7 Code 9 or 99 will always mean Not applicable or Nothing. Code 9 or 99 will always mean there was No Information or the Respondent did not answer. Code 8 or 88 will always mean Don't Know, unless other— wise indicated. In each case in the following pages the column to the left contains the column number of the IBM card; the second column contains the ”variable” number used in the computer program; the third column contains an abbrivated form of the item; and the fourth column contains the code within each column of the IBM card with an explanation of the code. Coder instructions always follow a line across the page and are clearly indicated. In some cases when codes are equal to others already used, they are not repeated eacha time, but reference is made to a previous code or the immediately previous code with ”samefl I 51 1—1 CARD 1 Column Variable Question Detail Code 1, 2 1 Nation 01 — Hawaii 02 - Guatemala 03 — Japan 04 — 99 As assigned 3 2 Location (City) — 9 AS assigned - No information — Central — West — East - South — No information 4 3 Region (Guatemala) 1 9 l 2 3 A 9 5, 6 4 Group Number 1 — J. Palma (Guatemala) 2 - A. Marure ”A” 3 — A. Marure "B" 4 — San Esteban 5 — E1 Ingeniero 6 — Sta. Elena 7 — Vado Hondo 8 — Jan Jacinto 9 — Sébana Grande lO — Chamac 11 — Palo Gordo 12 — Champollap l3 — Federacidn l4 — quuihuila 15 — El Salto "A” 16 — El Salto ”B” 7, 8, 9 5 Respondent Number 001 — 000 As assigned 10, ll 6 Deck or Card 01 Number 12 7 Project Director — Tanaka 1 2 - Gilbert 3 — Cessna 4 — 8 as assigned 9 — No information 52 1-2 Column Variable Question Detail Code 13 8 Year of Adminis- tration 3 — 1963 4 - 1964 5 - 1965 6 — 1966 7 — 1957 8 — 1968 9 - 1959 14, 15 9 Month of Adminis- tration 01 — Jan. 02 — Feb 03 — March 04 — April 05 — May 06 — June 07 - July 08 — August 09 — Sept. 10 — Oct 11 — Nov 12 — Dec. 16, 17 10 Day of Adminis- tration Ol - 31 18 11 Administered by 1 — Tanaka 2 — Gilbert 3 - Cessna 4 — 8 As assigned 9 — No information 19 12 Sex of Respondent l - Masculine 2 — Feminine 20, 21 13 Age of Respondent* O6 — 6 Years 07 — 7 years 08 — 8 years 09 - 9 years etc., i.e., years 22 14 Population of — Rural 1 — 999) — Town 1000—4999) City 5000—and over) — 9 As assigned Stratified Area 41‘ LAJIDH l *“Round” age to nearest year. 53 1-3 Column Variable Question Detail Code 23,24,25 15 Grade Point 0 Record Average** to Actual Guatemalan Sam 1e 100 Score (Range 9 — 100 26,27,28 16 Drawing Completion 000 — 480 Test Scores 29, 3O 17 Total Number of Dimensions (DCT) 00 — 24 31, 32 18 Number of Objects OO — 77 (DCT 2 33,34 19 Number of Integrated (DCT)3 Drawings (Range 0.0 - 8.0) 00 — 80 35,36 20 Number of Meaningful DCT)A Drawings Range 0.0 — 8.0) 00 — 80 37,38 21 Number of Proportional (DCT) Drawings (Range 0.0 — 8.0 00 — 8O **Instruction to Coder: Col. 23, 24, 25 Number grades from 0 to 100 are used. Record directly as given. Col. 29, 3O (DCT) The total number of dimensions in the eight (8) drawings constitute this total. Each drawing falls into one (1), two (2), or three (3) dimensional category. The sum of the dimensions constitutes the total score for these columns which can range from 0 to 24 (e.g., if the subject drew three (37 one—dimension drawings, two (2) two dimension drawings, and three (3) three—dimension drawings, his score would be 18; C01. 21, 32 (DCT) The total number of objects in the eight (8) drawings determines this range of O_to _7_7_. Two or more 2—D - ‘ ' drawings which are simple stick drawings ,aw score as only one object. Abstractions 7 and designs are not scored as objects. Col. 33. 3A (DCT) Col. 35. 36 (DCT) Col. 37. 38 (DCT) 54 The total of eight (8) frames in which the stimulus has been integrated deter— mines this score. Where partial or pseudo integration is the case, these are scored as halves (.5). Since parts must be considered, Column 33 will be used for wholes and Column 34 for parts, thus the range is from www- Each frame is scored for meaningfulness. Meaningfulness is defined as the ability of the drawing to convey something of a representational nature to the examiner. Objects, designs, or other constructions which fit this definition for each frame and scored as one (1), thus the range is from Qvto 8 Total number of proportional drawings is determined by the scoring of one (1) point for every frame that the parts relate to the whole. Since pro— portionality depends upon meaningful— ness, no drawing without this quality is perceived to be proportional. A range of O to 8 is used as each frame is scored—independently. APPENDIX E DATA TABLES 55 56 TABLE 4.——One way analysis of variance of Wartegg total scores: geographical regions. Sum of Degrees of Mean Source of Variance Squares Freedom Square F Between groups 58809.989 3 19603.330 20.453 Within groups 267404.018 279 958.437 Total 326214.007 282 57 TABLE 5.——One way analysis of variance of Wartegg total scores: classrooms. Sum of Degrees of Mean Source of Variance Squares Freedom Square F Between groups 132853.115 15 8856.874 12.230 Within groups 193360.892 267 72A.198 Total 326214.007 282 TABLE 6 .——One way analysis of variance of Wartegg total scores: urban, town, and rural schools. Sum of Degrees of Mean Source of Variance Squares Freedom Square F Between groups 48077 2 24038.5 24.20 Within groups 278135 280 993.0 Total 326214 282 58 TABLE 7 .——One way analysis of variance of mean—grades: geographical regions. Sum of Mean Source of Variance Squares d.f. Square F Between groups 10950.447 3 3650.149 11.502 Within groups 88536.302 279 317.334 Total 99486.749 282 TABLE 8 .——One way analysis of variance of mean—grades: classroom groups. Sum of Mean Source of Variance Squares d.f. Square F Between groups 23713.956 15 1580.930 5.571 Within groups 75772.793 267 283.793 Total 99486.749 282 TABLE 9 ,——0ne way analysis of variance of mean—grades: urban, town, and rural schools. Sum of Mean Source of Variance Squares d.f. Square F Between groups 10224 2 5112 16.07 Within groups 89262 280 318 Total 99486 282 59 TABLE 10.——Simp1e correlations of the Wartegg total and its parts: total sample. a >: U) 44 H +4 U) H CO -v—l G) H 4.) r-I C: C: “5 0 C0 0 u—i c: B c e H 5 o (D O 0 JJ (H -.—I r-l no "—4 :0 (U 60 4—> n w m ;.o h c a C0 (1) (D C: a) 0 no "—4 O -a u a o p w o a Q :4 :4 o E E h 4.) cc o CU (U C.) -r~I :3 .Q C. (I) S-« > 3:0 Q zco Fl 2 m 1. Wartegg Total Score 1.00° .07 .44° .32° .27° .27° 2. Dimensionality 1 00° .12° —.07 .14° .14° 3. Number of Objects 1.00° .05 .31° .26° 4. Integration 1.00°—.24° —.10 5. Meaningfulness 1 00° .83° 6. Proportionality 1.00o N=283 °Significant at the .05 level. 60 TABLE 11.-'Mu1tiple correlation coefficients between Wartegg total and its parts: total sample. R R2 S S2 .57 .32 27.98 782.64 N = 283 TABLE l2.-—Highest order partial correlation coefficients between Wartegg total and independent variables including mean grades: total sample. Partial Correlation Variable Coefficient 1. Dimensionality .03 2. Number of Objects .37 Integration .38 Meaningfulness .14 KIT—DUO Proportionality .01 61 TABLE l3.—-Beta weights of the Wartegg total and the independent variables: total sample. Beta Weights Variable (N2283) l. Dimensionality .02 2. Number of objects .34 3. Integration .36 4. Meaningfulness .22 5. Proportionality .02 62 TABLE 14.——Simp1e correlation coefficients between scores and re-scores 0f the Wartegg total test and its parts: sub-sample Of total sample. ’ Simple Correlation Variable Coefficients l. Dimensionality .66° 2. Number of Objects .400 3. Integration .95° 4. Meaningfulness .580 5. Proportionality .76° 6. Total .84° N = 50° Significant at the .01 level. 63 .Hosoa mo. can so secoaeasmam. mwmuz O00.4“ , aoHHmGOHoLOQOLm .s 0mm. coo.H mmoofisemcflcmoz .m OH.I oqm.u oo.H coflomamoooH .m com. cam. mo. coo.a noobpr a0 .oz .: .aH. can. so.- ems. .oo.a sonacsoansosao .m can. can. 0mm. gas. so. .oo.H Hoooo wmoonoz .m owa. omH. 03H. cam. HH. oma. coo.H ncombw :moz .H .d a n n n m an n a d e 1 . w o.u a J O u a s J.+ u I J I. on O U 38 _ B I. .3 B I. on J I o J l O 0 e a u n I q u .m P 8 TL O nL. B O a I u u e I .3 s I. e o I. e 1 S 3, 3, I A S 8 AA .oaoEmm Hence "mused mpH com HmBOp wwooemz com moomngCmoE o0 coflomHOLLOO OHQEHmII.DH mqm